计算机应用与软件2013，Vol.30Issue(2)：58-64,124,8.DOI:10.3969/j.issn.1000-386x.2013.02.015
基于正交设计的自适应ε占优MOEA/D算法研究
RESEARCH ON ADAPTIVE EPSILON-DOMINATION BASED ORTHOGONAL MOEA/D FOR MULTI-OBJECTIVE OPTIMIZATION
摘要
Abstract
MOEA/D is a simple, effective multi-objective optimization algorithm, however, when MOEA/D updates subproblems, it may lost some good individuals, and thus lead to reducing the convergence speed of the algorithm. In this paper, in order to remedy this pitfall, an improved MOEA/D based on the orthogonal design adaptive epsilon-domination is proposed. The new algorithm can be characterized as; (1) Using the orthogonal experimental design with quantization to initialize the population. (2) Proposing an improved epsilon-domination which can be self-adaption and Using it to update Archive population which retain the obtained non-dominated solutions. ( 3 ) The elitist strategy is introduced into MOEA/D, speeding up the convergence speed. The simulation results show that the new algorithm improves the speed of original MOEA/D' s convergence and the distribution of non-inferior solutions in multi-objective problem.关键词
MOEA/D/自适应ε占优/正交实验/多目标演化算法Key words
MOEA/D/Adaptive epsilon-domination/Orthogonal design/Multi-objective evolutionary algorithm分类
信息技术与安全科学引用本文复制引用
周攀,张冬梅,龚文引,李阳,刘凯伟..基于正交设计的自适应ε占优MOEA/D算法研究[J].计算机应用与软件,2013,30(2):58-64,124,8.基金项目
国家自然科学基金项目(40972206,61075063) （40972206,61075063）
中央高校基本科研业务费专项资金项目(1323520909). （1323520909）
