Using Optimal Control Principles to Adapt Evolution Strategies
Burcin Aktan, G.W. Greenwood, Molly H. Shor · 2006
It has long been known that adapting the strategy parameters in an evolution strategy during a run improves the quality of the search. Several adaption schemes have been proposed in the past. In this paper we present a new adaption scheme based on optimal control principles. Our preliminary results indicate this new method is quite effective. Results are included to show how the adaption scheme works on a multidimensional sphere model and an Ackley's function fitness landscape.