FACT SHEET

SAS/OR® Software Optimize business processes and address management science challenges with enhanced operations research methods

What does SAS/OR® software do? SAS/OR software provides a powerful array of optimization, simulation and project sched- uling techniques to identify the actions that will produce the best results, while operating within resource limitations and other relevant restrictions. Why is SAS/OR® important? Organizations can consider more alternative actions and scenarios, and determine the best allocation of resources and plans for accomplishing goals. Incorporating operations research analytics adds structure and repeatability to decision-making processes, lets you make the most of your analytic and BI investments, and delivers a competitive edge. For whom is SAS/OR® designed? SAS/OR is designed for people in any industry with operations research (or manage- ment science) experience who build decision-guidance models by applying operations research techniques to solve real-world problems. Adding well-designed user interfaces can open up these methods for use by people who interact with the problems on a business level.

Choosing the actions that produce the Benefits chosen. This allows problems to be best outcomes requires the ability to formulated and solved intuitively and • Apply a wide range of operations create, consider and evaluate alternate efficiently regardless of their specific research methods. SAS/OR offers the scenarios. SAS/OR® software helps model, mathematical form. broadest available spectrum of opera- solve and communicate the best solutions • Easily incorporate more data. With tions research modeling and solution to complex planning problems quickly SAS/OR it is easy to indicate where and techniques, and includes state-of-the- and effectively. The software brings how input data will be used in a model. art methods for mathematical optimiza- together essential optimization, simulation Data/model separation is maintained, tion. The depth of detail and realism in and scheduling solution capabilities in an which is critical when reusing mod- SAS/OR software’ modeling capabili- integrated and adaptable environment. els or model components. Users can ties, control of optimization, simulation select the aspects of the solution to be and scheduling processes, and inte- Organizations worldwide use SAS/OR reported and can control the form in grated approach to data access and software to solve planning problems and which they are reported. information delivery enable organiza- address business challenges such as: tions to identify and apply the best • Generate quicker, better answers. • Resource allocation and management. responses to complex planning SAS/OR includes analytic and solution • Production and inventory planning. problems. methods that are tuned to address large, complex, real-world problems. • Product mix and composition. • Build models interactively and experi- • Staffing allocation and scheduling. ment with data. SAS/OR lets you build models interactively, modifying Product Overview • Supply chain optimization. constraints or variables and experi- With SAS/OR software, modelers trans- • Capital budgeting, asset allocation and menting easily with the effects of form real-world scenarios into mathemat- portfolio selection. changes to underlying data. In math- ical models. When altering models to • Optimization of service contract ematical optimization, a specialized better reflect the key elements of business agreements. modeling language enables you to problems, they can consider various • Public utility capacity planning. work transparently and directly with options, leveraging essential modeling, symbolic problem formulations, and an • Optimal facility site location. optimization, simulation and scheduling appropriate solution method for the capabilities from within SAS. • Workflow productivity improvement. current problem can be automatically Most SAS/OR capabilities are surfaced with descriptive and , Mathematical Optimization within a common language and all use a and builds on those analyses to deliver SAS/OR contains sophisticated mathemat- common data format, which allows proactive decision guidance. ical programming techniques that can help analysts to seamlessly utilize , determine the best use of limited resources data cleansing, forecasting, experimental Combining power and accessibility with to achieve desired goals and objectives. design, Monte Carlo simulation or any of the SAS foundation of data management, the hundreds of statistical functions analytical (, forecasting, data offered by SAS Analytics, and avoid the mining, etc.) and reporting features, Algebraic, symbolic optimization hassles of dealing with multiple niche SAS/OR enables you to coordinate directly modeling language software packages. Operations research is with critical supporting and follow-on The OPTMODEL procedure provides a never performed in isolation; it is part of a activities as you build, use, maintain and rich optimization modeling language with continuum that begins with data integra- update a wide range of models. specialized syntax and constructs that tion, grows by informing decision makers enable problems to be represented directly and efficiently. This makes it easier to review models for initial validation, make subse- quent adjustments or run models with new data. This clarity is critical if optimization models are to be distributed for use across many departments or divisions, or if analysts are reassigned and pass planning models to their colleagues to carry on with imple- mentation and/or adaptation for alternate scenarios.

