A Hybrid Genetic Algorithm with Fitness Sharing Based on Rough Sets Theory

摘要

This paper presents a new method integrated sharing genetic algorithm (SGA), rough sets theory (RST) and bit-climbing algorithm for multimodal function optimization. We apply the SGA to complete the globe search and form niches which indicate the promising locations. Then, we utilize the strong qualitative analysis ability of rough sets to identify these niches. Finally, the bit-climbing algorithm is used to complete the local search and refine the solution in each of niches. The experiment has proved that using this approach to solve the multimodal function optimization is efficient both in robustness and in accuracy.