摘要
和许多经典的群智能算法一样,万有引力搜索算法在解决很多优化问题的时候,容易陷入局部最优解并且收敛精度不高。针对这样的情况,提出一种基于变异策略和反向评估机制的改进万有引力搜索算法。该算法通过引入反向评估机制和变异策略,显著地提高了万有引力搜索算法中粒子的局部寻优能力和全局探索能力。通过对三个标准测试函数进行仿真实验,表明其与基本的万有引力搜索算法、传统粒子群算法相比,提出的基于变异策略和反向评估机制的改进万有引力搜索算法在函数优化问题上具有更好的优化性能。
mLike many of the classic swarm intelligence algorithm,gravitational search algorithm is easy to fall into local optimal solution and convergence precision that is not high in solving many optimization problems. In this situation,this paper proposes an improved universal gravitation search algorithm based on reverse evaluation mechanism. The local optimization ability and the global exploration ability of the particle in the gravitational search algorithm are significantly improved by introducing the reverse evaluation mechanism and variation strategy. Simulation experiment of three standard test functions,show that improvement of universal gravitation search algorithm based on reverse evaluation mechanism on this paper has better functions in function optimization problem,compared with basic gravity search algorithm and the traditional particle swarm algorithm.
出处
《武汉轻工大学学报》
2016年第3期60-63,76,共5页
Journal of Wuhan Polytechnic University
关键词
反向评估机制
变异策略
万有引力搜索算法
reverse evaluation mechanism
mutation strategy
gravitation search algorithm
