Simulation Optimization

Welcome to this FlexSim feature spotlight. In this video we’re going to take a look at the Optimizer, an invaluable tool in finding the optimal configuration for a process. Jordan Johnson, who helped develop the Optimizer, is here with us to answer a few questions and shed some light on how the Optimizer works.

Markus: What makes the latest version of FlexSim’s Optimizer so great?

Jordan: FlexSim incorporates an optimization search engine called OptQuest, made by OptTek Systems. Once a model is ready to be optimized, the OptQuest engine runs simulations to explore the relationships between your model parameters and determine how to optimize your system.

M: Why OptQuest, why not some other optimization package?

J: Well, OptQuest is an industry-trusted optimization package. It uses the latest evolutionary algorithms to search a given design space intelligently and quickly, and can even change search algorithms during an optimization if it needs to. Also, OptQuest is built directly into FlexSim. If we used a standalone optimization program it would be much harder to transfer data to and from OptQuest.

M: What are the most important changes that you made to the Optimizer?

J: The biggest changes I made are the inclusion of the latest OptQuest engine, total integration with the Experimenter, and a complete overhaul to the user interface.

M: What features does the latest OptQuest engine provide?

J: The latest engine allows users to do multi-objective searches. For example, say you want to maximize revenue but minimize cost, and that for your process increasing revenue increases cost. OptQuest is capable of finding the Pareto front, which is the set of optimal solutions formed by those two objectives, allowing you to choose the best optimal solution for your process.

M: Why did you integrate the optimizer with the Experimenter?

J: Well, the Optimizer and the Experimenter do essentially the same thing. The Experimenter allows users to run multiple configurations of the same model and observe the results. The Optimizer does the same thing except that the OptQuest engine generates values for each configuration and uses the results to generate better and better configurations. Since the Experimenter was already multi-threaded, the Optimizer now shares that ability—it can run multiple simulations simultaneously, drastically reducing the overall optimization time.

Thank you for watching. For informative videos on other tools and features in FlexSim, check out the FlexSim Start Page or visit www.flexsim.com.