Logarithmic adaption crowding genetic algorithm for multimodal function optimization
Liu Wenta · Journal of Computer Applications · 2014
Crowding genetic algorithm can obtain multiple optima of multimodal functions, but it has low efficiency, and cannot get a higher precision in limited iterations. In order to obtain all optima of the multimodal function quickly, the crowding genetic algorithm based on logarithmic adaption was presented combined with niche crowding genetic and climbing operators. The algorithm computed the distance values of climbing operators by logarithmic adaption according to the iterations, which made the population maintain genetic diversity in the process. According to the experiments and comparative analysis of several one-dimensional and two-dimensional multimodal functions, the test results show that the algorithm can ensure both the solution accuracy rate and the convergence speed in the limited iterations, and obtain all optimal solutions more stably. It is proved to be an effective algorithm for the multimodal function problems.