Demand-Responsive Transportation Systems

DEMAND-RESPONSIVE TRANSPORTATION SYSTEMS

Special Report 136 Highway Research Board National Research Council National Academy of Sciences National Academy of Engineering

•. , __;• -•-, .- HIGHWAY RESEARCH BOARD 1973

Officers James A. Moe, Director, California Department of Public Works William L. Garrison, Chairman Elliott W. Montroll, Albert Einstein Professor of Jay W. Brown, First Vice Chairman Physics, University of Rochester Milton Pikarsky, Second Vice Chairman Milton Pikarsky, Commissioner of Public Works, W. N. Carey, Jr., Executive Director Chicago David H. Stevens, Commissioner, Maine Department of Transportation Executive Committee B. R. Stokes, General Manager, San Francisco Bay Area Rapid Transit District Ralph R. Bartelsmeyer, Federal Highway Administra- Robert N. Young, Executive Director, Regional Plan- tor, U.S. Department of Transportation ning Council, Baltimore (ex officio) Frank C. Herrmger, Urban Mass Transportation Ad- ministrator (ex officio) Henrik E. Stafseth, Executive Director, American Association of State Highway Officials (ex officio) Ernst Weber, Chairman, Division of Engineering, Na- tional Research Council (ex officio) Charles E. Shumate, Executive Director, Colorado Department of Highways (ex officio, Past Chair- man, 1971) Alan M. Voorhees, President, Alan M. Voorhees and Associates, Inc. (ex officio, Past Chairman, 1972) Hendrik W. Bode, Gordon McKay Professor of Sys- tems Engineering, Harvard University Jay W. Brown, Director of Road Operations, Florida Department of Transportation W. J. Burmeister, Executive Director, Wisconsin As- phalt Pavement Association Douglas B. Fugate, Commissioner, Virginia Depart- ment of Highways William L. Garrison, Edward R. Weidlein Professor of EnvirQnmental Engineering, University of Pittsburgh Roger N. Gilman, Director of Planning and 11evelop- meit, The Port Authority of New York and New Jersey Neil V. Hakala, President, Esso Research and Engi- neering Company Robert N. Hunter, Chief Engineer, Missouri State Highway Commission George Krambles, Operating Manager, Chicago Transit Authority A. Scheffer Lang, Office of the President, Association. of American Railroads Saunders Mac Lane, Max Mason Distinguished Service Professor, Department of Mathematics, The Uni- versity of Chicago Harold L. Michael, School of Civil Engineering, Pur- due University D. Grant Mickle, President, Highway Users Federation for Safety and Mobility John T. Middleton, Consultant, Washington, D.C. DEMAND-RESPONSIVE TRANSPORTATION SYSTEMS

Special Report 136

Proceedings of a conference held June 12-14, 1972, at Ann Arbor, Michigan

Subject Area:

84 Urban Transportation Systems

Highway Research Board Division of Engineering National Research Council National Academy of Sciences National Academy of Engineering Washington, D.C., 1973 NOTICE

The conference reported herein was held under the aegis of the National Academy of Sciences-National Research Council with the approval of the Governing Board of the NRC. Such approval indicated that the Governing Board considered that the problem is of national significance, that solution of the problem required scientific or technical competence, and that the resources of NRC were particularly suitable to the conduct of the project. The institutional responsibilities of the NRC were then discharged in the following manner: The members of the conference committee were selected for their individual scholarly competence and judgment, with due consideration for the balance and breadth of disciplines. Responsi- bility for all aspects of this report rests with the committee, except that opinions and conclusions attributed in the report to individuals are not necessarily those of the committee, the Highway Research Board, or the National Research Council. Although the reports of Highway Research Board committees are not submitted for approval to the Academy membership or to the Council of the Acad- emy, each report is reviewed by a second group of appropriately qualified individuals according to procedures established and monitored by the Academy's Report Review Committee. Such reviews are intended to determine, inter alia, whether the major questions and relevant points of view have been addressed and whether the reported findings, conclusions, and rec- ommendations arose from the available data and information Distribution of the report is approved, by the President of the Academy, only after satisfactory completion of this review process.

ISBN 0-309-02099-9 Library of Congress Catalog Card No. 73-2355 Price: $3.60

Available from Highway Research Board, National Academy of Sciences, 2101 Constitution Avenue, N.W., Washington, D.C. 20418. CONTENTS

FOREWORD, Leon M. Cole v DIAL-A-RIDE PROJECT IN ANN ARBOR: LEGALITY, Jerold Lax 68 INTRODUCTION, Daniel Roos 1

DIAL-A-BUS: GUIDELINES FOR DE- General Issues SIGN AND IMPLEMENTATION 3 DEMAND-RESPONSIVE TRANSPOR- TATION AS SEEN BY THE TRANSIT WORKER, John M. Elliott 75 Implementation of Specific Systems FORD MOTOR COMPANY'S ROLE IN DIAL-A-BUS EXPERIMENT IN BAY DIAL-A-RIDE DEVELOPMENT: 1972 RIDGES, John A. Bonsall 10 AND BEYOND, Karl W. Guenther 80

TELEBUS PROJECT IN REGINA, USER PREFERENCES FOR DIAL-A- Wallace G. Atkinson 16 BUS, Richard L. Gustafson and Francis P. D. Navin 85 B-LINE DIAL-A-BUS SYSTEM IN BATAVIA, Robert P. Aex 23

MODEL CITIES DIAL-A-RIDE SYSTEM Dispatching Systems IN COLUMBUS, OHIO, William C. Habig 27 COMPUTERS, TAXIS, AND GRASS ROOTS TRANSPORTATION, Robert C. MODEL CITIES JITNEY TRANSPOR- Cherry 94 TATION IN BUFFALO, Michael White 31 USE OF DATA PROCESSING IN TAXI- CAB CONTROL, John Davidson 99 DIAL-A-RIDE APPLICATION IN GREAT BRITAIN, Philip R. Oxley 33 GENERAL PURPOSE COMPUTER DIS- PATCHING SYSTEM, Nigel H. M. DIAL-A-RIDE DEMONSTRATION IN Wilson and Bernard Trevor HADDONFIELD, D. W. Gwynn and Higonnet 102 Anthony U. Simpson 38

DIAL-A-RIDE DEMONSTRATION IN Banquet Speech HADDONFIELD: PLANNING AND INI- TIAL OPERATION, Douglas M. ISSUES AND POLICY QUESTIONS CON- Medville 44 FRONTING PUBLIC TRANSPORTATION, Elbert C. Mackey 105 DIAL-A-RIDE PROJECT IN ANN ARBOR: DESCRIPTION AND OPERA- TION, Thomas Urbanik, II 53 Participants 110 DIAL-A-RIDE PROJECT IN ANN ARBOR: PUBLIC RESPONSE, Michael J. Berla 61 Sponsorship 113 FOREWORD Leon M. Cole Professor and Director, Graduate Program in Community and Regional Planning, University of Texas at Austin

Research on demand-responsive trans- of Technology and the HRB Committee on portation concepts began in the 1960's and New Transportation Systems and Tech- was first widely reported in "Tomorrow's nology were joint sponsors, and the Ann Transportation: New Systems for the Arbor Transportation Authority and the Urban Future," the summary report of Highway Safety Research Institute of the extensive studies on new systems of ur- University of Michigan also participated. ban transportation conducted by the U.S. More than 125 persons from transit and Department of Housing and Urban Devel- taxi industries, labor unions, manufac- opment and published in 1968. That re- turers, government, consulting firms, port identified demand-responsive trans- and universities attended. In addition to portation as the most promising new papers and discussions, field trips were concept that could be implemented in the scheduled to inspect the Ann Arbor relatively near future. demand-responsive system and also the The Massachusetts Institute of Tech- work being done by Bendix, Ford, and nology, a major contributor to the early General Motors. research of such concepts, sponsored The Committee on New Transportation the First Annual Conference on Demand- Systems and Technology is pleased to Responsive Transportation Systems in sponsor this publication of the conference the summer of 1970. That was a 1-day proceedings. It is the second in a series conference at which M. I. T. researchers of reports on demand-responsive systems presented their findings. In September and services. The first, "Demand- 1970, the Highway Research Board's Actuated Transportation Systems," Special Committee on New Transportation Sys- Report 124, contains the proceedings of tems and Technology and Purdue Univer- the 1970 Purdue Conference. We are sity sponsored a 2-day conference that aware of the widespread current interest brought together both researchers and in information about changing develop- operators of demand-responsive transrans-- ments in demand-responsive transpor- portation. By the summer of 1971, 3 tation. The material in this report, pre- demand-responsive systems had been pared by such a broad cross section of implemented and several other systems people who have been gaining experience were planned for implementation during and learning lessons in implementing the next year. Reports were given on innovative transportation systems con- these systems at the Second Annual Con- cepts, represents the most current in- ference on Demand-Responsive Trans- formation of its kind on demand-responsive portation Systems, a 2-day conference transportation installations. sponsored by M.I.T. in the summer of 1971. Because interest in demand-responsive transportation continued to increase and because conferences served as one of the best communication channels between operators and researchers, the Third Annual Conference on Demand-Responsive Transportation Systems was held in June 1972. This special report contains the presentations and discussions of that conference. The Massachusetts Institute INTRODUCTION

Daniel Roos Associate Professor, Department of Civil Engineering, Massachusetts Institute of Technology

Demand- responsive transportation publication, but others represent tran- systems provide flexible personalized scripts of oral presentations. The latter point-to-point service in response to in- were edited by the conference committee dividual travel requests. Typically, chairman and by the HRB editorial staff there are no fixed routes and schedules. for clarity and conciseness. Instead, a dispatching center receives Bonsai described the oldest imple- telephone calls from customers and as- mented demand-responsive system, which signs vehicles to service the customers. is in Bay Ridges, Ontario. That system The objective of the dispatching opera- is primarily intended to serve as a feeder tion is to provide efficient, direct ser- to the GO Transit commuter railroad, vice to each customer, to group cus- although more recently it has been ex- tomers with similar origin-destination panded in the off-peak hours to a general pairs on the same vehicle, and to reduce area-wide service. The system in Regina, the cost of service to each passenger. Saskatchewan, described by Atkinson, Demand- responsive transportation primarily serves a feeder function to a systems complement conventional fixed- fixed-route bus line but also provides a route and scheduled systems. In low- more general area-wide service. Aex and medium-density areas, they can pro- discussed the operation of the area-wide vide a total transportation service where system in Batavia, New York, that also conventional, fixed-route buses are not replaced a fixed-route bus system. The economically feasible. In high-density Columbus, Ohio, system, described by areas, they can provide feeder service Habig, and the Buffalo, New York, sys- to line-haul facilities and thus fill the tem, described by White, are examples of existing void between conventional transit demand- responsive transportation in Model and taxi service. Cities areas. Oxley reported on several Many different types of demand- demand- responsive systems that are operat- responsive service are possible. These ing in or proposed for cities in England. include dial-a-bus, dial-a-ride, demand Two papers are included on the federally jitney, demand-actuated road transit, sponsored demonstration project in Had- computer-aided routing systems, genie, donfield, New Jersey, a community served telebus, call-a-ride and taxi-bus. by the Lindenwold Hi-Speed Line. Gwynn The l'Ecole Polytechnique, under con- and Simpson discuss the operation of the tract to the Transportation Development system, and Medville describes a user Agency of Canada, is preparing a 5- survey conducted by the technical monitors volume manual to aid those interested in of the project. the design and implementation of demand- Three papers on the Ann Arbor sys- responsive systems. A short report was tem include one by Lax, the city attorney, given at the conference on the progress of on the legal aspects of demand-responsive that project, and portions of the first vol- transportation, and his discussion will ume are included in this Special Report. undoubtedly be helpful to others who are All but two of the other papers presented involved in implementing similar sys- at the conference are also included. They tems. The other two papers contain re- can be grouped into two categories: those ports by Urbanik on the operation of the that deal with implementation of specific system and by Berla on the public's re- systems and those that deal with general sponse to the system. issues concerning many systems. Some The general papers include one by of the papers were formally written for Elliot on the relation of the transit Daniel Roos

workers' union to demand-responsive transportation, one by Guenther on the The Fourth Annual Conference on Demand- role of private industry, and one by Responsive Transportation Systems will be held Gustalson and Navin on the results of in September 1973. Those interested in attend- surveys of user preferences. ing or presenting a paper should write Daniel All of the implemented systems are Roos, Room 1-181, Massachusetts Institute of currently manually dispatched, and for Technology, Cambridge, Massachusetts 02139. that reason the two reports by Cherry and Davidson are particularly interesting. They describe compute r- dispatching systems that have recently been implemented in their taxicab companies. Wilson and Higonnet further discuss the use of computer-dispatching of demand- responsive vehicles. The decision to publish the conference proceedings was made for several reasons. First, they contain the most current information on demand-responsive transportation systems. Second, the conference had an unusual cross section of participants and ideas representing many different organizations concerned with transportation. Third, the experiences gained by those who have implemented demand- responsive systems are valuable for the implementation of similar or new systems. The presentations were made by drivers, owners, and operators of various implemented systems. Their intent was not to present scholarly papers but rather to explain their systems from their perspectives. One of the principal objectives of the conference was to bring together people from many different ori- entations to focus on an area of common concern. Therefore, there is a wide variation in the approach used for each presentation. An attempt was made during editing not to destroy the original flavor that each author wanted to convey. DIAL-A-BUS: GUIDELINES FOR DESIGN AND IMPLEMENTATION

The dial-a-bus concept is no longer an This is part of a 5-volume manual that is being abstract idea. Since 1964, it has been produced to aid those interested in the design implemented in more than a dozen mu- and implementation of a dial-a-bus system. The nicipalities in North America and Europe. manual will contain information on all aspects of Though dial-a-bus is still in its experi- dial-a-bus from concept to operation and is being mental phase in Canada, Regina has not prepared by l'Ecole Polytechnique de Montreal been the only city to start such a system. under contract to the Transportation Develop- ment Agency of the Ministry of Transport, Bay Ridges, Ontario, has experimented Canada. This portion of the first volume is in- successfully with it, and this transit op- cluded in this report to provide background in- tion might well succeed in your commu- formation on demand-responsive transportation. nity.

DIAL-A-BUS IS DESIGNED TO In January 1971, W. G. Atkinson, FILL A GAP transit manager for Regina, Saskatche- wan, suggested to Harry Walker, Mayor Dial-a-bus is basically a hybrid be- of Regina, that some of the city's most tween a bus and a taxi system. It is de- pressing transportation problems could signed to fill a need not met by other be solved by a new mode of public transit available transportation modes. It com- that could substantially reduce the transit bines the door-to-door flexibility of pri- deficit. He pointed out that the new sys- vate automobile travel with the economy tem could improve transportation in of public transit, and it provides a areas of low population density, elimi- greater mobility to persons who would nate waiting outdoors for buses (Regina otherwise lack the opportunity to travel. winters often have subzero tempera- An additional advantage is that it can tures), pick up passengers at their door- be implemented as a primary system in steps, take them to their destinations small communities that do not yet have a quickly and comfortably, cost each pas- transit service. Or it can be grafted senger only 10 cents more than a regular onto existing systems either by replacing bus ticket, operate at low cost during some bus routes suffering from low pa- both peak and off-peak hours, cover most tronage or by adding more convenient of its operating expenses, require a rel- feeder access to other conventional lines. atively small capital outlay, and create Unlike less flexible transit modes, dial- more jobs than other transit alternatives. a-bus needs no fixed routes, and, be- By the end of 1971, after 4 months of cause small buses can be used, narrow experimental operation in southwest winding streets are not out of bounds. Regina, dial-a-bus surpassed many of the objectives originally established for it in the Regina telebus feasibility study. HERE'S HOW IT WORKS Passenger revenues are close to covering operation costs. The system has When a person wants to go somewhere, carred a winter average of about 1,200 he phones the dial-a-bus number and passengers daily and as many as 2,200 on gives a dispatcher his address, destina- some bitterly cold days. The summer tion, and the time at which he wants to be average is lower, about 650, but that is picked up. The dispatcher passes the in- to be expected. formation on to one of the bus drivers Transportation Development Agency

assigned to the area. The driver pre- Figure 1. Dial-a-bus service patterns. pares his route from a list of customers' many-to-one many-to-few many-to-many addresses and picks up each passenger at or near his doorstep. :Iv For a guranteed pick-up time, a pas- 5' senger can phone for a dial-a-bus up to, 4 \ X say, 20 minutes before his planned departure time. The amount of advance S I notice needed depends on the length of the I "I route. The bus arrives at the passen.- ger's door within a few minutes of the time set by the dispatcher. When he possible origin-destination pairs - one possible dynamic bus routing steps on board, there is usually a seat waiting for him. Regular subscribers who use the system are assigned to the drivers in ad- plicity: It is convenient and easy to un- vance and are given priority over casual derstand. From the planner's viewpoint, users. the system's attraction is its flexibility: A passenger can phone for a dial-a- Nearly all the elements are adjustable. bus with less advance notice if he is will- The buses can be small, medium, or tug to take a chance that a bus is en route large—anything from an 11-seat modified near his home. Each bus is radio- van to a 42-seat bus. They can be mod- equipped, and a driver can handile occa- estly appointed or furnished with luxuries sional last-minute requests radioed to such as indoor-outdoor carpeting, air him by the dispatcher, if they do not dis- conditioning, and music. Not only can rupt the route. the routes be planned to satisfy demand, Dial-a-bus is adaptable to a wide trip by trip, but the level of service can range of routing and scheduling possibili- be easily adjusted without the customer's ties. For example, the Ontario commu- being affected or even being aware of the nity of Bay Ridges has set up what is change. Dial-a-bus can operate a fleet called a many-to-one (many origins, one of 2 vehicles or 200. The dispatching destination) service between commuters' system can be manual at first and com- homes and a commuter rail station. puterized later as fleets become larger. Regina provides an example of a many- If there is a computerized dispatching to-few service, which connects directly and monitoring system, the customer, to schools, nursing homes, shopping instead of requesting a bus from the dis- centers, and main bus routes. A many- patcher, can phone the dial-a-bus num- to-many (many origins, many destina- ber, dial a few additional digits, then tions) service is offered in Batavia, New hang up. The additional digits signal the York. These service patterns are shown size of his party, the time at which he in Figure 1. wishes to be picked up, the address of his destination, and a user identification code. The call is routed directly to the FLEXIBLE BUT SIMPLE central dispatching computer. The computer determines which bus is best able From the customer's point of view, to serve the customer and assigns it to the attraction of dial-a-bus is its sim- pick him up. Transportation Development Agency

ADDRESSES URBAN SOCIAL NEEDS Figure 2. Population distribution in urban areas.

So far we have been enumerating cer- 0 FRINGE AREA 0 D tain advantages particular to dial-a-bus 0 CITY PROPER 0 systems. Although dial-a-bus is flexible, 0 simple to use, and somewhat superior to 0 conventional bus service, these are not z the only attributes that justify its imple- 0 I- 4 mentation. The real case for dial-a-bus —J emerges as one examines the social and 0. 0 transportation problems that now face the 0. contemporary urban community, for it 56 61 66 56 61 66 56 61 66 56 61 66 56 61 66 addresses social needs in a manner that 6 conventional modes are hard pressed to 1.0 match. The following 4 sections are de- .0 4' o' voted to a more detailed discussion of this subject. Figure 3. Population distribution in urbanized areas. Boot i r' 0 FRINGE AREA 17 600 CITY PROPER The Low-Density Fact 4 uJ --

The number of people living in low- 400 0 density fringe areas is constantly rising (U (Fig. 2). While the size of the city N proper (high density) stays nearly con- 200 E stant, low-density development covers an cc increasing percentage of the metropolitan 0 area (Fig. 3). About 80 to 90 percent of 5661 66 56 61 56 56 61 66 56 61 66 56 61 66 .st te Montreal, Toronto, and Ottawa is low 0' density. ço' Transportation often appears to be an Figure 4. Employment distribution. afterthought in the planning and development of low-density areas. There are a 2.0 1. number of reasons for this. Costly I 0 FRINGE AREA i transit systems developed for use in i CITY PROPER I large cities cannot easily be adapted to o the growing needs and limited budgets of o II smaller towns. An additional problem is I I that even complex transit systems cannot 1.0 maintain the quality of service offered if (U II they operate in sprawling suburban areas >. II O Such overextended systems are charac- -J V/A terized by inflexibility, high costs, and 0. r,,A unsatisfactory service, all of which have o(U V//A i pci made public transit a poor alternative to the private car- 1964 1969 1981 Transportation Development Agency

The Transportation Challenge Social Objectives and New Markets

The emergence of low-density areas Dial-a-bus has shown it can trim has significantly altered the shape and transit deficits incurred by serving low- the structure of the cities. Of particular density areas. However, it should not be interest are the new travel patterns that viewed solely as a means for reducing have in many cases made existing sys- losses. It may be equally important to tems awkward and incomplete. achieve objectives such as service im- The largest single demand made of lo- provements in transit and mobility for cal transit still occurs in the daily mass the captive riders as we have shown shuttle between home and work. Yet above. Another feature of dial-a-bus of current trends point less and less toward considerable social importance is that it trips to and from the downtown core (Fig. is labor intensive—a desirable charac- 4). Urban travelers are now likely to teristic in an era of spiraling unemploy- travel from home to work and back with- m ent. out ever leaving the suburban fringe. Obviously a large city's transportation Now that urban populations are becoming system cannot be totally converted to dispersed over a greater area, the down- dial-a-bus. The latter is intended as a town region has ceased to be the only em- collection and distribution system for ployment center. Most administrative low-density and low-demand areas. Yet, offices remain downtown, but new fac- even within a large city, its flexibility tories and other commercial establish- allows it to reach passenger markets that ments are appearing in the fringe areas. traditional transit modes could never The result is a complex and diffused pat- serve. It is adaptable to sporadic and in- tern of trips to and from work. frequent demands such as late night The same helter-skelter effect is re- travel. A nighttime dial-a-bus service flected in traffic patterns during leisure in downtown areas could attract custom- hours, as people make more trips to new ers who hesitate to step outside the door shopping and recreational facilities in because of the rising crime rate and the low-density communities. These dif- prospect of waiting on a street corner for fused trips are increasing in number as the infrequent bus service. Dial-a-bus people have more leisure time. could also cater to special occasions and The work, shopping, and pleasure needs such as ferrying people home from trips discussed above may be regarded parties—a solution to New Year's Eve as the noncaptive transit market. A travel! group that presents an even greater challenge consists of people having no In Step With Urban Planning alternative to public transit. Included here are the young, the poor, and the New transportation systems generate aged. At the present time, people who commercial, industrial, and residential are concentrated in places such as development. When transportation ex- homes for the handicapped or retire- pansion lags behind fast-paced urban de- ment villages are frustrated in their velopment, the result is haphazard and travel desires. Trips to visit friends unplanned growth along its route. Con- are inconvenient and expensive, and, in sequently, private developers often seem many cases, the work trip is an im- by default to determine the city struc- possibility. ture. However, even if the city were Transportation Development Agency

Figure 5. User locations related to fixed-route Figure 6. User locations related to service. demand-responsive service.

TERMINUS

X XX Xl X x x

XX XXX X XX

XX XX Xix

Ix Xl

XI X X Xix X X

given control of developments around workable in about a dozen municipalities transportation routes, existing modes of in North America and Europe. public transportation would inhibit the As early as 1964, the first demand- ability of planners to establish an urban responsive bus system operated in Pe- infrastructure. Dial-a-bus could be a oria, Illinois, as a weekday subscriber more malleable tool. Because it is flex- service to take workers from their ible enough to serve existing urban hous- homes to factories in the morning and ing patterns and can also help structure back again at night. new communities, it can promote the In 1969, in Mansfield, a town of 60,000 orderly expansion of cities. Examples of in Ohio, a private transit operator ex- user location related to types of transit perimented with a dial-a-ride service service are shown in Figures 5 and 6. that linked a residential neighborhood to the town's business district. Via a radio-telephone hookup, passengers whose homes were off the regular route A TESTED SOLUTION could arrange to be picked up by dialing directly through to the bus driver, who Dial-a-bus is no longer a figment of doubled as a dispatcher. some planner's imagination; it is a tested In July 1970, the Ontario government system that has already proved itself introduced a dial-a-bus experiment in Transportation Development Agency

Bay Ridges, a residential community of collar commuter, ridership is concen- 15,000. Dial-a-bus serves as a feeder trated in the rush hours. However, large system for the GO Transit commuter rail installation costs and high ridership re- line and has succeeded not only as a quirements make it impractical for feeder system but also as this communi- small- and medium-sized cities; low- ty's first internal public transit system. density growth areas on the outskirts of The experiment in Regina, Saskatche- large cities limit its practicalness there. wan's capital city, was sponsored jointly Even where rail is required to serve by municipal, provincial, and federal high-volume corridors, there remains governments. The city has a population the need to collect and distribute passen- of 145,000 and a typical low density of gers to and from the stations. 5,000 persons per square mile. Dial-a- Bus systems with sizable patronage bus was introduced in 1971 to replace can provide a high level of service at a some bus lines where sagging patronage low fare; capital costs comprise the pur- was creating large deficits. chase of vehicles and garage facilities. Both Bay Ridges and Regina have high Buses have a medium capacity, have levels of automobile ownership, and the fact some routing flexibility, and can travel that those 2 experiments work indicates on most city streets. However, in new that drivers may leave their cars at home low-density areas, winding street pat- if they are offered a good alternative. terns preclude systematic coverage by bus. More important, because demand is more spread out in those localities than CAN EXISTING SYSTEMS ADAPT? in the older high-density areas, service is less convenient. Many homes in The problem of declining population fringe areas lie so far from the sched- densities underlines the need for a trans- uled stops that the bus does not offer portation system flexible enough to serve residents a practical alternative to the areas with a scattered range of origin private automobile. As a result, its and destination points. It is, therefore, passenger volume suffers. clear that, if public transit is to succeed, Automobile travel is comfortable, ac- it must be able to evolve with the city's cessible, and reasonably fast. Because growth and remain responsive to its fu- its route is almost unrestricted, it offers ture needs. Yet most conventional the most flexibility. But it does not pro- modes fail to meet this criterion. It is vide a very economical solution to urban interesting to see in what respects exist- transportation problems: The average ing modes are inadequate. car in urban use carries only 1.4 persons Rail transit (either commuter rail or per trip. Many trips involving 2 passen- subway) can carry a high volume of pas- gers or more are made for persons other sengers at a low fare. Trains are com- than the driver. Children who are chauf- fortable, dependable, and fast. But they feured to and from school are in this cat- have no, routing flexibility: Their tracks egory. Forty-five percent of Canadians can neither shift with changing transpor- are nondrivers. These include those who tation patterns nor service more than a cannot afford cars and those who because few primary arteries of heavy traffic of age or physical disability cannot drive flow (usually in and out of the city cen.- them. For both groups, taxis provide ter). To the extent that railroads are little alternative because fares are too geared to meet the needs of the white- high to permit their regular use. Transportation Development Agency

No published data have been found on pete with the private automobile. Every the total number of taxis operating in driver has been subjected to the frustra- Canada. However, it has been estimated tions that set in when no one is moving. that some 25,000 are in service. Though But a traffic jam is nothing more than an it is a highly personalized service, the overflow of individual drivers acting au- high costs involved do not, in general, tonomously in a demand-responsive permit regular taxi travel. manner. Though dial-a-bus systems can offer Dial-a-bus is demand responsive in a similar services at a lower cost to the broader sense. By coordinating their user, the taxi business will not neces- requests through a dial-a-bus dispatcher, sarily suffer from this competition. commuters may ultimately do away with First, there will be individual transpor- the sort of traffic congestion that private tation needs that can only be met by the automobiles are responsible for. In ad- faster, more flexible, and more person- dition to reducing pollution and conges- alized taxi service. Second, even in tion, dial-a-bus can serve those sections small towns, dial-a-bus serves a market of the population that do not have access quite different from that of taxis because to private cars. it is designed to lure drivers out of their private automobiles and to replace inefficient bus routes. Further, taxi companies can act as operators of dial-a-bus systems. A number of shared taxi services already embody some aspects of dial-a-bus. Even this brief account has demonstrated the inadequacy of existing modes. The inflexibility of rail and bus lines, the high costs of taxis, and the shortcomings of the private automobile—all of these weaknesses argue for an alternative that incorporates the strengths and avoids the pitfalls of conventional transit options.

DIAL-A-BUS IS A DEMAND-RESPONSIVE SOLUTION

The private automobile is the mode over which the user exercises the most direct control. Though it is convenient, parking usually presents a problem, and, where the vehicle is shared by a family, it is not always accessible. Dial-a-bus is an attempt to create a form of public transit that is controlled by its users— transit that is demand responsive. It is for this reason that dial-a-bus can corn- DIAL-A-BUS EXPERIMENT IN BAY RIDGES John A. Bonsall Operational Planner, Ontario Ministry of Transportation and Communications

The Bay Ridges dial-a-bus experi- household survey in the community showed ment was planned and designed and is cur- that the average driver estimates his trip rently being operated by staff of the On- cost, excluding parking, to be about 82 tario Ministry of Transportation and cents, which is almost identical to the Communications. The experiment was train fare. For those drivers who do pay implemented in July 1970 to demonstrate for parking in downtown Toronto, the the potential of a demand-responsive same survey showed that the average transportation mode as a feeder service daily parking cost is about 80 cents. to a commuter rail facility. The deci- Bay Ridges itself is typical of many sion to implement a feeder service dem- modern Canadian middle-income subdi- onstration project may be traced to an visions; 94 percent of 3,000 dwelling units original recommendation of the Metro- are single-family houses serviced mainly politan Toronto and Region Transporta- by a street system of crescents and cul- tion Study, which was conducted during de-sacs and, therefore, lacking the through the mid- 19601 s. This study lead to the streets necessary for an efficient fixed- establishment in May 1967 of a com- route bus system. The gross area of the muter rail service (GO Transit) linking community including Frenchman's Bay is several suburban communities with about 1,200 acres, but the actual occupied downtown Toronto and to a limited ex- area is only 800 acres and houses a pop- perimental fixed-route feeder bus ex- ulation of approximately 14,000. It is not periment in Bay Ridges, the eastern ter- a sell-sufficient community, for there are minal of the commuter rail system. almost no job opportunities available in Bay Ridges was chosen as the site for the area. the feeder bus experiment largely be- The dial-a-bus service designed for cause it was a well-delineated community Bay Ridges comprises 2 basic compo- with no existing transit service. There- nents.. The most important is the many- fore, unlike most other communities to-one feeder service that operates on served by the commuter rail service, 20-minute cycles during peak periods and there was no existing transit franchise to 60-minute cycles during off-peak periods. complicate the implementation of the ex- These cycle headways, of course, are periment. governed by the commuter rail headways. This fixed-route feeder system was Hours of service are from 5:30 a. m. to operated from May 1967 to March 1968 1:00 a. m. the following day. The second when it was discontinued because of the service component is an off-peak many- low level of its patronage. There was to-few system that has a limited many-to- thus no internal transit system operating many capability centerd on the local shop- in the community during the 2 years im- ping center and operating from 8:00 a.m. mediately prior to the introduction of the to 4:00 p. m. This service was inaugurated dial-a-bus service. Ridership on the in February 1971 to utilize some of the commuter rail service, however, flour- spare system capacity during off-peak ished during this period; 52 percent of hours. The cycle time for this service is the 800 or so daily person work trips 30 minutes. originating in Bay Ridges and destined Neither the feeder nor the local ser- for downtown Toronto were made on the vices require very sophisticated dis- rail system. The 20-mile trip to down- patching techniques. In the case of the town Toronto takes about 40 minutes feeder service, a potential rider simply whether by road or by rail, and a recent phones the dispatcher a minimum of 1 Bonsall 11

hour before he requires service. His Figure 1 shows the total average request is recorded by the dispatcher in weekday patronage experienced by the a specially prepared log and then trans- experiment during its first 2 years of ferred either manually or by radio to the operation. Although typical seasonal map used by the driver. The driver variations and the influence of special uses this marked-up map to plan his events in Toronto are apparent, a con- route through his particular zone. tinuous growth in patronage is exhibited. People requiring service from the sta- The current annual average daily patron- tion simply board the bus at the station age is about 530 although a high of 980 and give the driver their destination was reached last March. The patronage addresses as they do so. Hence, as he of the local off-peak service, which leaves the station, the driver has in his represents about 20 percent of the total bus a map marked with both pickups and daily patronage, has been analyzed as to drop offs he must make before returning the percentage of true many-to-many to the station at the next train-arrival trips made. It would appear that only time. 24 percent of the trips can be so classi- In peak periods a total of 4 buses are fied. Overall operating statistics for the allocated to the service; each one covers experiment show that the average trip 1 of 4 zones. This equipment require- productivity per mile is 1.5. Interest- ment is reduced to only 1 vehicle in the ingly enough, this figure is fairly similar off peak except when the local service is to that exhibited by the Regina telebus operational. During this latter period service. the addition of 1 further bus allows the Inasmuch as a fixed-route service was operation of a 30-minute-cycle service operated in the community sometime that meets at both the station and the prior to the introduction of the dial- a- local shopping center. Each bus is as- bus, it is interesting to compare the signed to one-half of the community and operating characteristics of the 2 types sweeps its assigned zone every 30 of transit service. It should, however, minutes before returning to the shopping be noted that the fixed-route service was center where a passenger can make a was limited to peak service only. The free transfer to the bus allocated to the route system was a 1-way loop typical of other zone. The call-in time for this that provided in many low-density sub- divisions. The fare was initially 10 cents service is '/2 hour as opposed to the feeder service's requirement of 1 hour. Apart from this, the service's operation is essentially similar to that of the Figure 1. Weekday patronage. feeder service.

