摘要
为了提高高维多目标优化算法的收敛性和解集的分布性,提出了一种基于降维结果一致性检验结合模糊支配的高维多目标进化算法COPCA-FDNSGA-II。在NSGA-II的主成分分析算法模型基础上,利用模糊理论对算法中的支配关系进行改进,针对信息不完备及伪解干扰的情况,在进化前期,对用模糊支配优化算法得到的非支配解进行主成分分析,去除冗余目标,并对降维结果进行一致性检验。将该算法与其他算法在测试函数上进行对比试验,结果表明,该算法在收敛性和分布性上具有明显优势。
In order to solve the high-dimensional multi-objective optimization problem caused by increasing the target to bring selection pressure drop and the problem of difficult to converge to the Pareto optimal front. The paper proposes a multi-objective evolutionary algorithm(COPCA-FDNSGA-II) which is based on dimension reduction results consistency check with fuzzy dominance. On the model of algorithms based on the principal component NSGA-II, the dominance relation of the algorithm is improved by using the fuzzy theory, in the condition of incomplete information and pseudo solution, in the early stage of evolution, the non-dominated solutions obtained from the fuzzy control optimization algorithm are used to analyze the principal components, to remove the redundant targets, and to test the consistency of the reduced dimensional results. Finally, the proposed algorithm is compared with other algorithms on benchmark test problems. Simulation results show that the algorithm has obvious advantages in convergence and distribution.
作者
陈立芳
祁荣宾
CHEN Li-fang;QI Rong-bin(Key Laboratory of Advanced Control and Optimization for Chemical Processes,Ministry of Education,East China University of Science and Technology,Shanghai 200237,China)
出处
《控制工程》
CSCD
北大核心
2018年第12期2224-2231,共8页
Control Engineering of China
基金
国家自然科学基金面上项目(21276078)
国家自然科学基金青年项目(61403141)
上海市“科技创新行动计划”研发平台建设项目(13DZ2295300)
上海市自然科学基金(15ZR1408900)
