SAS® Simulation Studio
Graphical Modeling and Analysis Software for Discrete Event Simulation
Discrete event simulation is the modeling of dynamic processes where events that affect the processes occur at distinct points in time. Perhaps a large restaurant chain needs to understand the operational characteristics of a restaurant. How many cashiers, cooks, service lines and drink stations are needed? Or maybe a manufacturer is unable to meet goals for units delivered. There are bottlenecks affecting production, but where? Discrete event simulation software can help you build a model that mimics the behavior of complex, real-life systems so you can better understand and optimize them.
How SAS® Can Help
SAS Simulation Studio provides a full set of tools, including a graphical user interface, for building, executing and analyzing the results of discrete event simulation models. It provides extensive modeling and analysis capabilities suitable for both novice and advanced users, and is available as part of SAS/OR® software and as an add-on to JMP®. It also interfaces with JMP so you can easily create and analyze experimental designs as well as determine the most appropriate distribution model for the events to be simulated – a very important factor in discrete event simulation.
Graphical User Interface Elements
The GUI for SAS Simulation Studio uses a hierarchical structure to help you organize your work. The top level is the project – a collection of models and experiments that correspond to each process, or system, being studied. Within each Project window, you can create one or more Model windows where you build models.
Templates of blocks contain components that are used to build your simulation models (via drag-and-drop) in a Model window. The blocks provide an easy way to create and arrange multiple types of entities, such as queues, servers, and routing devices. Also, blocks provide for data input, data generation, numeric and graphical monitoring and reporting, the use of resources, and other functions. You can select groups of blocks in a model, assemble them into "compound blocks" and add the compound blocks to your templates for further use. Both individual and compound blocks may be copied and pasted within and between models and projects.
In each project, you can create one or more Experiment windows, which provide an organized way to initialize, plan and coordinate multiple runs of your simulation models. You can easily specify the start time, end time and number of replications for each version of a model. If you have defined factors (parameters to initialize) and responses (metrics to record) for your model, you can investigate the effects of varying factor values on responses.
SAS Simulation Studio GUI
Design of Experiments and JMP® Integration
Systematically varying inputs through designed experiments is a great way to improve physical processes. However, experimenting with models by varying one factor at a time is inefficient and limits your ability to understand the system. Instead, by carefully choosing settings for multiple experimental factors, you can vary many factors at once and gain more insights with fewer trials.
SAS Simulation Studio, through the Experiment window, supports both manual and automated design of experiments. In manual design of experiments, you create the experimental design points and determine how many replications should be run for each point. This is useful if you need to make a direct comparison of a small number of different versions of the model or if you want to carry out a highly specialized experimental design.
For automated design of experiments, SAS Simulation Studio integrates with JMP. A single command invokes the Custom Designer in JMP to create a set of design points that is automatically passed to SAS Simulation Studio. You can modify or augment this design using either JMP or SAS Simulation Studio. The results of the experiment can be passed directly to JMP for analysis and also can be saved for later analysis with SAS or JMP.
Stationary and Mobile Resources
SAS Simulation Studio provides stationary resources (represented by blocks such as queues, servers and delays) that have fixed locations in the model. It also supports mobile resources, which are a special type of entity created during the simulation run. They can flow through the model just as regular entities do. Mobile resources can carry attributes and are processed by the same blocks that process regular entities, and can be seized by other entities as required. Mobile resource availability levels and their operational status can be controlled with scheduling features in SAS Simulation Studio.
Monitoring the Execution of Discrete Event Simulation Models and Analyzing Simulation Results
In SAS Simulation Studio, several blocks are dedicated to producing graphical analysis of the simulation results, both as the discrete event simulation model runs and at the termination of a run. These displays can be useful when debugging or tuning a model, or for single runs of models. Blocks can also collect data during simulation runs so the data can be stored for later, more extensive analysis. You choose whether to store data by default as SAS data sets or as JMP tables. You can then use either SAS or JMP, respectively, to carry out further analysis.
- Supported for Windows 32-bit and 64-bit.
- Requires JRE 1.6.0 or higher.
- Requires JMP 9 or higher for automated experimental design and input analysis.
Ready to learn more?
Call us at 1-800-727-0025 (US and Canada) or request more information.