The buses used in Bay Ridges are BOO basically a special conversion of a Ford Econoline van costing about $7,500. 600 Perimeter seating for 11 passengers and a details such as the raised roof and lowered 400 entrance step are provided. Our experience in Bay Ridges suggests that the 200 11-seat vehicle used there is too small

and that a 15-seat arrangement would TIME have been more appropriate. Bonsall 12

but was raised to 20 cents, and both use which compares very favorably with that of fixed stops and flagging of the bus of the private automobile. In terms of were encouraged. A survey done at the coverage, dial-a-bus by its very nature time showed that 60 percent of the riders covers 100 percent of any community could be classed as captive to the sys- being served. For a fixed-route system. When a comparison is made be- tem, current thinking suggests that the tween the trip productivity per bus mile catchment area extends about '/ mile on of the 2 services, it is apparent that either side of a bus route. Applying this dial-a-bus often exceeds the productivity parameter to the situation in Bay Ridges of the fixed-route system. shows that the original fixed-route ser- Figure 2 shows probably the most vice was within this distance of the significant result of the Bay Ridges ex- majority of the community's residents. periment, which is the change in the use The considerable increase in transit of the 3 modes used to reach the GO riding on dial-a-bus, as compared to station. Seven months after implementa- that on the fixed-route bus, must, there- tion, dial-a-bus had become the most fore, be largely a function of the door- popular means of getting to the GO sta- to-door service provided. tion. The significance of this result is Analysis of the current ridership by underlined by the situation in Bay Ridges time of day and direction shows the very of free parking at the station and an peaked nature of the service demand. average dial-a-bus fare at the time of Approximately 64 percent of all -riding the survey of 22.5 cents. This same occurs in the morning and evening peak survey showed that some two-thirds of periods. The directional split shows the present dial-a-bus patrons can be greater riding from the station than to it. considered as noncaptives. Since the The household survey determined that it data shown in Figure 2 were compiled, was the availability of a ride in the 1 there has been a further increase in direction that accounted for the majority dial-a-bus riders. of this imbalance. Only 8 percent of the The magnitude of the impact of the 1-way riders stated that the inconve- dial-a-bus service on trip-making nience of making the phone call was the within Bay Ridges may be explained reason for their unequal use of the ser- partly by the high level of service vice. The implication of the directional offered. As an example, the average nature of the trip-making and also of the trip time on dial-a-bus is 6.7 minutes, fact that 80 percent of the trips to the station during the morning peak are actually prebooked on a weekly basis is Figure 2. Access mode to station. that there is considerable spare dispatch- hOED ROUTE DIAL-A-BUS DIAL-A -BUS ing capacity within the existing system. FEB./ 71 We have analyzed the dispatching demand and have determined that the practical limit for a 1-man system is about 50 calls per hour. This means that the present Bay Ridges dispatcher could handle a 50 percent increase in trip- making befora a significant increase in dispatching manpower was required.

DISTANCE FROM STATION Incidentally, our analysis of the dis- Bonsall 13

patching demand also showed that even some 31 percent of off-peak costs are at- after several months of service about 4 tributable to the dispatcher. Costs or 5 information calls on the average per average 80 cents per bus mile of which hour are still made. 24 percent is dispatching cost. The analysis of annual costs and Unfortunately the homogeneous nature revenues showed that about 50 percent of the Bay Ridges community has pre- of the costs are now being covered by cluded any investigation of the influence revenue. The pertinent data are sum- of socioeconomic factors on dial-a-bus marized below. ridership. The only factor of any significance that was determined from the Costs household survey was the predominance Fixed $ 2,000 of 1 car/2 driving license households Semivariable 13,000 among the population using. dial-a-bus Variable as compared to the total community pop- Drivers 42,000 ulation. Dispatcher 18,000 As a result of an evaluation of the dial- Vehicle 8,000 a-bus system that was carried out in July and August 1971, certain operational 83,000 and service changes were recommended. Revenue 41,000 These changes included the elimination $42,000 of Sunday service, the curtailing of evening dispatching and its replacement with The current deficit of $42,000 is for a an answering service, and the raising of 6-day-a-week, 20-hour-a-day service, the fares from 25 cents cash and 20 cents which is significantly more service than ticket to 30 cents cash and 25 cents communities of the size of Bay Ridges ticket. The changes were all imple- usually have. If evening service were mented in May 1972. Since then, costs curtailed after 8 p. m., the deficit could have obviously been reduced by the elim- be reduced to about $30,000, which is ination of the Sunday service and the not out of line with the deficits of many reduction of evening dispatching. Our of the smaller transit operations in initial impression of the changed dis- Ontario. These figures must be viewed patching system suggests that its effect with some caution, however. The on evening ridership is minimal. The driver's wage rate of $3.64 per hour is evaluation of the fare increase impact on below that of many public systems, and the other hand is complicated by the the impact of increasing this rate may be larger influence of seasonal ridership judged from the fact that 72 percent of variations. A comparison of the rider- all costs are attributable to the driver's ship since the fare increase with that wage rate. Further examination of the during the same period last year shows trip-making and associated costs by no decrease in the dial-a-bus patronage. time of day shows that, if all capital Further analysis is required to deter- and other fixed costs are charged to the mine what impact the fare increase might peak-period trips, the cost per trip have had on the overall growth in pa- varies from 50 cents for the peak to tronage. $1.36 for trips made on a Sunday; the The following initial conclusions can average is 55 cents. Better dispatcher be drawn from the Bay Ridges experi- utilization could reduce this figure, for ment. Bonsall 14

Dial-a-bus can compete success- Question: What happens if you do not fully with the private automobile for use the train? work trips in low-density areas; Answer: The time is slightly longer, Dial-a-bus is a premium service but you wait in your own home. And, if that will tolerate a premium fare in you are a regular user, you, of course, middle-income areas; and design your schedule around the regular For line-haul transit, dial-a-bus vehicle-arrival time. must be considered as an alternative or Question: What is the average wait supplement to park-and-ride service. per passenger? For the Bay Ridges design, the minimum Answer: We require a 1-hour calling average trip cost is about 50 cents. time, but on the local service we reduce that to '/2 hour. We do not advertise that we will accept people on a shorter calling time than that; but, if there is INFORMAL DISCUSSION capacity on the bus and it is in the right Question: Is weather a significant place, we can pick up people within a factor in dial-a-bus ridership? few minutes of the call-in time. Answer: The winter of '71 made Question: What is the average trip people start using the dial-a-bus system. time to the train station in the morning? But the important point is that once people Answer: About 2 or 3 minutes for used it they stayed with it. Two-thirds those who have booked regular service of the people who use it say that they are to meet the train. not captive to the system but have alternative transportation. Nearly 70 per- Question: Do the buses, then, essen- cent of the people who used the old fixed- tially pick up the same people every day route system were captive to it; that is, and carry them to the same place? they had no car. Why has this happened? Answer: During the morning peak, 80 One factor is trip time. The average trip percent of our riders are regular riders time on dial-a-bus is about 6 minutes and who are booked by the week. on the fixed-route bus was about 12 min- Question: How many hours is each bus utes. The latter time did not include ac- driver on duty? cess time. That 6 minutes compares Answer: We use some part-time quite favorably to the time it would take drivers for the peaks, but those who are you to get into your car and drive it from full time work for 8 hours. the station to your home. Question: What share of the house- Question: Is that the time in the holds have used dial-a-bus at any time? vehicle? Answer: Our household survey indi- Answer: That is the time from get- cated that about 70 percent of the house- ting off the train to getting off the ve- holds have tried the dial-a-ride system. hicle at your home. Question: Has any consideration been Question: Is any of that time spent given to a subscription service to elim- waiting for the vehicle? inate the dispatcher? Answer: No, the vehicles are there. Answer: We have discussed this but You get off the train, walk a few yards, feel that there is enough work for the dis- and get on the bus. patcher because there are enough trips Bonsall 15

being made that are not on a regular basis. Question: When the weather is bad and there is an extra heavy load of people so that the driver cannot make his rounds in time to catch the train, what happens? Answer: The dispatcher has the final say in controlling the buses, and he can, in those circumstances, reallocate his own boundary. For instance, a zone 1 bus might pick up some people in zone 2 if there is a heavy demand. If there is a heavy demand throughout the community, we do have a standby bus that we can use if necessary. Question: How many dispatchers do you have? Answer: We have 3; 2 are on duty during the day. Question: What is the average fare? Answer: Before the increase, the average fare was 221/2 cents. Now it is about 261/2 cents. TELEBUS PROJECT IN REGINA

Wallace G. Atkinson General Manager, Regina Transit System

Regina is the capital of Saskatchewan the driver in your area, and the driver and has a population of about 140,000. It is includes your address in his route. You located approximately 600 miles due north are picked up and delivered to your door of Denver and was built on the prairie; when you want to travel. every tree in this city was planted, for All kinds of people use telebus: there are no natural trees. The down- teachers, nurses, housewives, business- town is relatively new, and most of the men, and grade-school and high-school large buildings have been built in the past students— every segment of society. 10 years. For the first time, the aged, the handi- The Regina system and the one in Bay capped, and the young are able to travel Ridges are quite similar in that they are on their own without assistance from manually dispatched, and they provide friends or relatives. neighborhood service to few destinations Regina's version of telebus is de- (55 percent of the riders go to the central signed as a feeder system. Passengers business district). The difference is that are transported to a transfer depot where the Regina system is fully integrated with they board a full-sized bus to travel to an existing transit operation. The tele- their final destinations. They return to bus system carries about 5 percent of the the transfer depot by the connecting ser- total transit patronage in the city. In ad- vice and continue to their homes by tele- dition to CBD trips many other services bus. The answering service receives are provided; 35 percent of our passen- calls anytime during the day or night. gers go to local destinations, and about Regular customers may book service in 10 percent go to schools within the area. advance and thus eliminate the necessity So, it is roughly 10 percent many-to- of making daily phone calls. The tele- many, about 35 percent many-to-few, and bus then simply arrives at its specific about 55 percent many-to-one. Telebus time each day, and the service becomes serves about 6 major destinations in the as routine as getting out of bed. You find area—the government complex, 2 hospi- yourself getting to know the driver and tals, and 3 shopping centers. The Regina the other regular customers on your ridership curves are not quite so steep as route, and soon you are traveling with a those for Bay Ridges; there is heavier off-peak usage of the Regina system. On September 7, 1971, the telebus Figure 1. Regina telebus. pilot project was put into operation (Fig. 1). After 4 months of operation, serving

an area of about 3 square miles and a ---I------ - population of 18,000, the public's reaction to the tele-bus system had already - I guaranteed its future success. On Thurs- day, January 27, 1972, the pilot project LOX was slightly more than 4 months in exis- . tence, and no fewer than 2,200 people 7EI4Øjg1 traveled by telebus in a single day. By March, some 100,000 passengers had been carried on the new service. Telebus begins with a phone call from your home. The dispatch office radios Atkinson 17

group of friends rather than total strang- bus operations are given in Table 4. e rs. In summary, I would like to say that To transit management, telebus is an the public votes with their feet In Regina, exciting answer to the growing needs of we have 400,000 votes for telebus. In an expanding city. Because telebus re- Bay Ridges, and in Ann Arbor, there are places more costly fixed-route buses and similar responses. Cities that have put provides better service, it is receiving enthusiastic support from city councils. But the greatest response of all is from Table 1. Fares and ridership for 1971- the passengers-the people who have 72 winter. found their kind of transit service. They love it, and they will tell you so. Fixed Our total transit system carries about Item Route Telebus 8 million passengers per year. Usage in Fares, cents the city on the regular system is about 60 Adults 25 35 Students 15 25 rides per capita annually. This usage is Children considerably higher than that in a com- Cash 10 15 Tickets' 6.7 11.7 parable city in eastern Canada or in the Senior citizens 16 26 United States because of the past history Public passes 18 25 of transit service. The Regina system Avg 19.5 29.0 has been city-owned and -operated since Percentage of riders Adults 47.3 58.4 it started in 1911, and, with the exception Students 13.0 8.3 of a few years in the 19201 s and a few Children years in the 1940's, it has always been Cash 6.0' 11.1 Tickets 6.0 3.5 subsidized. As a matter of policy to keep Senior citizens 11.8 8.3 the fares as low as possible, it is sub- Public passes 15.9 10.4 sidized at the rate of one-third the cost Total 100.0 100.0 of providing the service, which roughly 'Plus 5 cents. covers the capital and administrative bSurcharged to cents on July 1. 1972, but not included here. overhead. Our fares are similar to the 'Estimated. Ontario fares but are lower than those in most large cities in Canada and in the United States. Because of subsidization Table 2. System operations for 1971-72 winter. policies in western Canadian cities, transit fares have been kept low and usage Item Fixed Route Telebus has remained high. On the telebus, an Operating vehicles 66 6 adult pays a 35-cent fare and gets a free Operating hours 164,000 16,200 transfer to the regular system. Coming Operating mileage 1,710,000 183,000 Annual passenger trips 7,575,000 305,000 back on the regular system, he pays 25 Population served 122,000 15,200 cents and another 10 cents when he gets Trips per capita 62 20 on the telebus with the transfer. So, Area served, square miles 2.75 Demands/bus hour 46 18.8 one can go anywhere in the city for 35 Demands/square miles/hour 21 cents. Statistics on operations, fares, Schedule speed 10.4 10.7 Passengers/mile 4.42 1.76 and costs per revenue-mile for both the Operating speed - 12.6 telebus and the fixed-route systems are Average vehicle tour, miles - 4.0 Capacity utilization, percent 80 50 given in Tables 1, 2, and 3. Additional Transfers, percent 22 54 statistics on initial and increased tele- Atkinson 18

Table 3. Costs per revenue mile. 19 72-73 winter in the existing area. We had had a tremendous amount of Fixed Item System Telebus publicity across Canada, and it is filter- ing into the United States. The public is Operators' wages 0.65 0.52 absolutely demanding this kind of service. Fuel and maintenance 0.18 0.18 Capital 0.17 0.12 They are asking for it in London, in Ontario, Overhead 0.21 0.37 in Winnipeg, in Vancouver, in Pasadena, Total 1.11 1.19 in Hollywood, and anywhere the public has seen any publicity. Sit down with any person that has never heard of this system before, and in 10 minutes you have a Table 4. Initial and revised telebus operations. wildly enthusiastic passenger. It is an automatic reaction: "1.1 we had that kind Winter Winter of bus service in our community, I would Item 1971-72 1972-73 ride." And they do. We proved that they Area served, square miles 2.75 do; many people who ride the telebus Peak 2.75 Off peak 5 were never on a bus in their lives before. Population served 15,200 We are getting people who could not start Peak 18,000 Off peak 32,000 their cars on a cold morning and took the Future addition 30,000 bus and then continued to take it. We are Service interval, minutes Weekday peak 15 nc getting the young children who had to be Weekday off peak 20 nc driven everywhere by their parents but Week night 40 nc who can now safely ride the telebus. Saturday 20 nc Hours, a. m. to p.m. There are children on the system who are Weekdays 6:30-11:35 6:00-11:35 going to day nurseries. How many 3- and Saturday 6:45-9:35 nc Number of vehicles 4-year olds can ride a regular transit a.m. peak 6 12 system safely? We have handicapped Off peak 3 8 people who cannot walk to a bus stop. p.m. 6 10 Noon 5 8 We have people with cystic fibrosis, mus- Night 1 2 cular dystrophy, asthma, and other handi- Saturday peak 5 7 Passengers per day caps so that they cannot walk to a bus Avg 1,200 2,000 stop, cannot stand in the rain, or cannot Maximum 1,955 3,500 Demands per bus hour 20 22 walk on a dusty day. We have a door- Avg fare, cents 29 32 stop service for these people. We are Vehicle tours, minutes also catering to all sorts of other special Peak 30 nc Off peak 20 nc needs. If the system has any trouble, we Nights and Saturday 40 nc can call the passengers at home and tell Scheduled speed, mph 10.7 nc them what happened. Note: nc = no change Our biggest problem at the moment is that we -cannot get the equipment manufacturers to build the type of equipment in an integrated system have had similar we need. This applies to both the buses successes and are still on a growth curve. and the communications systems. This I estimate that Regina has only tapped 50 is where the big hold-up is going to be in percent of the market in the initial sys- expanding these systems. The software tem. We are expecting to go to about has been developed, but the hardware is 3,000 passengers per day during the limited. I hope that we can put as much Atkinson 19

pressure on as many city councils, poli- Question: What difference is there betic ians, and equipment manufacturers as tween your peak and off-peak service in- we can so that they will get busy and de- tervals and your peak and off-peak vehi- velop and implement the system that the cle tours? public is demanding. I think the public Answer: In the peak period, the ve- has already voted for this system. The hicle tour is 30 minutes. We lap 2 ve- revolution is coming And if transit hicle tours 15 minutes apart, and so there managers do not get involved, it will roll is 15-minute service. In the off-peak, we right over them like a steam roller. run a 20-minute vehicle tour, and there is That is why we got involved, and that is just 1 bus in each area; we provide 15- why other transit people are getting in- minute service in the peak period and 20- volved. minute service in the off-peak period. The main-line buses run every 15 minutes in the peak period and every 20 in the off- INFORMAL DISCUSSION peak.

Question: In the peak hour, does one Question: Apparently, you have a va- have to walk to a station? What is the riety of equipment. Can you say anything walking distance? about what good luck or what bad luck you Answer: People have the option of have had with this system? walking to the main-line station or being Answer: We operated the system all picked up by tele-bus at their doors. winter and carried heavy peak loads with There are about 1,500 apartments within a standard 42-passenger transit bus 3 blocks of the main line to which tele- equipped with radio. And about all we did bus passengers transfer. We have people was paint a green stripe on it. Now to do who ride as few as 2 blocks on a rainy this was pure hell, because it went down day and are quite willing to pay 35 cents to 40 below zero on January 27, and we to do so.. could hardly steer the bus. The normal transit bus is not equipped with power Question: What are your thoughts on steering. It became almost impossible to carrying school children on the telebus get up and down the crescent streets. But system? we operated all winter long and carried the Answer: Carrying school children peak loading. It was not until March 1972 puts a peak load on top of the normal that we received our small vehicles. The work-trip peak load, although shifted success of the system in severe weather somewhat. The work peak occurs at was with the big equipment. We think we 8:00 in the winter and at 7:30 in the sum- will have greater success now that we mer, when we go on earlier work hours. have the small buses. We have two 14- School trips peak between 8:30 and 8:45 passenger Dodge Maxi-Vans. They are (school starts at 9:00), and we can ac- about a foot and a half longer than the commodate that peak with buses that have buses used in Ann Arbor and in Bay finished carrying work trips. But serv- Ridges. We have in addition a 23- ing the school peak load is expensive. It passenger Flxette in service and more requires more equipment, and, of course, 14- and 18-passenger vehicles on order. the children's fares are only 15 cents. We are also using a unibus that seats 23, But we think we have to provide a total similar to the one used in Bay Ridges. It service to the community. has a good body shell, but it has been Atkinson 20

mounted on atruck chassis, and so it rides Question: How many months in the like a truck. We have had to completely winter do you have weather so bad that rebuild the suspension system because the an automobile would have trouble steering people in the back seat were being thrown a or traveling through snow? foot in the air' If the driver hits abump, Answer: The write-up we got in Time up they go! The kids love it; it is usually magazine indicated that it was 20 below full of kids in the back seat. We have had in Regina all year round! Actually we endless problems—broken springs and have very severe weather from about shock absorbers that keep tearing loose January 10 to the end of February when from the frame. I will not mention the temperatures are 35 to 40 below at night supplier, because these problems apply and up to maybe 0 in the daytime. This to every piece of truck equipment in presents problems in driving cars. How- transit service today. They are all lousy ever, I would not want you to believe that We do not have a decent, air-ride, small a demand on this system is created be- bus for this kind of service. We expect cause the weather is cold; any bad to buy 7 buses by this winter. Frankly, weather—rain, wind, dust, heat, and so we do not know what to buy. I think I can on—affects ridership. If we had a rain safely say, at this moment, there is not a shower this afternoon in Regina, the pas- sufficiently reliable small transit bus in senger level would go up by 100. When the 20-passenger range available. I may the women get their hair done, they do get some frowns from the manufacturers, not want to walk in the wind, and so they but we have looked at 7 different com- get on this system. panies and have not found an acceptable small transit bus. We think there is a Question: Is that another way of say- real technology gap between the type of ing that if the weather were nice all the equipment that we need and the type of time, like it is in Miami, you would not equipment that is available. If anybody have so many passengers? can come up with a transit bus in the 20- Answer: Not at all. People do not to 24-passenger range, we will be very like to walk in hot weather either but happy. We think there is a big marketfor prefer to ride in air-conditioned buses. it. The weather is anything but nice for walking around in Miami. And, of course, it rains down there, too! Question: What is the ratio of buses required to population served? Question: How do the drivers feel Answer: Our original feasibility study about this system? Are they bidding for estimated 1 bus for 700 dwelling units, this work? but we think that can go up with demand. Answer: The drivers do bid for this Present demand is 21 passengers per work. There is a certain amount of square mile per hour. We think that 1 status in being involved in the new sys- bus can handle a square mile at a demand tem. We did not preselect drivers. We level of about 35 per hour. We think that let them bid in the order of seniority, we can run 20-minute vehicle tours and but we reserved the right to remove a carry about 12 persons per tour, which driver if he were unable to handle the would be about 36 per hour. At those run. Some of them have asked to get loading capacities, we would break even off after they tried it because they did on operating costs. not like the barrage of information com- Atkinson 21

ing over the radio. Our drivers come In October 1973, we are going to tn- from the middle third of the seniority pie the system. We are going to have list. The oldest drivers will not touch a new communication system—fancy it (they do not want to learn anything new, radio-telephone dispatching, direct phone for they are waiting for retirement), and to the bus driver in the off-hours, and a the youngest drivers do not get a chance full super-deluxe system with provision to get at it. for teleprinters and all kinds of computer equipment. We expect to have 60,000 Question: What is your street pattern residents on the system by October. like? Answer: The street pattern in Regina Question: Would you elaborate a little is almost identical to the one in Bay on who is riding the system? Ridges—crescents, cul-de-sacs, and bays. Answer: Everybody is riding. We In some cases we cannot get into those have children as young as 3 years, senior bays in winter, and so we phone the people citizens, and business people. The heav- and ask them to come out to the end. But jest riding is coming from the high- we have operated under zero visibility, income families (average home $40,000 under 40-below conditions, and under and average income $25,000). Over half heavy blizzards. We got a bus stuck in a of our passengers come from families snowbank one night and had to take every- that own 2 or 3 automobiles. The other body's telephone number and radio them half comes from 1-car families. I guess to the dispatcher who telephoned their the average car ownership in the area is homes to say, "Your daughter (or your about 1.6 to 1.7 per household. We have wife or your husband) is stranded on such- taken pictures at 10:00 a.m. on a weekday and-such a street, andwe'il get her home at a 3-car home where the garage doors when we can." This is a very personal- were opened and 2 cars were in the gar- ized service, and you can do this sort of age and 1 was on the street. The cars thing. stay home unless there is some special need to go across town. I might add that Question: Is the driver self-routed? parking in Regina is fairly cheap (about Answer: Yes, the driver is completely $ 15 per month), and it is free at the self-routed and, in some cases, after 2 government buildings. weeks, he does not even need a map. He keeps it all in his head. Question: When you take away the fixed-route buses, which are reasonably Question: Your services are directed cheap, and put on the market expensive to shopping centers. Is there any special telebus, what about the low-income telephone for people to use? people and their attitudes toward this? Answer: No, we do not have any, but Answer: Maybe I should have started we have considered installing special out by saying that in Saskatchewan we live telephones. In the heavy-use areas, there in a "welfare state". Furthermore, it is a are 4 nursing homes: One has 350 em- myth, you know, that transit is for the ployees, another has 200, and a couple are poor. Certainly, there are poor in our small. We have considered putting direct community, and there are people who need lines in those. Our system was pretty assistance. The people on welfare, the crude last winter. We just used normal people who really cannot afford the fare, telephones and had no fancy equipment. and the handicapped people are given a Atkinson 22

pass that is paid for by the state. They pay nothing on the system, but we get the full value for the pass from the state. The in-between areas where there is high demand from low-income groups will have both fixed-route service and telebus because the demand is high enough to provide both. People in those areas will have a choice of fixed-route service or deluxe service. Anyone in those areas unable to walk to a bus stop will get a telebus pass paid for by welfare funds. We think this is the way to go.

Question: What is the bus-driverwage rate? Answer: Driver wage rate is among the highest in the country—about $4.00 per hour. B-LINE DIAL-A-BUS SYSTEM IN BATAVIA

Robert P. Aex Executive Director, Roche ste r- Genesee Regional Transit Authority Executive Vice President, Batavia Bus Service, Inc.

Batavia, New York, is a small com- SERVICE munity located half-way between Roches- ter and Buffalo. It has a population of Batavia is not much different from about 18,000, a wide main street, and most communities. Automobiles fill city considerable automobile traffic. streets, and traffic congestion is bad at On June 3, 1971, newspapers in Roch- times. In an urban renewal project, ester, where the Rochester-Genesee cleared land has had to be used on a ratio Regional Transportation Authority is of 3 to 1 for automobile parking space and located, and in Batavia, where the bus for new buildings. Our goal is to break system operates, reported that the au- the almost complete dependence on the thority was taking over the bus system. automobile. The Batavia papers also carried stories B-Line service is designed to fully for 2 or 3 days about a new concept, dial- meet the transportation needs of people a-bus, that was to replace the old fixed- in the community. It provides transpor- route system. tation from iny point within the city to The buses in the old system were any other or a 20-minute schedule. about 13 years old and had traveled about Door-to-door transportation to and from 650,000 miles each. The seats had been day-care ceHters, schools, colleges, and reupholstered many times. During 1969- work elimin4es the need to drive an auto- 1971 the vehicles were frequently out of mobile or bédriven in one. B-Line ser- service because of breakdowns. Rider- vice also in.ides the delivery of small ship and revenues were decreasing, and packages wi't1in the city. expenses were increasing. The annual Special seJvices are provided for operating deficit was about $7,000 even senior citiçns who live in a high-rise though the vehicles were fully depreciated. apartmen1 project on Main Street. Dial- The B-Line, the name given to the a-bus picks up passengers at the entrance new dial-a-bus service, was further an- of the pro")',~cl and takes them downtown or flounced in a descriptive brochure mailed wherever tfie want to go. Two days a to every address in the city of Batavia. week there isree bus service on a Shop- The brochure described the manner in pers Special paid for by a local retailer. which the B-Line would operate, the A local bank also pays for free bus trans- fares, the operating hours, and so forth. portation to banks. The bus drivers help In addition, newspaper advertising and the elderly to get on and off the buses and posters further described the B-Line and provide any other assistance needed. the individual services. Dial-a-buses serve public housing Dial-a-bus service was installed on projects and make it easy for women to October 11, 1971, when many community go shopping accompanied by their children. leaders were invited to meet at one of the The buses also pick up and deliver pas- local shopping centers, view the buses, sengers at the front doors of the several and ride in them in a parade. The old hospitals in Batavia. buses were allowed to operate for 14 Service has been expanded several days after the start of the new system. times since it began on October 11, 1971. Then, on October 22, it was announced When service was extended to LeRoy and that the old loop bus was "going, going, Stafford, 2 nearby communities, the size gone" to "bus heaven" after 13 years and of the bus fleet was increased. We started 650,000 miles. with three 23-passenger Flxettes and Aex 24

have added a fourth 23-passenger Flxette ing the use of this equipment. It is hoped and a 10-passenger Ford Courier. that digital communications will reduce We recently announced a Bankers Spe- voice messages by at least 50 percent and cial, chartered by a local bank on sched- will permit reduction in dispatching per- uled dates, on which passengers may sonnel required. ride free to do banking in the downtown Paperwork is kept at a minimum. area. Free bus service in cooperation Drivers carry a vehicle log on the bus with the local merchants was also offered while on dial-a-bus service. An inside for shopping trips downtown between the office form is used as a reservation list hours of 10 a.m. and 5 p.m. on Thursday, for home-to-work or home-to-school ser- Friday, and Saturday. vice. A route sheet is used by the driver when he is on subscription service. A small form is used by the telephone op- COMMUNICATIONS AND erator on which to record requests for RECORD KEEPING service. A few operating reports are important. Bus drivers take all their commands One shows the number of hours each ve- from dispatchers by radio. The dis- hicle operates each day and the hours on patcher uses a 6- by 4-ft control board, each type of service. The totals are car- radio, telephone, and magnets. Our ried out for the day and for the week. A productivity goals are 10 passengers per weekly ridership report tells management vehicle per hour for general service and what is happening in the various catego- 17 for subscription service. ries of service and establishes the level Communication between the base sta- of vehicle productivity for each week. tion and the vehicles has been by radio. A 20-ft antenna on top of a 100-ft tower enables signals to be sent and received RIDERSHIP throughout the service area. Vehicles are equipped with 2-way mobile units: 1 Ridership increased from 944 in the frequency to receive messages and 1 fre- first week to more than 1,500 at the end quency to send messages back to the base of the fifth week and to more than 2,000 station. at the end of the sixteenth week. The Recently, it was decided to experiment fixed-route system carried about 75,000 with digital communications. An appli- riders per year, and the B-Line is ex- cation has been prepared to be submitted pected to carry about 100,000 to 110,000 to the Federal Communications Commis- during its first year. sion for permission to install and use Productivity for dial-a-bus service radio telemetry equipment. Initially, a has ranged from 7.0 to 12.0 and has av- 30-day experiment will be conducted eraged 9.0. Productivity for subscrip- through the use of telemetric equipment tion services has ranged from 12.0 to at the base station and in 2 vehicles. The 17.0 and has averaged 15.0. results will be evaluated, and a decision will be made to order equipment for the base station and all vehicles on a lease- COSTS AND REVENUES purchase basis for a 6-month trial. If the trial is undertaken, a second evalua- Vehicle operating costs, exclusive of tion and a decision will be made regard- driver wages, have been computed to be Aex 25

15 cents per mile to cover maintenance mailed to government and industry lead- and operation including gas, oil, tires, ers throughout the area. repairs, insurance, and depreciation and We have had national publicity in the excluding administration and debt service. syndicated column of James J. Kilpatrick, Driver wages are computed at $3 per in Nation's City magazine, and in stories hour plus 20 percent for fringe benefits. written about other dial-a-bus operations Total system costs per vehicle-hour have throughout the country. been computed to be $13.54 per hour to cover all costs including administration and debt service. Total system costs per INFORMAL DISCUSSION vehicle-mile have been computed to be 62.34 cents per mile to cover all costs, Question: Can you say anything more including administration and debt service. about your proposed use of digital com- After 7 months of operation, income munication? from all sources for the month of April Answer: We are now controlling these 1972 covered all of the direct operating vehicles by the use of radio, telephone, expenses and part, but not all, of depre- and a control board. We have under con- ciation. Additional income has been de- sideration a proposal for digital com- veloped from advertising, package de- munication— radio and teleprinting. I livery, and dial-a-bus charters. hope we do try digital communications, Fares are as follows: but I cannot indicate at this time what action the board will take. To work and return, 10 trips per week $4.00 Question: Do you have union drivers? To school and return, 10 And, if so, how do you get around having trips per week $3.50 to take the packages on and off buses for To college $4.00 your customers? 11 tickets per trip $0.36 Answer: We do not have union drivers Dial-a-bus, per trip $0.60 yet. How do we get the drivers to do this Day-care service, per trip $0.25 and that? Our approach to both union and Charters, per hour $8.00 nonunion people is that, with bus ridership decreasing all over the country (we are losing 6 to 8 percent of our riders a PROMOTION AND PUBLICITY year in Rochester) if we do not do something pretty soon, there are not going to We ran coupon ads in the paper that be any jobs for anybody. We must not read, "The B-Line wants to know .... " only expand but also provide service of a Posters were also used to give details of new kind—personalized service to people. the different services such as those to Drivers, if there will be any drivers' work, to school, and to college. jobs, will have to fill the needs just like Various promotional gimmicks were the company will have to. developed to increase the awareness of the new service. A very successful pro- Question: How much of a deficit do motion was in the form of "I Ride the B- you have, and how do you make it up? Line" buttons. We also offer free tickets Answer: At the present time, the in- to people who have not ridden the B-Line. come from all sources—fare box, adver- We also prepared a brochure that was tising, package delivery, and charters— Aex 26

is meeting all of our direct operating ex- ship 30 percent over what it was under expenses and part, but not all, of depre- the old fixed-route system. ciation. We believe that we will not be very far into the next fiscal year before Question: How many hours per day are we will be able to close the gap. you in service and how many days per week is there service? What is the aver- Question: How do you make up the age time for pickup and delivery? deficit? Answer: We have a parameter of 20 Answer: At the present time, we are minutes that we try to use-20 minutes covering all of our direct operating ex- from the time a person calls, we try to penses, and therefore we have a positive pick him up, and 20 minutes from the cash flow. The deficit that now exists is time he gets on the bus, we try to deliver confined to depreciation. There is no him to his destination. We are averaging deficit to be made up; it simply means about 10 minutes. We operate from 6 a.m. that, if we cannot close the gap by the til 6 p.m., Monday through Friday. time these vehicles have to be replaced, we will have to go somewhere else to get Question: Will you explain again how the money for new vehicles. We are you pay for your capital and interest? hopeful that we will be able to do so. in Answer: It is paid for out of income other words, there is a positive cash from all sources. If there is anything we flow; $200,000 was borrowed to buy the are sure of, it is that the fare box alone old bus company and to add new equip- (at these fares) will not cover the load. ment. That money is being paid back with interest, and we expect the income Question: What are you doing about to do that also. At the present time, I working for lower fares in the inner city? think we are well along toward our goal, Answer: When we first examined the but I would be misleading you if I told you Batavia situation, we found that one of the that we had completely made it. greatest sources of taxicab revenues was the welf are department. We are convinced Question: If this service has rates that the primary consideration is not fares higher than those of the fixed-route op- but service. Is it service that somebody eration it replaced, will you have any needs? And is it service that is depend- problem with residents over the higher able and reliable? For the kind of ser- rates? vice we are giving in the city of Batavia, Answer: The previous fixed-route it is the judgment of the citizens—not system suffered from a very rapid and ours—that our service is reasonably steadily decreasing ridership. The own- priced. ers had no capital to replace the equipment, nor did they have the heart to do it Question: What is the wage rate for because it was a losing venture. The fare the drivers? on the old fixed-route system was 25 Answer: Our drivers are averaging cents. If you compare the fixed-route $3.00 an hour plus 20 percent fringe ben- and dial-a-bus fares, it is 25 cents ver- efits. However, we expect that that will sus 60 cents; for work service, it is 25 be changed shortly. If not by us, then by cents versus 40 cents; and for school somebody else service, it is 25 cents versus 35 cents. In 6 months, we have increased the rider- MODEL CITIES DIAL-A-RIDE SYSTEM IN COLUMBUS, OHIO

