New Fuzzy Self-Tuning Genetic Algorithm

To overcome the drawback of“early maturing”of the classical genetic algorithm (GA) and improve its global convergency and convergency speed, a new fuzzy self-tuning genetic algorithm was proposed. In the new algorithm, the overall population is divided into several sub- populations and each sub-population has its own operators. Fuzzy reasoning is applied to give effective operators more opportunity to search under the condition of keeping the overall population size unchanged. The fuzzy reasoning can sense the contributions of these operators and then decides their population size. Simulation result of function optimization shows that with the proposed algorithm, the phenomenon of the“early maturing”can be effectively overcome, and a satisfying optimization result can be obtained.