Variation experiment comprises a series of Simulation experiments, which are based on a single scenario with varying its one or several parameters.
Variation experiments are used in two main cases:
- You want to see how changing a certain parameter value in a scenario affects the supply chain. Instead of creating multiple scenarios holding different parameters' values, running a Simulation experiment for each of them and comparing the results manually, you can set up a Variation experiment, which will automatically vary the required parameter, run a simulation experiment for each variation and present you with the results that can be easily compared.
- You want to see how a certain stochastic parameter in a scenario affects the supply chain. You can set up a Variation experiment that will repeatedly run a Simulation experiment, produce statistically significant results, calculate statistical data basing on the received data and provide the final result in the form of a table.
You should create at least one scenario to run the Variation experiment. Prior to running the experiment you should specify the scenario to work with, define the variations and select statistics to collect the data for. When you launch the experiment, it passes through a series of iterations, one for each variation. After the experiment is finished, you can compare the values of the statistics collected for each variation to view how the differences in the initial conditions affect the model behavior.
How can we improve this article?