Sholtes, Robert M., "Optimizing System Behavior using Genetic Algorithms", 1994
ua435
This paper explores the use of geneticalgorithms (GAs) for optimizing system dynamics models. System dynamics offersa unique and powerful approach to identifying the most successful policies formanaging complex problems. Unfortunately, policy makers too often avoid the useof models because of high level of experience required to operate the modelsand the time and expense which results from trail and error testing of amultitude of policy options in order to discover the best policies. The role ofsystem dynamics models as decision-makers tools would be greatly strengthenedif model users could simply identify the goals for the system being modeled andhave the system dynamics model identify the best management actions. Current analysisand optimization techniques used with system dynamics models are not capable ofautomatically determining which policies most nearly produce the desired systembehavior. One emerging optimization technique, (GAs), offers great promise inautomating the identification of the best policies for selected system goals.