William C. Habig Executive Director, Mid-Ohio Regional Planning Commission

The model cities transit project, outside the Model Cities area but does which has been in operation since Octo- link up with CTC radial routes, 6 of which ber 11, 1971, is a part of the Columbus run east-west through the neighborhood. Model Cities Program, which is in its We designed the fixed-route service to go second action year ending June 30, 1972. to many key points within the area. Three We serve an area on the near-east side senior citizen high-rise buildings in the of Columbus that is 21, square miles in neighborhood provide our most frequent size and has about 37,000 people accord- riders. There are 3 major hospitals, 97 ing to the 1970 Census. Of that total, 85 churches (which run us ragged on Sun- percent are black. Automobile ownership day), and 5 major supermarkets. There is very low. There is a high need for is a well-established route between 1 se- mobility, and so it is a natural place to put nior citizen building and this supermarket. in a dial-a-ride system to give all resi- A library and a YMCA serve as major dents of the area similar mobility. recreational facilities in the Model Cities Before we started dial-a-ride service, area. we ran fixed-route service for 7 months At the outset of the fixed-route sys- from March 1 through September 30, tem, we found that the 45-passenger bus 1971. We had originally proposed dial-a- was not maneuverable on the narrow ride as the transit project for Model streets and alleys over which we had to Cities, but we were not able to implement operate. Therefore, we started using a it because of institutional problems, which 19-passenger Fixette vehicle not only as I will not discuss. We started the fixed- a solution to that problem but also as a route service with 45-passenger GMC vehicle that met the requirements of dial- coaches. All of the buses were furnished a-ride service. by the Columbus Transit Company (CTC), The dial-a-ride service consists of a the regional private operator. In this basic loop with 21 time checkpoints that case, we paid charter rates ($9.50 an our dispatchers and call-takers use to hour) for the bus, driver, and supervision. schedule the service. This loop is run All we had to do was develop the routes, both clockwise and counterclockwise. manage the project, keep track of the rid- During peak periods, we operate 4 buses, ership, and adjust the service. We had 2 in each direction. The 21 checkpoints originally hoped to begin dial-a-ride on are marked on a map, and each tour July 1, 1971, but because of contract de- around the line must pass those points but lays we lost another 3 months, and it was can deviate from the route anywhere be- October 11 when we finally started the tween a pair of them. The checkpoints new service. are marked in the field with the logo and The fixed-route service was provided the phone number indicating that this is a by 6 buses operating on 4 routes for 102 dial-a-ride checkpoint and a person may bus-hours per weekday. These buses wait for the bus at that location. We en- were painted a special color, and they courage people to call, however, so that carried about 600 persons during peak we can give them an estimated pickup hours of an average weekday. There time and keep better track of operations. were some 50 bus stops marked with the CTC bought 3 new air-conditioned logo "Ride On" and bus route numbers so Fixette buses at our request, and we have that people could easily travel throughout a lease-purchase contract with them. the neighborhood. This project is an in- These buses cost about $16,000 apiece, traneighborhood service. It does not go and at the end of the project the city of Habig 28

Columbus can purchase them. We also on the driver, but we find that this is the purchased a fourth bus (used) to provide best way we can do it under our circum- additional service. The buses are radio- stances.) The driver records checkpoints equipped, and we now use a handset (mi-. at which he received a call, the time, the tially, we used a loudspeaker on the bus, pickup or drop-off location, and the type but that proved too difficult because com- of fare paid. munications were heard by passengers). We have had problems with people not CTC stores the buses at its main garage. being available when the bus arrived. Any resident in the area can dial the Many people will not be waiting outside number of the dispatch center and re- the building. One woman came out of the quest service. We operate from 6 a. m. house, put her baby on the bus, and went to 10 p. m. weekdays, from 8 a. m. to back into the house for awhile. She left 8 p. m. on Saturday, and from 8 a. m. to the baby to reserve the bus so that it 7:30 p. m. on Sunday, a total of 51 bus- would not go away. Many similar situa- hours per weekday and fewer on week- tions have occurred that really hamper ends. At the dispatch center, located in the operation. When service was origi- the center of the neighborhood, there are nally designed, we estimated only about 4 phone lines on a rotary. The level of '/ to 1/2_ minute average wait per pickup; work during the peak period requires 3 actually, the wait is 1 minute and more call-takers and 1 dispatcher; as many as in many instances. The driver is in- 50 calls per hour have been received structed to go up to the pickup location, during this period. We are on a very blow his horn once if there is no one out- limited budget and do not have type- side, wait no more than a minute, and writers or any other kind of equipment. then proceed onward. The clockwise or counterclockwise The fares were set at 20 cents for direction of the bus is indicated on a call adults and 10 cents for children at the card. The call-takers record name of request of the Model Cities Agency. We the person, number in the party, pickup think the fare is far too low and should and drop-off addresses, phone number, be 50 cents. However, because Model date, and time. Cities funds are involved, we follow the The dispatch board has separate maps agency's request to maintain fares at this for the 2 directions, the maps are divided level. Because of the low fare, the ser- into 4 segments. The cards are posted vice is heavily subsidized. The adult on a drop-off or a pickup peg in the 4 fare will be 25 cents in the next year of segments on each map. The call-takers operation. give the cards to the dispatcher, who The loads on the buses vary—our puts them on the hooks and then, as the average load is about 9 passengers for a buses operate, traces their paths by the given tour on a 19-passenger bus. There 21 checkpoints on the map with magnetic were constant complaints that the old markers. The dispatcher calls a batch of 45-passenger buses were always empty. cards to the driver as a bus approaches The productivity of the buses is about a particular sector on the map. We rely 7.6 paid fares per bus hour. Although on both the dispatcher and the driver to there are certain key focal points such as schedule a tour. The driver is required hospitals and supermarkets, there is a to record a series of drop offs or pick- many-to-many system in whichthebuses ups on his log sheet. (We debated this go from any origin to any destination because of the additional burden it puts within the 2.5-square mile neighborhood. Habig 29

The dial-a-ride system also ties into drivers rebid all jobs, and we have a the main system, which is operated by a real training problem on our hands when private company and is losing riders at that occurs. the rate of about 17 percent a year. We The ridership is now averaging 350 lost a transit levy in the primary elec- per day on a 7-day basis and about 400 tion, so that the outlook for public transit per weekday. It grew steadily until Feb- in Columbus is rather dim. Had the levy ruary, leveled off, and then began to de- passed, radios could have been installed cline slightly during the spring period in the large buses so that they could be when the weather improved. We feel that linked up better for transferring. As the kind of service now provided with 4 requested by the Model Cities Agency, buses and the budget constraints make it anyone using a dial-a-ride bus can pur- impossible to build any more ridership chase a 45-cent ticket and use that to because our waiting times are too long to board the large bus without paying any do so. additional fare. So, passengers never We are now going into the third year pay more than the city fare, and, in ef- of the Model Cities Program in Colum- fect, we carry them free. Only about 5 bus, and we are requesting $300,000 to percent of our riders transfer, however, carry on the dial-a-ride project and to because so many routes that they can improve the level of service. We have walk to run through the neighborhood. had many breakdowns that have forced us One of the conclusions that we have to operate only 2 buses during some of drawn from this project is that the fare our peak periods, and that has caused is much too low, especially for a system problems in rider confidence. that is to continue beyond the Model We will continue to have the same set- Cities Program. It is our feeling that, up with drivers. We are paying $8 per if we are funded for another year, that driver-hour to the bus company, in- year may be the last year of this project. cluding wages, fringes, overhead, and We have, therefore, encouraged the city profit. In addition, we are paying the and the residents to raise the fare sub - company lease costs, storage, and main- stantially and to begin to tailor the ser- tenance. Our operating cost that we are vice so that it can be continued as a paying out to the bus company is, there- neighborhood service. We have not been fore, more than $12 per bus-hour; a able to convince them to do so, and the normal charter rate that we paid last low-fare service to residents will con- year was $9.50. One of the reasons for tinue during the next 12 months. the high level of subsidy is the high costs All bus drivers are employed by CTC that we incur by dealing with a private and are union members. We found that transit company. they learned the system very quickly and During the next year we plan to do a have assisted us by keeping track of des- very expensive evaluation of data. We tinations far and above what we expected. have the call cards (nearly 400 a day), Drivers can operate several tours that which we are going to convert so we can have regular riders without any guidance more readily process them in the com- at all from the dispatch center. The dif- puter. We are getting the first reliable ficulty, though, is that "extra-board" transit origin and destination data for drivers have to be trained each time they innercity residents in Columbus that come on the system. Every 4 months have ever existed. We think that the data Habig 30

are going to be quite valuable not only for travel on the dial-a-bus project but for overall travel in the neighborhood.

INFORMAL DISCUSSION

Question: What has been the response to the dial-a-ride? Answer: The response from the neighborhood in general has been good. In fact, a private firm queried residents, and our PrOject rated No. 1 of 33 Model Cities projects that included housing, health care, and everything else. The reason for the good response is that the service is very visible to the residents. We have had some problems with breakdowns, and some waiting times are longer than we like; but, in general, I think the residents view this service very positively.

Question: How large is the service area? Answer: The population is 37,000.

Question: What is your level of service? Answer: The scheduled headways were running 4 every 25 minutes. Our average waiting time is about 25 minutes, and our average driving time is about 25 minutes. This is based on preliminary data and would have to be verified.

Question: Might a taxicab company operate the service? Answer: Well, it might. We are doing it the way we are because of the taxi franchise and the bus company franchise. At one time, we tried to use taxicabs because they have more flexibility, but we got into real tough political and legal problems and wound up with this kind of an operation. MODEL CITIES JITNEY TRANSPORTATION IN BUFFALO

Michael White Monitor, Buffalo Model Cities Jitney Transportation

The jitney program was first proposed though at the time we did not have a no- by 3 senior citizens and 1 staff member ticeable number of people riding, because from the Model Cities Agency. These 4 the area was just too large to cover with people became a committee to bring about 3 buses. the Model Cities Jitney Program. On March 3 we put 3 new buses into The first proposed jitney program was service, and our total ridership went up supposed to provide transportation for approximately 2,000 people that first everyone in the Model Cities Neighbor- month. Our return times went from 30 hood area and to be a paid system. to 20 minutes to less than 15 minutes on The next proposal was to provide a the average. Our riders increased more free transportation service to any per- than 25 percent, and we were able to sons living in the neighborhood who were serve many more group trips. We took blind, handicapped, elderly, or members 2 of the 3 new buses and used them to of any club, group, or organization. The pick up return trips, carry long-distance franchised company offered no major op- trips, help out when the regular bases position to this proposal, and arrange- fell behind, and help with group trips. ments were made with an established cab The third bus was used for emergencies. company in the area to operate the new We find that practically everyone wants system. to ride between the hours of 7 and 11 a. m. The proposed alms of the jitney pro- Between 11 a.m. and 5p.m., most of the gram are to decrease the isolation of the people who went out in the morning return; elderly, the blind, and the handicapped and from 5 to 10 p. m. ridership decreases. and to help them get to doctors, hospi- Although the period from 10 to 12 p. m. is tals, shopping, recreational activities, slow, we keep our buses operating mainly and back home again safely. The pri- because of the contract and also because mary goals are to improve mobility of we have a few people who are returning the elderly in general by offering a free from work and churches. door-to-door personalized service that We serve any Model Cities Neighbor- can be obtained simply by a telephone hood area club, group, or organization call. The senior citizens call a day in that makes reservations at least 3 days in advance to make arrangements for pick- advance; we do not serve anyone who is ups the next day. At the same time they aged 59 or under, not handicapped, or not may also make the return trip arrange- blind unless that person is in one of the ment, or that can be made the next day reserving groups (The exception is that when they are ready to return. senior citizens may be accompanied by The jitney program started on Decem- someone to give them help or a small ber 15, 1970, in the Model Cities Neigh- child with whom they are babysitting.) borhood area. The area has a population Group trips may be as far as 50 miles of 62,000 of which 7,000 (according to the away. 1970 Census) are aged 50 and older. Only those aged 59 and older are eligible to Buses From To use the jitney service. We began our operation with 4 buses. 2 7 a.m. 3 p.m. Three were to be used on a day-to-day 2 8 a.m. 4 p.m. basis, and one was to be left in the shop 1 9 a.m. 5 p.m. for emergencies. However, we found 1 10 a.m. 6 p.m. that we used all 4 buses every day, even 1 12 noon 8 p.m. White 32

From 7 a.m. to 3 p.m., all 7 buses are in service, and we have no emergency vehicle. At 3 p.m., 2 drivers get off, and 1 driver itarts the second shift. At 4 p.m., the last shift starts with 1 driver. Be- tween 4 and 6 p.m. we have only 4 drivers and after 8 p.m. we have 2 drivers. DIAL-A-RIDE APPLICATION IN GREAT BRITAIN

Philip R. Oxley Transportation Planning Associate, Ford of Europe, Inc.

Following the development and dem- MAIDSTONE onstration of demand-responsive public transport systems in North America, a Operator: Denis Hire Car number of authorities in Great Britain Company expressed interest in the system. British Service: Many-to-two, off peak, Railways in particular would like to es- Monday to Saturday, 2 tablish whether such a system can be ef- vehicles plus 1 standby fective as a means of inducing additional Fares: 20 pence flat fare and rail use. In the London area this means 10 pence for children off-peak services only because during 3 to 14 peak times the railways in this area are Status: Application granted by already at or close to absolute capacity. South Eastern Area Away from this areathe service is viewed Traffic Commissioners by British Railways as a feeder to its May 3, 1972; service main lines, which are under increasing scheduled to commence competition from road-based trans- August 30, 1972 port as more motorways are opened. A number of subsidiaries of the Na- tional Bus Company have also become Figure 1. Dial-a-ride study locations as of May 1972. interested, particularly for small towns. At present these are usually served by longer distance routes that pass through them, but the patronage available is not sufficient to justify the use of conven- S INVESTIGATIONPROPOSAL tional vehicles (40+ seats) on in-town services. Finally, a number of taxi and hire-car firms are also considering the use of dial-a-ride, particularly since the Maid- stone application by a taxi company was approved. CASTLEFORO None of the services proposed has yet become operational, and the success or otherwise of demand-responsive trans- LOL?GH portation still remains problematical in SDOROUt$I the United Kingdom. However, judging by the response to the publicity that occurred RUGflY when the first 2 proposals (Maidstone and I4ARLCW Abingdon) were approved, there seems 8?PDOW * little doubt that demand-responsive BRISTOL LONDON S DATI4 * transportation will spread if the early MAPOSTONE services are successful. The following sections summarize the 6 dial- a-bus studies that, to date, have progressed the furthest. Locations are shown in Figure 1. Oxley 34

The service operates between the res- Fares: 8 pence flat fare idential areas of Loose and Coxheath Status: Application granted by (jopulation 6,000) on the south side of South Eastern Area Traf- Maidstone and the center of the town. fic Commissioners April The 2 drop-off points in the town center 26, 1972; service starts are at the western end of the main shop- June 4, 1972 ping street and at Maidstone East Rail- The service operates from the resi- way Station. The two pickup points are dential areas of Abingdon to the town at the railway station and a midway point center on Tuesdays and Sundays between on High Street. The hours of operation 9 a.m. and 9 p.m. On the otherdays of the are 9 a.m. to 4 p.m. and 5:20 to 11 p.m. week the bus operates on a fixed-route, Monday to Saturday; there is no Sunday fixed-schedule basis with a flat fare of 5 service. pence between the hours of 9:50 a.m. and The timing of the dial-a-ride bus ar- 3:30 p.m. For the purposes of the dial- rivals and departures at the railway sta- a-ride service, this small country town tion is designed to meet all off-peak of approximately 18,000 population has London trains because one of the looked- been divided into 3 zones, each of which is for results of the service is the increased serviced in sequence. use of these off-peak trains. They run The control office is situated in the on headways of approximately 30 minutes. main shopping precinct in the center of Passenger requests for journeys from the town, and 1 of the 2 pickup and drop- the service area to the town center may off points is immediately behind it. The be made by telephone to the control of- other dial-a-ride stop is on the south side fice, which is situated at Maidstone East of the shopping area by the bus stops used Railway Station. Passengers must call by longer distance bus services. at least 60 minutes prior to the departure Requests for service into the town times of the trains they wish to catch, center are made by telephone to the con- although later requests will be accepted trol office (a phone recorder is usedwhen if service permits. For the reverse the office is not manned). in the reverse journey passengers may either make their direction, passengers register their re- requests direct to the control office (the quests with the bus driver who does his bus standing bay is immediately outside own scheduling on the outbound tours. the office) or telephone if they wish to be picked up on High Street. A free direct- BRISTOL line phone is available by the High Street pickup point. The vehicles are 15- Operator: Bristol Omnibus Com- passenger buses built to normal PSV pany, Ltd., under con- regulations and fitted with 2-way radios. tract to British Rail- ways ABINGDON Service: Many-to-one, Monday to Friday, 2 to 3 ve- Operator: City of Oxford Motor hicles Services, Ltd. (Na- Fares: 40 pence provisionally tional Bus Company) Status: Application to the Traf- Service: Many-to-two, Tues- fic Commissioners' day to Sunday, 1 ve- Court to be heard in hicle plus 1 standby August 1972 Oxley 35

The service centers on Bristol Park- eration for a government way, a new main line railway station that grant was opened to traffic on May 1, 1972. Most trains run from South Wales to The Harlow- Kingsmoor service is London, but there are also a small num- proposed to operate between the Kings- ber that run to North West England. moor residential neighborhood (popula- The service is designed to meet all tion 5,000), a local shopping center (Sta- train arrivals and departures to and from ple Tye), and the Harlow Town Center London and to serve the northern sector (the High). Dial-a-ride vehicles operate of Bristol. The station itself lies about on a fixed route through the neighborhood 1 mile beyond the built-up area and at with deviations from that route in represent has only 1 scheduled (30 minutes) sponse to requests for doorstep service. bus service. There are no fixed routes In some places of the area where it is not for the dial-a-ride buses, but 2 or 3 spe- possible to provide actual doorstep ser- cific pickup points may be designated at vice because of the physical layout of the which the buses always call. Otherwise, area, collection points for small groups of use of the service to get to the station is dwellings have been identified. It is pro- by telephone request made to the station posed to discontinue the existing scheduled booking office. The ticket clerk doubles bus that serves Kingsmoor after the first as a dial-a-ride dispatcher. In the re- few weeks of dial-a-ride operation. Be- verse direction, passengers purchase cause of the need still to provide a ser- tickets at the booking office, board the vice that can be used by people who do not bus, and give their destinations to the have access to a telephone, the dial-a- driver. ride service will operate on fixed head- At present, 3 share taxis are being ways. operated under a temporary license by The old Harlow service serves the High, the Bristol Omnibus Company pending a local shopping center (the Stowe), and the introduction of the buses. A flat fare the Harlow Mill Railway Station (off- of 40 pence is charged for this service, peak times only). Unlike the Kings- but it is possible that this may be re- moor proposal, the service is routed in duced when the buses are brought into use. response to requests, and fixed scheduling occurs only during off-peak times HARLOW NEW TOWN when London train arrivals and departures are served at the Mill station. Operator: London Country Bus The control office for both services is Services, Ltd. situated in the center of Harlow and is in Service: Many-to-few, 7 days contact with the vehicles via mobile tele- per week, 17 hours printers with radio communications re- per day, 6 vehicles served for emergency use. The question Fares: 3 to 12 pence sug- of the use of a computer for backup to the gested range manual dispatch system is currently under Status: Proposal approved by consideration. the Harlow Urban Dis- trict Council and at CHELMSFORD present lodged with the Department of the En- Operator: Eastern National Bus vironment for consid- Company, Ltd. Oxley 411

Service: Many-to-few, 7 days per Traffic Commissioners' week, 17 hours per day, Court July 20, 1972; ser- 4 to 5 vehicles vice starts September 25, Fares: Not yet determined 1972 Status: Proposal under consideration by the operating The service, known as "Chauffeur company Coach," runs among 5 residential areas and the shopping center of Harrogate. The service operates between a resi- Each residential area has 2 inbound and dential area (Galleywood) on the outskirts 2 outbound buses per day; for example, of Chelmsford and a number of stops in the Knox area service is served by in- the center of the town including the main bound buses leaving the service area at bus station, the railway station, and the 9:10 a.m. and 2:10 p.m. and by outbound principal shopping street. It is a fully buses leaving the town center at 11:05 demand-responsive service in response a.m. and 4:05 p.m. to telephoned requests for service. Minor The service is designed specifically changes are proposed to the regular buses for the shopper, and within the town cen- that operate in or close by the service ter the buses follow a fixed route but will area including turning short the one ser- pick up and drop off passengers where- vice that penetrates into the area. ever requested. In the residential ser- The service area has a population of vice areas, the buses follow a fixed route approximately 6,000 and a very wide but will deviate in response to requests range of housing types. Second-phase for doorstep service. There is no pre- plans have been proposed for extending mium fare for the doorstep service. Pas- the service area to cover an adjoining sengers wishing to be picked up attJeir suburb (Great Baddow) and for provid- doors may telephone their requests to a ing a local many-to- many service within control office; in the reverse direction, the 2 areas. Such a service involves they simply hail the bus and tell the driv- the use of as many as 14 vehicles. er where they wish to be dropped off. So that stop time will be minimized in the HARROGATE town center, fares from outbound passengers are taken when they leave the Operator: West Yorkshire Road bus and not when they board it. Car Company, Ltd. Service: Many-to-one, Monday, OTHER PLACES Tuesday, Wednesday, a.m., Thursday, Fri- Further dial-a-ride applications that day, Saturday; 2 to 3 are still in the study stage include East vehicles Kilbride New Town, Basildon New Town, Fares: 10 pence for adults (15 Eastbourne, where the proposal was re- cents return) and 5 cently approved by the Borough Council, pence for children 3 to and London. 14; short-ride single fares: 5 pence for INFORMAL DISCUSSION adults and 3 pence for children Question: Telephones are not so uni- Status: Proposal approved by versally distributed in the United Kingdom Oxley 37

as in the United States. Does the lack of telephones give you any problems? Answer: It is a problem. The average telephone ownership in the United King- dom is about 40 percent of the households. Some of the areas we are servicing, of course, are significantly higher than that. In some of the better residential areas it is nearly 100 percent. Because there are no telephones, we are cutting off, in effect, a part of the population, or at least making it difficult for them. One possible way around this is to make provision for people to book well in advance; the other way is, of course, to do what we are doing with the Kingsmoor, and that is to introduce some element of fixed routing and fixed scheduling. Telephone ownership is increasing at the rate of, I think, 6 or 8 percent per year so that one must realis- tically look at dial-a-ride as a service that, if it is going to succeed, will only succeed in perhaps 5 or 10 years. For the moment, we have to accept that we are, in fact, cutting off some of the people.

Question: Will the systems cover costs out of the fare box? Answer: I think you will have to ask me that in about 3 months. I really do not know yet. The general principle that is being followed in Abingdon is that service ought to pay for its operating costs out of the fare box, but I think the bus company, in this case, would expect that it would not have to cover its capital costs as well. Whether it will cover its operating costs I do not know. It has operated only 1 day so far, to my knowledge, and it did not cover its operating costs on its first day. DIAL-A-RIDE DEMONSTRATION IN HADDONFIELD

D. W. Gwynn Director of Research, New Jersey Department of Transportation Anthony U. Simpson President, DAVE Systems, Inc.

The Haddonfield Dial-A-Ride Demon- Hill, Barrington, and Lawnside. The stration Project is sponsored by the area is approximately 51/2 square miles New Jersey Department of Transporta- and has more than 25,000 people. This tion under a grant from the Urban Mass relatively small initial area was chosen Transportation Administration. Total to minimize any start-up problems and funding for the project is approximately yet provide capability for controlled ex- $1.7 million; about 80 percent is pro- pansion at a later date. vided by the federal government, and 20 Work started in Haddonfield in August percent is provided by the state. Reve- 1971, and service commenced on Febru- nues from operations are returned to the ary 19, 1972, after a 6-month prepara- project and are used to offset costs. tion phase. The drivers went out on The transportation department has strike on March 1; so, there were only primary contracts with 3 organizations: 10 days of initial operation. The strike Highway Products, Inc., to supply 11 occurred during negotiation of a new con- standard-vehicles and 1 specially de- tract between Transport of New Jersey signed vehicle for invalids; Transport of and the Amalgamated Transit Union (it New Jersey to supply drivers and main- had nothing to do with the dial-a-ride tenance for all dial-a-ride vehicles; and project). The strike lasted 76 days, and LEX Computer Systems and DAVE Sys- service was resumed on May 15. To tems to design, manage, and operate the date, we have completed approximately 5 project. In addition, MITRE Corporation weeks of operation. is under contract to UMTA to evaluate The principal objective of the Haddon- the program. This evaluation includes field Demonstration Project is to deter- analysis of survey data, analysis of oper- mine the public's attitude toward and its ating and revenue data, and preparation acceptability of dial-a-ride service; a of computer software. second objective is to determine the eco- The Haddonfield project is similar to nomic feasibility of manually controlled earlier dial-a-ride projects, and, al- dial-a-ride systems. though bigger and more complex, it We provide 4 principal types of ser- draws extensively on their experiences. vice: many-to-many, many-to-one, one- In particular, several of the features of to-many, and shuttle. The many-to-many the Haddonfield project reflect the work service operates 24 hours a day, 7 days a that has been done by M.I.T., Ford Mo- week, primarily because the Lindenwold tor Company, and GO Transit. Line also operates 24 hours, 7 days a Haddonfield, New Jersey, a suburban week. The many-to-one service is de- community, is located about 7 miles east signed to gather commuters in the morn- of Philadelphia and has a population of ing and deliver them to Line-haul stations approximately 15,000. The PATCO line or their work places. The one-to-many (also locally called the Lindenwold Hi- service is used primarily to deliver pas- Speed Line) runs through Haddonfield; a sengers from the Haddonfield station to station is located in the center of town. their homes in the evening. The shuttle The trip from Haddonfield to central service was instituted to handle high de- Philadelphia takes 15 minutes on the Lin- mands for service between 2 primary lo- denwold Line. cations in the service area, e.g., from The initial demonstration area in- the station to the Music Fair. It is not a cludes the entire municipality of Haddon- scheduled shuttle but implemented when- field, portions of the boroughs of Cherry ever this type of operation is the most Gwynn and Simpson 39

Figure 1. Dial-a-ride bus. intensity reading lights, perimeter seating, carpeted floors, and sizable storage racks. Their short wheelbases and tight turning radii are valuable features for maneuvering on side streets. One vehicle has been specially constructed for the riniI,fl handicapped (Fig. 2). The vehicle has a i manually operated ramp in the rear of the vehicle and can accommodate 3 wheel- chair passengers in addition to regular passengers. The vehicles have been filled to capacity several times during the first 5 weeks of service, and public reaction to the vehicles has been favorable. The dial-a-ride office is staffed by controllers who have 4 principal func- efficient method of handling the demand. tions: receive calls, schedule, dispatch, The buses (Fig. 1) seat 17 people and and perform clerical and accounting du- provide room for 10 standees. The vehi- ties. Each controller is trained in sev- cles are air-conditioned and have high- eral functions so that the system can be

Figure 2. Bus for passengers in wheelchairs. Figure 3. Typewriter for preparing trip tickets.

k ' TIIi Gwynn and Simpson 40

operated by 1 or more persons depending pickup ticket to indicate the time that the on demand; the maximum required to customer expects to be picked up, and date has been 3 controllers. other critical data are indicated, such as When a request for service is re- whether the invalid vehicle must be used ceived, a trip ticket is typed in duplicate or whether an important appointment on a specially constructed electrical must be kept. These magnetic holders typewriter (Fig. 3). Information re- are then placed on an enlarged map of the corded on the trip ticket includes pickup service area (Fig. 4), oriented to show address, delivery address, telephone the vector from pickup to delivery point. number, status and number of persons Colored markers are later placed in the riding, estimated pickup time, and pas- pickup and delivery holders to indicate senger's name. We take the telephone which vehicle will be used. Each bus has number so that we can inform passengers its own color code, and the colored mark- of any significant changes in their sched- ers trace the route for each bus. As the uled pickup times. bus travels along its route, the holders One copy of the trip ticket, referred are taken off the board and are given to to as the pickup ticket, is placed in a the dispatcher (Fig. 5), who relays the black magnetic holder with a pointer. information to the driver. The other copy, referred to as the de- The scheduler makes the decision livery ticket, is placed in a white mag- about which vehicle to assign for each netic holder. A code is marked on the passenger on the basis of a number of algorithm and pattern recognition techniques. In essence, he is trying to max- imize vehicle utilization while meeting Figure 4. Scheduling map. the constraints of each passenger's ex- pectations. Communications are usually initiated by the drivers when they call in by 2-way

Figure 5. Dispatcher's console.

•

1_) , Gwynn and Simpson 41

radio at each stop. The driver is then predetermined time and location. A card given and records the next stop and the is made, and each day the request is au- next-but-one stop. Thus, if the driver tomatically processed. gets to his next stop but cannot get through Service at the Haddonfield station does on the radio, he can go ahead without de- not always require a call to the control lay to his next-but-one stop. At all times, room. During the hours from 4:15 p.m. the control center knows which passen- to 6:45 p.m., a commuter may simply gers are on each vehicle and the approx- look at a zoned map on the station wall imate position of each vehicle. (Fig. 6), determine the zone in which he The system provides an exceptional lives, board the bus that has his zone level of security as a result of instant marked in its front window, and inform driver comnrnnications that can be re- the driver of his destination. The driver layed to police, identified passengers, makes up a tour from the addresses given position of each vehicle known to within a him by the passengers that board his ve- few blocks, locked fare vaults, and door- hicle and in this instance does not have to to-door service. In fact, at night drivers receive instructions from the control use high-intensity flashlights to illumin- center. The dispatcher does, however, ate paths to a customer's door if re- often insert many-to-many trips into the quested. driver's tour, especially on his way back The drivers have between 20 and 35 to the station. years of experience as bus drivers. We Recently, one of our buses participated were concerned that there would he retraining problems with drivers who had this amount of seniority. But that did not happen; they learned very quickly. They Figure 6. Service area and zone map at station. like their work, perform it well, and, in fact, have become our best public relations and sales staff. Numerous letters and phone calls have been received com- plimenting drivers on their politeness and considerate actions. There are basically 3 fares: 60-cent cash fare, 50-cent discount ticket sold in books of either 10 or 40 tickets, and 40- cent senior citizen or group and family ticket sold in books of 10 tickets. Group and family tickets may be used by 3 or more persons traveling together to and from the same pickup and delivery points any hour on Saturday, Sunday, and holi- days and between 9:30 a.m. and 4:30 p.m. and 7:00 p.m. and 6:30 a.m. on weekdays. Other tickets may be used any day and any hour. Subscription service is also provided. A customer may call and make arrange- nients to be picked up each day at a Gwynn and Simpson 42

in an interesting contest between public we had a great number of inquiries. Al- transportation and the private automobile. though we were able to adjust to the high Two reporters from a Philadelphia news- demand, the wait times in some cases paper held a contest to see which one were as high as 1 hour. (The local taxi would arrive at the office in Philadelphia company at the same time had wait times first, the one using public transportation significantly longer.) We also had a lot of or the one using a private automobile. false calls during the first few days of We had no knowledge that this race was operation. The youngsters found out how taking place. Both participants started the system worked much faster than the out at the same time (midafternoon, non- adults did. Teenagers would place a re- rush hour) from an apartment building in quest that sounded like a valid order. the dial-a-ride service area. As one re- Later, control personnel started recog- porter left in his car, the other placed a nizing voices and the pattern of the calls call from a public phone for a dial-a-ride and called back to verify the trip request. bus pickup, waited 11 minutes for the bus, Now the problem is under control, but rode to the Haddonfield station, waited 6 there are still a few false calls. Radio minutes for the train, and then rode to interference was also a major problem Philadelphia. Although the reporter during the initial start-up. The frequency using the car won the race by about 3 assigned to the project was shared with a minutes, his costs were a dollar more trucking company in New York, 80 miles in terms of bridge toils and parking. In away. The same weather problems were addition, he had to fight traffic, whereas also shared. Though interference was the reporter on public transportation was easily tolerated on ordinary days while able to organize his notes for the article drivers were being trained, when the en route. snow storm hit the area the trucks from The manual scheduling system that we the trucking company had serious prob- have devised can be replaced by a com- lems and used the radio for long periods. puter if the service area is expanded. This created a critical situation for the The typewriter can be replaced by a tele- dial-a-ride project. During the strike, a typewriter or cathode ray tube terminal, frequency change was requested and ap- and minimal retraining of personnel will proved, and the problem is now entirely be required. The dispatcher console now resolved. displays information similar to the type The familiarity of the drivers with this of information that will be presented by a new type of system also created a prob- cathode ray tube. We expect to have a lem. The bus drivers were used to driv- computer later this year. It will be ing along a standard route. They had to tested, debugged, and run in parallel with be taught not only dial-a-ride procedures the manual system before being placed on but names of all streets in the service line. After that, the computer will do the area. The house numbers in the area are controlling and decision-making, and the not marked on the curbs and, in many in- manual system will serve as a backup stances, are not marked at all. Thus, system. the drivers had to learn many individual During the first 10 days of operation in addresses. They were provided with February, there were a number of signif- maps that had all the streets identified icant problems. The first day of fare and all the house numbers in blocks. collection occurred in the middle of the During the initial start-up, there were first serious snow storm of the year, and some problems with the drivers getting Gwynn and Simpson 43

lost, and this resulted in the dispatcher Answer: We anticipate that there will be having to guide some of the drivers around a computer-operated demonstration that by radio. To assist us in picking up and will continue beyond this initial phase. delivering passengers at night, the customers were requested to leave the lights Question: What do you tell someone on at their homes so that the pickup ad- who asks whether service will continue dress could be identified easily. after the demonstration? Ridership has been increasing stead- Answer: We evade the question where- ily. To date, there is no indication that ever possible. We imply that it is a con- the demand is leveling off. Weekly rider- tinning operation, and we try to imply ship at the end of the fourth week was that as long as possible so as to en- 2,436 passengers. Demand has been pri- courage people to change their modes. marily for many-to-many service. Shop- We do not lie, however; if we are asked pers and senior citizens have used the the direct question, "How long are you service far more than commuters. As a funded for?" we tell them. result, the advertising, which was initially aimed at the community as a whole, Question: Do you use a zone scheme has now been shifted more toward the for routing buses? commuter. It is anticipated that many Answer: No, we try to route the buses commuters will continue to use their completely without regard for zones ex- automobiles until they are faced with the cept, of course, when we provide one-to- need to replace them, and then dial-a- many service in the evenings. Then, of ride, by eliminating a major capital outlay, course, the vehicles are assigned tempo- will look much more attractive. This ef- rarily to a zone. fect, however, will take several years before it develops fully. Question: Do you take deferred and periodic requests? Answer: We take both deferred and INFORMAL DISCUSSION periodic requests. The deferred requests are put on a board and are processed a Question: Do you not offer more ser- half hour before the actual pickup is vice than you have a demand for? needed. The periodic calls are put on a Answer: At this time we are trying to cardex system, and the tours are made ensure that capacity is ahead of demand. up in the morning. The overall funding for this demonstration—the $1.7 million—assumes no revenues. All revenues are put into a separate account and returned to the government separately. Therefore, even if we did not take in a dollar, we are still fully funded. That is one reason why we can have the service level somewhat ahead of what the demand is. And we hope to keep it that way.

