An Alternative Preference Relation to Deal with Many-Objective Optimization Problems
An Alternative Preference Relation to Deal with Many-Objective Optimization Problems
复制标题
DOI：
10.1007/978-3-642-37140-0_24
复制
发表时间：
2013-03
期刊：
影响因子：
--
通讯作者：
Antonio López Jaimes;C. Coello;A. Oyama;K. Fujii
中科院分区：
文献类型：
--
作者：
Antonio López Jaimes;C. Coello;A. Oyama;K. Fujii
为了提高多目标进化算法（MOEA）在多目标优化问题中的可扩展性，提出了一种将成就函数与ε-指标相耦合的替代偏好关系.使用Deb-Thiele-Laumanns-Zitzler（dtlz）和Walking- Fish-Group（wfg）测试套件评估所得算法。我们的实验结果表明，我们提出的方法有一个很好的性能，即使使用大量的目标。关于dtlz测试问题，他们的主要困难被发现在于存在的显性抗性解决方案。相比之下，wfg问题的难度并没有因为增加更多的目标而显著增加。
In this paper, we use an alternative preference relation that couples an achievement function and theε-indicator in order to improve the scalability of a Multi-Objective Evolutionary Algorithm (moea) in many-objective optimization problems. The resulting algorithm was assessed using the Deb-Thiele-Laumanns-Zitzler (dtlz) and the Walking- Fish-Group (wfg) test suites. Our experimental results indicate that our proposed approach has a good performance even when using a high number of objectives. Regarding thedtlztest problems, their main difficulty was found to lie on the presence of dominance resistant solutions. In contrast, the hardness ofwfgproblems was not found to be significantly increased by adding more objectives.