|
|||||
|
|
| 一种针对高维决策空间的进化多目标优化方法 | |
| 引用本文： | 向勇,TANG Chang-jie,曾涛,ZHU Ming-fang,邱江涛,WANG Rong. 一种针对高维决策空间的进化多目标优化方法[J]. 系统仿真学报, 2008, 20(9): 2329-2333 |
| 作者姓名： | 向勇 TANG Chang-jie 曾涛 ZHU Ming-fang 邱江涛 WANG Rong |
| 作者单位： | 1. 四川大学计算机学院,四川成都,610065;成都电子机械高等专科学校计算机工程系,四川成都,610031 2. 天津师范大学计算机与信息工程学院,天津,300387 3. 四川大学计算机学院,四川成都,610065 |
| 基金项目： | 国家自然科学基金，四川省教育厅资助项目，成都电子机械高等专科学校科研项目，天津师范大学校科研和教改项目 |
| 摘 要： | 进化算法可并行处理多个解的特性使得它特别适合解决多目标优化问题。针对高维决策空间,将基因表达式编程引入多目标优化,设计了新的个体结构和操作,提出了一个进化多目标优化算法EMOGEP。实验结果表明,新算法在低维决策空间是可行和有效的;在高维决策空间中,表现出了比传统进化多目标优化算法更好的性能;多模态情况下,新算法能很好的逼近理论Pareto前沿。 |
| 关 键 词： | 高维 多目标优化 进化算法 基因表达式编程 |
Evolutionary Multiobjective Optimization Method for High Dimensional Decision Space |
|
| XIANG Yong,TANG Chang-jie,ZENG Tao,ZHU Ming-fang,QIU Jiang-tao,WANG Rong. Evolutionary Multiobjective Optimization Method for High Dimensional Decision Space[J]. Journal of System Simulation, 2008, 20(9): 2329-2333 | |
| Authors: | XIANG Yong TANG Chang-jie ZENG Tao ZHU Ming-fang QIU Jiang-tao WANG Rong |
| Abstract: | Evolutionary algorithms are particularly suited for Multiobjective Optimization Problem (MOP), because they process a set of solutions in parallel. Aiming at high dimensional decision space, an innovative algorithm named Evolutionary Multiobjective Optimization Gene Expression Programming (EMOGEP) was presented. A new structure for individual was designed and some genetic operators were proposed. Experiment results show that EMOGEP is feasible and effective in low dimensional decision space. Especially in high dimension,EMOGEP performs better than traditional EMO algorithms. In the circumstance of multimodality, EMOGEP approximately converges to the true Pareto-optimal front. |
| Keywords: | high dimension multiobjective optimization evolutionary algorithm gene expression programming |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |