Use the experimenter tool in FlexSim
Use the experimenter tool to run and analyze multiple simulation scenarios for your model.
Transcript
00:03
In FlexSim, the Experimenter tool automates the process of running different inputs within your model and collecting the results.
00:11
This allows you to run, analyze, and compare multiple simulation scenarios easily,
00:18
enabling informed decision-making to optimize system performance.
00:22
To prepare for using the Experimenter, first you need to create a model that is set up with inputs,
00:28
or parameters in a parameter table, as well as performance measures to gather results for specific aspects of your simulation.
00:35
Among other important terms, a scenario is a specific set of parameter values under which a simulation is run,
00:42
and a replication is one instance of running a scenario.
00:46
You decide how many replications per scenario you wish to run.
00:51
Also, a job defines a strategy for generating Scenarios within the Experimenter tool.
00:57
When an Experimenter job runs, the model runs once for each replication of a scenario and records the results in a results database file.
01:06
In this example, a model, parameter table, and performance measurement table are already created to demonstrate the Experimenter.
01:15
Here, in the parameter table, Parameter1 controls the location of Processor1, and Parameter2 controls the location of Processor2.
01:26
Notice that if you adjust the Value for these parameters—here, to “7” and “23”—and Reset the Model,
01:34
then the location of each Processor changes in the Model.
01:38
The Performance Measure shown here is connected to the sink and is set to measure the throughput of the entire model.
01:44
Now, to open the Experimenter, in the Toolbar, select Statistics > Experimenter.
01:55
In the Experimenter dialog, you can see that Experiment1 is set up as the default.
02:01
Click Add.
02:05
You can add as many jobs as you wish, and there are three types of jobs that you can select from:
02:11
An Experiment job allows you to specify a set of scenarios and how many replications to run of each.
02:18
For an Optimization job, you specify an objective,
02:22
and the Experimenter automatically generates scenarios that attempt to optimize that objective.
02:28
Finally, a Range Based job allows you to specify a set of parameters and a range of values for each,
02:34
and the Experimenter automatically generates scenarios for each permutation of the parameter value.
02:40
In this case, select Experiment to return to the default Experiment1 job.
02:46
To select the parameters you want to adjust in this job, expand Parameters, and in this case, select both parameters.
02:56
Then, use the arrows to adjust the number of Scenarios.
03:02
Here, add three more scenarios for a total of four.
03:06
Next, adjust the parameter values for each scenario.
03:11
In this example, Parameter1 can have a location value between 7 and 13,
03:17
and Parameter2 can have a location value between 17 and 23.
03:22
With that in mind, change the value of Parameter1 for Scenarios 3 and 4 to “13”.
03:29
Then, change the value of Parameter2 for Scenarios 1 and 3 to “17”, so that each Scenario has a different combination of values.
03:39
On the Jobs tab, you can also adjust the Name, Warmup Time, and Stop Time, as well as the Replications per Scenario.
03:49
Here, leave this set to “5” and all the other defaults.
03:53
Click the Run tab, and then click Run to begin the experiment.
03:60
In the table shown, each scenario is a row, each replication of a scenario is a box,
04:06
and a key indicates the meaning of each color—proposed, submitted, running, recording, or complete.
04:22
In this case, all scenarios are green and complete.
04:26
The results database file path is listed at the top of the Experimenter dialog.
04:31
Use Default Path is currently selected, which uses the file path of the model.
04:37
You can also deselect this option to specify a custom path.
04:42
Click View Results to see the results based on the performance measurement that you set up.
04:47
You can also choose to view the Data, Box Plots, or Mean.
04:53
Expand Replications Plot to view alternate plot options—
04:58
for example, you can select a Frequency Histogram or Data Summary to view your data in different ways.
05:05
Click Close to close the Performance Measure Results.
05:09
Back in the Experimenter dialog, to delete the file, click Delete Results File.
05:16
To add additional jobs, simply return to the Jobs tab and follow the previous process.
05:23
You can create multiple jobs and multiple job types.
05:27
The Experimenter is a powerful tool that allows you to run and analyze multiple simulation scenarios,
05:33
optimizing the outcomes and performance of your model.
