Improvement and Application of Multi-objectiveParticle Swarm Optimization

• Main Authors: Shih-Yuan Chiu, 丘世元
• Other Authors: Tzung-Ying Sun
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/jpdwxm

碩士 === 國立東華大學 === 電機工程學系 === 95 === Recently, there have been several studies to extend Particle Swarm Optimization (PSO) algorithm to handle multi-objective problems (MOPs). This thesis presents the density based proportional distribution method for multi-objective particle swarm optimization (PD-MOPSO) which includes the external archive, cluster, disturbance and proportional distribution scheme. These four schemes are used to retain the non-dominate solution set, increase the difference of solutions and faster the calculate speed of algorithm, increase the search depth and enhance the maximum spread of our algorithm. Six benchmarks and a multi-objective task schedule application were adopted to verify the proposed method, and compare with Sigma method, NSGA-II algorithm and SPEA2 algorithm. From the results, the proposed method is better than other three in General Distance, Diversity Metric, Maximum spread and Error Ratio can be observed. The proposed method is more robust than other three methods in solving multi-objective problems (MOPs).