Levels of evolution for control systems
John J. Grefenstette · Institution of Engineering and Technology eBooks · 1997
Evolutionary algorithms (EAs) are general purpose search and learning methods that can be applied to a variety of problems relating to control systems. This chapter focuses on the range of representation levels at which evolutionary algorithms can be applied to control systems, including evolving control parameters, evolving complex control structures and evolving control rules. The discussion also outlines the use of evolutionary algorithms for testing intelligent control systems. In this case, the EA is used to identify weaknesses in a control system by searching for challenging test cases.