00:03
In FlexSim, the Experimenter tool automates the process of running different inputs within your model and collecting the results.
00:11
This allows you to run, analyze, and compare multiple simulation scenarios easily,
00:18
enabling informed decision-making to optimize system performance.
00:22
To prepare for using the Experimenter, first you need to create a model that is set up with inputs,
00:28
or parameters in a parameter table, as well as performance measures to gather results for specific aspects of your simulation.
00:35
Among other important terms, a scenario is a specific set of parameter values under which a simulation is run,
00:42
and a replication is one instance of running a scenario.
00:46
You decide how many replications per scenario you wish to run.
00:51
Also, a job defines a strategy for generating Scenarios within the Experimenter tool.
00:57
When an Experimenter job runs, the model runs once for each replication of a scenario and records the results in a results database file.
01:06
In this example, a model, parameter table, and performance measurement table are already created to demonstrate the Experimenter.
01:15
Here, in the parameter table, Parameter1 controls the location of Processor1, and Parameter2 controls the location of Processor2.
01:26
Notice that if you adjust the Value for these parameters—here, to “7” and “23”—and Reset the Model,
01:34
then the location of each Processor changes in the Model.
01:38
The Performance Measure shown here is connected to the sink and is set to measure the throughput of the entire model.
01:44
Now, to open the Experimenter, in the Toolbar, select Statistics > Experimenter.
01:55
In the Experimenter dialog, you can see that Experiment1 is set up as the default.
02:01
Click Add.
02:05
You can add as many jobs as you wish, and there are three types of jobs that you can select from:
02:11
An Experiment job allows you to specify a set of scenarios and how many replications to run of each.
02:18
For an Optimization job, you specify an objective,
02:22
and the Experimenter automatically generates scenarios that attempt to optimize that objective.
02:28
Finally, a Range Based job allows you to specify a set of parameters and a range of values for each,
02:34
and the Experimenter automatically generates scenarios for each permutation of the parameter value.
02:40
In this case, select Experiment to return to the default Experiment1 job.
02:46
To select the parameters you want to adjust in this job, expand Parameters, and in this case, select both parameters.
02:56
Then, use the arrows to adjust the number of Scenarios.
03:02
Here, add three more scenarios for a total of four.
03:06
Next, adjust the parameter values for each scenario.
03:11
In this example, Parameter1 can have a location value between 7 and 13,
03:17
and Parameter2 can have a location value between 17 and 23.
03:22
With that in mind, change the value of Parameter1 for Scenarios 3 and 4 to “13”.
03:29
Then, change the value of Parameter2 for Scenarios 1 and 3 to “17”, so that each Scenario has a different combination of values.
03:39
On the Jobs tab, you can also adjust the Name, Warmup Time, and Stop Time, as well as the Replications per Scenario.
03:49
Here, leave this set to “5” and all the other defaults.
03:53
Click the Run tab, and then click Run to begin the experiment.
03:60
In the table shown, each scenario is a row, each replication of a scenario is a box,
04:06
and a key indicates the meaning of each color—proposed, submitted, running, recording, or complete.
04:22
In this case, all scenarios are green and complete.
04:26
The results database file path is listed at the top of the Experimenter dialog.
04:31
Use Default Path is currently selected, which uses the file path of the model.
04:37
You can also deselect this option to specify a custom path.
04:42
Click View Results to see the results based on the performance measurement that you set up.
04:47
You can also choose to view the Data, Box Plots, or Mean.
04:53
Expand Replications Plot to view alternate plot options—
04:58
for example, you can select a Frequency Histogram or Data Summary to view your data in different ways.
05:05
Click Close to close the Performance Measure Results.
05:09
Back in the Experimenter dialog, to delete the file, click Delete Results File.
05:16
To add additional jobs, simply return to the Jobs tab and follow the previous process.
05:23
You can create multiple jobs and multiple job types.
05:27
The Experimenter is a powerful tool that allows you to run and analyze multiple simulation scenarios,
05:33
optimizing the outcomes and performance of your model.