Question: What will happen after the initial demonstration? DIAL-A-RIDE DEMONSTRATION IN HADDONFIELD: PLANNING AND INITIAL OPERATION Douglas M. Medville Technical Staff Member, Mitre Corporation

The objectives of this paper are to de- model to forecast demand for DAR); (c) scribe some of the results of the house- to assess reactions to DAR; and (d) to hold survey conducted in the Haddonfield, obtain data that could be used to establish New Jersey, area by the MITRE Corpor- initial DAR operating parameters. ation before the dial-a-ride (DAR) demonstration project started; to indicate how some of these survey results were SURVEY RESULTS used in reaching decisions concerning the initial operating parameters; and to give Trips some of the initial operating results of the demonstration. The distribution of work locations is The survey was conducted in October as follows: 1971 in 4 Camden County communities: Haddonfield, Barrington, Lawnside, and Location Percent part of Cherry Hill township. The com- b munities are located about 7 miles east Home Within service area 9 of Philadelphia and are linked to Phila- Outside service area delphia by a high-speed rail facility, the 13 Lindenwold Line; a station is located in Use Lindenwold Line Do not use Lindenwold Line 72 Haddonfield. The surveyed area has a population of about 33,000 people. Inter- The following modes are used for views were conducted in about 500 of the work trips: 10,000 households. Socioeconomic characteristics of the surveyed communities are given in Table 1. Mode Percent The objectives of the survey were (a) Use Lindenwold Line to measure trip characteristics in the Use automobile to get to initial service area (i.e., where people station 78 went on local trips, how many trips were Walk to station 18 made, and which modes were used); (b) Take bus or taxi to station 4 to determine modal split and characteristics, such as travel time and cost (in- Do not use Lindenwold Line formation concerning modal attributes Use automobile 85 and their ratings was used to develop a Walk or take bus or taxi 15

Table 1. Socioeconomic characteristics.

Item Haddonfield Cherry Hill Barrington Lawnside All Population 13,118 8,646 8,422 2,757 32.943 Black population, percent 1.7 0.5 1.3 90.4 9.0 Area, square miles 3.3 2.5 1.6 1.5 8.9 Households 4,345 2,648 2,484 669 10.146 Mean income/1966 IRS return, dollars 10,365 9,819 6,977 5,410 - Automobiles/household 1.5 1.7 1.4 1.4 1.5 Workers/household 1.1 1.4 1.2 1.3 1.2 Licensed drivers/household 1.9 2.1 1.8 1.9 1.9 Persons/household 3.1 3.6 3.4 4.0 3.2

Medville 45

Of the estimated 47,630 round trips made portance ratings obtained generally agree to work each week, it is estimated that with the results of other such studies. 15,400 can be made completely or par- People in 250 of the households surveyed tially via DAR. The remainder of these were asked to rate the importance of trips are to destinations that cannot be these attributes for work trips, and people served by DAR or by DAR combined with in the other 250 households were asked to other public transportation. rate the importance of these attributes For nonwork trips, 92 percent of the for nonwork trips. Results are given in trips within the survey area and 97 per- Table 3. Both groups rated trip cost as cent of the trips to Cherry Hill Mall, a being least important and reliability as large regional shopping center about 3 being most important. miles north of downtown Haddonfield, were made via automobile. User Reaction to DAR For those people who use the Linden- wold Line, distributions of modal attri- About half-way through the interview, butes for existing work trips were ob- respondents were given a description tained. It was felt that practically all of of DAR and asked whether they could these trips could be diverted to DAR. use the service. As expected, 75 percent Time and cost of work and nonwork trips of the people who currently use the Lin- by automobile and ratings of modal attri- denwold Line for work trips stated that butes are given in Table 2. Comfort, con- they could use DAR for trips to the sta- venience, and reliability were measured tion. Of those people who were not using by the use of the semantic differential. the Lindenwold Line, however, only 25 Respondents were asked to rate the mode percent thought they could use DAR. Most on a scale from 1 (very uncomfortable) of these people stated that DAR, either by to 7 (very comfortable). itself or in combination with another mode could not take them to their work destina- Importance Ratings tions because they were outside of the envisioned service area. (Of those inter- After attribute values were obtained, viewed, 72 percent did not use the Linden- respondents were then asked to rate the wold Line.) importance of these attributes. The im- Based on the current volume of work trips made in the service area and on the reactions to DAR obtained in the survey, it is estimated that about 3,000 round trips a day could be diverted to DAR. This Table 2. Time, cost, and attribute rating for automobile mode.

Round- Rating Trip Table 3. Mean attribute rating for existing Relia- Time Cost Com- Con- modes. Type of Trip (mm) (cents) fort venience bility

Work trips Attribute Work Trips Nonwork Trips To station 7.4 35 6.0 6.3 6.4 Nonwork trips Time 6.1 5.2 Within sur- Cost 4.8 4.7 vey area 9.2 48 6.3 6.2 6.4 Comfort 5.7 5.3 To Cherry Convenience 6.3 6.2 Hill Mall 17.2 65 6.3 6.1 6.3 Reliability 6.6 6.4 Medville Ki

is the maximum size of the work-trip Table 4. Mean attribute rankings for DAR. market, and it is not expected that this trip volume will be achieved. Attribute Work Trips Nonwork Trips About half of the households surveyed Waiting time 1.7 2.1 indicated that they would make additional Fare 2.8 2.5 Door-to-door service 2.8 2.5 trips if DAR were available. The mean Travel time 3.1 3.4 number of additional trips per household Comfort of vehicle 4.7 4.4 was 2.0. Information concerning the Attractiveness 5.8 5.8 amount of latent demand that actually is being satisfied by DAR will be obtained in future on-board surveys. The most frequent reason given (42 the relative importance of various DAR percent) by those who thought they could attributes for work and nonwork trips not use DAR was that trip destinations (Table 4). Fare is ranked quite high for were outside of the proposed service DAR but was ranked as being the least area. Thirty-six percent of those who important for existing modes. Thus, did think that they would use DAR stated there may be some inconsistencies in that they were satisfied with their current people's thinking with respect to their mode (usually the automobile). attitudes toward the cost of various Those respondents who thought that modes. they would use DAR were then asked to indicate a reasonable fare, travel time, initial DAR Attributes and waiting time for various types of trips. The mean for work and nonwork Partially as a result of the household trips is as follows: survey, the initial DAR service area was established in all of Haddonfield, that part Non- of Cherry Hill Township that was sur- Item Work work veyed, and those parts of Barrington and Lawnside that are north of Interstate 295. 1-way fare, cents 25 49 Although it is somewhat smaller than the Travel time, min 12.5 13.3 survey area, this initial service area is Wait time, mm 15.0 about 5.4 square miles in size and contains 8,200 households. In addition, sev- These values were considered in estab- eral local shopping malls lying at or be- lishing initial fares and times. yond the perimeter of the service area About 10 percent of the people living are also served by DAR. Respondents in in the service area are aged 65 and older. the household survey expressed a strong For persons in this category, respon- interest in having DAR service to these dents indicated mean fares to be 53 cents shopping areas. for those under age 65 and 41 cents for Based on survey data, initial DAR those over age 65. The randomization fares and times were set as follows: test for 2 independent samples revealed that these distributions were significantly Item Amount different at the 0.05 level. For this reason, the DAR fare for older citizens was Fare, cents set 10 cents lower (40 cents versus 50 Cash 60 cents) than fares for other people. 10-ticket book 50 Respondents were also asked to rank Persons over 65 40

Medville 47

Item Amount Figure 1. Weekly ridership (May21 to July 28, 1972). Fare, cents Children None Travel time, min 10 to 15 Wait time, min 10 to 20

INITIAL OPERATING RESULTS

Ridership

Figure 1 shows weekly ride rship during the 11-week period since service began on May 15. DAR has served 27,435 passengers during this period. Figure 2 shows average ridership on weekdays, Saturdays, and Sundays. The average number of weekday riders was 495 during this period; the peak was 580 passengers on June 23. The greatest number of riders was generally carried on Fridays, and the lowest number was carried on Sundays. Saturday ridership exceeded the average for weekdays during the first week of operation and again during the week ending June 11, when it reached a daily peak of. 402 passengers. Since that time, Saturday ridership has Figure 2. Average ridership on weekdays. leveled off somewhat, with a mean of 297 Saturdays. and Sundays (May 21 to July 28, passengers per day. Sunday ridership 1972). during the 11-week period has remained relatively constant, increasing from a low of 92 passengers per day during the first week of operation to a mean of 136 passengers per day for the period from June 11 to July 30. The hourly ridership on a typical weekday is shown in Figure 3. (Datashownin Figures 3 through 9 were obtained on 5 weekdays: June 8 and 16 and July 10, 18, and 26.) Trips between 9 a.m. and 4p.m. account for 45 percent of the daily total. About 20 percent of daily trips are made between 4 and 7 p.m., and about 7 percent are made between 7 and 9 a.m. The last percentage may be low because Medville 48

Figure 3. Average number of riders. per hour or day. It depends on the number of requests for service received and the number of vehicles available to service these requests. Vehicle productivity, however, should not be viewed as the only measure of system performance. For a given demand rate, a large num-. ber of vehicles in operation will result in both a relatively low vehicle productivity and a high quality of service because waiting and riding times will be short. As the number of operating vehicles is reduced, productivity will increase, because each vehicle will carry more passengers, and quality of service will decrease, because there will be fewer vehicles to service customer requests. Average hourly weekday vehicle dis- many service area residents live within tribution and average hourly vehicle walking distance of the station or because productivity are shown in Figures 4 and those who travel to the station by auto- 5. During the first 11 weeks, the number mobile have not yet changed their com- of requests for service was low with muting habits. respect to the number of vehicles in op- The average number of riders, stan- eration. Consequently, the average pro- dard deviations, and percentage of weekly ductivity was 4.09 passengers per total from June 19, 1972, through July vehicle-hour. One reason for this is the 30, 1972, (excluding July 4) are given in experimental nature of the project. An- Table 5. other is that the hourly vehicle distribution during a 24-hour period was based Vehicle Productivity on the typical hourly transit demand

Vehicle productivity is defined as the number of requests received per vehicle Figure 4. Average number of vehicles in use.

Table 5. Average ridership and standard deviation by day of week.

Standard Percent of Day Mean Deviation Weekly Total

Monday 446.3 26.26 15.19 Tuesday 491.8 48.23 16.73 Wednesday 525.7 34.78 17.89 Thursday 512.8 36.61 17.45 Friday 528.3 27.65 17.98 Saturday 299.3 21.44 10.18 Sunday 134.7 9.65 4.58 Total 2,938.9 100.00 Medville 49

Figure 5. Average vehicle productivity nod of the day for the 5 representative weekdays are given in Table 6. Vehicle productivity is highest in the 2 off-peak periods-9 a.m. to 4 p.m. and 7 to 11 p.m.—reflecting the nonwork trip use that dial-a-ride is receiving in these periods.

Quality of Service

The quality of service greatly influ- ences the number of riders attracted to dial-a-ride. Although quality of service may connote a combination of factors, it is considered here to consist only of distribution, and this has not been the time—average wait time, pickup time case. deviation, and ride time. Productivity on a typical weekday be- Three types of dial-a-ride trips are tween 5 and 11 a.m. reflects the relatively offered: immediate, deferred, and pe- small use being made of dial-a-ride by riodic. An immediate trip is one that the commuters and the relatively high num- passenger wants to start as soon as pos- ber of vehicles in service. Productivity sible. A deferred trip is one that the pas- between 11 a.m. and 4 p. m. is above av- senger wants to start at some specified erage because the vehicle supply dips time in the future. A periodic trip is one during this period and ridership continues that is made regularly, at the same time, to rise. Between 4 and 5 p.m. productivity to the same destination, on the same day dips sharply but then rises and remains or days of the week. One telephone call above average from 5 to 10 p.m., reflect- initiates periodic service. ing a fairly good fit between vehicle sup- Wait time is the time elapsing from ply and requests for service. A better the end of the telephone call requesting match between supply and demand in the service to the time the vehicle arrives to other hours of the day is expected after pick up the customer. Of the three types the driver rebid in September 1972. of trips serviced by dial-a-ride, wait Average hourly ridership, vehicles in time is a measure of quality of service service, and vehicle productivity by pe- for immediate trips only. Three factors contribute to the average weekday wait time of 12.64 mm (Fig. 6). The first is Table 6. Average ridership, vehicles in service, and the time required to execute the control productivity by time period. room procedures, i.e., assigning thetrip to a vehicle, notifying the driver of the Vehicles location of the pickup, and recording data. in Produc- Period Riders Service tivity The second factor is the time required for the vehicle to travel from its current 7:01 a. m. to 9:00 a. m. 23.00 6.35 3.62 location to the pickup point. A third fac- 9:01 a. m. to 4:00p.m. 31.54 7.30 4.32 4:01 p.m. to 7:00p.m. 32.73 7.83 4.18 tor is a delay by the passenger who is not 7:01 p.m. to 11:00p.m. 15.40 3.45 4.46 ready when the vehicle arrives. 3.97 1.39 2.86 11:01 p.m. to 7:00 a.m. Beginning with the early morning hours, Medville 50

Figure 6. Average wait time. Figure 7. Average pickup time deviation.

I '''e°'°''° 11 ','' t,1? oo?i

pickup time deviation, 3 are the same as those that contribute to wait time. The fourth is the accuracy with which the control room staff estimates the promised pickup time. This estimation process takes into account the average pickup time deviation of the previous half hour, the number of buses available, and the loca- the mean wait time gradually increases tion of the origin of the current trip. as the number of requests for service in- Ride time is the time the passenger creases and reaches a maximum of 16.5 rides on the vehicle from pickup to de- min during the 3 to 7 p.m. peak period. livery. The average weekday ride time The relatively high wait times from 4 of 9.95 mm (Fig. 8) is close to the aver- to 5 a.m. and 8 to 9 p.m. are not typ- age automobile ride time of 9 min and ical because on the days sampled very implies that passengers usually travel few trips were made during these directly from origin to destination. periods. As with the mean wait time, the mean Pickup time deviation is the difference ride time increases on an hourly basis between the pickup time promised to the until peak ridership is reached during the passenger at the time the trip is re- midday period. At this point, the ride quested and the actual pickup time. Un- time reaches and remains at a level of like wait time, pickup time deviation can be measured for immediate, deferred, and periodic trips. Figure 7 shows that the mean pick- Figure 8. Average ride time. up time deviation is -2.13 mm, indicating that on the average vehicles arrive about 2 min earlier than promised. This deviation, which usually falls within a narrow range from -4.5 to + 3.0mm, is desirable because customers are not likely to be satisfied with the service if vehicles arrive excessively early or late. Of the 4 factors that contribute to the Medville 51

about 11.5 mm. Consequently, between Figure 9. Average wait and ride times and vehicle 2 and 7 p.m. when the greatest num- productivity. ber of passengers are carried, the typical dial-a-ride trip will take about 24 mm (13 min wait time and 11 min ride time). Table 7 gives average wait and ride times and average pickup time deviations by period of the day for the days sampled. As the day progresses and ridership increases, the average wait and ride times also increase. As vehicle productivity (requests per vehicle per hour) increases, so do the average wait and ride times because each vehicle will have more passengers to pick up and drop off. The relation between vehicle productivity and unit-ride time is shown in Fig- ure 9. The data show that, as the day I I, progresses and ridership increases, both vehicle productivity and customer wait and ride times also increase. The correlation between the sum of wait and ride times and the vehicle productivity was found to be +0.52. That is sufficiently high to indicate that a statis- Thus, if vehicle productivity is 4.0, then tically significant positive correlation average wait and ride times will be 24.21 exists between these 2 variables. When mm. Indeed, when actual vehicle produc- one is plotted against the other, it is tivities average about 4.0, the actual wait found that the regression equation de- and ride times are about 24 mm. scribing the relation is

Wait time + ride time = 12.89 + [2.83 x vehicle productivity] As of August 1972, dial-a-ride vehicles in the Haddonfield area carried more than 3,000 one-way trips per week and between Table 7. Average wait and ride times and pickup 500 and 600 per day. This does not rep- time deviations by time period. resent a steady-state demand level because usage of the system has been Pickup steadily increasing since the demonstra- Wait Ride Time Period Time Time Deviation tion project began. The most notable feature of demand for DAR has been the 7:01 a. rn to 9:00 a. rn 9.86 7.29 -3.39 9:01 a. m. to 4:00p.m. 12.98 10.15 -3.07 relatively high usage made of the system 4:01 p.m. to 7:00p.m. 14.29 11.36 -1.11 during the off-peak periods for many-to- 7:01 p.m. to 11:00p.m. 16.26 10.86 +1.61 11:01 p.m. to 7:00 a. m. 15.76 8.02 -1.75 many nonwork trips. The household survey provided a great Medville 52

deal of information concerning the attitudes toward and preferences of potential DAR users. Through this survey, residents of the service area had a voice in the selection of system operational parameters. The 60-cent cash fare, reduced rates for senior citizens and for those pur- chasing books of tickets, service to local shopping areas, 15-min average wait time and 10-min average ride time are all based on statements of survey respondents regarding dial-a-ride attributes that they preferred or felt were reasonable. As far as possible, the service has been designed to meet the needs of those whom it serves. DIAL-A-RIDE PROJECT IN ANN ARBOR: DESCRIPTION AND OPERATION

Thomas Urbanilc, II Dial-A-Ride Project Coordinator, Ann Arbor Department of Traffic Engineering and Transportation

The purpose of the Ann Arbor dial- a- Figure 1. Service area. ride project is to test the market response and economic viability of door-to- door, dynamically routed and scheduled public transportation in a Michigan community. The project is sponsored by the Michigan Department of Commerce, Bureau of Transportation, and the Ann Arbor Transportation Authority (AATA). The Transportation Research and Plan- ning Office of the Ford Motor Company is technical consultant. This report describes the project and gives a summary of the operation from September 20, 1971, to February 29, 1972. The information was assembled by the consultant and project staff.

SERVICE There are a maximum of 3 vehicles in service; the number is set according to The residential service area (Fig. 11) the demand. contains approximately 3,300 households. The fare structure is 60 cents cash Within this service area, customers may (exact fare) for a 1-way trip. Tickets request doorstep service by telephoning may also be purchased in strips of 10 for the dispatch center. The primary down- $5. A $15 unlimited-use pass is avail- town destination points to be served have able for a calendar month. Any number been organized into a loop. Passengers of members of a family may travel be- may request pickup or delivery at any tween the same 2 points on 1 pass. A point along the loop. At the 2 extremes special unlimited-use pass that is good of the loop, key points have been estab- only between 9:00 a.m. and 3:00 p.m., lished where there are free direct Monday through Friday, and all day Sat- telephones to the dispatch center. The urday is available for $10. loop is merely a convenient way to provide a many-to-few mode of operation. Vehicles can enter the loop at any point VEHICLES and, in actuality, rarely complete a trip around it. The entire dial-a-ride fleet consists of In addition, service is provided to the existing vehicles from the AATA fleet ex- University of Michigan, St. Joseph's cept for a Ford Courier (10-passenger Hospital, and University Hospital. At the van with roof conversion to provide a dispatcher's option, callers are provided stand-up interior) on loan from Ford's point-to-point service within the residen- Transportation Planning and Research tial service area. Office. The other 2 vehicles making up The service hours are from 6:30 a. m. the basic 3-vehicle fleet are 1969 Ford to 6:00 p. m., Monday through Friday, and club wagons (school bus package) that from 8:00 a. m. to 6:00 p. m. on Saturday. were refurbished and repainted for dial- 54 Urbanik

a-ride service. As many as 3 GMC 3301, patcher radios a driver and gives the 28-passenger vehicles have been used as driver a sequenced list of pickups in the backup vehicles for breakdowns or acci- service area. dents. During times of heavy demand, the dispatcher may transmit an unordered list. COMMUNICATIONS When he arrives at the downtown loop, the driver calls the dispatcher. (The The dial-a-ride system shares the point of entering the loop is usually de- existing AATA radio channel and trans- termined by the driver unless he is pre- mitter. All equipment is from existing viously instructed by the dispatcher.) inventory. The radio frequency is The driver circulates on the downtown 44,520 MHz (low band) and suffers skip loop (or executes the hospital and univer- from distant stations during summer sity leg), simultaneously dropping off in- months. The telephone communication bound passengers and picking up outbound consists of 3 public lines (group hunt) passengers according to the list of stops and 2 direct lines servicing the 2 free transmitted by the dispatcher. The telephones previously mentioned. These driver radios the dispatcher when he lines are in addition to 3 existing for the leaves the loop. regular AATA service. All outbound distributions are performed in sequence so that an outbound subtour is completed at the farthest point CUSTOMER SERVICE in the service area from downtown. No passengers are picked up until all pas- The dispatcher answers all incoming sengers are dropped off. The driver telephone calls and records the following radios the dispatcher after all passengers information: time of call, pickup ad- are dropped off. If there is insufficient dress or point, and delivery address or demand to dispatch another tour, the point. The dispatcher then gives the driver is instructed to park either in the customer an estimate of the pickup time. service area or downtown depending on For walk-in customers, the driver anticipated demand. radios the information given above to the dispatcher so that a complete record is made of each vehicle tour. JUDE RSHIP At all times, there is only 1 dispatcher on duty. During peak times, the head During the 24-week period from Wed- dispatcher for the line-bus operation an- nesday, September 22, 1971, through swers the phone and places the customer Saturday, March 4, 1972, 23,541 passen- on hold for servicing by the dial-a-ride ger trips were provided. Total weekly dispatcher. ridership has exceeded 1,200 trips 5 times during this period; the average for the last 11 weeks was 1, 176. Weekday DISPATCHING SYSTEM ridership has leveled off at approximately 214 trips per day, and Saturday ridership The central concept in dispatching the has declined to 108 trips from a preholi- Ann Arbor dial-a-ride system is the ve- day average of 148. (Standard deviation hicle tour. The tour beings when the dis- of week total is 53 trips; standard devia- Urbanik

tion of mean weekday average is 18 trips.) Figure 2. Hourly demand. Day of week trends have been estab- VEHICLES IN SERVICE lished as follows (days do not include 13.0

Thanksgiving, Christmas, and New Yeart s 12.0 Day): 11.0

Day Percent 10.0

Monday 17.3 9.0 Tuesday 17.8 0 Wednesday 17.5 8.0 rhursday 17.5 7.0 Friday 19.0 >- ..j 6.0 Saturday 10.9 ° 5.0

Figure 2 shows the hourly demand 4.0 pattern experienced on 3 typical days in December. The times are when service ( was requested and not when passengers 2.0 were picked up; therefore, some smooth- 1.0 ing of peaks in actual service is experienced. The dominant morning peak reflects travel to work and to Slauson Junior High School. The 10:30 to 11:00 HALF HOUR TIME INCREMENTS a.m., 1:00 to 1:30 p.m., and 4:30 to 5:00 p.m. peaks result from Slauson School shift changes. Rationale for the assignment of 2 vehicles from 8:30 a.m. to TRIP GENERATORS 2:30 p.m. is clear from the data shown in this figure. Table 1 gives the daily ridership total Approximately 93 percent of all dial- since the project began. Through De- a-ride trips are to or from points on the cember 18, 1971, 822 households have downtown loop, the hospitals, or the uni- used the dial-a-ride service at least versity area. The remaining 7 percent once, accounting for a 'tpenetration" of are intraservice area trips such as from approximately 25 percent. Throughout one address to another. The relative that period, new households were trying popularity of the several destination points the service at the rate of approximately available to dial-a-ride users is approxi- 63 per week (11 per day average). Al- mately as follows: though household use has not been tabulated subsequent to December 18, it is Point Percent almost certain that by February 29, 1972, Downtown loop 67 more than 1,000 households, or fully one- Slauson Junior third of those eligible, have used dial-a- High School 17 ride service at least once. However, Hospitals 12 most of these households are infrequent East University users. stops 4 Urbanik

REVENUE The revenue increase resulted from the higher pass cost. Although pass sales Table 2 gives a summary of revenue reached a peak of 92 in January, the pass - and ridership for the 5 pass periods since revenue peaked at $1,110 during Febru- the beginning of the project. Also given ary. The increased pass cost appears to are average pass fare,, as derived from have caused some users to switch from pass ridership and pass revenue, and pass to ticket or cash fares. The in- average fare for all users. The pass creased pass cost raised the average fare charge was $10 per month through Jan- from a low of 40.4 cents during the first uary; it was increased to $15 per month pass period to 49.4 cents during Febru- in February. ary. The increase was due in part to Even though the December ridership fewer rides per pass. During the first was lower than that for November, the period there were 38.9 rides per pass, revenue- was higher. This is because but there were only 34.4 during the Feb- more passes were sold, but were used ruary pass period. less, and because more rides were paid for with cash. Revenues for January and December were about the same, but Jan- SERVICE TIMES uary ridership was higher, making a lower average fare. The February revenue Detailed time studies have been con- *as $300 higher than that for January, ducted on 4 separate occasions; results but the ridership was almost the same. are given in Table 3. Ridership for the

Table 1. Daily ridership.

Week Total Beginning Monday Tuesday Wednesday Thursday Friday Saturday 268 9-20 - - 45 70 90 63 89 92 70 90 107 81 529 9-27 573 10-4 94 109 99 86 105 80 93 129 114 118 147 96 697 10-11 707 10-18 106 130 130 134 134 73 143 113 119 136 165 130 806 10-25 1,123 11-1 154 191 235 192 208 143 222 219 207 221 162 1,230 11-8 199 1,145 11-15 204 191 173 186 228 163 200 193 200 -. 147 131 871 11-22 1,200 11-29 199 205 202 192 238 164 208 208 199 209 155 1,192 12-6 213 1,197 12-13 205 238 201 193 242 118 216 185 170 128 97 -. 796 12-20 -. 621 12-27 129 128 137 136 91 191 195 194 192 201 102 1,075 1-3 1,140 1-10 198 202 185 214 234 107 238 232 207 233 250 89 1,249 1-17 1,229 1-24 184 214 226 242 234 129 226 194 188 220 212 119 1,159 1-31 1,225 2-7 234 209 237 223 227 95 2-14 203 229 209 199 209 113 1,162 2-21 174 225 208 228 223 111 1,169 2-28 205 213 207 219 227 107 1,178 'Holiday U rbanik' 57

study dates was as follows: In general, service, as measured by service times, improved between December Date Number 10, 1971, and February 18, 1972. Most 10-15 and 10-19 277 notable improvements were the reduc- 12-10 209 tions in average times and variances of 2-18 209 inbound waiting, inbound riding, and total inbound travel. The inbound service improved to both the downtown ioop and the Table 2. Monthly revenue and ridership. hospital-university area. There has been some deterioration in service times for Riders outbound travel from downtown, but out- Type Avg of Revenue Num- Per- Fare bound service times from the hospital- Fare Month (dollars) her cent (cents) university area have improved. Pass September- October 480 1,867 52.1 25.7 November 813 2,664 55.8 30.5 December 882 2,399 52.2 36.8 PRODUCTIVITY January 920 2,781 56.5 33.1 February 1112 J 51.5 43.6 The productivity of equipment and labor Total 4,205 12,257 53.8 34.3 used in a dial-a-ride system is the most Cash September- important determinant of cost per ride for October 835 1,392 38.9 November 1,033 1,722 36.1 a given set of prevailing local conditions. December 1,098 1,830 39.7 Productivity depends on overall demand January 1,040 1,734 35.3 February 1,046 1,744 35.5 levels, demand density (passengers re- Total 5,052 8,422 37.0 questing service in a given area), service TicketS September- area density and network, and character- October 131 262 73 istics of the dispatching system. November 154 308 6.5 Typical system productivity for Feb- December 148 297 6.4 January 162 323 6.6 ruary ridership levels was as follows: February 268 536 10.9 Total 863 1,726 7.6 Item Amount Otherb September- October 59 1.6 Vehicle-hours/day 27.2 November 79 1.6 Labor-hours/day December 76 1.6 January 81 1.6 Driver, regular 27.7 February 81 1.6 Driver, overtime 1.5 Total 376 1.6 Dispatcher 12.0 Total September- Total 41.2 October 1,446 3,580 40.4 Average passenger November 2,000 4,773 41.9 December 2,128 4,602 46.2 trips served 214 January 2,122 4,919 43.1 Passengers per February 2,424 4,907 49.4 _ . vehicle-hour 7.85 Total 10,120 22,781' 444 Passengers per 'Used by customers. Through observation period, 83 percent of tickets driver-hour 7.40 sold were actually redeemed; therefore, revenue from ticket sales was higher than amount Shown. Passengers per bNot accounted for in other fare types. Number was tabulated precisely dispatcher-hour 17.8 clur,ng first month and that same percentage was used for other months. 'Does not agree with total of 23,541 because 4 days (March 1 through 4) Passengers per arenogincluded. labor-hour 5.20 Urbanik 58

During the period from 4:30 to 6:00 p. m. bound tours) at the 1 percent confidence on high demand days, average productivity level and is accurate within ±12 min in often reaches 10 to 11 per hour. Although more than two-thirds of the cases studied. the question of ultimate system capacity remains unanswered, there is no doubt T = 25.36 mm + (3.17 min x N1 ) that a potential for growth exists during + (2.15 min x N0) many of the midday hours. The problem is one of maintaining service times. where During busy hours, stops are closer to- T = tour, gether, and consequently a given vehicle N1 = number of inbound stops, and may serve more calls per hour than dur- No = number of outbound stops. ing periods of slack demand. Produc- tivity could be increased during slack The tour times calculated by this equation hours only by forcing longer waiting and will be valid for the present demand den- riding times for the users during that sity. However, the objective of increas- period. Therefore, an increase in off- ing demand and thus reducing distance peak demand appears to be a highly de- between stops will force a recalibration sirable goal to build productivity levels. of the coefficients. The degree of ride sharing achieved in the Ann Arbor dial-a-ride system has resulted in an average of 5.25 passengers/ SURVEY OF CUSTOMER vehicle-tour. A detailed analysis of tour REACTION times has allowed the development of a simple linear relation for tour times, Two separate surveys have been con- which is statistically significant (H2 = ducted to measure customer reaction to 0.805 for inbound tours and 0.710 for out- dial-a-ride service. The first was a 3-

Table 3. Time study results.

Wait Time Ride Time Total Obser- Direction Date vations Minutes S.D. Minutes S.D. Minutes S.D. To downtown 10-15 and 10-19 80 8.0 6.2 11.2 4.0 19.2 6.8 12-10 51 11.5 8.3 14.9 5.1 26.4 8.7 2-18 45 9.9 7.9 11.7 4.0 21.5 8.4 From downtown 10-15 and 10-19 101 5.0 6.1 11.7 4.3 16.7 7.2 12-10 45 10.1 10.2 13.0 4.9 23.1 12.2 2-18 . 42 11.8 8.3 15.2 8.4 27.1 10.8 To hospital-university 10-15 and 10-19 12-10 14 16.1 10.0 18.4 8.1 34.5 9.1 2-18 23 9.1 7.4 15.0 4.6 24.1 8.5 From hospita1-university 10-15 and 1019 12-10 8 9.8 7.2 21.1 9.8 30.9 9.4 2-18 15 10.1 9.8 17.0 6.7 27.1 11.9

Note: Service times are not weighted by multiple calls for the same trip. Each datum is an observation of waiting, riding, and total travel time for a specific pickup and delivery address, independent of multiple riders; i.e., service times for 2 persons traveling together are recorded once, not twice.

