摘要
粒子群优化(PSO)算法在求解复杂的多峰问题时极易陷入局部最优解,通过分析种群多样性与局部最优解间的关系,提出一种基于动态邻居拓扑结构的粒子群算法。该算法在运行过程中,每间隔若干代,根据粒子间的距离更新每个粒子的邻居,该策略增加种群的多样性,进而提升粒子跳出局部最优解的能力。实验结果表明,该算法比其他PSO算法具有更好的性能。
Particle Swarm Optimization(PSO) algorithms may easily get trapped in a local optimum,when it solves complex multimodal problems,by analyzing the relationship between swarm diversity and local optima,this paper presents an improved particle swarm optimizer based on dynamic neighbor topology(DPSO for short).In DPSO,the neighbor of each particle is dynamically constructed at several iterations,which increases the swarm diversity and improves the ability to escape from local optima.In benchmark functions,the DPSO algorithm achieves better solutions than other PSO algorithms.
出处
《计算机工程》
CAS
CSCD
北大核心
2011年第8期210-212,共3页
Computer Engineering
基金
山东省科技攻关计划基金资助项目(2009GG10001008)
广东省自然科学基金资助项目(9451806001002294)
深港创新圈基金资助项目(200810220137A)
贵州教育厅社科基金资助项目(0705204)
关键词
粒子群优化
动态邻居
种群多样性
函数评价
Particle Swarm Optimization(PSO); dynamic neighbor; swarm diversity; function evaluations;