欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
首先分析DRS（dominance resistant solutions）多目标优化问题的特点，证明基于Pareto-支配关系的多目标优化问题算法求解该类问题很难收敛．然后，提出一种新的基于ε-支配关系的进化算法-ε-支配进化算法（EDMOEA），给出该算法框架和详细流程．最后，将ε-支配进化算法和NSGA-Ⅱ算法应用于求解一组典型的DRS多目标优化问题和常用的多目标优化测试问题，基于算法的收敛性和Pareto最优解集分布性进行评价和比较分析，表明ε-支配进化算法的有效性．
英文摘要：
The DRS（ dominance resistant solutions）existing in the multi-objective optimization problems （MOP） is characterized, and it is proved that the evolutionary algorithms fail to find the true Pareto fronts for the DRS MOP. A new evolutionary algorithm based on ε-dominance relationship, called the ε-dominance MOEA （EDMOEA） , is proposed to solve the DRS MOP, and algorithms procedure is designed. Then it is proved that the EDMOEA is able to fix the DRS issue of the DRS MOP. Finally, a set of DRS MOP and other MOP commonly used in literatures are adopted to test and compare the EDMOEA and the NSGA-Ⅱ, and the experiment results illustrate that the EDMOEA is superior to the NSGA-Ⅱ on the MOP of both the DRS and others.
同期刊论文项目
同项目期刊论文
期刊信息
位置：