Octobor waiting times are not properly shown because direct phones were not in and many persons did not call ahead. "Walkons" and "hail stops" were recorded as 0 wait times. b5rvice not operating. Urbanik 59

day on-board survey in which 298 valid What aspect of dial-a-ride service has returns were obtained. The second was not been attractive or satisfactory to you? a telephone survey of households who Users were asked, What improvement in used dial-a-ride service only once or dial-a-ride service would encourage you twice during the first 3 months to deter- to use it more frequently? The number mine what improvement might induce of respondents and their responses are these occasional users to use the service are given in Tables 5 and 6. more frequently. In addition to these The responses indicate that waiting surveys, others have been conducted by time is the most critical variable from a Berla and are reported in the next paper user's point of view. A greater variety in this Special Report. of available destinations seems to be the General information obtained in the surveys is given in Table 4. Because of the great concern over irregular riders, Table 5. Number of survey respondents. all surveys contained questions designed to urge service area residents to "sound Respondents off" about dial-a-ride service. In gen- To Service eral, it has been rather difficult to elicit Percent Questions any negative remarks about the service. Num- Non- Num- Per- People who have never used the service Survey ber Users users or who have stopped using it were asked, her cent On-board 298 100 255 86 Telephone 102 52 48 36 35 Home interview Table 4. Characteristics of users and Users 204 111 55 their travel. Nonusers 463 99 21

Item Percent

Trip purpose Table 6. Percentage of respondents and responses Work 35 regarding service improvement. Shop or personal business 33 School 23 Other 9 Home Inter- Former mode view Survey Automobile 50 Tele- Walk 19 On-Board phone Non- City bus 13 Response Survey Survey Users users Induced travelS 5 Other 13 Shorter wait Ride once a week or less 55 times; more Ride one-way only 50 accurate wait- Automobiles/household time prediction 38.0 19.5 0 0 None 8 More available One 47 destinations 22.2 25.0 62.2 49.5 Two or more 45 Lower fare; Licensed drivers 56 ability to make Men 69 change 23.2 11.1 18.0 12.1 Women 31 Shorter ride Age time 10.9 11.1 0 0 Under 18 39 Extended ser- 25to34 21 vice hours 0 8.3 9.9 11.1 Over 65 1 More inconvenient than Tripo that were not made before and would not automobile 0 11.1 9.0 14.1 have been made had dial-a-ride service not been Other 5.7 13.9 0.9 13.2 available. Urbanik 60

single improvement that would induce be approached as a long-range policy and nonusers to try dial-a-ride or to encour- educational issue by the Ann Arbor Trans- age occasional users to travel more fre- portation Authority. quently. Lower fares are simply not a As part of the research conducted by realistic option at this point (from an Berla, 886 citizens in the city's fourth Ann Arbor Transportation Authority ward (which is included in the dial-a-ride policy point of view). Experience with service area) were asked to express their extended service hours on Friday even- attitudes toward public transport. The ing has shown that ridership response is following responses were made to the not adequate to cover costs, at present question, Do you agree or disagree that budget levels. Ann Arbor really does not need a public Table 7 gives the results of a telephone transit system? survey that was conducted to determine why persons who tried dial-a-ride Response Percent only once or twice did not ride again. Disagree strongly 75 This survey showed that only 3.9 percent Disagree somewhat 13 of the respondents had had a bad service Disagree somewhat and experience; conversely, more than half of agree somewhat 5 them used dial-a-ride again, and 8.8 per- Agree somewhat 2 cent have become regular users. The Agree strongly 3 problem identified by this survey is that Do not know or no most families view dial-a-ride as a back- answer 1 up or secondary mode of transportation that is used only when the automobile is They were then asked how transit service not available. No evidence indicates that could best be improved in Ann Arbor and poor service is responsible for the rela- responded as follows: tive infrequent use most families make of dial-a-ride. This type of attitude toward Response Percent public transportation is impossible to address at the operating level and must Improve bus lines 8 City-wide dial-a- ride only 25 Table 7. Responses regarding user's experience Combined dial-a- with service. ride and line service 61 Respondents (percent) No bus service 1 Response 0 Other, such as Totally dissatisfied and will not use again 3.9 rapid transit 3 Neutral or supportive but automobile sat- isfies travel needs and cannot perceive of Do not know 2 any possible future need for service 18.6 only moderately dissatisfied or supportive and might use again under extreme cir- There appears to be substantial support cumstances, i. e., if automobile is dis- for public transport in general and for abled 15.7 Dissatisfied with service but use on irreg- dial-a-ride in particular. The Ann Arbor ular basis 0.0 Transportation Authority can be encour- Satisfied with service and use on irregular basis 52.0 aged by this result and should develop Satisfied and use regularly 8.8 policies that will turn this general support No response 1.0 into ridership on the system. DIAL-A-RIDE PROJECT IN ANN ARBOR: PUBLIC'RESPONSE Michael J. Berla Research Associate, PhD Program in Urban and Regional Planning, University of Michigan

Ann Arbor, like most other U.S. the technological system that effects the cities, is almost totally dependent on movement and the human beings who use private automobiles for the movement of it. people. Those who have access to auto- During the past decade, there has been mobiles have certain obvious advantages: much research on transit technology in They can go wherever and whenever they the areas of systems, hardware, and want to; they can choose those who oc- cost-effectiveness relations. In economic cupy their vehicles; and many of the terms, the supply side of the transporta- costs of their driving are heavily sub- tion equation has been intensively studied. sidized by members of the nondriving Much less attention has been paid to public. the demand side of the equation: who uses Conversely, there are strong disad- a given system? How much? What are vantages when a transportation system is the factors that encourage and discourage so heavily dependent on the automobile as the use of alternative modes? One rea- its only type of vehicle. For individuals, son, it seems to me, for a de-emphasis it is relatively expensive in terms of of the demand side of the transportation total cost per passenger trip, and it is equation has been the difficulty in analyz- inefficient in terms of the number of ing the demand factors as precisely as hours per day, week, or year that the au- the supply factors have been analyzed. tomobile is actually operated. For a For example, although it may be rela- community, it is wasteful of scarce urban tively easy to develop cost analyses for land necessary for streets and parking various alternative transportation modes, facilities, it generates more air pollution it is much more difficult in analyses of per trip than other modes, and it ex- consumer behavior to quantify the com- cludes or seriously disadvantages those ponents of demand. who are unable to operate or afford pri- In simplest terms, the supply cost of vate vehicles. a unit of transportation includes annual- Any effort to provide a more diverse ized capital costs plus operating costs transportation system in Ann Arbor im- per unit moved. As long as we are will- plies that serious efforts will have to be ing, as we have been traditionally, to made to shift a significant proportion of ignore the many externalities associated those trips now being made by private with transportation enterprises, we are automobile to other modes, primarily able to calculate relatively simply the public transit. To state this necessity, supply prices for various transportation however, is a long way from beginning to alternatives. bring about such a change. The change Unfortunately, the same has not been that is envisioned in the discussion that true in the analysis of demand for trans- follows will be extremely difficult to ac- portation. We have simply made the as- complish precisely because, for so many sumption that some set of preferences individuals and households, the present, somehow related perceived cash cost, single-mode system seems to be working travel time, convenience, reliability, and so well. However, the Transportation comfort in some way to produce individ- Authority believes that Ann Arbor has ual demand functions for each consumer. much to gain from the effort. These demand functions can somehow be A transportation system may be char- aggregated to produce a single demand acterized as a method for moving people curve for a given transit service. When and goods. The 2 major components are that curve is intersected with the relevant Berla MM

supply curve, an equilibrium price and computerized registered voter file. quantity are produced. Comparison with a list of households that But what, in fact, are the components had been created by a census of the pro- of the demand function for transportation? posed dial-a-ride service neighborhood Take cash cost as one example. Accord- revealed that approximately 90 percent of ing to Lansing and Hendricks (1), whose all households in the neighborhood were work was based on data gathered in 1963 discovered by a computer canvass of the and 1965 in an extensive national survey, registered voter file. fewer than 1 out of 3 drivers has ever A second survey was conducted during calculated what it costs to make the jour- a 1-month period, beginning during the ney to work by automobile; even fewer thirteenth week after the system went in- agree on those costs. Yet economic to operation. The sample for this survey theory is based on the assumption that consisted of 40 percent of all house- consumers are knowledgeable about the holds in the original dial-a-ride service costs of options they face and that they neighborhood— approximately 840, of tend, at least in the aggregate, to make which 675 were actually interviewed. rational decisions among them. Finally, a third major source of data Again, even if consumers were aware about consumers consisted of the dis- of the cash costs of various transporta- patch records kept. Ideally, every single tion choices, how would they trade these passenger trip made on dial-a-ride is costs against noncash factors such as assigned to a specific household. The travel time, convenience, comfort, and other end of the trip is also recorded as privacy? We really do not have the be- well as the number of persons in the ginning of a set of comprehensive data household making the trip and the meth- regarding these crucial questions—let od of payment (cash, ticket, or flash alone a theory that would relate these pass). data if they existed. One problem that developed in data In Ann Arbor, when it became clear from dispatch records, was the inability that there would be an experimental of the system to distinguish particular demand-responsive transit system im- apartment numbers within a multiple unit plemented in a portion of the city, we de- street address. For this reason, trip cided that, in addition to the usual work data assigned to 96 households at on cost factors, we would place some multiple-family addresses have been de- emphasis on studying the demand for leted from the following data. The bal- such a system. We began by conducting ance of this paper, then, consists of a a mail survey in the political subdivision discussion of several of the items col- of the city in which the experiment was lected from each of the remaining 579 to be conducted. Because there had been single-family households located in the little publicity regarding dial-a-ride original dial-a-ride service neighbor- prior to the completion of the survey, we hood. Particular attention will be given felt that the responses would provide a to significant differences between using reasonable set of "before-experiment" and nonusing households and to differ- data, which could later be compared with ences among various levels of use among "after-experiment" data. user households. What we are attempt- Some 6,300 four-page questionnaires ing to do is to isolate those factors most were consequently mailed to an address strongly associated with and, therefore, list generated from the city clerk's best predictive of levels of household use Berla 63

during the first 13 weeks of the experi- to receive the marketing treatment and mental project. some were not. In fact, however, this One aspect of the experiment that was distinction was not made by the U. S. under control of the researchers had to Postal Service; consequently, households do with the paid advertising for the new that were part of multiple-unit complexes system. Given an unrealistically low were all assigned missing data codes to item for marketing in the project budget, indicate their status on the newsletter the Transportation Authority decided that mailing list. paid advertising for the service should be We are then left with 579 single-family limited to direct mail inasmuch as all units out of a total of 2,066 units in ser- other forms of advertising would go in vice area A for which we have both sur- large proportion to households not in the vey data and reliable trip and marketing service area. (As originally constituted, data. Of these 579 units, 410 or 71 per- the service neighborhood comprised less cent never used the dial-a-ride service than 7 percent of all households in the during the monitoring period, and 169 city of Ann Arbor; this represented an or 29 percent used the service at least even smaller fraction of the circulation once. delivered by local printing or broadcast- Table 1 gives relevant variables, ing media.) where available, for the 675 survey The direct-mail format consisted of a households, the 306 single-family units series of 8 two-page or four-page news- that received newsletters, and the 273 letters mailed to 80 percent of the house- single-family units that were not mailed holds in the service area during the newsletters. course of the 15-week period beginning First, consider simply the following about 2 weeks before service was ini- distribution of the 579 households by tiated. The 20 percent of households that number of trips made during the study did not receive the newsletters consisted period. of half of the total 40 percent sample that was surveyed in December and January. Use Percent For the marketing impact experiment, Did not use 71 the 20 percent constituted the control Did use 29 group, whereas the 80 percent that did itrip 7 receive the newsletters were the treat- 2-3 trips 7 ment group, 25 percent of whom were 4-8 trips 7 drawn into the survey sample. 9 or more trips 7 As with trip data, we also found that treatment data (i. e., information as to The maximum number of trips by a single whether the household had received the household was 140. The maximum num- newsletters) were unrealiable for multi- ber of trips generated by any family in family units. This resulted from the the entire dial-a-ride service area dur- fact that in apartment units the typical ing the 13 weeks under consideration was method of distributing bulk mail is to 345, or about 41, percent of all trips leave a pile of the pieces in some central made by the almost 2,100 households in place near the apartment mailboxes the neighborhood. rather than to distribute one piece to From data collected in almost 600 home each mail box. In many apartment com- interviews, 'what can we say about differ- plexes, some households were supposed ences in households, given differences Berla 64

Table 1. Variables for survey households.

Households

Received Did Not Receive Variable All Newsletters Newsletters Percentage using system 1 or more times in first 13 weeks -. 30.0 28.4 Use during first 13 weeks Mean - 2.76 3.68 Standard deviation - 9.57 16.23 Mean number in household 3.14 3.5 3.45 Median income class $15,000 to 20,000 $15,000 to 20,000 $15,000 to 20,000 Median education of male head of household Bachelors degree Bachelor's degree Bachelor's degree Mean number of automobiles 1.72 1.76 1.83 Mean number of licensed drivers 2.07 2.10 2.18 Mean number of full- or part-time employees 1.37 1.36 1.44

Da1a not available. in use of the service? First of all, we Ann Arbor would bebetter off if know that use or nonuse of the service part of the central business district were was related to the following items that closed to private automobiles. were collected in the survey data: the Many familes in Ann Arbor would respondent's attitude toward the inter- be better off if they could spend less of relation between the automobile and their incomes on owning and operating the city environment; the number of private automobiles. teenagers in the household; the ratio of Ann Arbor's future is seriously number of persons over age 5 in the threatened by the growth of private auto- household to number of automobiles reg- mobile ownership. istered to household members; and edu- My family would be willing to re- cational attainments of male and female place some of our trips by private auto- heads of households. mobile with trips by public transit. Perhaps the most interesting item in The scores on these 6 items were this list is that dealing with attitudes to- summed and averaged. The more a re- ward the automobile and the city's en- spondent agreed with these statements, vironment. Thirteen items were con- the more likely was it that the household structed to attempt to probe this relation. was one of those using dial-a-ride. We Factor analysis was then performed on are now in the process of analyzing these the scores on these items, and from them data to produce an algorithm for predict- 6 items were selected that seemed to ing the probability that a given household capture most of the information they con- will use this kind of service. tained. Respondents were asked to eval- Within the smaller group of 169 house- uate whether they agreed or disagreed, holds who used the system one or more along a 5-point scale that was provided, times and on whom we have reliable use with the following 6 statements: and survey data, there is a rather in- There are too many private auto- teresting distribution of the variable mobiles in Ann Arbor today. "number of trips." The distribution is More people in Ann Arbor should negative exponential for the values be- get out and walk or ride bicycles instead tween 1 and about 30 trips. Beyond this of driving their cars. limit, because of the existence of extreme Berla 65

cases, the distribution becomes curvi- range of potential ridership than tradi- linear even when expressed in logarithms. tional aggregated modal-split and origin- We have broken the set of user fami- destination models have done in the past. lies into the 156 households that used the As a result of using these aggregated system between 1 and 29 times, the models, the weekly ridership per house- straight-line exponential portion of the hold predicted in our application for state curve, and the 13 households that used demonstration grant funds was 6 times the system more than 29 times during the what the actual figure turned out to be 13 weeks under study. (Incidentally, after the system stabilized and allow- these 13 households, representing only ances were made for changes in the scope about 7 percent of all user households in of the service area from what had initially the sample, generated collectively al- been projected. most exactly half of all trips recorded One other aspect of our survey seems by this sample.) worthwhile of attention at this time. We Once we have identified a user family, asked all respondents whether, if the another set of variables appears to ex- localized dial-a-ride system could be plain the actual number of trips that any expanded to a city-wide service that given household will actually make, given would cost the taxpayers about $5 per that it will make 1 or more trips. Here person or about $500,000 per year, they the strongest variables on which we col- would vote for or against such a tax. lected data are ratio of number of per- Of 883 respondents, 64 percent answered Sons in the household to number of auto- "Yes", 23 percent answered "No", and mobiles; length of time the family had 13 percent answered "Don't know." Sup- lived at its present address; number of port was about as high among those house- residents in the household; household in- holds who had used the system one or come; number of full-time employees; more times as among those who had not. educational attainment of the heads of We were interested in predictors of household; and number of licensed driv- this variable measuring political support ers in the household. Interestingly, the for funding such a system. Table 2 gives set does not include the index of atti- a 2-dimensional comparison of house- tude toward the automobile and the holds: those in the service area and not environment. in the service area and those sent the It should be possible to develop a pre- newsletter and not sent the newsletter. diction formula, based on regression Initially, we predicted that the effects of analysis, for this subset of the sample. these 2 variables on level of support for How we will handle the extreme 8 per- a city-wide system supported by a half- cent of cases that have, in this particu- million dollars in taxes would be both posi- lar sample, accounted for approximately tive and additive. That is, we hypoth- half of the total trips is not yet clear. esized that people in the service area Perhaps the best first approximation will would, on the average, exhibit more sup- be simply to use a multiplier on the num- port for the service than those not in the ber of trips predicted out of the straight- service area; and we hypothesized that, line portion of the sample. in either case, people who received the 8 In any case, it is hoped that predic- newsletters —would be more favorable to tions generated from this kind of house- to a tax-supported city-wide system, re- hold survey analysis can do a far more gardless of where they lived, than those accurate job in the future of setting the not so treated. Berla 66

Table 2. Support for dial-a-ride subsidy by service and marketing treatment.

Service No Service All

Marketing Treatment Response Number Percent Number Percent Number Percent No newsletter Yes 185 61 39 66 224 62 No 77 25 12 21 89 25 Don't know 43 14 6 11 49 13 Total 305 100 57 100 362 100 Newsletter Yes 179 67 39 66 218 67 No 54 20 12 20 66 20 Don't know 36 13 8 14 44 13 Total 269 100 59 100 328 100 All Yes 364 63 78 67 442 64 No 131 23 24 21 155 23 Don't know 79 14 14 12 93 13 Total 574 100 116 100 690 100

We found that, however, 63 percent of difficult to wean any large number of in- the households receiving service and 67 dividuals away from deeply entrenched percent of those not receiving service travel behavior patterns, particularly supported the imposition of a tax to pro- when travel modes are strongly reinforced vide city-wide dial-a-ride service. The by existing institutional arrangements. data also showed that 62 percent of house- The American city is currently based holds that did not receive the newsletter on the premise that most people will and 67 percent of those that did sup- make intraurban trips by privately ported the tax. Again, the difference is owned and operated motor vehicles. And marginal. a person's automobile is widely recog- These variables proved to be inter- nized as a symbol of his wealth and status. active rather than additive. That is, the Further, because most of the cost of setting on one variable predicted a dif- automobile ownership and operation is ferent rate of change, and indeed of not related to the number of miles the direction of change, on the other variable. vehicle is driven, being heavily dominated Whereas, in the service neighborhood, by the fixed costs of depreciation, in- support for the city-wide system subsidy surance, and licensing, the marginal cost went up, as predicted, with application of of any given trip by private automobile is the newsletter marketing, it went down, relatively low. under the marketing treatment, in those The ability to bring about significant households that were not in the dial-a- trip diversion, then, is heavily dependent ride service area. The differences are on the rate of automobile ownership in a relatively small, so that a good deal of community. In Ann Arbor, that rate is caution must be exercised both in accept- about 1.75 per household, or 7 cars for ing and in interpreting these results. every 4 households. As long as there is What preliminary conclusions can be any significant proportion of desired trips drawn from Ann Arbor's dial-a-ride ex- that can only be made by private automo- periments to date? First of all, and per- bile, most affluent families will choose to haps most important, it is extremely own enough automobiles to be able to Berla 67

make those trips. Yet, because of the propose to city government and the voters low marginal cost of any trip, once fixed a specific policy that calls for major costs of ownership have been met, many efforts to increase the scope of public of the trips that could have been made by transit in the city. That increase will public transit are more cheaply and con- come in the form of a city-wide system veniently made by private vehicle. of neighborhood dial-a-ride feeders offer- Yet, Ann Arbor seems to be in the ing coordinated transfers to an express process of reevaluating its transportation trunk-line system for interneighborhood preferences and habits. Evidence of this transportation and to major trip genera- is still scattered, out it is sufficient to tors such as the university, hospitals, encourage the Transportation Authority shopping centers, and the downtown cen- in its efforts to expand the percentage of tral business district. trips made on public transit vehicles. Thus, we project a reversal of the Last year, for the first time in recent downward trend in public transit that has memory, a road improvement bond issue characterized urban life in the United was voted down by about a 2 to 1 major- States since World War II. That reversal ity. Most of the opposition came from is apparently under way in Ann Arbor and groups and individuals who said that the in other cities throughout the country and city needed to rethink its transportation indeed throughout North America. That patterns and should explicitly attempt to dial-a-ride has a major place in the plan- divert trips from private to public ning for this reversal now seems assured. transportation. In March, 1972, the question of supporting a city-wide dial-a-ride system at a REFERENCE cost to the taxpayer of approximately $5 per person, or $500,000 per year, was in- 1. Lansing, J. B., and Hendricks, G. cluded in a survey of voter attitudes in Automobile Ownership and Residen- a sample drawn from the entire city. tial Density. Institute for Social Re- Here again, the favorable response won search, Aim Arbor, Mich., 1967, majority support, 56 to 35 percent, with Ch. 4. 8 percent undecided. Finally, during the past 2 years that the city budget has been under consideration, the dial-a-ride project has had additional money voted for it, i. e., more than that in the recommended budget of the city administrator. These decisions came during an era when the city's fiscal condition has been more precarious than at any other time in the past 2 decades. These and other scattered pieces of evidence indicate, at least to members of the Transportation Authority, that the public and itselected representatives are rethinking basic transportation issues in Ann Arbor. Within a few months, the authority will DIAL-A-RIDE PROJECT IN ANN ARBOR: LEGALITY

Jerold Lax City Attorney, Ann Arbor

Prior to the commencement of dial-a- The city responded to the complaint of ride service in Ann Arbor in September the taxicab companies by filing a motion 1971, rather clear indications existed for summary judgment, in which the that the local taxicab industry regarded Ford Motor Company joined. In its mo-. the program with fear and suspicion and tion, the city answered the principal con- and would give serious consideration to tentions of the plaintiffs as follows: instituting litigation to prevent the program from coming into being. In an ef- (1) The alleged necessity for compliance with the fort to cooperate with the taxicab industry Taxicab Ordinance. The Ann Arbor Taxicab Ordi- and to alleviate its fears, the Ann Arbor nance (City Code, Chapter 85) defines "Taxicab" as Transportation Authority specifically de- follows: signed the dial-a-ride program so that taxicab companies could bid to become 7:151(1) "Taxicab" shall mean and include any motor vehicle the operators of the system. No bids operated solely ormainly within the public streets and quasi- were received1 however, and the author- public places of this City, accepting passengers for transporta- ity proceeded with plans to operate the tion for hire on call or demand, between such points as may be system itself. To no one's great sur- directed by the passenger or passengers. The term taxicab shall not include vehicles furnishing mass transportation service, prise, a lawsuit denominated Kon et at. such as motor buses which operate over fixed routes or on a v. City of Ann Arbor et al. (Washtenaw fixed schedule or between definite termini; buses employed County Circuit Court, No. 5967) was com- solely for transporting school children; chartered buses; or menced by Ann Arbor's 2 major taxicab motor vehicles used solely for funerals, weddings, christening,, companies just a few days prior to the and similar events. scheduled commencement of service. The principal relief requested in the suit Plaintiffs assert that the dial-a-ride system, if estab- was an injunction against the operation of lished at all, must be established in conformity with the dial-a-ride system. the Taxicab Ordinance, but it is clear from the very The taxicab companies contended that definition of taxicab that the ordinance is inapplicable the establishment of dial-a-ride would be to dial-a-ride. unlawful for several reasons. First, the ordinance states that taxicabs will operate "between such points as may be directed by Dial-a-ride vehicles were really the passenger or passengers." The dial-a-ride vehi- taxicabs and were, therefore, required cles, however, are not subject to the specific direc- to obtain licenses under the Ann Arbor tions of the passengers. As described in the dial-a- taxicab ordinance; ride work program, the vehicles will pick up passen- The granting of licenses to exist- gers at their homes but will be permitted to drop ing taxicabs by the city constituted an im- these passengers off only along a loop surrounding plied agreement by the city that it would part of the central business district. As the work program makes explicit, "No stops will be made on not engage in a competing business or, in streets off the loop." the alternative, that if it did engage in Second, the ordinance specifically exempts the fol- such a business it would do so on terms lowing from the definition of taxicab: "vehicles fur- identical to the terms under which the nishing mass transportation service, such as motor taxicab industry operates; and buses which operate over fixed routes." The dial-a- Ford Motor Company (which was ride vehicles will be operated under the auspices of the sued as a co-defendant) was being greatly Ann Arbor Transportation Authority, a body corpo- enriched by the program without giving rate duly organized under P.A. 55 of 1963 [M.S.A. adequate consideration in return, and the Section 5.3475(1) et seq.] for the specific purpose of public was thereby defrauded. operating a mass transportation system. Each vehicle Lax 69

will provide transportation for 12 persons at one time, fully contend that the city agreed not to commence a and no passenger will have the power—as he would in dial-a-ride system, which, by definition, is not even a a taxicab—to limit the number of passengers to be taxicab system. Plaintiffs state only that their li- transported. Moreover, as previously indicated, the censes are implied contracts, and, under the clear-cut dial-a-ride vehicles operate at least in part along a decisions of the U.S. Supreme Court, the exercise of fixed route, namely, the central business district loop. the municipal police power in the public interest can- Hence, because the dial-a-ride vehicles are not to be not be relinquished by mere implication. subject to the specific directions of their passengers, Thus far, plaintiffs' contentions have been treated because they will furnish mass transportation service, purely as matters of law and have been shown to be and because they will operate over fixed routes, these untenable. It might be added, however, that plain- vehicles are simply not taxicabs under Chapter 85 of tiffs' arguments are also factually unsupportable. The the City Code and,therefore, need not conform to the dial-a-ride system, in its initial phases, will serve only provisions of that chapter. 2,100 of Ann Arbor's 31,000 households and only (2) The alleged unfair competition, breach of con- between 6:30 a.m. and 6:00 p.m. on Mondays tract, and "deprivation" of property without due pro- through Thursdays and between 6:30 a.m. and 11:00 cess. The individual plaintiffs are municipal taxicab p.m. on Fridays and Saturdays, and is thus clearly in- licensees, and they claim that this status gives them capable of competing—let alone competing unfairly— standing to prevent the city from instituting the dial- with the taxicab industry, which serves the entire a-ride system. A remarkably similar contention was community 24 hours a day. Thus, even if the cre- advanced by the operators of private streetcar systems ation of a municipal transportation system gave plain- which had been municipally franchised when the city tiffs a legal foundation for their claim—which it does of San Francisco proposed to construct a municipal not—their "damages" in this case would be entirely system; the battle progressed through the federal too speculative to justify equitable relief. courts and up to the U.S. Supreme Court, and at all (3) The alleged illegality of cooperating with Ford levels the power of the municipality to create its own Motor Company. Plaintiffs argue that, because Ford transportation system was upheld [United Railroads Motor Company stands to benefit from its agreement v. San Francisco, 239 F. 987 (ND. Calif. 1917); af- to cooperate regarding the development of the dial-a- firmed, 249 U.S. 517 (1918) (Holmes, J., for a un- ride system, commencement of the system is illegal. animous Court)]. The short answer to this contention is that, if benefit Before the U.S. District Court, the operators of the to a private party invalidated a government contract, private streetcar systems argued—much as plaintiffs do a government could almost never enter into con- here—that the creation of a municipal system would tracts—a patently absurd conclusion. breach contractual obligations created by their fran- Plaintiffs appear also to advance the related argu- chises and would deprive them of property without ment that the consideration received from Ford is due process of law. The court rejected this argument "woefully inadequate" and that the agreement is quoting with approval from the decision of the U.S. therefore invalid. This argument, particularly in light Supreme Court in Knoxville Water Co. v. Knoxville of the facts of the situation, is entirely without merit. (200 U.S. 22): It is first to be noted, as a matter of law, that a municipal contract is presumed to be valid (1). More- A municipal corporation, when exerting its functions for the over, the adequacy of consideration is not generally a general good, is not to be shorn of its powers by mere implica- matter of judicial concern (1). In this case, it is the tion. If by contract or otherwise it may, in particular circum- considered judgment of the members of the Ann stances, restrict the exercise of its public powers, the intention Arbor Transportation Authority that the arrangement to do so must be manifested in words so clear as not to admit with Ford Motor Company is more than fair to the of two different or inconsistent meanings. city and its residents and very much in the public interest. In the instant case, plaintiffs do not contend—nor While it is true that Ford may obtain data which could they truly—that the city of Ann Arbor in grant- will be useful to it in developing dial-a-ride systems in ing taxicab licenses explicitly agreed not to enter the other localities, the information to be obtained from taxicab business. A fortiori, plaintiffs cannot success- the Ann Arbor experiment will be public information, Lax 70

usable not only by Ford but by all other interested companies on appeal were essentially the parties. Furthermore, Ann Arbor is under no obliga- same as those made at the trial level, tion whatever to obtain future vehicles from Ford if with a slight shift in emphasis. Rather the system proves successful. Ford, in exchange for than complaining principally of a pur- this information, is devoting numerous hours of ex- ported violation of an implied agreement pert manpower to developing a system which is ex- not to compete, the companies contended pected to be of great long-term benefit to the citizens they were being denied equal protection of Ann Arbor in meeting their transportation needs; of the laws because they were governed additionally, Ford will, at no cost, lend a vehicle to the Transportation Authority for use in the initial by standards different from those applied phases of the program. Even if the court is inclined to dial-a-ride concerning such matters to consider the question of adequacy of considera- as rates and licensing. The city, in its tion, there can be no doubt that the citizens of Ann brief, answered this contention as fol- Arbor are being treated fairly in the instant situation. lows: Summary: What was true in the case of the San Francisco streetcars in 1917 is even more true in to- While appellants phrase their constitutional argu- day's crowded urban environment: The municipality ments both in terms of due process and equal protec- must be permitted to further the public interest by tion, it appears that these arguments are based on a improving the system of public transportation. Par- single premise, namely, that similarly situated activ- ticularly in a case like the present one, where the pro- ities are being treated in an unlawfully dissimilar posed improvement is experimental in nature and manner. As appellees have demonstrated in the pre- covers only a small part of the city, the speculative ceding portion of this brief, it is simply not the case fears of the taxicab industry provide no basis for that dial-a-ride and taxicabs are similar activities; equitable relief. moreover, appellants' suggestion that taxicab rates are somehow forced upon them by a malevolent city government is simply untrue. Hence, appellants' The city further contended that the premise is false, and their argument is without sup- taxicab industry was precluded by the port. However, it is worth going on to point out that legal doctrine of estoppel from obtaining even if dial-a-ride and taxicabs were virtually identical injunctive relief. The city argued that, in their operations and even if the city did force the at the urging of the taxicab industry, the taxicab industry to charge particular rates, appellants' city had gone out of its way to make it constitutional claims would be invalid, for the follow- possible for the taxicab industry to being reason: Dial-a-ride is an activity of a governmen- come the operator of the dial-a-ride sys- tal agency, performed in the interest of public health, tem and that it would therefore be ineq- safety, and welfare, and this fact would make it con- uitable for the taxicab industry to be per- stitutionally permissible to govern dial-a-ride by stan- mitted to keep the system from coming dards different from those applied to the taxicab in- into being. dustry. Following a hearing, Washtenaw In Springfield Gas and Electric Co. v. City of County Circuit Judge Ross W. Campbell Springfield [257 U.S. 66 (1921)], the U.S. Supreme granted the city's motion for summary Court considered, and unanimously rejected, the judgment, thereby dismissing the lawsuit. claim that it constituted a denial of equal protecticn A copy of the judge's opinion is included of the laws for a state to require private utilities to be in the Appendix. regulated by a utilities commission while allowing a municipality to set the rates for a utility owned by it. Subsequent to the issuance of the Cir- Mr. Justice Holmes explained the Court's conclusion cuit Court order, the plaintiffs filed an in the following way (257 U.S. 70): appeal with the Michigan Court of Ap- peals. The matter was argued in April The private corporation, whatever its public duties, is orga- 1972. The arguments of the taxicab nized for private ends, and may be presumed to intend to make Lax 71

whatever profit the business will allow. The municipal corpo- elected not to appeal the decision of the ration is allowed to go into the business only on the theory Court of Appeals to the Michigan Supreme that thereby the public welfare will be subserved. So far as Court. Hence, it would appear that the gain is an object, it is a gain to a public body, and must be used legal basis for Ann Arbor's dial-a-ride for public ends. Those who manage to work cannot lawfully make private profit their aim, as the plaintiff's directors not system has been firmly established. only may but must. Reference Appellants cite an A.L.R. annotation and a handful of cases which purportedly demonstrate that it makes 1. McQuillin. Municipal Corporations, no difference that dial-a-ride is a governmental project 3rd Ed. Sections 29.96 and 29.02. and that dial-a-ride and taxicabs must follow the same procedures. These authorities, however, are utterly irrelevant. All that these authorities indicate is that, if a government provides a service, it cannot unreason- Appendix ably discriminate among users of the service. For example, the Ann Arbor Transportation Authority could TRANSCRIPT OF OPINION OF not discriminate unreasonably among dial-a-ride riders. The authorities in no way suggest that there is WASHTENAW COUNTY CIRCUIT JUDGE anything unlawful about a governmental agency en- ROSS W. CAMPBELL gaging in an activity which competes with a private activity, even if rate structures are different and even THE COURT: Gentlemen, I apologize if such competition is detrimental to the private ac- for being much longer than I had antici- tivity. pated, but in deference to the amount of Indeed, the propriety of governmental agencies en- work which counsel have put into the case, gaging in activities potentially competitive with pri- the numerous serious questions pre- vate business has been upheld by the highest courts sented and their complexity required on both the state and federal level, and the poverty of more time to decide the malter than I had appellants' position is best indicated by the fact that anticipated, and I wanted to be able in appellants have consistently ignored these controlling rendering my decision to make the opin- decisions throughout these proceedings [Springfield ion as detailed as the complexity and Gas and Electric Co. v. City of Springfield, supra; number of issues required. United Railroadsv. San Francisco, 249 U.S. 517 First of all, I would like to comment (1918); Detroit v. Wayne Circuit Judges, 339 Mich. that this, indeed, is a most unfortunate 62 (1954); Andrews v. City of South Haven, 187 situation. The public through its duly Mich. 294(1915)]. To summarize, it is not true that dial-a-ride is the elected officials and government is trying same sort of transportation system as the taxicab in- to develop and improve less expensive dustry, nor is it true that any governmental agency systems of transportation for the people of forces the taxicab industry to charge rates higher than the community, and the changes they are those of dial-a-ride; however, even if either or both of attempting to introduce, at least experi- these allegations were true, appellants would have mentally, necessarily compete with and failed to state any valid constitutional claim. threaten the livelihood of those who are established in providing additional ser- On June 2, 1972, the Court of Appeals vice. The situation is somewhat remi- rendered a unanimous decision upholding niscent of the dislocation that we know the Circuit Court and affirming the legal- accompanied the advent of the industrial ity of Ann Arbor's dial-a-ride system. A revolution many years ago, a process copy of this decision is also given in the which is still in evolution. But this case Appendix. The taxicab companies have is not so much a conflict between the mu- Lax 72