Linear, integer, mixed integer, nonlinear and With SAS/OR, you need to learn only one set of statements and commands to build and solve a wide range of optimization models. Optimization models often evolve The SAS Simulation Studio graphical interface provides interactive model building and during the implementation process. As experimental design capabilities. analysts adjust their formulations to address changes in requirements, constraints and/or the objectives can change from linear to nonlinear expressions and vice versa. There’s no need to worry about switching modeling environments or employing different syntax to use appropriate solution algorithms.

Powerful optimization solvers and presolvers SAS/OR provides a suite of solvers that is streamlined for simplicity and tuned for the best performance when finding optimal solutions. This enables you to tackle even larger enterprise problems and solve them more quickly. Optimization solvers include primal and dual simplex, network simplex, interior point, branch-and-bound, and nonlinear solvers that are especially suited to handle large, sparse problems. Sample interface demonstrates the use of SAS/OR to optimize workforce allocation, factoring in shift hours, pay rates and the opportunity cost of unmet demand. Decomposition algorithm Key Features The decomposition algorithm (linear and mixed integer linear optimization) exploits Mathematical optimization a structure often found in optimization • OPTMODEL procedure: models — blocks of constraints, each • Flexible algebraic syntax for intuitive model formulation. involving an exclusive set of decision vari- • Transparent use of standard SAS functions. ables. After these blocks have been identi- • Direct access to linear, network, mixed integer, quadratic, nonlinear, and constraint fied, the algorithm solves the resulting programming solvers. component problems in parallel, coordi- • Support for the rapid prototyping of customized optimization algorithms, including nating with the solution of the entire named problems and subproblems. problem and ultimately reducing overall • Ability to run other SAS code within PROC OPTMODEL with the SUBMIT block. solution time significantly. • Ability to execute solver invocations in parallel with the COFOR loop. • Aggressive presolvers to reduce effective problem size. Network optimization • Multithreading in underlying technologies for improved optimization performance. • solvers, including primal simplex, dual simplex and network The OPTNET procedure provides algorithms simplex; and interior-point with crossover. for investigating the characteristics of • Parallel branch-and-bound mixed integer programming solver with cutting planes networks and solving network-oriented opti- and heuristics. mization problems. Input data sets are • Option tuning for mixed integer programming. designed to fit network-structured data. • Decomposition algorithm for linear and mixed integer programming. Algorithms include minimum-cost flow, • General nonlinear optimization solvers, including primal-dual interior point, primal- shortest path, traveling salesman problem, dual active set, and multistart capability. connected components, cycle detection, • Covariance matrix output available for nonlinear optimization. and several others. • Multiple network diagnostic and optimization algorithms. • Parallel hybrid global/local search optimization, including multi-objective optimization. Multistart algorithm helps • Constraint programming capabilities with scheduling and resource features. identify better solutions Many nonlinear optimization problems can Discrete-event simulation be classified as nonconvex. In such cases, • Versatile, graphical modeling capabilities; create and save custom components. the optimization problem might have many • Model both static and mobile resources. locally optimal solutions that are not • Automated experimental design and input analysis via integration with JMP®. globally optimal. To increase the likelihood • Drive models with historical data in SAS data sets or JMP tables. of identifying a globally optimal solution, • Integrate with SAS or JMP for analysis of results. the multistart algorithm selects multiple • Support for large models and large experiments. starting points and begins optimization in • Search facility enables search of all blocks in model. parallel from each. The best solution found • Hierarchical modeling: compound blocks and submodel blocks. among all starting points is reported.

