摘要
论文提出了一种基于粒子群的多目标优化算法,该算法采用Pareto支配关系来更新粒子的个体最优值和局部最优值,用存储池保存搜索过程中发现的非支配解;采用聚类算法裁剪非支配解,以保持解的分布性能;采用动态惯性权重法来平衡粒子群对解空间的局部搜索和全局搜索,以提高算法的全局收敛性能。实验结果表明,论文算法是有效的,能有效的求解多种多目标优化问题。
This article presents a Particle Swarm Optimization（PSO） algorithm for muhiobjective optimization problems. PSO is modified by storing nondominated solutions externally,preserves population diversity using the Pareto dominance relationship,and incorporates a clustering procedure to reduce the nondominated set without destroying its characteristics. Several benchmark cases are tested and show that the method can efficiently find multiple Pareto optimal solutions.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第18期40-42,78,共4页
Computer Engineering and Applications
关键词
粒子群优化
多目标优化
演化计算
particle swarm optimization,multi-objective optimization,evolutionary computation
