Function Optimization Using Evolutionary Programming with Self-Adaptive Cultural Algorithms

Self-adaptation can take place at several levels in evolutionary computation system. Here we investigate relative performance of two different self-adaptive versions of Evolutionary Programming(EP). One at the individual level adaptation proposed by Schwefel and Saravanan & Fogel and one at the population level using Cultural Algorithms. The performance of the two versions of self-adaptive EP are then compared to each other for a set of selected unconstrained function optimization problems. For most optimization problems studied here, the pooling of information in the belief space at the population level improves the performance of EP.