摘要
为了改善粒子群算法易早熟收敛、精度低等缺点,提出一种多尺度协同变异的粒子群优化算法,并证明了该算法以概率1收敛到全局最优解.算法采用多尺度高斯变异机制实现局部解逃逸.在算法初期阶段,利用大尺度变异及均匀变异算子实现全局最优解空间的快速定位;随着适应值的提升,变异尺度随之降低;最终在算法后期阶段,利用小尺度变异算子完成局部精确解空间的搜索.将算法应用6个典型复杂函数优化问题,并同其他带变异操作的PSO算法比较,结果表明,该算法在收敛速度及稳定性上有显著提高.
To deal with the problem of premature convergence and low precision of the traditional particle swarm optimization algorithm, a particle swarm optimization (PSO) algorithm based on multi-scale cooperative mutation, is proposed, which is guaranteed to converge to the global optimal solution with probability one. The special multi-scale Gaussian mutation operators are introduced to make the particles explore the search space more efficiently. The large-scale mutation operators can be utilized to quickly locate the global optimal space during early evolution. The small-scale mutation operators, which are gradually reduced according to the change of the fitness value can implement the accuracy of the solution at the late evolution. The proposed method is applied to six typical complex function optimization problems, and the comparison of the performance of the proposed method with other PSO algorithms is experimented. The results show that the proposed method can effectively speed up the convergence and improve the stability.
出处
《软件学报》
EI
CSCD
北大核心
2012年第7期1805-1815,共11页
Journal of Software
基金
国家自然科学基金(61074076)
国家博士后科学基金(20090450118)
中国博士点新教师基金(20092304120017)
黑龙江省博士后科学基金(LBH-Z08227)
关键词
粒子群算法
早熟收敛
多尺度
协同变异
适应度
particle swarm optimization
premature convergence
multi-scale
cooperative mutation
ftness