摘要
将基于Pareto支配关系的局部下山算子和差分算子引入免疫多目标优化算法之中,提出了一种求解多目标问题的Memetic免疫优化算法(Memetic immune algorithm for multiobjective optimization,简称MIAMO).该算法利用种群中抗体在决策空间上的位置关系设计了两种有效的启发式局部搜索策略,提高了免疫多目标优化算法的求解效率.仿真实验结果表明,MIAMO与其他4种有效的多目标优化算法相比,不仅在求得Pareto最优解集的逼近性、均匀性和宽广性上有明显优势,而且算法的收敛速度与免疫多目标优化算法相比明显加快.
A Memetic immune algorithm for multiobjective optimization （MIAMO） is proposed by introducing two types of local search operators. These operators are the Pareto dominance based descent operator and the differential evolution based operator. In MIAMO, the position and spatial relations between antibodies in the decision space are used to design the two heuristic local searching strategies with the assistance of which the efficiency of the immune multiobjective optimization algorithm can be improved. Experimental results indicate that, comparing with the other four efficient multiobjective optimization algorithms, the MIAMO performs better in approximation, uniformity, and coverage. It converges significantly faster than the immune multiobjective optimization algorithm.
出处
《软件学报》
EI
CSCD
北大核心
2013年第7期1529-1544,共16页
Journal of Software
基金
国家重点基础研究发展计划(973)(2013CB329402)
国家自然科学基金(61072139
61001202
61272279)
高等学校学科创新引智计划(B07048)
教育部长江学者和创新团队发展计划(IRT1170)
国家教育部博士点新教师基金(20090203120016
20100203120008)
中国博士后科学基金(20090461283)
陕西省自然科学基础研究计划(2011JQ8010
2009JQ8015)
关键词
多目标优化
人工免疫算法
MEMETIC算法
multiobjective optimization
artificial immune algorithm
Memetic algorithm
