In the simplest terms, simulation is a method for describing the probability distribution of an outcome variable, given a set of input variables. Sometimes those input variables include one or more decision variables, and the ultimate goal is to determine the best values for those inputs. But simulation, by itself, offers no assistance in identifying an *optimal*, or even good, set of decisions.

Ideally, we would like to marry optimization's power to identify the best decision variables with simulation's power to describe outcome distributions. Unfortunately, the optimization approaches that we covered in Chapters 8–12 are based on the premise that the objective function can be measured deterministically. But in simulation models the objective function is not deterministic; instead, it is the expected value or some other function of a random variable. In many cases, Analytic Solver Platform can optimize these models as well. However, in the context of simulation models, a number of issues arise that do not arise when optimizing deterministic models.

We begin with a simple but common example of a simulation problem containing a single decision variable, and we illustrate how to optimize such models using simulation sensitivity. We use the same example to introduce the optimization capabilities in Analytic Solver Platform as they apply to simulation models. We then show how to use Solver in more complex problems involving more than one ...

Start Free Trial

No credit card required