Interactive modeling Project and resource scheduling and solution environment • Critical path method and resource-constrained scheduling. In the OPTMODEL language you can • Calendars, work shifts and holidays for determining resource availability and schedules. modify your optimization model interac- • Full support for nonstandard precedence relationships. tively, dropping or restoring constraints, • Versatile reporting, customizable Gantt charts and project network diagrams. fixing decision variables at specified values, • Earned value management analysis for project execution tracking. or altering the underlying data. This enables • Decision analysis: you to try out different versions of the same • Create, analyze and interactively modify decision tree models. model and experiment easily with the • Calculate value of perfect information (VPI) and value of perfect control (VPC). effects of changes. You can also define and • Bill of material (BOM) processing: name multiple models to solve individually • Read from standard product structure data files and part master files, or combined files. or as part of a larger solution strategy. Inter- • Produce single- or multiple-level bills of material, including indented and mediate solutions can be saved for use in summarized BOM. future optimizations. All aspects of interme- • Produce summarized parts, listing items and quantities required to meet the diate and optimal solutions are fully acces- specified plan. sible for examination, analysis and reporting. Global/Local Search Project and Resource compare all possible outcomes. In input data sets you describe the problem struc- Optimization and Scheduling ture, the probabilities of various outcomes Constraint Programming SAS/OR software’s project scheduling and the associated payoffs. SAS/OR capabilities give you the flexibility to plan, analyzes the decision problem, incorporates SAS/OR includes two options for those manage and track project and resource utility functions and attitudes toward risk, confronting some of the most challenging schedules through a single, integrated and identifies an optimal decision strategy. optimization-related problems. PROC system. The software handles compli- OPTLSO applies multiple global and local cated situations involving multiple project search algorithms in parallel to solve optimi- Bill of material processing record keeping, resource priorities, zation problems that include difficult (nons- Bills of material are used in manufacturing project and resource calendars, substitut- mooth, discontinuous, nondifferentiable, to show the relationships linking parts and able resources with skill pools, multiple etc.) functions, and can also solve especially materials, subassemblies, assemblies and and nonstandard precedence relation- difficult types of problems such as mixed finished products, and can also be used to ships and activity deadlines. You can integer nonlinear optimization. Constraint explore the roles of multiple levels of create and update single- and multiple- programming with PROC CLP solves subsidiary tasks in major activities. SAS/OR project schedules, incorporating struc- constraint satisfaction problems (optionally performs bill of material processing, tural, time, and resource constraints. adding an objective function) using reading product and component structure Inputs to the scheduling process include powerful consistency algorithms, tailored data and composing the information into hierarchical project structures, resource for specific classes of constraints, along with single-level, multiple-level and indented requirements, and work shift/calendar/ a choice of search strategies. Each bills of material. Summarized reports show holiday information for activities and approach can be useful for problems that quantities of all items needed to fill orders resources. Both replenishable and are difficult or impossible to formulate or for finished goods. These capabilities can consumable resources are supported, solve with standard optimization methods. work in conjunction with SAS/OR software’s and resources can be assigned in teams project scheduling features to determine as needed. Extensive control over the Discrete-Event Simulation the impact of parts availability on produc- scheduling process is provided. Output tion and delivery schedules. SAS Simulation Studio features a GUI that includes detailed project schedules and requires no programming and provides all profiles of resource usage and availability the tools needed for building, executing across timelines. Graphics include Gantt TO LEARN MORE » and analyzing discrete-event simulation charts and network diagrams. models. A broad array of modular blocks, SAS/OR takes advantage of the SAS®9 each with customization options, enables Earned value management analysis engine, part of the SAS Platform. Many SAS procedures have been enhanced so code you to build detailed, realistic simulation SAS/OR includes earned value manage- launched from SAS 9 can run in SAS Viya™, models. You can model resources in static or ment capabilities that enable you to track, the SAS Platform’s new distributed, mobile form, further increasing the models’ analyze and predict the cost and schedule in-memory engine. For more information, realism. Experimental design (manual and performance of projects in progress. A set of visit sas.com/platform. automatic) facilitates what-if experimenta- metrics based on comparing actual versus tion and more extensive exploration of how planned progress and costs detects devia- To learn more about SAS/OR, download system controls and operating conditions tions from the schedule/budget early in the white papers, view screenshots and see affect key performance metrics. SAS Simula- project, providing a factual basis for targeted other related material, please visit tion Studio can integrate with JMP for exper- corrective action. imental design and input analysis, and with sas.com/or-software. JMP and SAS for source data and analysis of Decision analysis simulation results. Decision trees help structure sequential decision-making processes under uncertain conditions by enabling you to examine and

To contact your local SAS office, please visit:sas.com/offices

SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Copyright © 2018, SAS Institute Inc. All rights reserved. 101429_G91746.1018