改进的约束多目标粒子群算法  

Improved constrained multi-objective particle swarm optimization algorithm

作　　者：凌海风[1,2] 周献中[1] 江勋林[2] 萧毅鸿[1] 

机构地区：[1]南京大学工程管理学院,南京210093 [2]解放军理工大学工程兵工程学院,南京210007

出　　处：《计算机应用》2012年第5期1320-1324,共5页journal of Computer Applications

基　　金：国家自然科学基金资助项目(90718036)

摘　　要：在约束优化问题搜索空间分析的基础上提出了一种改进的约束多目标粒子群算法(CMOPSO)。提出一种动态ε不可行度许可约束支配关系作为主要约束的处理方法,提高了算法的边缘搜索能力和跨越非联通可行区域的能力。设计了一种新的密集距离度量方法用于外部档案维护,提高了算法的效率;提出了新的全局向导选取策略,使算法获得了更好的收敛性和多样性。数值仿真实验结果表明约束多目标粒子群算法算法可得到分布性、均匀性及逼近性都较好的Pareto最优解。An improved Multiple Objective Particle Swarm Optimization（MOPSO） algorithm for solving constrained multi- objective optimization problems （CMOPSO） was proposed based on the analysis of the characteristics of the multi-objective search space. A processing method taking dynamic ε unfeasible degree allowable constraint dominance relation as the main constraint was brought forward in this paper, which aimed to improve the algorithm＇s ability of edge searching and crossing unconnected feasible regions. A simple density measuring method was put forward for external archive maintenance, which intended to improve the efficiency of the algorithm. A new global guide selection strategy was put forward, which brought better convergence and diversity to the algorithm. The computer simulation results show that the CMOPS0 algorithm can find a sufficient number of Pareto ootimal solutions that have better distribution, uniformity, and approachability.

关 键 词：多目标优化 多目标粒子群 距离量度 档案维护 全局向导选取 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]