nicipal and private enterprises as a mat- fixed route. Now these are not the only ter of mutation and experimental change indicia, but it would be difficult to con- in the form of public transportation ser- ceive of a taxi, at least within our tradi- vice, as I see it. tional concept of a taxicab, in which a Let us assume for a moment, without passenger did not have those two rights. deciding, that dial-a-ride is a taxi ser- Now even if these vehicles are other- vice under Chapter 85, Section 7.151 (12), wise classed as taxicabs or even if they of the Ann Arbor City Code, and that, if are ordinary taxi vehicles which the city operated by a private person or a corpo- should choose to utilize for this purpose ration, it would fall within Chapter 85, (I do not understand that they do but as- Section 7.161, of the code, which requires suming for the purpose of this argument a certificate of public convenience. or this opinion that the city were to uti-, The Court does not interpret Chapter lize ordinary types of vehicles like those 85, Section 7.161, of the code as applying used as taxicabs), I would find that they to the city itself. It would be patently are vehicles which are within the words useless and circular to require a city to of the ordinance furnishing mass trans- obtain from itself a certificate of public portation service, and the furnishing of convenience and necessity before it could mass transportation service is not de- operate a taxicab service itself. The pendent upon the configuration, the ge- provision of the code was obviously in- ometry, size, number of seats, or the tended to apply only to persons other than color of the vehicle which is used for that the city itself. So, I find first that pro- purpose. As such, I find that these vehi- vision of the city code does not apply to cles are expressly exempted from the the city itself should it undertake to op- definition of taxicab under Chapter 85, erate a taxi service. Section 7.151(12), of the Ann Arbor Second, viewing the complaint in the City Code. My third finding, then, is manner most favorable to the plaintiffs, I that these vehicles are not taxicabs find that there is no estoppel operating within the definition of this section of against the defendants. the code. Third, there is no allegation that an Now, there is no question but that the individual passenger in the vehicles dial-a-ride system will compete with the which the transportation authority would plaintiffs, but does it constitute unfair be operating would have the power, as competition, within the technical defini- they would in a taxicab, to limit the num'- tion of that phrase, as grounding an action ber of passengers who could be in those under the law? To do so, there must be vehicles, again viewing the complaint in traditionally a passing off or pawning off the manner most favorable to the plain- the goods or services of one person as tiffs; that is, there is no allegation that those of another. It is not every com- the passengers or any one passenger petition, no matter how hard it may be could hire the entire vehicle with one fare on the person who is not used to that and deprive other persons or other mem- competition, which falls within the legal bers of the public from occupying empty definition of unfair. seats in it. There is no allegation here of any There is further no allegation in the passing off or pawning off of the services complaint that the vehicle could be hired provided by the proposed transportation to take any particular route that the pas- authority as those of any of the plaintiffs, senger wishes; instead it must follow a individual or corporate; and, accord- Lax 73

ingly, my fourth finding is that there is court opinion growing out of the case no unfair competition within the legal which I just cited (United Railroads v. definition of such a phrase as capable of San Francisco, 249 U. S. 517, 520) that grounding a cause of action. a covenant by a city not to grant to any Do the city licenses issued to plain- other person or corporation a privilege tiffs constitute a contract which prevents similar to that granted to the covenantee the city from going into the taxi business does not restrict the city from itself ex- itself? If so, such a contract exists only ercising similar power. by implication. I would quote from Mr. Crippen has well made his point United Railroads of San Francisco v. City here that the city originally put this sys- and County of San Francisco (249 U. S. tem contract out for bids and might very 517, 993) as follows: well have contracted with a private agency for this purpose. But that is not the question before us here, and we will In the construction of legislative enactments and of not address ourselves to that. What we ordinances and of contractual relationships which directly concern the public, the doctrine which con- have here is a case where a municipal trols is as announced in Knoxville Water Co. v. Knox- authority itself will be operating the ville, 200 U.S. 22; 26 Sup.C. 224, 50 L.Ed. 353: "A transportation system. Accordingly, the municipal corporation, when exerting functions for fifth finding of the Court is that the fran- the general good, is not to be shorn of its powers by chise issued by thecity to the plaintiffs mere implication. If by contract or otherwise it may, does not constitute a contract by the city in particular circumstances, restrict the exercise of its not to compete. public powers, the intention to do so must be mani- The plaintiffs complain of deprivation fested by words so clear as not to admit of two dif- of property without due process. The ferent or inconsistent meanings." kind of damage which constitutes depri- This general rule is but another form of stating a vation of property without due process principle that statutory grants by way of franchise or and grounding of an action on that basis property, in which the government or public has an is damage which results from conduct, interest, are to be constructed strictly in favor of the like taking or appropriation, that would public, and whatever is not unequivocally granted is be tortious in and of itself, unless in withheld. Nothing passes by implication. proceedings in eminent domain or under some other law authorizing it on the con- I find nothing in the law making such a dition that damages be paid. In this con- franchise as was granted to the plaintiffs nection I would again cite the United in this case an exclusive one pro tonto, Railroads case, atpage 521: "Mere com- and under these circumstances I must petition alone does not ground such a right accept the reasonable interpretation of or claim for damages. Mere competition the language used in the ordinance under alone is not such a tortious taking as to consideration here as not showing any ground such an action." Accordingly, my deliberate purpose to make a surrender sixth finding is that there is no violation of the city's rights, nor as a conferring of the constitutional provision forbidding of such an exclusive right to the plaintiffs the taking of property without due pro- as against the city as would enable them cess, viewing the allegations of the com- to ground this action, even on the theory plaint in their most favorable light. of a covenant or contract by the city not The complaint further alleges that the to compete. Ford Motor Company is giving the city a I would point also to the appellate free vehicle and technical services in ex- Lax 74

change for the city permitting the Ford law through the operation of dial-a-ride Motor Company to do certain things. The as proposed by defendants. decision as to the adequacy of considera- Chapter 85, Section 7.161, of the city's tion is in the first instance one for the taxicab ordinance reads: "No person duly elected representatives of the city to shall operate any taxicab in the city of determine. Their decision and the terms Ann Arbor without first having obtained a of the contract in this case do not appear certificate of public convenience and ne- to the Court to be so inadequate as to be cessity from the board authorizing such evidence of fraud or to shock the con- operation." The language of the ordi- science of the Court. My seventh finding nance precludes its application to defen- is that I do not find the consideration in- dant city [United Railroads of San Fran- adequate nor any evidence whatsoever of cisco, 249 U. S. 517; 39 Supreme Court fraud from the face of the complaint. 361 (63 L.Ed. 739, 1919)]. For the same reasons that I have The basic premise from which plain- hereinbefore stated, my eighth finding is tiffs advance their due process and equal that I find no denial of equal protection to protection arguments is rights they as- the plaintiffs. Ninth, I do not find that sume they have as licensees. We find the actions of the city constitute an un- that basic premise to be false. Def en- reasonable, arbitrary, or capricious ex- dant city has reasonable control of its ercise of police power. streets (Const. 1963, Article 7, Section For the reasons stated, the motion for 29). Plaintiffs have no right to use the summary judgment is granted. Court is streets without the consent of the city adjourned. [Melconian v. City of Grand Rapids, 318 Mich. 397 (1922)]. The licenses plaintiffs rely on are nothing more than a OPINION OF THE COURT OF privilege to do what is prohibited without APPEALS such licenses [C. F. Smith Co. v. Fitz- gerald, 270 Mich. 659 (1935)]. Plaintiffs appeal from the trial court's In establishing and operating dial-a- grant of summary judgment [GCR 1963, ride, defendant city is doing what the 117.2(1)] in favor of defendants. We af- mass transportation authorities act, firm. supra, authorizes. Plaintiffs are licensed by defendant Affirmed but without costs. city under its ordinance to operate taxicabs in the city. Under authority of the mass transportation authorities act [MCLA 124.351 et seq.; MSA 5.3475(1) et Quinn, Brennan, and Targonski, JJ. seq.], defendant city has instituted and operates an experimental transportation system known as "dial-a-ride." Plain- tiffs' action sought to restrain defendants, individually or collectively, from establishing and operating dial-a-ride. On appeal, plaintiffs contend that dial- a-ride is subject to the city's taxicab ordinance; that plaintiffs are denied due process of law and equal protection of the DEMAND-RESPONSIVE TRANSPORTATION AS SEEN BY THE TRANSIT WORKER John M. Elliott President, Amalgamated Transit Union

Few challenges are more important to only competes with public transportation transit workers and their union repre- but causes the traffic jams that stall our sentatives than to find ways and means of public transit vehicles. Meanwhile, fed- revitalizing public transportation in eral transit-aid funds have been kept to a urban areas. A strong public transit small fraction of the federal highway-aid system is essential to the economic and funds, and even those transit funds appro- social health of cities. In addition, it priated have been misused for capital im- represents to the transit worker his only provements that offer little or no near- chance for a secure job, earnings ade- term benefit to the riders of bare-bones quate to provide a decent standard of transit systems. living, and the protection of a reasonable On a number of public occasions, the pension when his working years are over. Amalgamated Transit Union has gone on Accordingly, for some years the Amal- record in support of a dramatic restruc- gamated Transit Union has eagerly turing of our industry, based on better searched for a remedy or remedies service to the public and equitable cost giving promise of rejuvenating public sharing by all those who benefit from transportation as an economically viable transit. We have urged that public trans- institution. portation be operated on a completely We know that the real cause of the fare-free basis, with the costs prepaid transit industry's ever-worsening eco- primarily by the local taxpayer. This nomic position has not truly been the new form of universal public transporta- skyrocketing of labor costs in an infla- tion that is supplied by and for the entire tionary era, although certainly those community served by the system at abso- costs must be expected to climb more lutely no user charge to the passenger is, rapidly in a labor-intensive industry such we believe, the single best hope of pro- as transit. The real economic difficulty, viding every urban citizen an efficient, however, lies elsewhere: in the declining convenient, and attractive alternative to productivity of a fixed-route transit sys- the private automobile. tem that carries an ever-decreasing On the other hand, we are firm in our number of passengers for every mile or conviction that revitalization of our indus- hour of service operated as fares in- try must also include improved service crease and service deteriorates. through innovations such as express bus We have been frustrated, especially lanes and, perhaps even more important, in the last several years, by what we demand-responsive doorstep service that consider the failure of government and will make public transportation available transit-industry management to respond to everyone in the entire community dynamically and effectively to the chal- served by the transit system. lenge presented by this problem of de- Thus, the ATU has looked with favor clining productivity. For many years, on dial-a-bus as an attractive improve- millions of dollars of taxpayers' money ment, offering jobs and economic prog- have been spent in building freeways and ress to our membership and increased in providing downtown parking facilities, ridership and productivity to the transit both of which encourage urban sprawl, system. For several years in public dispersed trip origins and destinations, statements, we have urged that dial-a-bus and more and more reliance on the auto- be given a much higher priority in the mobile in direct competition with our federal transit-aid programs, which have industry. The automobile, in turn, not focused, we feel, far too much on Elliott 76

capital-intensive remedies, such as proper operating project to determine highly automated-rail and fixed-guideway whether they would improve public transit systems designed primarily for service service. to and from the downtown areas. We Two years later the first only federal suggested then that the dem and- activated dial-a-ride demonstration project—a concept of dali-a-bus holds more prom- manually dispatched service with a fleet ise of attaining a total system of reliable of 12 vehicles—finally was launched at low-cost public transportation to the Haddonfield, New Jersey. This service entire community. We feel that the dial- is provided by drivers and maintenance a-bus concept offers to the transit indus- personnel belonging to our union and em- try a real opportunity to open new mar- ployed by Transport of New Jersey. The kets in the lower density areas and demonstration service was held up by an wherever trip origins and destinations unrelated labor dispute and resumed fol- are too widely dispersed to permit ser- lowing a long and difficult strike for a vice by conventional line-haul transit. new working agreement. The operating Frankly, in the years since 1968 when results available to date, which tend to the new-systems studies of the U.S. De- show considerable ridershipin the off-peak partment of Housing and Urban Develop- and weekend hours, must be regarded as ment recommended dial-a-bus for preliminary and probably affected at special study and demonstration because least to some degree by the labor dispute. of its near-term potential and limited Under the best of circumstances, as we development costs, practically nothing have pointed out to the department, there has been accomplished to advance this can be no possible relevancy of the man- concept in an operational setting. In ual test to the ultimate success or failure June 1970, the Urban Mass Transporta- of a computer operation. tion Administration wrote to us stating We predict that if there is no change that dial-a-bus would be given "a proper in the present attitudes of the Congress demonstration" because it was regarded and the executive branch of the federal as "one of the few near-term new system government, this still promising new- solutions for public transportation prob- system concept called dial-a-bus will be lems, particularly in lower density resi- sunk without trace, in apparent deference dential ,areas." Unfortunately, IJMTA to those who believe in mass expenditures decided to demonstrate this demand- for new capital equipment, automated responsive service by using manual dis- rapid transit systems, people movers, patch of vehicles, although only a reli- and the like. The proponents of the so-' able computer-dispatch capability, with called "capital-intensive" approach would its memory bank, could guarantee opti- have us believe that, because as much as mal performance of the dial-a-bus sys- 80 percent of all transit operating costs tem in terms of speed, reliability, con- at present are labor costs, the only way venience, and cost. In August 1970, we to solve the industry's economic problem wrote an open letter to the U.S. Depart- is to eliminate labor. As recently as ment of Transportation inquiring why it April 12, 1972, UMTA stated to the House should take so long and be so difficult to Appropriations Committee that any addi- make use of dial-a-bus techniques, sys- tional investigation into dial-a-ride tech- tems', and equipment that had already nology in fiscal 1973 been developed and laboratory tested and needed only to be demonstrated in a ...depends on what we learn from Haddonfield and Elliott 77

from approximately 10 other non-federally supported We have had such limited experience projects similar to Haddonfield that are in operation. with dial-a-bus in an operational setting If we conclude that dial-a-ride's economic character- that it is difficult as yet to appraise the istics are such that virtually no communities in the nature and extent of its impact on col- Nation are willing to support it, then there will be lective bargaining in the industry or on no additional technological development. If, how- the needs and desires of our membership. ever, we conclude that there are a substantial number We believe, however, that demand- of communities willing to support dial-a-ride, then we responsive services should improve the plan to (1) test our first generation computerized con- convenience, reliability, and speed of trol system in Haddonfield, using the existing manual transit and thus generate a greatly in- control system as backup; (2) commence extension creased patronage base. The increased of the first generation computer system into a second ridership, inspired by this more respon- generation system, one that does not rely upon man- sive service structure, should enable the ual control for backup; and (3) search for a new site transit system to function more produc- to conduct a second dial-a-ride demonstration. tively, measured in terms of the number It seems to us that UMTA proposes to of passengers carried per vehicle-hour reject computerized dial-a-bus from any and mile operated, and thereby to reduce future federal funding on the basis of a the overall cost per ride. manually dispatched operation and, even Moreover, there is every reason to worse, because the experiment proved believe that dial-a-bus transit might, for "uneconomic" in terms of its inability to the first time, enable the industry to tap support itself from the fare box. substantial ridership from the off-peak Such outmoded and inequitable con- market, which, typically, has far lower cepts of fare-box financing have long demand densities and dispersed origins proved unworkable as applied to conven- and destinations. Coventional line- haul tional public transit. Secretary of transit serves 5 to 10 times more people Transportation Volpe, himself, expressly during the peak hours than during the rejected them, stating that the fare box average midday period. Penetration of should not be expected to cover all the the off-peak market should offer the in- costs of providing essential transporta- dustry substantial labor and other cost tion services. Why should dial-a-bus be economies and provide increased rev- differently treated? To the extent that enues as well. Dial-a-bus should help dial-a-bus simply provides new service stabilize the number of jobs in the indus- in currently unserved areas, replaces try, reduce the need for split-shift fixed-route service, or functions as a schedules, and otherwise provide a means collector -feeder system between line - of achieving higher labor productivity haul services and lower density residen- without eliminating jobs. tial areas, it seems to us no different We fully expect that a public transpor- from traditional forms of public trans- tation system, using a proper mix of portation that have, in many cases, re- dem and- responsive and fixed-route tech- ceived the financial support of the com- niques, can succeed in replacing the pri- munity at large. Of course, to the vate automobile as the preferred means extent that dial-a-bus is used to provide of transportation for many urban trips. a true premium or luxury type of service Such a user- and demand-oriented system, on a convenience basis, the individual because of its increased patronage and user may reasonably be expected to pay productivity, should be far more econom- his full way without support by public ically viable than conventional route- funds. oriented transit. Whether or not it can Elliott 78

fully pay its way, such a system will bet- worked out on a consensual basis by the ter serve the community and, in our view local management and union bargaining is, therefore, more deserving of tax committee in terms of the services to be support. In any event, we are convinced provided and the needs of the parties. At that demand-responsive transit will pro- this early point in our experience with vide the transit worker with better job dial-a-bus, we would urge that demand- security and the potential for greater responsive operations be integrated into earnings. Demand-responsive transit is, the regular service with only such mini- therefore, an attractive opportunity to mum revision of normal compensation, the worker, can help stem the industry's seniority, and working conditions as is economic decline, and can, at the same clearly essential and agreeable to both time, provide new job opportunities, bet- parties. As in any collective bargaining ter wages, and more adequate pensions, situation, we would expect management to health and welfare, and other benefits propose to do this work under terms most and conditions of employment. In other favorable to itself, while the worker, as words, we see no reason why demand- usual, will be more impartial and sacri- responsive techniques should present any fice at least some portion of his interests special collective bargaining problems to the greater benefit of the community! for our members or for the industry. Presumably it is on that basis, rather Under no circumstances, however, than self-interest, that has led our mem- should the city transit worker who pro- bership in Haddonfield and Rochester to vides dial-a-bus service, as distin- agree to certain restrictions on their guished from regular line-haul service, normal picking rights, based on seniority, be asked to accept lower wages or more as they apply to dial-a-bus assignments restrictive working conditions in order and the right to bump into and out of this that dial-a-bus can be made to pay its special work. way or that lower fares can be charged. Perhaps we should close this presenta- We have always taken the position that it tion by stating that transit labor can only is not an answer to the industry's eco- be counted on to look with favor on nomic problems to reduce wages and demand-responsive transit as long as the labor costs to the lowest possible level policies and programs for its implemen- consistent with the need for an adequate tation are sound and as long as adequate supply of manpower. The suggestion levels of employee protection are provided that demand-responsive services be pro- to those who may be adversely affected by vided at substandard wages and working such innovation. Under such conditions, conditions, at least until they prove suc- the worker's response to the introduction cessful, is no less acceptable than any of dial-a-bus systems should, in general, other request that the worker subsidize be friendly. conventional transit operations whose As a final caveat, we might add that in true costs neither the employer nor the our judgment the transit industry has been community as a whole is prepared to pay. very slow to revise its fixed-route struc- As we see it, any special labor impli- tures and to take advantage of demand- cations of demand-responsive service, responsive concepts. A continuing failure which may require adjustments in wages, in this regard may well lead to the unnec- hours, and working conditions, are prop- cessary introduction of competitive sys- erly left to the local collective bargaining tems pledged to more dedicated service process. These can and should be to the economic and social life of the Elliott 79

community. We feel strongly that this the management has not given us the time should not happen, and we urge transit to give effective service. management to make a greater commitment to the earliest possible introduction Question: A problem that we have of these demand-responsive services in encountered during the Haddonfield dem- the interests of better public transporta- onstration is that every time drivers tion to the community as a whole. choose their shifts we wind up with many new drivers who require training. Can we not get a permanent set of drivers INFORMAL DISCUSSION assigned to the service? What is the position of the Amalgamated Union? Question: During the course of your Answer: Well, as long as you envision formal comments you made a reference dial-a-bus as a short-term solution, I to the split shift. Will you elaborate on think that will continue to be a problem. that? If I understood correctly, you I think that, if there is any value to dial- indicated that there might have been a a-bus, it is not only that it has the impact move to eliminate the split shift in work- of being an immediate, short-term value ing conditions, and it seemed to me that but that it will be there in the long haul. you concurred in that. Is that correct? On that basis then, it serves your pur- Answer: You put the emphasis in the poses as operators of a transit system to wrong place. We saw that this was an train as many people as you possibly can— opportunity that would bring about the just as you have done and just as the elimination of the split shift. By using industry did when it trained streetcar demand-responsive techniques, we could men to be bus drivers or gasoline me- fill in the gap where we now have unpro- chanics to be diesel mechanics or bus ductive and unpaid manpower for periods drivers to be charter or sightseeing of 2 to 5 hours in midday. I meant that guides. We do this because we want as we could fill in that gap with the demand- many people in the labor pool as possible. responsive needs. The lesson that you should not rely on too few people to staff an operation has been Question: Do you foresee conflicts demonstrated during this conference. between labor and management on dial- Two people are now absent because they a-bus? were the key people in their local situa- Answer: 'I see no conflict. The con- tions. You do not want to have to depend flict—if there is one—usually appears on a small group of people in your dial-a- because of the demands of management bus system. and the demands of the public to get on with the job. The bus driver very much Question: But has anything been re- prefers to give courteous, safe, and solved on this issue at the national level? complete service but finds it difficult in Answer: I tried to maybe go around face of constant demands such as "Come the back door to tell you that this prob- on, make this traffic light" or "You're ably is not so big a problem as you think. going to be half minute late, and a half But if you think it is, then it has to be minute late means you're going to have resolved at the local bargaining table. 25 percent of the people to carry, and This union is formed of fairly autonomous that will delay everybody." It is not that local unions. We give leadership, but not we do not desire to give the service but direction. FORD MOTOR COMPANY'S ROLE IN DIAL-A-RIDE DEVELOPMENT: 1972 AND BEYOND

Karl W. Guenther Dial-a-Ride Program Manager, Ford Motor Company

Ford Motor Company recently an- early 1960's. We can now confidently nounced its entry into the transportation predict 20 operating systems in North systems business. The first product of America by the first quarter of 1973 this new venture is ACT, a driverless, (Fig. 1). computer- cont rolled vehicle that oper- Our own research and experience with ates on lightweight low-cost aerial guide- operating systems (Table 1) has led to ways. ACT is designed to transport peo- the following conclusions: ple and goods within congested activity centers, to connect these activity cen- The operation of dial-a-ride pub- ters, and to provide access from low- lic transportation service is totally fea- density suburban areas to central loca- sible from a technical point of view; tions. The patronage of the public will be We realize that one type of system at a substantially higher level for this new cannot provide for all transportation kind of service than for the conventional needs, however. In his announcement of fixed-route, fixed-schedule service; and the company's entry into the transporta- There is a significant diversion of tion systems business, Henry Ford, II, travel to dial-a-ride from the private stated: "Ford is constantly looking for automobile. new ways to improve personal mobility and to transport goods. Improved mobil- The credibility for these 3 statements ity and better urban planning in the cities is found in the detailed descriptions of of today and tomorrow will require not operating systems presented in other one new system but a whole new family of papers in this report. It is now generally flexible, convenient, and efficient trans- accepted that dial-a-ride is a technical portation systems. We are working on success and that market response is fa- other new systems that will be comple- vorable. However, these conclusions mentary to ACT." One such comple- fall to address the all important cost- mentary system is dial-a-ride. revenue questions. Is Dial-A-Ride an Dial-a-ride provides transportation in economic success? smaller communities and low-density suburban areas; in contrast, ACT serves high-density areas. Designed to serve relatively low-volume, diffused trip- making patterns, dial-a-ride offers the Figure 1. Existing and projected number traveler doorstep pickup in response to a of dial-a-ride installations. telephone request. He is taken where he wants to go, when he wants to go. Dial-a-ride has been carried from the concept stage into practical applications. Several operating dial-a-ride systems have demonstrated conclusively that there is a better way of moving people in these low-density situations than with conven- tionally operated buses. Ford's dial-a-ride program dates back to 1969; however, research on demand- responsive systems dates back to the 1995 1970 1975 Guenther 81

Table 1. Dial-a-ride systems in operation.

Service Trips Trips Area Service Riders per per Trip Avg Funding Square Area per Vehicle- Labor- Cost Fare Location Miles Population Weekday Hour Hour5 ($) ($) Agency Percent

Ann Arbor 1.36 10,000 214 7.7 5.4 1.35 0.50 Local 37 State 63 Bay Ridges 1.34 13,700 463 11.3 7.6 0.60 0.25 Province 100 Batavia 4.75 17,300 455 6.7 6.6 0.61 0.50 Local 100 Columbia 6.00 17,300 54 4.4 3.0 N.A. 0.50 Local 100 Columbus 2.50 55,000 485 9.5 6.3 N. A. 0.20 Federal' 100 Haddonfield 6.50 16,000 333 2.6 1.8 N. A. 0.50 SLate 20 Federal' 80 Regina 2.75 18,000 1,200 15.0 11.5 0.54 0.32 Local 100

Note: Data given represent a snapshot of 7 operating systems for the winter of 1972 and cannot be interpreted as an accurate representatton of any system today. Most of the systems represented have since changed, ridership has increased, service has been expanded, and in some cases productivity has improved 'Winter 1972. 5 tocludes dispatcher. 'Department of Housing and Urban Development. doepartment of Transportatton.

To answer, we must first look at sys- predictions made in the M. I. T. research tem productivity, which is the key to work of 19 68-69. This relation tells us dial-a-ride economics. Figure 2 shows very simply that dial-a-ride is more ef- how actual productivity in the field, as ficient if the distance traveled between experienced in the winter operating sea- stops is decreased, that is, if more peo- son of 1972, relates to demand density. ple request service in a given geQgraphic Because dial-a-ride costs (largely labor) area. Therefore, to operate at ma.xi- are most accurately accounted on an mum efficiency, dial-a-ride must gener- hourly basis, the number of passenger- ate a relatively high demand-more than trips served per vehicle-hour is a direct 10 trips/square mile/hour. Only the and very meaningful method of comparing 2 Canadian systems have regularly systems. In this case, the actual expe- achieved this productivity, although Bata- rience coincides very closely with the via, Ann Arbor, and Columbus have all done so on occasion. The problem here is not a technical but rather a marketing issue: how to, Figure 2. Dial-a-ride operating experience. generate more demand from a given

20 area. Most existing dial-a-ride systems do have the capacity to produce produc- tivities in excess of the averages shown. S We thus observe that, although almost >x every operating dial-a-ride system has S demonstrated a greater ability to attract

S passengers than conventional bus ser- . S vice, in some cases the demand has not

S been high enough to produce really effi-

S cient operation. The second element of economic via- I I I I I I I 0 10 20 3t bility of dial-a-ride is actual costs, DEMAND DENSfTY (TRIPS/Ml.'/fIR.) which are made up primarily (60 to 75 Guenther 82

percent) of driver and dispatcher labor. should remain a local issue for each The actual cost per ride is obtained by community to address in its own way. dividing total hourly system cost by total We also must reflect on fare struc- hourly productivity. Because hourly tures (it may be desirable to price dial- labor rates vary so greatly among com- a-ride service to more closely cover munities, no generalizations on cost are costs) but this again is a local decision. possible. We do find differences among Our principal findings from pursuing systems in their use of dispatching labor, dial-a-ride research, then, are listed however. below. So that dispatching and overhead costs can be kept in line, low-cost manual dis- The concept is becoming readily patching techniques have been developed. acceptable to transit operators, and more Some manual dispatching systems for small-scale systems go on line every small projects are "overdesigned" at the day. The "missionary work" has been expense of operating efficiency. There accomplished; dial-a-ride has credibility is no need for 2 dispatchers when 1 will and is becoming an important part of the do. Elaborate visual dispatching aids transportation scene. (colored maps, lights, and markers) are The operation of door-to-door, impressive to visitors, but they can ac- dynamically dispatched public transpor- tually get in the dispatcher's way. The tation is totally feasible from a technical responsibilities of the dispatcher and his point of view. interactions with the driver deserve The public is responding favorably careful study and, when properly defined, to this type of service, from a standpoint can make the dispatching operation much of patronage, willingness to pay (directly more efficient. This is one area where a and indirectly), and diversion to public qualified professional with experience transport from automobiles. can be of great value in designing a new At the present scale of small sys- system. tems and at present fare levels, dial-a- Data given in Table 1 do not show that ride cannot be operated on a break-even even one small-scale dial-a-ride opera- basis, i. e., fares cover all costs. This, tion is covering all of its costs. (Batavia however, has not discouraged many com- is coming close and with added revenue munities from planning expanded sys- sources does hope to break even.) Dial- tems. a-ride is a very important service improvement and is highly attractive to the As a correlative area of research, we public, but it is not a short-term cost at Ford have used dial-a-ride as a test cutter at the present scale of operations. case for learning about the process of The actual level of subsidy required in a innovation in public transport. We have given community depends on local labor found in this research that rate, fare structure, and demand for the service. An important observation here Inertia and resistance to innova- is that many communities have elected to tion are greater than we originally envi- provide dial-a-ride service at public ex- sioned; penses, and many are considering expan- A sponsor, an operator, and tech- sion of services on the basis of results nical support are required to implement obtained so far. The matter of public a system successfully; funds for support of public transport One can identify those character- Guenther 83

istics of a community that produce a fa- flexibility to use dual-mode vehicles, in- vorable climate for innovation; cluding dial-a-ride buses, that are oper- Design and implementation of dial- ated on conventional surface streets and a-ride system can be a "do-it-yourself" highways but are specially equipped to project, providing proper tools are avail- enter the guideway where they are auto- able (such as Ford's computerized sys- matically controlled, register a destina- tem design models) and proper attention tion, merge into the main guideway, and is given to all elements of the design; and then leave the guideway to operate on Each dial-a-ride system must be surface streets. Intuitively, the dual- custom-designed to meet specific local mode concept is very appealing. One political and social needs. can imagine, for example, a future system where no transfer is necessary We have identified and are pursuing either to get on a line-haul mode or to several promising areas for ongoing re- get from the transit station to the final search. destination. Whether dual-mode systems can really work better than properly in- Economies of scale in larger sys- terfaced single-mode systems is a ques- tems. We can conclude that spreading tion that our future research will attempt dispatching costs and overhead over more to answer. vehicles will produce some benefits, but, Although it is the largest single more important, will larger systems in- component of total freight costs, urban duce higher demand densities? goods movement has received very little Automated dispatching. Our own study. There may be substantial spin- studies to date have indicated possible off from some of our dial-a-ride re- application in the automation of informa- search into the field of local truck dis- tion flow as opposed to actual decision- tribution systems, and we expect to making. launch a substantial research effort in Digital communications. A pilot this field. test is scheduled for the fall of 1972. Dual-mode dial-a-ride (Fig. 3). In An obvious area of Ford concern has this system, the guideway and controls been vehicle development. We all ac- for the Ford ACT system provide the knowledge that currently available small buses have design compromises that make them less than perfect for dial-a- Figure 3. Dual-mode vehicle system. ride application. The present low- volume market does not justify our tool- ing to produce a special vehicle, nor does the potential future market appear to warrant extensive research and development expenditures. Therefore, we expect to continue working with specialty manufacturers, basing small buses on standard product lines. We are making a Ford Econoline bus conversion, which is explicitly engineered and manufactured to meet transit system needs, available to the industry in the fall of 1972. This Guenther 84

conversion embodies strength and safety features that have not been available in small buses. Ford will continue to actively promote the dial-a-ride.concePt as part of an overall commitment to better public transportation in urban areas. The present program calls for spreading design- implementation knowledge throughout the transportation community, making a stronger, safer vehicle available, and continuing research in larger, more sophisticated systems.

INFORMAL DISCUSSION

Question: Are there other sources of revenue for dial-a-ride besides the fare box? Answer: Yes. Some of these sources are being developed in Batavia.

