The application of one improved niche genetic algorithm for Elman recurrent neural networks
Zhiyong Liu, Mandan Liu, Feng Qian · 2004
During optimize the weights of Elman recurrent neural networks by genetic algorithm, if use simple genetic algorithm, two problems appear: constringency too early and the speed of search too slow in the anaphase. Aim at these problems, an improved Niche genetic algorithm is presented in this paper. Simulation results show that this algorithm can achieve perfect effect.