The Experimenter is used to define, run and analyze experiments on defined model scenarios. See the Experimenter Example for an example of how to use the Experimenter.
Topics
Variables/Scenarios
To create an experiment run, variables are added to the experimenter. These variables are things in the model that you want to change as part of a given experiment. They may be simple values in labels or global tables or they may be the number of Operators for a given team or the position of a FixedResource object. You create and edit Experimenter Variables in the Scenarios Tab. Each variable is a row in the Scenarios table.
Experiment scenarios are associated with experiment variables. A scenario is a specific configuration of the set of variables that you have defined. Variables must have at least one scenario, which is a numeric value that is assigned to the associated variable. In the image above, the scenarios are Center, All Left, All Right, etc. and specify the x position of Processor2 and Processor 3. When the experimenter runs, it will run a defined number of replications for each scenario.
You can have any number of variables and scenarios in an experiment.
Default Reset Scenario
The default reset scenario option allows you to set up your model based on a given scenario. Select a scenario from the drop down and then hit the Reset button to reset the model. The values assigned to the different variable will be applied to the model.
Performance Measures
Performance measures allow you to get statistical data from your model and use it to assess and compare the results of your scenarios. The statistics may include things like throughput, average wait time, or values from dashboard widgets, etc.
Analyzing Results
Once the experiment is finished (all bars in the Experiment Status window are green) you can analyze the results of your performance measures by clicking the View Results button of the Experiment Run tab. This will open the Performance Measure Results window.
The data in this window allows you to compare the results of different scenarios. There are several options for how to display the data including a Replications Plot, a Frequency Histogram, a Correlation Plot (for examining correlations between multiple performance measures), a Data Summary, and a Raw Data view.
This window also displays the output of each replication and results from dashboard widgets. Performance measure results and dashboard widgets may then be exported to an HTML format for distribution.
