An improved gradient-based niching genetic algorithm and its application

A new gradient-based niching genetic algorithm is proposed in this paper to solve the problems of premature convergence and slower searching ability near a point in conventional genetic algorithm. The best point of genetic algorithm is selected as the start point of gradient to improve the ability of finding the local best and restricted competition selection niche algorithm is used to ensure the population diversity. At last the proposed method is applied to estimate the parameters of nonlinear system. Results show that the proposed algorithm can not only improve the ability of finding the local best but also solves the problems of premature convergence.