引力搜索算法中粒子记忆性改进的研究

Analysis on improvement of particle memory in gravitational search algorithm

针对引力搜索算法(GSA)对一些复杂问题的搜索精度不高的问题,特别是高维函数优化性能不佳、优化过程容易出现早熟的现象,因此考虑将粒子群优化(PSO)算法中关于局部最优解和全局最优解的概念引入引力搜索算法中,对引力搜索算法中粒子的记忆性进行改进,这样使得粒子的进化不仅受空间中其他粒子的影响,还受到自身记忆的约束,以此来提高算法的搜索能力。通过对选用的10个基准函数测试,证明了该方法的有效性。

As the gravitational search algorithm plays bad performance in search accuracy of the complex issues, especially the poor search quality of standard Gravitational Search Algorithm （GSA） in the high dimensional function optimization. It is easy to get into premature convergence in the optimization process. Therefore, the idea of the particle swarm optimization algorithm was introduced to gravitational search algorithm, which was used to improve the memory of particles. The particle evolution is not only influenced by other particles in the space, but also by its own memory constraint, which is used to improve the ability of exploitation. The test of the 10 benchmark functions confirms the validity of the method.