基于孤立点策略的多目标粒子群算法  

Multi-Objective Particle Swarm Optimization Based on Acnode

作　　者：陈刚[1] 张绍德[1] 武因培[1] 

机构地区：[1]安徽工业大学电气信息学院,安徽马鞍山243002

出　　处：《工业控制计算机》2012年第12期60-62,共3页Industrial Control Computer

摘　　要：为了提高和改善多目标粒子群算法(MOPSO)的收敛速度和Pareto前端的分布性,基于传统MOPSO算法,在速度更新中引入孤立点,用以增强粒子的全局探索能力和提高粒子的多样性。其次,提出并采用一种源于空间矢量内积的方法更新外部非支配解集。最后进行算法的Matlab仿真实验。通过对三组典型的基准测试函数计算表明,该算法不仅具有较好的收敛速度和分布性能,且算法复杂度也不高。In order to enhance and improve the MOPSO＇s convergence and diversity of solutions in Pareto optimal front,the con- cept of acnode is used in the velocity＇s update formula,which is based on the traditional MOPSO algorithm to enhance the global exploring.Secondly,a simplified approach on the basis of vector inner product is proposed and used in the update of external archive of non-dominated solutions.The last is the simulation experiment of the algorithm by Matlab.The experi- mental results indicate that the proposed algorithm has better performance not only in convergence and diversity,but also the complexity.

关 键 词：多目标优化 粒子群算法 PARETO支配 孤立点 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]