Chapter 9 covered simulation output analysis for obtaining accurate results from a single simulation model. This can generate descriptive analytics if we are modelling the current state of a system that already exists or predictive analytics if we are modelling a new state of a system or a system that does not exist. This chapter covers comparing the results across two or more versions of our simulation model. Scenario analysis (also termed what-if analysis) involves changing input parameters or the model design and observing the results. This generates predictive analytics, the ability to predict future performance to help plan for the future, and is the type of analysis most associated with simulation. We can also define targets for output measures and run the simulation with varying input parameters (scenarios) in order to meet this goal. This generates prescriptive analytics in that the simulation recommends a choice of action to reach a goal from predictions of future performance.

If we have many combinations of input parameters to explore, then we need to search the solution space (the number of possible scenarios required). If we are conducting predictive analytics, we can search the solution space to find the best options using approaches such as experimental design. If we are conducting prescriptive analytics, we can use approaches such as multi-comparison or optimisation software to find the best option for our target performance. We may identify a number of potential best options, and in this case, we may analyse them in more detail using scenario analysis.