A Study on Genetic Algorithm for Multiobjective Optimization Problem

• Main Authors: Wei-ting Lee, 李韋廷
• Other Authors: Wei-mei Chen
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/dv7jqj

碩士 === 國立臺灣科技大學 === 電子工程系 === 95 === An efficient genetic algorithm for the multiobjective optimization problems is proposed. To reduce the computational cost, a variant of k-d tree which is called marked k-d tree is used in our approach. And we also use the unconstrained archive to preserve all non-dominated solutions. Finally, we compare the performance of our proposed algorithm with those of using C metric on the final Pareto set for simple test problem and 0/1 knapsack problem. Our experiments demonstrate that the algorithm we proposed outperforms the other popular multiobjective genetic algorithm, especially for the higher dimensional cases.