Question: How important is public attitude in general in convincing politicians to spend public money for a dial-a-ride system? Answer: Sociological researchers tell us that, if a community is oriented in a public-spirited way toward one.thing, it will tend to be oriented that way toward all things. If a community supports day- care centers, it will probably also support public transportation. There is a coalition in Ann Arbor right now between what I call the bicycle freaks and the transit freaks. They have motivated a large bloc of the community to stop additional road improvements in this current budget year. This is a fantastic development. It suggests how important public attitude is and that the people in the community must want something before politicians cause it to happen. If the people are indifferent, forget it. USER PREFERENCES FOR DIAL-A-BUS

Richard L. Gustafson Research Economist, General Motors Research Laboratories Francis P. D. Navin Vice President, R. H. PrattAssociates

This paper discusses the results of a proportion of people earning more than survey of user preferences for the dial- $15,000 in Columbia and about the same a-bus transportation system in Colum- proportion in the lower income brackets. bia, Maryland. The analytical techniques The respondents have a higher educa- used in the Columbia survey are based on tional level. The population is more those in surveys conducted in the cities dense but much smaller; Columbia has a of Warren and Center Line, Michigan (1, population of 10,000 as opposed to 200,000 5, 6). in Warren. The residents of Columbia The present research study sought to have seen a dial-a-bus system in opera- achieve 2 objectives: tion. The dial-a-ride service was truly a To determine user preferences for demand-responsive transportation system dial-a-bus in an area where an actual providing many-to-many service for the system existed, and Columbia residents (1). The fare was To evaluate similarities and dif- $0.25 or 10 tickets for $2.25. To request ferences between the results of the sur- service, the resident called the dispatcher, veys in Warren and in Columbia. who checked the location of the vehicles and then assigned the caller to one of the Two attitudinal surveys conducted in vehicles if the estimated pickup time was different cities can provide further in- agreeable. The dispatcher also took sights into the transportation system calls to reserve service in advance. characteristics that users regard as important. A survey conducted in one city is difficult to generalize to other areas. SURVEY DESIGN Preconceived notions of transportation and prejudices often force the analyst to The methods of paired comparison and reserve judgment of preferences to the semantic scaling were selected as the case study city. The attitudinal surveys measurement devices for the home inter- in Warren and in Columbia afford an op- view. The method of paired comparison portunity to compare the preferences of was used to establish a scale of prefer- 2 different populations in dissimilar en- ences for various system characteristics. vironinents. The semantic-scaling technique was used Warren has primarily blue-collar to evaluate design alternatives for a workers; most residents have only a high number of system characteristics; A school education. The household incomes more detailed discussion of the techniques are concentrated in the $10,000 to and the questionnaire design is given in $15,000 range. The amount of public another report (4). transportation available is limited; and The paired-comparison questionnaire most important, a dial-a-ride system originally had 32 system characteristics. was a completely new concept to these For the Columbia survey the number was people. The system was explained reduced to the following 15: thoroughly during the home interview, but the respondents were forced to rely Arriving at your destination when on their imagination or their own percep- you planned to, tion of public transportation. Making a trip without changing Columbia has a more diversified pop- vehicles, ulation than Warren. There is a greater Spending a shorter time waiting Gustafson and Navin 86

to be picked up, questions assessed the respondent's Paying a lower fare, previous experience with public transpor- Spending less time walking to a tation systems. An additional page was pickup point, added to the Columbia survey asking the Spending a shorter time traveling respondent whether he used dial-a-bus in the vehicle, frequently, occasionally, or not at all. Being able to take a direct route, with fewer turns and detours, Having small variation in travel DATA COLLECTION time from one day to the next, Being assured of getting a seat, A home interview survey was conducted Calling for service without being for 2 populations: (a) all residents of delayed, Columbia and (b) users of dial-a-bus. Having more protection from the The general population survey sample weather at public pickup points, was selected as follows: From the alpha- Being able to select the time betic list of residents for each village, a when you will be picked up, name in the first 10 names was randomly Having a convenient method of selected and then every tenth name there- paying your fare, after was selected. The address was lo- Having freedom to turn, tilt, or cated on a map and assigned to a sample make other adjustments to the seat, and survey area. The users of dial-a-bus Having a greater chance of being were selected from the records of 1 able to arrange ahead of time to meet and week's calls to the dispatcher. One of sit with someone you know. the first 10 names was randomly selected and then each tenth name thereafter. This list was merged with the general The semantic - scaling questionnaire eval- population list, and the names were ran- uated the desirability of design alterna- domly assigned to 1 of 6 interviewers. tives for the 15 characteristics. Those Interviews were conducted during the day characteristics that were common to both and evening. Two call-backs were made Warren and Columbia surveys had ex- before the name was removed from the actly the same wording and accompanying list. All members of the household over illustrations in both surveys. The same the age of 14 were interviewed. The statistical operations were performed on paired-comparison survey yielded 131 both sets of data. respondents, and the semantic-scale sur- In the paired-comparison question- vey had 100 respondents. naire, not all of the paired choices were included in the survey. Three subsets of characteristics were established: levels ANALYSIS OF WARREN AND of service characteristics, convenience COLUMBIA RESULTS factors, and vehicle design characteristics. Lower fare, assurance of a seat, The preferences derived from the and shorter travel time were paired with paired-comparison surveys in Warren and all of the characteristics. This allowed in Columbia are shown in Figure 1. Only the development of a general scale from the 15 system characteristics common to the relation of the 15 characteristics to both surveys are indicated on the scale. the selected 3. The results from the 2 surveys are quite In the original surveys in Warren, no similar in that dependability is most Gustafson and Navin 87

Figure 1. Responses to paired-comparison for system information but was seldom questionnaires in Warren and Columbia. used. Therefore, many potential users were unable to contact the dispatcher. WABMEN (786 RESPONDENTS) - JMOIA >131 RESPONDENTS> Consequently, calling without delay had a ARRIVING WHEN PLASNED lOG WHEN PLABNED higher preference ranking from the Co- NU B/A DELAY lumbia residents. The service was well WARING A'SEAT WAIT TIME received, and often vehicles were unable to serve the demands, and wait times were NV TRANSFER TRIP TRAVEL TIME CALLIRG W/O OELAT PICK-UP TIRE as high as 60 minutes. The overload also SHELTERS AT PICK-UP ANSFER TRIP LESS WAIT TIME caused increases in the travel time. CHOOSE PICK-UP TIME EMS AT PICK-UP LOWER FARE U A SEAT LESS WALK TO PICK-UP. DIRECT HAUTE SHORT TRAVEL TIME Table 1. Means of semantic scales for MERE OIRECT ROUTE DAULE TRAVEL TIME S WALK TO PICK-UP Warren and Columbia respondents. FARE EASY FARE PAYMENT Warren Columbia Characteristic (813) (100) DEFENDABLE TRAVEL TIMES Importance of fare 5.7 4.8 EASY FARE PAYMENT importance of travel time 5.5 5.2 ADJUSTABLE SEATS Assurance of a seat 5.2 4.7 ADJUSTABLE SEATS Waiting time at pickup 5.9 5.7 ABILITY TO MEET FRIEND ON TEA.. ABILITY TO MEET FRIENDS ON VEH. Pickup location Place of call 6.1 6.4 Nearest corner 5.5 5.1 Neighborhood 4.9 4.2 Nearest major street 4.1 4.3 Facilities at pickup location important to both populations. Charac - None 4.0 4.2 Curbside phone 4.6 5.0 teristics relating to time and cost are in Enclosed shelter 5.5 5.7 a cluster below the most preferred char- Overhead shelter 5.4 5.7 Waiting time, mm acteristics. Then, well below these 5 6.1 6.4 characteristics are those concerned 10 5.8 6.1 15 4:9 5.0 with convenience and vehicle design. 20 3.8 3.9 The Warren respondents have only a Early arrival, mm traditional transit system as a frame of 5 . 6.1 6.4 10 6.1 6.2 reference, and that is probably reflected 20 4.2 4.0 by the high preferences for having a seat 30 2.7 2.6 Interior design and a no-transfer trip. The Columbia Standard 5.1 5.6 residents did not rank those characteris- Grouped seats 3.0 . 3.8 Deluxe 5.2 5.4 tics so high because dial-a-bus does not One-way fare, dollars have transfers or very many standing 0.40 . 5.7 4.1 passengers. 0.50 5.7 3.3 0.60 4.5 2.2 The Columbia residents have indicated 080 2.9 1.6 through their preferences some of the 0.90 3.0 1.4 1.00 1.7 1.4 shortcomings of dial-a-bus, particularly Fare collection those experienced when the service was 20-trip ticket 4.2 4.3 Credit card 3.3 2.9 initiated. The dispatcher was averaging Monthly pass 3.7 4.0 2 minutes on the phone per request be- Tokens 4.1 4.3 cause he had to supply system informa- Exact fare only 4.2 4.6 Cash with change made 5.3 5.9 tion. A separate number was provided Gustafson and Navin 88

This problem is indicated by the prefer- Figure 2. Waiting-time sensitivity of Warren and ences for shorter travel time and depend- Columbia respondents.

able travel time. WAITING TIM.! SENSITIVITY- SEMANTIC SCALE The fare for dial-a-bus was $0.25, and the fare for the fixed-route system in Warren was $0.35. The Warren residents would most likely believe that any new system would cost even more. Be- cause of the existing fare levels, one would hypothesize that the relative preference for lower fare would be lower in Columbia than in Warren, which is the case. Table 1 gives the means of semantic scales for both the Warren and the Co- UN DE S t B A B L C lumbia surveys. The means of the desirability of system alternatives closely paralled each other. The fare importance is lower, which is consistent with the Figure 3. Early-arrival sensitivity of Warren and paired- c omparis on results. The desira- Columbia respondents.

bility of pickup at the place of call is EARLY UBRIVM SENSITIVITY- SEMANTIC SCLC higher for Columbia, and the desirability DESIRABLE 7 decreases more rapidly as the pickup

gets farther away from the respondent's ---I- place of call. Figure 2 shows the respondent's sensitivity to changes in the amount of wait- 4 -s' ----- ing time. The curves for both Warren -- and Columbia are horizontal up to a 10- SNR7V INT - minute wait, at which point the user's satisfaction diminishes more rapidly as the waiting time increases. The early- UNDESIRABLE arrival sensitivity demonstrates the same 10-minute threshold (Fig. 3). Not EARLY ARRIVAL IN MINUTES only is the threshold similar for both waiting for the bus and arriving early (another form of waiting), but the rate of change of acceptability up to 20 minutes attributed to the existing public transpor- is approximately the same. tation: - fares. A $0.40 fare is consistent The shape of the fare-sensitivity with existing public transit fares in War- curves shown in Figure 4 is similar for ren, but it represents a 60 percent in- both populations. Also the "knee" of the crease in Columbia. curve occurs at $0.65 and $0.75, and this may represent an upper limit of ANALYSIS OF SUBGROUPS IN fares for those potential patrons with a COLUMBIA choice of transportation modes. The different levels of satisfaction can be The paired-comparison results strat- Gustafson and Navin 89

Figure 4. Fare sensitivity of Warren and characteristics throughout the scale, and Columbia respondents. those seldom using the system have a

TAUT SENSITIVITY - SEMANTIC SCALE tendency to group the characteristics. DESIRABLE 7 - - The nonusers indicate high preferences for system characteristics that have been a problem with the dial-a-bus service. Calling without delay and less wait time are just below arriving when planned - on the preference scale for nonusers. Calling without delay is rated the same by 3 -.___ - Warren USD1 both occasional users and nonusers, but less waiting time is rated lower by the nonusers. Less waiting time is rated the CAunA1IJUSI same by frequent and occasional users, UNDESIRABLE 40 .50 .60 .70 .00 .90 1.00 and calling without delay is rated higher FARE IN DOLLARS by the frequent user. Sensitivity to waiting time distinguishes nonusers from users, and a preference for the calling- without-delay characteristic distinguishes Figure 5. Responses to paired-comparison the frequent user from the other two user questionnaires in Columbia by user group. groups. Two other characteristics de- PAIRED COMSIJLTS STRATIFIED BY MBIA) crease in preference from nonuser to

BOIl USER (47) OCCASIONAl. USER (51) FRERAKIYT USER (26) frequent user: no-transfer trip and lower fare. ARRIVING MIEN PLANNED ARRIVING USER PIJARIED ARRIVINGWREN6/0 PUJW(ED DELAY CALLING RIO DELAY CALLING RIO DELAY Users of dial-a-bus are more willing LESS WAIT TIleD - CALLING to accept the inconveniences that accom- LESS WAIT TIRE NO TRANSFER TRIP LESS WAIT TileD pany a public transportation system. CRUSOE PICK-OP TIRE- - CRUSOE PICK-IS TIRE Most of the characteristics receive a 54261 TRAVEL TIRE OROOT TRAVEL TIRE SAELTERE AT PICK-UP 00 TRANSFER TRIP lower preference rating by the users than HAYING A SENT SAGRT TRAVEL TIRE. SITELTEID AT PICK-EM CRUSOE PICK-UP TIRE RARING A SCAT by the nonusers except for those charac- ROOK DIRECT ROUTE teristics related to the advantages of dial- LESS WALK TO PICK-UP OBELIERE AT P1CC-UP RORE DIRECT ROUTE DEPENDABLE TRAVEL TIRE ROOK DIRECT ROUTE NO TRANSFER TRIP a-bus over a conventional bus system. LOWER FARE DEPENDABLE TRAVEL TIRE WAVING A SENT DEPENDABLE TRAVEL TIRE To the frequent user, choosing pickup LOWER FARE LESS WALK 10 PICK-UP time and arriving when planned are im- - LESS WALK TO PICK-UP portant characteristics that are currently LOWER FARE ABILITY TO being satisfied by dial-a-bus. EASY FARE PAYSROYT REEl FOIERUS ON VON. ABILITY TO IEET P01(1(05 OIl VON Table 2 gives the means of semantic ABILITY TO EASY FARE PAYRENT scales by frequent users, occasional REET FRIENDS GO YES. EASY FARE PAYRENT ADJUSTABLE SEATS. ADJUSTABLE SEATS ADJUSTABLE OEATS_ users, and nonusers in Columbia. The importance of fare is consistent with the paired-comparison results. The "knee" ified by nonusers, occasional users, and of the curve for the seldom and occasional frequent users of dial-a-bus are shown in user is approximately $0.65. The fre- Figure 5. The groups indicate similar quent user indicates a high mean desira- ordering of preferences for most of the bility for the 20-trip ticket, which is al- characteristics. The frequent users ready being used and is apparently indicate a more even distribution of popular with the, frequent user. The Gustafson and Navin 90

Table 2. Means of semantic scales for Columbia effect is 0.55, which indicates the means respondents by user group. of the groups are not significantly different. The interaction of the fare level and Occa- the groups (slope of the curves) has an F- Non- sional Frequent user User User value of 1.33, which is not significant at Characteristic (29) (47) (23) the 5 percent level of confidence. From Importance of fare 4.6 5.2 4.0 the data available, one is unable to reject Importance of travel time 5.1 5.0 5.5 the null hypothesis that there is no differ- Assurance.oI a seat 4.4 4.8 4.6 Waiting time at pickup 5.8 5.6 5.9 ence in the preferences of the 3 groups. Pickup location The semantic-scale values on waiting Place of call 6.3 6.5 6.4 times are shown in Figure 7. Satisfaction Nearest corner 5.8 5.6 5.1 Neighborhood 4.9 3.8 4.0 Nearest major street 5.1 4.2 4.0 Facilities at pickup location None 4.1 4.5 4.0 Figure 6. Fare sensitivity of Columbia Curbside phone 4.7 5.2 5.1 respondents by user group. Enclosed shelter 5.5 5.8 6.0

Overhead shelter 6.0 6.0 5.3 FIRE SENSITIVITY - SEMANTIC SCSI ICOLLWBIA1 Waiting time, mm DESIRABLE 7_ 5 6.3 6.8 6.4 10 6.3 6.2 6.3 15 4.8 5.1 5.7 20 3.5 3.8 4.9 w Early arrival, mm -J ..-..s0 0 5 6.3 6.5 6.4 10 6.1 6.4 6.0 20 4.2 4.1 3.8 C.) 30 2.5 2.9 2.1 zB- Interior design EN Standard 5.6 5.5 6.1 El) Grouped seats 3.8 3.9 3.8 Deluxe 5.7 5.6 5.0 One-way fare, dollars 0.40 4.0 4.3 3.8 UNDESIRABLE !!' 0.50 3.3 3.5 3.3 0.60 1.8 2.3 2.3 FARE IN DOLLARS 0.80 1.5 1.6 1.9 0.90 1.3 1.3 1.6 1.00 1.3 1.3 1 1.6 Fare collection 20-trip ticket 4.1 4.3 5.1 Figure 7. Waiting-time sensitivity of Columbia Credit card 3.3 2.6 3.2 respondents by user group. Monthly pass 3.9 4.0 4.4 Tokens 4.0 4.7 4.4 WAITING TIME - SEMANTIC SCAt! CAtIMBIVI Exact fare Only 4.6 4.7 4.6 flENIRADI U 7 - Cash with change made 5.6 6.3 5.9

EN -J sensitivity of the respondents to various C-) fare levels (Fig. 6) is consistent with the I- z preferences from the paired-comparison w • iu questionnaire. A 2-factor mixed-design C', analysis of variance was performed on EMEMMIMME the semantic-scale data to determine whether the 3 groups are significantly UNDESIRABLE IMMMMMIMMMM 'U different (3). The F-value for the group WAITING TIME (MINUTES)

Gustafson and Navin 91

diminishes more rapidly for the nonusers scales for the households with 1 or 0 than for the occasional and the frequent automobiles available and the households users. The 2-factor mixed-design anal- with 2 or more automobiles available are ysis of variance was performed on the 3 shown in Figure 8. The respondents with groups to determine whether differences 2 or more automobiles available have are significant. The F-value for the higher preferences for certain advantages group differences is 0.39 (not significant provided by the automobile: arriving at the 5 percent confidence level), which when planned, no-transfer trip, and short indicates that the mean values over all travel time. Those respondents also in- the waiting times are not significantly dicate a higher preference for calling the different for the 3 groups. The inter- system without delay and, understand- action of the waiting time satisfaction and ably, are more sensitive to the incon- the groups (slope of the curve) has an F- veniences of public transportation. value of 2.74, which for 6 and 288 de- The fare sensitivity and the waiting grees of freedom is significant at the 5 time sensitivity are shown in Figi.ires 9 percent level of confidence. The null hy- and 10. Both graphs demonstrate results pothesis that there are no differences in similar to those shown in Figure 1. The the interaction of the user groups and households with more automobiles avail- waiting time satisfactions can be rejected. able can substitute the automobile for the The nonusers are more sensitive to public transit alternative more easily. changes in waiting times than the users. If the system causes inconvenience (trans- The paired-comparison preference fers, long waits, or long travel time), the 2-automobile household is less apt to be a continuous user of the system. The 1- or 0-automobile households often have Figure 8. Responses to paired-comparison no alternative transportation, so they are questionnaires in Columbia by automobile- willing to endure some of the inconve- ownership group. niences incurred in using public transportation. ONE OR ZERO AUTO HOUSEHOLDS (53) __-I--- IWO OR MERE AUTO HOUSEHOLDS (56) ARRIVING WHEN PLASHED W/O DELAY Figure 9. Fare sensitivity of Columbia WHEN FLASHED respondents by automobile-ownership group. LESS WAIT TINE ARRIVINGCALLING H/U RELAY I LESS WAIT TIMEJj[NO TRANSFER TRIP FBA( SZNSITIVIT'( SEAWIIIC SCAlE SHORT TRAVEL TIME DESIRABLE 7 PICK-UP TINE O100SE PICK-UP TIME _!_I_l_RE0SE NO TRANSFER TRIP III SHELTERS AT PICK-UP SHORT TRAVEL TIME III SHELTERS AT PICE-UP+I+ HAVING V SCAT US HAVING A SEAT -J uuuui•ui MERE DI RECT ROUTE DEPEIIDABLE TRAVEL TINE MERE DIRECT ROUTE 04 DEPENDABLE TRAVEL TIME11[LESS iuuiuu.uu•• LESS WALK TO PICK-UP WALK TO PICK-UP LOWER FARE z23 LOWER FORETIT Ui AR ...... EASY FARE PAYMENT NONE EASY FARE PAYMENT ABILITY TH MEET FRIENDS ON yEN. ABILITY TO MEET FRIENDS ON VEH UNDESIRABLE 1 ADJUSTABLE SEATS..J4.L..ADJUSTVDLE SEATS FARE IN DOLLARS Gustafson and Navin 92

Figure 10. Waiting-time sensitivity of Columbia of using the dial-a-bus system. Phoning respondents by automobile-ownership group. the system and the longer waits were

WDITING LIMB SENSITIVITY- SEMANTIC SCALE I CSELSETIBI more important to the nonuser than to the DESIRABLE user. No transfer and fare have some effect in differentiating users and nonusers. w. This study provides important information to the designers of dial-a-bus systems. Some reservations were ex- I-z pressed in previous studies concerning w the applicability of surveys. The same Cl) technique was applied to 2 different areas with similar results. It, thus, represents an important step toward obtaining UBDESIBABLE

TO CO relevant information concerning the de- WAITING TIME (MINUTES) sign of demand-responsive systems.

CONCLUSIONS ACKNOWLEDGMENTS

The Columbia survey provided some The study is a joint effort by the staff valuable information concerning user of the Transportation Research Depart- preferences for the dial-a-bus system. ment of the General Motors Research Specifically, it validated a previous sur- Laboratories and Francis P. D. Navin. vey conducted in Warren, Michigan, and The authors wish to acknowledge the con- it provided more detailed information on tribution of Robert Bartolo and the Co- the preferences of the users of the sys- lumbia Association who started and op- tem. erated the dial-a-bus system and allowed Columbia respondents closely paral- the surveys to be undertaken. The Uni- leled their predecessors in Warren. The versity of Minnesota provided the funds differences in the results are mostly re- necessary to undertake the surveys in lated to the particular problems that Columbia. dial-a-bus encountered in the implementation process. These characteristics received higher preferences in the Co- REFERENCES lumbia survey because the respondent had been inconvenienced by that charac- Bartolo, R. C., and Navin, F. De- teristic of the system. The similarities of mand Responsive Transit: Columbia, the results are surprising given the dif- Maryland's Experience With Call-A- ferences in the 2 populations. Applying Ride. Paper presented at the annual results from one community to a com- meeting of the American Institute of pletely different community has been a Planners, San Francisco, 1971. problem. The results of this study Bauer, H. J. Case Study of a should increase confidence in the case- Demand-Responsive Transportation study approach. System. HRB Spec. Rept. 124, 1971, The differences indicated by users and pp. 14-39. nonusers are related to the inconvenience Brunning, J. L., and Kintz, B. L. Gustafson and Navin 93

Computational Handbook of Statistics. Scott, Foresman and Co., Glenview, 111., 1968. Golob, T. F., Canty,.E. R., Gustaf- son, R. L., and Vitt, J. E. An Anal- ysis of Consumer Preferences for a Public Transportation System. Transportation Research, Vol. 6, 1971. Gustafson, R. L., Curd, H. N., and Golob, T. F. User Preferences for a Demand-Responsive Transportation System: A Case Study Report. High- way Research Record 367, 1971, pp. 31-45.

INFORMAL DISCUSSION

Question: Who supported the surveys you described? Answer; The Warren survey was supported solely by General Motors, and the Columbia survey was supported by the University of Minnesota.

Question: Did the responses of the Warren survey nonusers correlate with those of the Columbia survey? Answer: That is a subject for future research. We have not had a chance to analyze those data, but we certainly intend to. COMPUTERS, TAXIS, AND GRASS ROOTS TRANSPORTATION

Robert C. Cherry President, Royal Cab Company of Davenport, Iowa

Royal Cab is a 40-year-old company. In July 1967, 5 or 6 months later, I When we purchased the company in 1967, told them that in noway were we respond- we had 6 cabs running and 11 available. ing to the public's need. "We're in a rut. The idea of people sharing a ride in our Nobody's creative, and nobody's caring taxicabs started when we began operating about the public. It's the public be damned, the company, and so we have never had a and I would like to change our image. I problem about sharing a ride in a taxicab. would like to start being responsive to Our problems have been some that I be- what a person wants when he calls for pub- lieve those in demand-responsive transit lic transportation." We then started using will be facing shortly. the Checker Cab, which looks like a 1927 When we bought the cab company, we box. It is a 7-passenger vehicle that had wanted to serve the public and meet the easy access and complemented our shared demand for transportation. In our city of rides. We started to keep records on the 125,000 people, there is 1 bus company, number of "water haul" passengers (that which is subsidized by the city, and 1 is a passenger who is not there when the taxicab company. cab arrives). It was amazing how many As you know, if one has the attitude we had. that something can't be done, it won't be At this point I started to look for a good done. When we bought our company, dispatcher. We now have 3 dispatchers, nothing could be done. I was told by the who have a very nerve-racking job and an drivers, "You put more than 8 cabs on unbelievable capability to remember the road, and we'll quit you. If they want where a vehicle is. They are a rare, cabs, they'll wait." I was told that this rare breed, and I knew we would have was not a taxicab town, and, to tell you trouble finding any more like them. I the truth, I very nearly believed them. knew, too, that, if our company was going On the first day in business, I got to grow and if we were going to be respon- there at 4 o'clock in the morning, exu- sive to the public's transportation needs, berant about our new company. We had our dispatching program had to be imple- 3 time orders (a time order is an order mented. I had no idea how. that is on the books); we missed 2of them Each of our dispatchers has a different because we did not get the cab there on personality and moods. When he comes time. We were really bad, and, conse- to work after a lovely chat with his wife quently, nobody cared about Royal Cab, and a good breakfast then everything goes and nobody cared about public transpor- well. If he comes from an unpleasant tation. Fortunately, all transportation situation, then all hell cuts loose, and, in our area was poor. as a result, our service breaks down. I I was told that people would not ride felt that there had tobe continuity: the cab cabs when the weather is warm, so when must arrive on time, every time; the warm weather came we cut back even driver working the cab must be treated farther. I thought that these people sure the same every time; a new driver must had more knowledge about transportation get the trip, and time must be taken to than I had. I said that I would go along "talk him through" to an out-of-the way and listen to them, but after a period of street. We had to have this kind of con- time I would be able to show them statis- tinuity, and I think it is also going to play tics that would prove them to be wrong. a large part in the success of the demand- They said, "Great! Go ahead!" responsive transportation program. Cherry NO

I felt that we also had to deploy vehi- make so much money. And these things cles better and earlier. I have heard happened: We thought there must be many questions about how long after a some way to eliminate this kind of dic- call is made does it take the bus to tatorship over the drivers who are out arrive, or how long after a person gets there in the cabs trying to make a living. on a dial-a-ride bus does it take to get to We wondered whether a computer could his destination. We have had this prob- dispatch taxicabs. lem for years, and it has been solved by I contacted the computer company, the ingenuity of the dispatcher who can and, when its representative arrived, he route a cab driver through a maze of looked at me (my office is just the size 29.5 square miles of roads so that he of a good latrine) and thought, "There does not go more than 4 blocks out of his must be a sale here for a typewriter and way as he moves through and picks up the company has sent over a systems various passengers going in the same di- engineer." I said, "come on in and sit rection. If the passenger is going to down." There was the most startled look work and cannot be late, the vehicle has on this man's face as though he were to go sometimes with fewer passengers thinking, "Good God, what can you pos- and in a more direct route. Our time sibly want from me ?" I said, "Do you from stopping in front of a house until the think you can dispatch taxicabs with a passenger gets out is much longer than computer?" "Oh yes What's your prob- the time reported for dial-a-buses. We lem ?" feel very fortunate to get a person out in Well, we had problems for him that he 1'/2 minutes. never knew existed. And all of a sudden I also felt that, if we could deploy our he went from a very quick "I can handle vehicles rather than "home" them (bring it" to "Hold on a minute, we had better all of our vehicles to the downtown area), have some meetings." We had meetings we could increase our efficiency. Today, and more meetings and discovered that after the last passenger is delivered, the this dinky littie cab company that had 21 driver stands by and gets further instruc- cabs had this big computer company com- tions to go on. This greatly increases pletely mystified. gross revenue per mile and also makes I have a friend who is an industrial the response time to the customer shorter. psychologist. I asked him to come over We needed to grow with some kind of and observe our dispatchers. After he profit program, and that required that we had watched them work, his remarks look at the structure under which we were, "They're machines! They don't worked. We had a taxicab driver on a make that kind anymore, and don't look commission rate taking all of his orders for their replacements." The computer from a dispatcher on an hourly rate. company representative also observed the Ninety-nine percent of all of our trips dispatching. His reaction was, "This are radio dispatched; we have no cruis- can't be true It is abnormal." Isaid, ing and pickup situations. So, if the "It may be abnormal, but that's what dispatcher did not like the way the driver happens day in and day out." parted his hair, if he mumbled a little A man that comes off a shift of dis- bit over the radio, or if the driver did patching taxicabs, whether it be 10 or 21, not hear the dispatcher right away because is like a zombie. He really is talking to of a bus or truck going by, the dispatcher himself. He has to settle down before he might get mad and the driver might not can discuss problems with his shift. A Cherry 96

dispatcher is under tremendous strain white that his gross revenue per mile is that becomes greater as business in- down. He will not be able to say, "The creases. There comes a point when he dispatcher ran me in circles all day can no longer handle it. What we wanted long." So with the computer, we get con- the computer people to do (if they could tinuity, we get good deployment of our and my money did not run out) was to vehicles, and we are starting to get a combine the best of our manual operation, management program together that at- the best of the zone operation, the best of tracts other people. the manual dispatching, and the best of We had a very difficult time convincing meter control. the computer people that we required a Interpretation is very difficult as fairly large computer. The type of equip- changes are made. I have always found ment we started with requires 17 over- it to be extremely difficult and continue lays. We do not have instantaneous re- to find it so. First of all, I am a little sponse. The wait period is too long, and younger than most of my drivers, and we realize we are going to have to go to they know transportation much better a larger computer. We have alienated a than I do. It is very difficult for them to lot of people in our break-in time. You realize that we went from 6 cabs to 21 can't imagine what a lady says to you cabs and are now heading for 40 and that when she calls and you very politely ex- they are making more money than they plain that she is being dispatched by a made when we had 6 cabs. Then if we computer. She is very much aware that throw a computer operation in on top of her cab has not arrived and does not want that, they will fight it because they do not any jazz about a computer. We have understand what is happening. made a lot of mistakes, but we have not We did an in-depth study with our backed off from the program because we drivers. An in-depth study with a taxi- know it is the best program for public cab driver is beautiful. Before doing so, transportation. We handle people; we you have to learn points 1, 2, and 3 and handle packages; we handle anything that learn them well, for if you skip point 2 is involved in transportation. The dif- and go to 3, he will bring you right back ference between a taxicab in Davenport, to it. No matter what is being discussed, Iowa, and a bus is the spelling. he wants to know what it does to him and We keep our gross revenue per mile his, pocketbook and why. He may also be up because we do not allow a cab to afraid of losing the dispatcher because cruise. A taxicab driver sits until there many times he may blame his doing a is another trip for him or the dispatcher poor job on the dispatcher and he does deploys him to another position. Now, not want that excuse taken away. We, the computer has to be the dispatcher therefore, had to have a good rap session and be able to relocate cabs. with all the people who are running and I would like to say something about directing our organization so that they complications of routing with the com- would realize that the computer is not a puter. If the street runs diagonally, a replacement for them but an implement- dispatcher is aware of it and does not try ing tool that will take a tremendous load to circumvent a driver around a lake or off of them. The computer does its job something to pick up a trip just because steadily day in and day out, and the per- he is in the area. Most of our 940 streets formance rate of the driver can then be are grid streets, but others go off at an evaluated. He can be shown in black and angle. We have been able to program Cherry eyj

them all into the computer by calling the Question: Do your drivers get any angle streets exception streets. pay for waiting at stations, or do they Let's take a look at what really hap- just get paid for trips? pens. The order is typed into the com- Answer: Yes, he gets paid for wait- puter, including information such as ing at stations. We charge 10 cents a origin, destination, number of people, minute after a 3-minute wait. We also and special messages such as wheel charge 25 cents for more than 3 packages chair or package. The computer of groceries. To make sure the driver searches a file for a taxicab closest to carries those bags, we allow him to keep this pickup point, that is either going or the quarter. We share waiting time with can go in that direction. The profitability him because it is taking our time, but criterion then determines which of sev- anything that is received for manual labor eral cars heading in that general direction is his. will gather more revenue per mile by taking this trip, and that cab gets the Question: Do you have figures on trip. The computer has a load factor computer benefits versus costs? built into it, and so we never overload a Answer: Yes, I have some gross rev- cab. enue figures. In 1967, we were getting I would like to close by saying that we 18 cents gross revenue per mile. That felt compelled to try to put together a is a loser any way you look at it. We computer program that other cab com- brought that up to 32 cents with manual panies operating 15 to 150 cabs on a 24- dispatching, and our last computer fig- hour basis could use and also afford. We ures were 44 cents. are open 7 days a week, 24 hours a day. We need to have continuity so that the Question: What is your average trip man who closes his tavern at 2 o'clock in cost? the morning gets the same response, the Answer: As a zone company, it is same ride, and the same time delay, if approximately $1.18. We hope to lower there is any, as the man who closes his that by increasing fares, and only in the shop at 4 o'clock in the afternoon. You taxi industry can you do this. We have have to have a people awareness and a sat around so long and done nothing that supply and demand awareness to make a price increase can look like a price de- this work. crease if handled properly. In the com- I really believe this is grass roots puter, we went from a base zone of 75 transportation. It is not the airlines. cents to a zone anywhere in the city of It is not the trains. It is grass roots 50 cents. Yet, our gross revenue per transportation. The problems can be mile picked up. So, needless to say, the solved and the demand can be handled people were happy. They were riding with the right management. cheaper. The ones that were paying more were the ones that went a long way.

INFORMAL DISCUSSION Question: How many cabs can the dispatcher handle? Question: What percentage of your Answer: I think the number of taxi- overall cost is allocated to dispatching? cabs the dispatcher handles can be quite Answer: We hope it is 5 cents per large, but the number of trips per hour trip cost for computer use. bogs him down. We find that, if there are Cherry 98

more than 110 trips an hour, he is lost because he cannot double trips.

Question: How many trips can a computer handle? Answer: It depends on the size of the computer. In simulation, we have run as high as 200 trips an hour.

