A dynamic sharing scheme-based multimodal niche genetic algorithm
Qing Hua Lu, Changyong Liang, Enqiao Zhang · 2008
A dynamic sharing scheme-based multimodal niche genetic algorithm is proposed. In the algorithm, niches are identified by a method of dynamic boundary identification. After the identification, individual sharing fitness is calculated on the basis of the identified niches, and the sharing scheme of dynamic niche is applied in the algorithm. Meanwhile, a concept of crossover selection rate is introduced, which divides the crossover into two operators, the in-niche crossover and the out-niche crossover. The two crossover operators can not only maintain the population’s diversity but also make individuals converge onto different peaks. Experiments show that the algorithm can solve multimodal function optimization problems with fast convergence rate, high precision and little computation.