Abstract:

In order to improve the performance of the gravitational search algorithm (GSA) in solving single objective optimization problems, a new GSA with mixed improved strategy is proposed. According to the evolution situation, the individual evolution rate strategy is proposed which is applied to enhance the rate of convergence. And a kind of variation strategy is adopted to balance the ability of global searching and local exploiting which avoid the possibility that the population fall into local optimum. Simulation experimental results on benchmark functions show that the GSA with mixed strategy has a good performance in avoiding premature convergence. Compared with GSA and other improved GSA, the new algorithm has a good performance not only in convergence rate but also in convergence precision.