Question: Do you have a good backup system? Answer: Yes, it is called manual. We have a large map that is incremented out, and we feel we could manage if we should have that breakdown. USE OF DATA PROCESSING IN TAXICAB CONTROL John Davidson Vice President, Yellow Cab Company, Los Angeles

The operation of demand-response ve- the population density, the presence of hicles (in this paper conventional, radio- one or more "walk-up" traffic generating equipped taxicabs) may be likened to the points (such as an airport, bus station, operation of a tramp steamer. The tramp or major hotel), and the topography of the steamer moves from port to port—not service area. having a fixed route or schedule, not I am closely connected with taxicab knowing whether its next trip will be long operations that vary in size from 10 to or short, and not knowing whether its 700 units, in service area size from 6 to next revenue trip will originate at the 480 square miles, in population density termination of its present revenue trip or from 1,000 to 25,000 people per square whether it must deadhead from the mile, in topography from seashore to planned termination point to the origina- desert to inhabited canyons, and econom- tion point of its next revenue trip. ically depressed areas to "millionaires' So it is with a taxicab operation, and row". therein lies the problem of utilization of The statement relative to dispatching manpower and vehicles. This is espe- difficulties is verified by experience. cially true of an operation in which a Small fleets and service areas where there master - servant relation is maintained are no major walk-up traffic points re- between employees and managers and in quire a system that depends on coopera- which employee benefits, taxes, and the tion between the controller and the vehicle like must be paid by the "master." If driver, and this cooperation is forthcoming such operations are to continue to exist, because the vehicle driver must depend on the dual problem of manpower and vehicle the controller for 98 percent of his busi- utilization must be solved in order to ness. Operations that encompass larger offer reasonable service to patrons and geographic areas, have customers with to secure this profitable utilization. varying economic situations, serve many This problem does not exist for those 'walk-up' points, and receive more than taxicab operations that are conducted by 10,000 requests in a 24-hour period, re- the individual entrepreneuer, such as the quire an entirely different system. The independent New York cabbie who cruises taxicab driver tries to "do the whole thing" or secures loads at fixed pickup points, on his own, and there is difficulty in or for those companies that lease vehicles handling the volume of requests and in equipped for taxicab use but do not main- matching the requests with available taxi- tain a service organization. cabs when they are also sought after at However, those operators of demand- walk-up points. response vehicles, be they taxicabs, There is, however, a consistent thread jitneys, dial-a-ride, or group-loaded ve- running through all of these divergent hicles, that offer transportation service, operations. That thread is a recognition which includes receiving telephone re- that the systems used, even though they quest for service at a specific address, are continuously updated and improved, find this problem of matching vehicles are still deficient in achieving the desired and customers to be one of the most goals of improved service and efficiency. vexing. It is suggested that the following steps In the taxicab industry this system— are necessary to bring about the needed called dispatching—varies by company. improvements. It is believed that they It depends on the size of the fleet being are both technically and economically dispatched, the size of the service area, feasible: Davidson 100

Data processing equipment (defined traffic conditions that varies from 5 as a conventional computer system with minutes in central city core to 8 minutes conventional cathode ray tube on tape in- in less densely populated areas, have no put and output) must be used for (a) re- cross topographical barriers such as ceipt, assignment, and retention of orders range of hills or freeway, and have no for service; (b) receipt, assignment, and more than 1 major walk-up traffic gen- retention of units available for service; erating point in each grid. and (c) matching of orders and available All streets in a grid are defined by be- units. ginning and ending street numbers in that Digital communications equipment particular grid. Wilshire Boulevard, for must be used for (a) transmission of example, is in 11 grids. Major inter- messages from mobile units to base, sections, hotels, stores, and points of giving vehicle identification and status interest are all listed in each grid. without driver input, and location by driver More than 22,000 street, intersection, or input; and (b) transmission of messages specific points are listed in the address from base to mobile units. file, each coded to the grid to which it is Automatic vehicle monitoring assigned. equipment must be used to relieve driver Automatic call distribution system of the responsibility of vehicle location equipment is used to assist in the internal input. handling of the requests for service. As orders are received from patrons, they Our company has moved into step 1 and are punched by conventional keyboard, is conducting discussions with appropriate verified by instant readout, and entered parties relative to steps 2 and 3. into central processing core. The central The area selected for trial use of the processor assigns the appropriate grid data processing equipment is the Los number to the order and stacks the orders Angeles metropolitan area. It encom- by grid number and by time of receipt. passes the service area of the Los Angeles Call-backs from patrons who have not re- Yellow Cab as defined by its franchise ceived service within the quoted time ret- from the City of Los Angeles and from ceive priority in the order stacks. the City of El Segundo; taxicab opera- As available taxicabs call in on radio tions were also permitted in the cities of channels that are dedicated to vehicle lo- Beverly Hills, Burbank, and certain cation, the cab identification and loca- specified areas in Los Angeles County. tion are also entered into the central The land area involved is approximately processing core by means of the conven- 480 square miles, and the population is tional keyboard. The appropriate grid more than 2,500,000. number is assigned to the available taxi- During the preliminary planning, cab, and then stacks are made in order various concepts and system designs were of receipt. considered. The final configuration was The central unit then processes the 2 a service area divided into a number of stacks—requests for service and available districts or grids (the term grid was used cabs—and displays on the screen the ad- although the areas are not all equal in dress at which service is desired and the size or population density). Grid param- identification number of the first available eters were that grids must not include taxicab in that grid. If a unit is not avail- areas in more than one political entity, able in the grid of the order, the central have a vehicular travel time in normal processing unit searches adjoining grids Davidson 101

for the first available unit and displays exceeds the limits that can be handled on its number. existing radio frequencies. The use of When the taxicab driver has acknowl- automatic vehicle monitoring equipment edged the call and the order, a copy is will complete the system for service printed of the order information including (patron to company to vehicle to patron). order-taker number, time of receipt, The use of digital communications be- service time quoted, dispatcher number, tween base and vehicle will further en- vehicle number, time of dispatch, and hance the use of radio frequencies but is whether the unit dispatched was found in not of vital importance. the grid in which the order was located, in another grid, or was located by voice search by the dispatcher. The program has not yet performed perfectly because the volume of vehicle- to-base communications cannot be handled in the limited amount of air time. The program has been modified so that less reliance is placed on vehicle location input and the orders are displayed on the screen with their identifying grid numbers. The progress to date has shown that with the volume handled—peaks of more than 600 orders per hour that require more than 2,400 radio contacts per hour— digital communications equipment is needed between, vehicle and base to handle vehicle identification, status, and location. The test area has 4 clear pairs of radio frequencies in the 150 MHz ban assigned for taxicab use, but these were not adequate to handle the flow of vehicle input and dispatch output. Steps 2 and 3 that will insure the proper flow of information from vehicle and proper identification and location are still in the future. Digital communications equipment for transmission of messages from base to mobile units has the lower priority of the 2 factors in step 2. The utilization of data processing equipment in the disptach of demand- response vehicles is technically and economically feasible. The utilization of digital communications between vehicle and base is needed when order volume GENERAL PURPOSE COMPUTER DISPATCHING SYSTEM

Nigel H. M. Wilson Assistant Professor, Department of Civil Engineering, Massachusetts Institute of Technology Bernard Trevor Higonnet Research Engineer, Department of Civil Engineering, Massachusetts Institute of Technology

Is a computer necessary and desirable so that personnel costs are reduced; and for a dial-a-bus operation? We would more than 1 quality of service so that like to submit that, in certain instances, people can choose according to the price it is. It can certainly provide more ef- they are willing to pay. fective routing in 2 respects. First, The first and foremost computer re- higher vehicle productivity can be quirement is reliability. This can be achieved through computer decision- achieved in a number of ways. In initial making than through manual decision- systems, reliability is ensured through making. Hence, higher dispatching cost some sort of manual backup mode. Look- or higher control center costs can be ing farther ahead, we can expect duplica- justified on the basis of a reduced overall tion of certain elements in the computer transportation cost because of the very system and perhaps completely duplexed high percentage cost that is reflected in systems to provide the level of reliabil- the driver and vehicle components of the ity required. This, of course, has im- system (about 70 percent of total costs). plications for the economics of the dis- Second, the computer provides a very patching system. consistent, effective dispatching service, Other requirements for computer- the quality of which does not correspond based systems include address -to - to the good days and bad days a dis- coordinate translation so that one can patcher may have, as Cherry observed. type in street addresses but have the Gustafson pointed out the importance to algorithm operate on coordinates or per- patrons of on-time arrival at their desti- haps zones and coordinates; error- nation and the variability of waiting time handling capability because initially op- before obtaining service. With a com- .erators and dispatchers interfacing with puter system, one can maintain much the computer will create errors, and more effective controls over these very typing errors will always be made; capa- sensitive service quality parameters. bility to handle unusual events because Computers can possibly reduce the people will want to cancel their trips, overall cost of the control function be- vehicles will break down, and people will cause personnel costs will be reduced. not show up after they have requested This can be effected in 2 ways: The first service; graphic-display capability to is by introducing digital communication enable a supervisor to maintain control and thus reducing the need for dispatch- of the system and influence the quality of ers, and the second is by allowing people transportation service being provided; to bypass the telephone operator and ac- and capability to receive standing re- tually input their request for service. quests, i. e., requests for service at the Costs can ultimately be reduced. same time every day from the same ori- Underlying all of this is the fact that gin to the same destination, and advance computer systems have higher capacities requests, i.e., requests for service sev- than manual systems. When systems eral hours after the call is made. Less grow to 30 vehicles or more, manual de- necessary but nonetheless desirable fea- cisions will become increasingly difficult. tures are automatic billing, standard There are other features that can be trips, and digital communications. implemented through computer dispatch- When implementation options are con- ing. These include automatic billing so sidered, it is useful to define systems in that people do not have to pay when they terms of machine size. Selection of a par- get on the vehicle; use of standard trips ticular machine is largely dependent on Wilson and Higonnet 103

time and development costs, which will computer programmed in a high-level vary with computer type. For example, language such as FORTRAN or PU. The the decision may be between using an ex- reason for using a high-level language is isting operating system or developing that it can be easily developed. Ability from scratch a real-time operating sys- to build a computerized dial-a-ride system. This is a very significant decision tem quickly is important. Flexibility is and one that will affect the amount of time also important. Ability to expand and to needed to get a completely operational modify is necessary so that errors in the system that provides transportation ser- initial system design can be corrected. vice. Another important decision is that The medium-sized computer is governed of the programming language to be used. primarily by the requirement of the high- Will the software be written in a high- level language, which tends to presup- level language, such as FORTRAN or pose such a computer. That such com- PL1, or will machine language be used? puters are easily available is also an im- The machine language will provide a portant consideration. much more effective and efficient com- In terms of the information to be gath- puter system but when implemented will ered at this stage, one should first de- be harder to modify (and consequently termine the feasibility of computer dis- less adaptable to a third party not famil- patching. Until recently, there has been iar with that machine) and will probably significant doubt in the minds of many take more time to develop. people that computer dispatching can Operating costs are clearly a very really work. Beyond this, field testing important point in deciding what com- of assignment algorithms must be per- puter system should be implemented. formed. Are proposed decisions rules Adaptability is one of the most impor- effective in an operational environment? tant considerations in computer system A third area of information is the utility choice and design. It is desirable to and the necessity of features included in have a computer system that can adapt to the dial-a-ride system. handling an increasing number of vehi- Computer dispatching with a minicom- cles as the demand increases within the puter system would have the feature of service area or as the service area is low capital and operating costs and thus expanded. Limited core storage and be more economical. Because its devel- computing time are significant restraints. opment is based on experience gained in If it should happen in the future that testing of the medium-sized, high-level dial-a-ride requires a certain language system described above, it computer or software feature, then could be quickly installed and operated adaptability of this kind is also re- without the operator knowing much the quired. machine language in which it was pro- An important step in the phasing of grammed. dial-a-ride experiments is operating with A further stage in the development of a manual system. This phase provides a computer dispatching systems is the pro- significant market test for the dial-a- vision of more powerful systems with ride concept and can be important in de- features such as standard trips and dig- cisions on the role the federal govern- ital communication interfaces to reduce ment should play. personnel costs and the overall dispatch- In the next phase, the dispatching sys- ing costs. tem would be operated by a medium-sized The system developed at M. I. T. is Wilson and Higonnet 104

programmed in a high-level language and costs play in the total cost of the system. fits in the medium-scale computer cate- So, it is conceivable that a 7-person gory. It evolved from the research into manual-dispatching system could be con- assignment algorithms and the simulation verted to a computer system with no per- model constructed to test various dis- sonnel reduction if productivity could be patching policies and economic feasibil- increased by 25 percent or more. This ity. This formed the basis for a very is the big unknown: Can we achieve in- simple unsophisticated, real-time dis- creases of 25 percent? patching system that was itself the basis There are 2 ways to find the answer to for the final medium-scale FORTRAN this question. The first is through product. The programs can operate on studies of systems such as that of Royal any IBM 360 or 370 computer with more Cab in Davenport, Iowa. The indications than 220K bytes. It is programmed in given in Davidson's paper are encourag- FORTRAN, operates under the disk oping in this respect. However, those in- erating system; can be operated either in creases were achieved under dynamic dedicated or in multiprogrammed mode management when the company was (e. g., dedicated on a 360/30 or parti- growing rapidly, and it is difficult to iso- tioned on a 360/50); and has provisions late the part played by the computer. for teletypes, 1050's, 2741's, and an ad- Nonetheless, it provides an important vanced remote display station for graph- data point. The second way is through ical input and output. We feel that this the Haddonfield Project. When comput- system can handle 20 to 30 vehicles. To erized dispatching is under way in that use this system, one must first code the project, we will have a basis for com- street network so that the address-to- paring manual and computer dispatching. coordinate translation scheme can work effectively. The travel-time prediction formula used by the algorithm must be calibrated, and the algorithm must be tested and modified as a function of the characteristics of the particular area being served. The key questions for the future are, How much will a computer improve service provided by a transportation system? and How much will it increase productivity? These questions are now largely unanswered. Looking at the overall costs of dial-a-ride, we see that perhaps 70 percent of the costs relate to the drivers and vehicles and that the remaining 30 percent relate to control and dispatching. If productivity of the transportation service can be increased by, say, 25 percent as a result of using computer dispatching, then it would be reasonable to increase control costs by as much as 55 percent because of the greater weight vehicle ISSUES AND POLICY QUESTIONS CONFRONTING PUBLIC TRANSPORTATION

Elbert C. Mackey Deputy Director, Michigan Office of Economic Expansion

local transit operations. This legislation Mr. Mackey was the speaker at the confer- basically involves creation of a new de- ence banquet. Although his remarks are partment of transportation and highways not specifically oriented toward demand- and a 2-cent increase in the state gaso- responsive transportation, they are included line tax. Of that, 0.5 cent is to be used as part of the conference proceedings. for an urban transportation discretion- ary fund, which would receive about $22 million per year for public transportation The Michigan Bureau of Transporta- and related purposes. Half of this amount tion was created in 1970 and was given is to be distributed to local governments the authority to finance projects in 4 gen- or authorities for operational assistance eral phases of a public transportation on a population and vehicle-mile of ser- improvement program: Economic and vice basis. The remaining funds are to Technical Feasibility; Research and be allocated to specific projects; primary Technology Development; Engineering emphasis will be on capital assistance. and Design for New Systems; and Dem- I would now like to present what I feel onstration of Improved or Expanded Pub- are some of the major issues and policy lic Transportation Services, Facilities, questions that now face not only the pub- and Equipment. lic transportation industry but society as In general, the purpose of the bureau's well. programs is to enable the demonstration Public transportation has not been re- and the feasibility of expanded and im- sponsive to the mobility needs of most proved public transportation service. urban residents. We have not had inno- Some of the specific projects that are vation in operating practices and in the funded by the state and directly related to development of transportation concepts. this objective are as follows: Departure from existing means of serving mobility needs has been and still is Purchase and operation of electric often viewed as a threat and not as a and propane-powered buses; means of offering more competitive Study and evaluation of new tech- transportation service. nology systems; I do not wish to dwell at length on the Operation of special buses for the history of urban transportation. How- aged and for intercity residents to em- ever, my point on the issue of conserva- ployment opportunities; tism or lack of innovation in public trans- Site-specific feasibility study of portation can be brought into focus by a new technology systems for urban core brief examination of the situation that mobility, large university campus access prevailed in the railway and streetcar and mobility, and activity center circula- operations and that also exists in the tion and connections; systems now being proposed for many Feasibility and planning studies of metropolitan areas. rapid transit in southeast Michigan; and Prior to the early 1900's, companies Study and evaluation of dial-a-ride operating streetcar and rapid transit in Ann Arbor. systems were prosperous and frequently represented monopolies that served at- With passage of pending legislation, most all urban passenger travel. In our role will be considerably more in- many instances, ridership increased at a volved with the operational aspects of faster rate than population until about the Mackey 106

end of World War I. applicability. Most operators were not aware of the Increasingly, urban land use and impending competition. Some even an- travel patterns reflect the dominance of ticipated that ridership would continue to the automobile. This trend is so obvious increase as long as population increased. that I would not mention it except that it Technological obsolescence was simply has special significance. The travel de- not anticipated until it was too late. One sires related to work, shopping, per- must appreciate not only that the compe- sonal business, recreation, and so forth tition was keen but also that the industry are becoming increasingly decentralized. was not subsidized as its rapidly devel- Residential sprawl is widespread. As a oping competition was. The situation consequence, radial travel desires (the was a difficult one to respond to. kind best served by line-haul operations) A significant factor contributing to this are becoming significantly less impor- situation was that the industry failed to tant. Of the demand that does exist, a put aside research and development large portion is peak-hour oriented. Con- funds. We need only look at the invest- sequently, it is difficult to find travel de- ment private industry makes each year mand that can be efficiently and effec- for research and development to know tively served by high-capacity line-haul how important this element is to survival. systems. As a result of lower densities, Of the total invested by private industry trip lengths have also increased. As a last year, excluding residential construc- consequence, those without accessto tion, 15 percent, or $15 billion, was for automobiles have and will continue to research and development. The Urban have difficulty getting to their destina- Mass Transportation Administration's tions either by walking or by public research and development fund in 1972 transportation. With these trends so was only about $53 million, and that was readily apparent, it seems rather probably the total for this multibillion- strange that most methods proposed to dollar industry. solve the urban transportation problem Many of the conditions that existed in are all too often improvements to, or the early 1900's exist today. The ability replication of, existing bus, subway, and to respond and serve transportation needs commuter rail facilities. in an ever-changing competitive market Most public transportation proposals has not developed. I maintain that the in major metropolitan areas focus the development of this ability is one of the majority of funds on rail facilities linking important issues confronting the public high-income suburbs or airports with transportation industry today. downtown centers and on new equipment. A great deal of attention is now being Wohl ()points out that the subway or rail given to the development of high-capacity rapid transit systems now being built and line-haul systems to solve the transpor- proposed are more related to the late tation problems in large metropolitan nineteenth and early twentieth century areas. In my opinion, these systems fall suburban commuter railroads than to in- in many respects to respond to the real tercity subway systems of the same pe- travel demand. A brief look at the dis- riod. tribution of population, the characteris- Wohl points out that the lines for these tics of travel patterns, and the charac- systems usually range far into the sub- teristics of these line-haul systems urbs and concentrate most of the stations raises some serious questions about their within the suburbs. Stations are spaced Mackey 107

far apart and thus provide poor linkages troit. There are 11 stations in the sub- between the station and the traveler's urbs, and 12 within the central city, 4 of origin or destination. Travelers must, which are in the CBD. Clearly, an at- therefore, use a feeder bus or private tempt has been made to respond to the automobile to get to and from the station. need for balance between commuter and To illustrate the inconsistency between local service needs. the technology applied or proposed and The ridership projections for the line the demands of the travel market, he are of particular significance, especially made the following observations on sta- when they are related to the incremental tion spacing: construction cost (Table 1). The percentage of riders on the suburban portion of the line (beyond the Detroit city limits City Miles Apart to Pontiac) and on the central city portion of the line (from the city limits to down- Los Angeles 1'/3 town), by access mode is given in Table Washington, D. C. 2. These data were compared in a de- San Francisco 2 tailed analysis that the Bureau of Trans- Chicago 2/3 portation made of the rapid transit pro- Boston 1/ posal for Detroit. It seems apparent New York 1/3 from this analysis and recent state-of- the-art developments that a second look may be appropriate at the type of transit Wohl also observed that the newly built technology proposed for the Woodward or proposed system lines are long. San corridor. Francisco's lines will all exceed 20 miles. This limited functional capability of a Those for Washington, D. C., will aver- rail rapid system to serve suburban ac- age about 12 miles but some will approach 20 miles. By contrast, Boston's longest line is about 15 miles, Chicago's is 16 miles, and New York City's is 15 Table 1. Ridership and costs. miles. The rigidity of these systems (the requirement of high densities) also Cost appears contrary to or inconsistent with 9-Mile Ridership Million today's life styles. Increment (percent) Dollars Percent In light of Wohl's comments, I would First 83 315 52.5 like to review the proposed Woodward Next 15 185 30.8 2 100 16.7 corridor rapid rail line in Detroit, which Last has been developed under planning studies Total 100 600 100.0 funded by federal, state, and local agencies. The specific proposal is to build a 27- mile, $600 million rail rapid transit line Table 2. Access mode. from downtown Detroit to the city of Pon- tiac. The line is to have 23 stations with Mode Suburb Central City a 1.2-mile average spacing. Of the line's Automobile 55 9 27 miles, 18 are located in suburban Feeder bus 27 54 Walking 18 37 communities and 9 are in the city of De- Mackey 108

cess as well as central city mobility was Of equal significance is the fact that dramatically questioned during recent many individuals are unable to use the consideration of the Governor's trans- automobile because it is at the wage portation bill. Members of the black earner's place of work or because they caucus were opposed to the steel rail are not licensed to drive. system because it did not meet the needs The issue, however, is how to define of their constituents. this market in terms precise enough for I think an appropriate question regard- an evaluation to be made of the costs of ing the application of this technology is, accessibility restrictions. This issue is Where are these facilities located in re- a complex one because market demand lation to workers or others having the and cost are also related to transporta- greatest potential need? Clearly, the nation system design. We are now well ture of the travel market and the need along in the development of new systems have not been examined closely in the de- that have the potential of improving mo- velopment of these systems. It is readily bility at a lower cost. However, we have apparent that, if we are to overcome the not yet really examined the mobility dis- shortcomings of the steel rail systems, parities and need differences among va- innovative efforts must be made to de- rious classes of society, let alone placed velop systems that satisfy both suburban a cost figure on them. We know very and central city transit needs. I do not little about the effect alternative trans- mean to imply that we are against rapid portation strategies might have, for ex- transit systems. Our concern is with the ample, on the opportunity and cost of appropriate application of technology to housing or public services such as water the problems that exist. and sewerage distribution systems and The question of the applicability of the land for recreation. We do know, how- rail rapid transit systems raises a general ever, from our various tax bills that the question of what the public transportation present policies being pursued at the market consists of and how we can re- local level are expensive. We need to be spond to that market with a competitive more concerned about optimizing our service at an acceptable cost. capital investments. It does little good Identifying potential markets and de- to pour money into one area if, at the fining benefits and costs of an applicable same time, policies or programs exist public transportation technology are not that will defeat our objective. easy tasks. Many people are prone to In many instances, I doubt whether say that there is little need for public present cost figures are representative transportation services based on patron- of true costs. We frequently assume that age trends and public transportation's because we have special funds for certain share of the market. I do not believe that purposes total project costs are cov- the demand for this service has ever been ered. This is not always the case. An adequately assessed. A tremendous need example in point is the Lodge Express- exists for low-cost transportation way in Detroit. Our analysis of antici- throughout this country. pated revenues raised by this facility in- Our automotive evolution during the dicates that it will only pay for about a past few decades has given many Ameri- fourth of its cost. cans unprecedented mobility. However, There are indications that the cost of for those who lack access to the automo- governmental services will increase in bile, a severe mobility handicap exists. the face of tremendous competition for Mackey 109

resources. Therefore, the need is cru- and the design of the automobile provide cial to define more precisely the mobility the basis for both flexibility and accessi- disparities among segments of society, bility. One must also be aware of the the cost of these disparities, and the comfort and convenience of the automo- benefits and costs of alternative trans- bile. Many homes do not have the com- portation strategies. fortable seats, air conditioning, stereo, This issue is also directly related to and privacy that the automobile has. Yet, another important issue, and that is, the amount of time spent in the home—at What kind of an urban environment do we least for most areas and for most age really want? At the present time, we do groups—is incomparable to that spent in not have local, regional, or state growth the automobile. policies that reflect strategies that have A question frequently asked about been carefully thought out and analyzed demand-responsive service is, Does it as to benefits and costs. We appear to cost more or less than line-haul service? be content with an increment of haphazard Although I believe the question of cost is sprawl each year. I think we need to ex- an important one, it needs to be consid- amine these policies and explore what ered in proper context. Certainly, con- role public transportation has in develop- cern for the market served is an impor- ing the kind of urban environment we tant consideration. In Ann Arbor, we are want. serving a substantially different market I would like to say something about the by the demand-responsive system than by demand- responsive transportation project the line-haul system, and we want to in Ann Arbor. We feel strongly about the broaden that market. A valid comparison need for supporting the Ann Arbor proj- of costs must also consider alternative ect. It has potential for solving a prob- objectives. If, for example, reductions in lem in one area as well as for providing air pollution, in downtown traffic conges- needed research regarding people's at- tion, and in parking costs are objectives, titudes and travel needs that cannot be then these must also be evaluated in gained from traditional research methods terms of the system's ability to achieve that lack the operational element. Cer- them. tainly, developing, managing, and operat- We can learn much more from demanding a demand-responsive project require responsive service projects than what the a substantial degree of innovation. operating limitations are. These proj- The people in Ann Arbor deserve con- ects provide the means of obtaining val- siderable credit for undertaking this uable data concerning user attitudes, project. Inevitably, there are higher travel preferences, and trip generation risks involved in the introduction of new rates. Such research information is es- service concepts. This type of service sential to the development of competitive entails a whole new set of operating cir- transportation systems. cumstances, and it may have the potential to go a considerable way toward achieving the desirable service characteristics REFERENCE provided by the automobile. Speed, flexibility, and accessibility 1. Wohl, M. Urban Transport We Could attributes of the automobile are the most Really Use. Technology Review. difficult to duplicate in new system operations. The extensive street network PART IC IPANTS

Leonard G. Abraham, GTE Laboratories, Eugene T. Canty, General Motors Re- Inc., Waltham, Massachusetts search Laboratories, Warren, Robert P. Aex, Rochester -Genesee Michigan Regional Transit Authority, F. Dawson Catton, Kates, Peat, Marwick Rochester, New York and Company, Agincourt, Ontario Wayne C. Allinson, Allinson, Inc., War- Robert C. Cherry, Royal Cab Company of wick, Rhode Island Davenport, Inc., Iowa J. B. Annand, Public Utilities Commis- Ed Cummings, Model Cities Transit sion of the City of Oshawa, Ontario Project, Columbus, Ohio Wallace G. Atkinson, Regina Transit James P. Curry, De Leuw, Cather and System, Saskatchewan Associates, San Francisco

George Bacalis, Ann Arbor Transit Au- Jack L. Dais, University of Minnesota, thority, Michigan Minneapolis Brent Bair, Institute of Urban and Re- John Davidson, Yellow Cab Company of gional Research, Iowa City California Robert C. Bartolo, The Rouse Company, Allan Davis, Frederic R. Harris, Inc., Columbia, Maryland Stamford, Connecticut Herbert J. Bauer, General Motors Re- Joe Diehl, Bus Builder, Loudonville, search Laboratories, Warren, Ohio Michigan William D. Drake, Ann Arbor, Michigan Daniel E. Benson, North American Rock- David A. Duffy, Florida Department of well, Chicago Transportation, Tallahassee Michael J. Berla, Ann Arbor Transpor- tation Authority, Michigan John M. Elliott, Amalgamated Transit Walter Bierwagen, Amalgamated Transit Union, Washington, D. C. Union, Washington, D. C. Alan Bingham, AC Transit, Okiand, Ray Feiler, Terre Haute, Indiana California Donn Fichter, New York State Depart- David Bone, Grand Rapids Transit Au- ment of Transportation, Albany thority, Michigan John J. Ford, Dave Systems, Inc., Had- John A. Bonsai, Ontario Ministry of donfield, New Jersey Transportation and Communications, Frederick Frye, Creighton, Hamburg, Downsview Inc., Delmar, New York Robert Brown, Allinson, Inc., Warwick, Rhode Island Jean Granger, Ecole Polytechnique, Edwin W. Bickhart, Mass Transit Sys- Montreal tems, Harrisburg Karl Guenther, Ford Motor Company, William E. Burton, Columbia Associa- Dearborn, Michigan tion, Maryland Richard Gustafson, General Motors Re- search Laboratories, Warren, Michigan James W. Caccamise, Rochéster-Genesee Regional Transit Authority, Rochester, New York William C. Habig, Mid-Ohio Regional J. H. R. Campbell, Winnipeg, Manitoba Planning Commission, Columbus Participants 111

Ruth Hartzler, Mitre Corporation, William E. Marshall, Twin Cities Metro McLean, Virginia Transit Commission, St. Paul Lance W. Haus, Cornell University, Philip R. McBride, Greater Lafayette Ithaca, New York Public Transportation Corporation, Albert Hayes, Yellow Cab Company, Indiana Pittsburgh H. T. McClure, Caterpillar Tractor J. Downs Herold, University of Michigan, Company, Peoria Ann Arbor William McCormick, University of Mich- Robert Hess, University of Michigan, igan, Ann Arbor Ann Arbor Phil McGuire, Institute of Urban and Bernard Trevor Higonnet, Massachusetts Regional Research, Iowa City Institute of Technology, Cambridge Jack Mcllroy, University of Michigan, Antoine G. Hobeika, Purdue University, Ann Arbor Lafayette, Indiana Douglas M. Medville, Mitre Corporation, Jerrold Hornstein, Chrysler Corporation McLean, Virginia Space Division, Titusville, Florida Daniel D. Morrill, Southeastern Michigan Transportation Authority, Detroit Joseph W. Mott, Port Authority of New Howard Johnson, Canadian Ministry of York and New Jersey, New York Transport, Montreal J. R. Mumford, Ford Motor Company, Dearborn, Michigan Roy E. Murphy, Lex Systems, Palo Alto, Edward J. Kazenko, Michigan Department California of State Highways, Lansing John F. Kennedy, Michigan Bureau of Transportation, Lansing Frank Navin, R. H. Pratt Associates, Mark L. Kermit, Martinez Public Works Garrett Park, Maryland Department, California Larry J. Newman, University of Iowa, R. Neil Kravetz, Rochester-Genesee Re- Iowa City gional Transit Authority, Thomas A. Niskala, Cleveland Transit Rochester, New York System, Ohio Richard Kunz, Transport Central, Charles B. Notess, State University of Chicago New York at Buffalo Yutaka Kuwahara, Hitachi Sales Corpo- ration, New York P. Oxley, Ford of Europe, Inc., Laindon, Essex, Great Britain John Lawson, Call-A-Bus, Fort Walton Beach, Florida Warner Paige, City of Terre Haute, Indiana Jerold Lax, City of Ann Arbor, Michigan George M. Peace, University of Mich- Terry L. Linger, Michigan Bureau of igan, Ann Arbor Transportation, Lansing Rodney P. Plourde, Fay, Spofford and Thorndike, Inc., Brighton, Massa- chusetts Elbert C. Mackey, Michigan Office of Richard W. Prather, Atlanta University Economic Expansion, Lansing Urban Transportation Project Participants 112

R. C. Rand, Johns Hopkins University, Phillip J. Sulentich, Metro Regional Silver Spring, Maryland Transit Authority, Akron, Ohio James E. Reading, Rochester-Genesee Regional Transportation Authority, Nadeem Tahir, Department of Transpor- Rochester, New York tation, Chicago James L. Roach, Michigan Department of Malcolm S. Tanton, N. D. Lea and Asso- State Highways, Lansing ciates, Ltd., Oakville, Ontario Kenneth Roberts, Mitre Corporation, Egons Tons, University of Michigan, McLean, Virginia Ann Arbor Daniel Roos, Massachusetts Institute of Technology, Cambridge Kenneth E. Underwood, Michigan Depart- ment of State Highways, Lansing Gilbert T. Satterly, Purdue University, Thomas Urbanik, II, Ann Arbor Depart- West Lafayette, Indiana ment of Traffic Engineering and James W. Schmidt, De Leuw, Cather and Transportation, Michigan Associates, San Francisco John B. Schnell, American Transit Asso- ciation, Washington, D. C. W. Vanderschmidt, Rochester-Genesee Frederick W. Schroeder, Columbia Park Regional Transportation Authority, and Recreation Association, Maryland Rochester, New York James A. Scott, Highway Research Board, Washington, D.C. James Wargo, McGraw-Hill, Detroit Bernice Shetterly, Lansing Model Cities, Lawrence Wasson, Jr., Citizen's Con- Michigan gress, Inc., Lansing, Michigan Curtis 0. Siegel, Washington, D. C. Anthony U. Simpson, DAVE Systems, Daniel Weisberg, Rockland County Planning Board, New York Inc., Haddonfield, New Jersey Robert E. Skinner, Jr., Alan M. Voorhees Michael White, Model Cities Jitney and Associates, McLean, Virginia Transportation, Buffalo, New York Herbert S. Wilson, Cornell University, Louis James Slade, Barton-Aschman As- sociates, Washington, D. C. Ithaca, New York J. R. Wilke, Lynn Model Cities, Robert Sloane, Boston Transportation Massachusetts Planning Preview Nigel H. M. Wilson, Massachusetts Insti- Haldon L. Smith, University of Michigan, tute of Technology, Cambridge Ann Arbor Joy Wooten, Lansing Model Cities, Lawrence M. Smith, GTE Laboratories, Michigan Waltham, Massachusetts Julien R. Wolfe, Southeastern Michigan R. B. Smith, Oshawa Public Utilities Transportation Authority, Detroit Commission, Ontario James Stoner, University of Iowa, Iowa Edward R. Wujcik, City of Pontiac, Michigan City Paul Stokes, Lansing Model Cities, Michigan John C. Zielinski, Ann Arbor Transpor- James Strichartz, Ford Motor Company, tation Authority, Michigan Ann Arbor Marcel J. Zobrak, Urban Mass Trans- Ling Suen, Transportation Development portation Administration, Washington, Agency, Montreal D. C. SPONSORSHIP

GROUP 1 —TRANSPORTATION SYSTEMS PLANNING AND ADMINISTRATION Charles V. Wootan, Texas A&M Univer- sity, chairman

Committee on New Transportation Sys- tems and Technology Leon M. Cole, University of Texas at Austin, chairman William H. Avery, Robert U. Ayres, George J. Bacalis, Eugene T. Canty, Arthur A. Davis, Donn Fichter, Donald Friedman, William F. Hamilton, II, M.D. Harmelink, Kenneth W. Heathington, William S. Jewell, Harold W. Merritt, Theodore F. Morf, Robert A. Olmsted, C. Kenneth Orski, Herbert A. Richardson, Daniel Roos, Albert J. Sobey, Charles E. Taylor, J. William Vigrass, Lucius Walker, Jr., Foster L. Weldon

Conference Committee Daniel Roos, Massachusetts Institute of Technology, chairman George Bacalis, Eugene Canty, Karl Guenther, Donn Fichter, and James A. Scott

James A. Scott, Highway Research Board staff THE National Academy of Sciences is a private, honorary organization of more than 800 scientists and engineers elected on the bascs of outstanding contributions to knowledge. Established by a congressional act of incorporation signed by Abraham Lincoln on March 3, 1863, and supported by private and public funds, the Academy works to further science and its use for the general welfare by bringing together the most qualified individuals to deal with scientific and technological problems of broad significance. Under the terms of its congressional charter, the Academy is also called upon to act as an official—yet independent—adviser to the federal government in any matter of science and technology. This provision accounts for the close ties that have always existed between the Academy and the government, although the Academy is not a governmental agency and its activities are not limited to those on behalf of the government. The National Academy of Engineering was established on December 5, 1964. On that date the Council of the National Academy of Sciences, under the authority of its act of incorporation, adopted articles of organization bringing the National Academy of Engi- neering into being, independent and autonomous in its organization and the election of its members, and closely coordinated with the National Academy of Sciences in its ad- visory activities. The two Academies join in the furtherance of science and engineering and share the responsibility of advising the federal government, upon request, on any subject of science or technology. The National Research Council was organized as an agency of the National Academy of Sciences in 1916, at the request of President Wilson, to provide a broader participation by American scientists and engineers in the work of the Academy in service to science and the nation. Its members, who receive their appointments from the President of the National Academy of Sciences, are drawn from academic, industrial, and government organizations throughout the country. The National Research Council serves both Academies in the discharge of their responsibilities. Supported by private and public contributions, grants, and contracts and by voluntary contributions of time and effort by several thousand of the nation's leading scientists and engineers, the Academies and their Research Council thus work to serve the national interest, to foster the sound development of science and engineering, and to promote their effective application for the benefit of society. The Division of Engineering is one of the eight major divisions into which the National Research Council is organized for the conduct of its work. Its membership includes representatives of the nation's leading technical societies as well as a number of members-at- large. Its Chairman is appointed by the Council of the Academy of Sciences upon nomination by the Council of the Academy of Engineering. The Highway Research Board is an agency of the Division of Engineering. The Board was established November 11, 1920, under the auspices of the National Research Council as a cooperative organization of the highway technologists of America. The purpose of the Board is to advance knowledge of the nature and performance of transportation systems through the stimulation of research and dissemination of information derived therefrom. It is supported in this effort by the state highway departments, the U.S. Department of Transportation, and many other organizations interested in the development of transportation. HIGHWAY RESEARCH BOARD NON-PROFIT ORG. NATIONAL ACADEMY OF SCIENCES—NATIoNaL RESEARCH COUNCIL U.S. POSTAGE 2101 ConEtitotiol, Avonoo Wo,hington, D. C. 20418 P A I D WASHINGTON, D.C. ADDRESS CORRECTION REQUESTED PERMIT NO. 42970

V) >- z a I—I LD —I I

U. z Do' 0

QN () LLJ , 0 -UJ OcOw OLL 0 00 0_J i-cj..cO