Experiment Parameters

The following parameters can be varied in an experiment:

The experiment parameters define how your model is executed during a simulation run. These settings control aspects such as which parameter set is active, how long the experiment runs, and what results are generated. Adjusting these values allows you to tailor each experiment to the scenario you want to evaluate, whether you are testing steady-state performance, exploring variability, or comparing alternative configurations. The table below describes each available parameter.

Parameter Description Applicable Experiments
Name The name of the experiment All
Seed The initial random number value. This can be used with the same parameters to repeat an experiment and have the same exact results given the same conditions. All
Duration The quantity and units of time to simulate in the experiment. All
Experiment Type The type of experiment to run. This can be single run, multi run, buffer tradeoff, rate, or gain/loss. These are explained in further detail in the table below. Some of these options may not be visible if the correct conditions are not met, for example you cannot create a new gain/loss experiment without interrupts to examine. Selected here
Replications The quantity of times to repeat the experiment. Any except single run
Min Capacity The minimum capacity allowable for a buffer of interest. Buffer tradeoff only
Max Capacity The maximum capacity allowable for a buffer of interest. Buffer tradeoff only
Step Size The incremental difference between each simulation in a buffer tradeoff experiment. Buffer tradeoff only
Experiment Buffer The buffer to be varied in a buffer tradeoff experiment. Buffer tradeoff only
Min Limit % The minimum percentage of the specified rate to test in a rate experiment. Rate only
Max Limit % The maximum percentage of the specified rate to test in a rate experiment. Rate only
Step Size % The incremental difference in tested rates as a percentage in a rate expiriment. Rate only
Interrupt Study Which interrupts to consider in the experiment. This can be all nodes, none, or single node. Another selection will appear titled "Node" when single node is selected. This is used to select the node to be considered Rate only
Limit Node Which node will be limited by the percentages specified in a rate experiment. This can be all nodes or any individual node within the model. Rate only

Different experiment types are designed to answer different kinds of questions; from examining a single operating scenario to comparing multiple runs or analyzing tradeoffs. The following table outlines the available experiment types and when each is most useful, helping you select the approach that best matches your analysis goals.

Experiment Type Description
Single Run A single simulation of the system with given parameters.. Useful for determining a lot of data about a model over a specified time period.
Multi Run Multiple single runs with the same settings, but different initial seeds. Useful for assessing the variability or volatility of a model.
Buffer Tradeoff Asseses the impact of a specific buffer at multiple incremental values between a minimum and maximum. Useful for assessing the value of buffers as well as the impact on the model when a buffer's capacity is increased or decreased.
Rate Examines the impact of varied rate limits on efficiency and production. This can examine all rates or a specific rate of interest within the model.
Gain/Loss Determines the impact of various interrupts on the model overall. This can be incredibly valuable to find real, sometimes unintuitive, areas to focus on improvements.