一种基于粒子群算法的改进多目标文化算法  

An improved multi-objective cultural algorithm based on particle swarm optimization

作　　者：吴亚丽[1] 徐丽青[1] 

机构地区：[1]西安理工大学自动化与信息工程学院,西安710048

出　　处：《控制与决策》2012年第8期1127-1132,共6页Control and Decision

基　　金：陕西省自然科学基础研究计划项目(2010JQ8006);陕西省教育厅科学研究专项基金项目(2010JK711)

摘　　要：提出一种基于粒子群算法的改进多目标文化算法并用于求解多目标优化问题.算法中群体空间采用多目标粒子群优化算法进行演化;信念空间通过对形势知识、规范化知识和历史知识的重新定义使之符合多目标优化问题;信念空间和群体空间的交互通过自适应的接受操作和影响操作来实现.若干多目标标准测试函数的仿真结果表明,改进多目标文化算法能够在保持Pareto解集多样性的同时具有较好的均匀性和收敛性.An improved multi-objective cultural algorithm based on particle swarm optimization（PSO-IMOCA） is proposed to solve multi-objective optimization problem. Population space evolves with the improved multi-objective particle swarm optimization strategy. Three kinds of knowledge, situational knowledge, normative knowledge and history knowledge, are redefined to accordance with the solution of multi-objective problem in belief space. The interaction between belief space and population space is implemented by the adaptive accept function and influence function. Simulation results of the benchmark test functions show that the improved multi-objective cultural algorithm can possess good uniformity and convergence as well as maintain the diversity of Pareto optimal solution.

关 键 词：多目标优化 文化算法 粒子群算法 PARETO最优前沿 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]