Introduction Into Simulation Modeling for Business Applications

AnyLogic pedestrian simulation projects for public buildings in France: museums, terminals, big shops

Vladimir Koltchanov PhD, Director ANYLOGIC EUROPE

AnyLogic Conference 2014 San Francisco, CA November 12 and 13

• Some information about Pedestrian simulation activity in France • Why use micro scale agent based models? • What is agent based pedestrian displacements model? • Touristic and cultural pedestrians flow simulations (Louis Vuitton Foundation, Louvre, Grande Palais, Eiffel Tower) • Railway and Metro Stations (Eurostar terminal, Montparnasse…) • Airport terminals (Orly Sud terminal, Nice airport terminal) • Big Paris Shops (Printemps) • Some future projects

Pedestrian Models (active developments started since about the 1980s)

Macroscopic Microscopic (flows are simulated, (every pedestrian is presented as a specific matter) no individual pedestrians are modeled)

Gas kinetic model Waiting lignes Continuous space Discrete space (cellular automata)

Social Force Model (geometrical, 1998 first French + Agent-based model, pedestrian “intelligence” simulation software PX4)

Macroscopic pedestrians flow models are not well situated for:

• Buildings or spaces with exceptional architecture • Dynamically changes in the displacement conditions • Origin/Destination data is often missing • Extreme conditions such as escape need to be evaluated • Need to take in account the individual behaviour of moving persons

• Easy local rules • Fast-to-calculate ? • Can be well-calibrated • Poor animation • See Blue & Adler

target • Newton mechanics • Realism • Relatively slow calculations Driving • Very realistic animation force

• Extended with higher level Resultant force decision making logic Inclination for attractors • See Helbing& Molnar Repulsion from other pedestrians Repulsion from walls and obstacles

Sum of social forces dx (t) (1) i  v t Random individual variation dt i Pedestrians forces dvi (t) (2) m  f t  t Walls forces i dt i i v0 te0 t v t (3) f t  m i i i  f  f i i (i j) ij w iw  i

Individual force

D.Helbing, I.J.Farcas, P.Molnar, T.Vicsek, Simulation of Pedestrian Crowds in Normal and Evacuation Simulations. In: Pedestrian and Evacuation Dynamics, Springer, 2002

• Railway stations • Metro stations In general all the facilities where the • Airports arrangement of physical space for

pedestrians affects throughput transport • Car parks capacity, quality of service, and safety • Pedestrian passageways • Shopping malls • Museums •

“attractions” Amusement parks • Stadiums • Concert halls

events • Worship facilities • Street events (festivals, rallies, demonstrations) • As well as production, warehouse and even movements of personnel in a kitchen…

1 Facility plan/drawing 2 Space markup

References to markup elements

3 Process description

Walls Services (service points) and queues

Target lines / pedestrian appearance lines

Waiting areas / target areas

Virtual corridors (pathways)

Acceleration / deceleration areas

PedSource Creates pedestrians on PedService Sets servicing a line, at a point or in parameters (where an area with a given is a delay, the selection rate, according to a of a queue, etc.) time schedule, etc.

PedGoTo PedWait Sets waiting Sets up an objective or parameters (where to a route wait, in relation to time, until an event) PedSelectOutput Divides a passenger PedSink flow Deletes passengers from the model

• Metrics typical for discrete event models ̶ Queue lengths ̶ Waiting time ̶ Time in a system ̶ Utilization of service points (services) • Metrics specific for pedestrian models ̶ Flow characteristics: the total number of passenger having passed through PedFlowStatistics a section per a unit of time, the same quantity per a unit of length ̶ Density in a certain area: PedestrianDensityMap the number of passengers per square meter (average per a unit of time); density charts

Model elements Manifestations program, foundation's facilities (controls, checkrooms, restrooms, restaurants, …), attractiveness of exhibits, obstacles, circulation space, individual and group behavior, non predefined displacements in the building

The criteria of project evaluation: global and local density, time of visit, individual visitors comfort

Background The more visited monument in the World The current throughput of the tower is insufficient to handle visitor flows The bottleneck is in the two upper levels of the tower where visitors spend too much time, partially because of poor flow management Goals The model was ordered by Eiffel Tower Operation Society to test several alternative scenarios of visitor flow management to determine if the traffic capacity of the Tower can be improved Model elements • Floor plan • Visitors routing policies • Elevator number, capacities and travelling times, • Visitors behavior (time spent at various attraction Eiffel points) Tower Demo

© ANYLOGIC EUROPE/EMSYSS | www.anylogic.com 23 Metro station project simulation. Tickets Control level. Controls transaction time, circulation space, obstacles, individual and group behavior, queue lengths and waiting times

Background A high-speed passenger rail service connecting Paris with London. All trains traverse the Channel Tunnel between Britain and France The Paris terminal of EUROSTAR is located at Gare du Nord The terminal space is limited and increased demand is significant Goals Find out how many trains the terminal can handle during rush hour (in particular, can two trains depart with 3-4 minute interval?) Optimize the number of control lines and their distribution by the types of passengers Optimize queue management policies

• Waiting time • Terminal plan including • Queue lengths check points • Number of late- • Time schedule coming passengers • Passenger flow • Check point loading distribution Simulation model (business/economy, EU/non-EU, etc.) • Passenger arrival time distribution • Check point time of passage • Passenger behavior in a waiting lounge • Queue management policies

Non-EU and B/C Waiting hall A Waiting hall B

Non-EU Waiting hall for B/C Customs EU Business Class

B/C UK Passport control EU FR Passport control EU Non-EU Ticket control EU Ticket control non-EU and Business class (priority)

All passengers wait here before registration is open

2D animation and statistics

INPUT (EXCEL Import) • The terminal layout with control lines • Train timetable • Passenger distribution (business/economy, EU/non EU, etc) • Passenger arrival time distributions • Control passing times • Controls number • Passenger behavior in the waiting area • Queue management policies

OUTPUT (EXCEL export) • Waiting times • Queue lengths • Number of passengers who missed their trains • Control lines utilization EUROSTARDemo

ORLY Sud Demo

Paris-Charles de Gaulle Terminal 2E passenger traffic simulation