|
|||||
|
|
| 一种新的正交多目标最优化遗传算法 | |
| 引用本文： | 刘海林,王字平,刘永清.一种新的正交多目标最优化遗传算法[J].计算机工程与应用,2002,38(11):27-29,131. |
| 作者姓名： | 刘海林 王字平 刘永清 |
| 作者单位： | 1. 华南理工大学自动控制工程系,广州,510640;广东工业大学应用数学系,广州,510090 2. 西安电子科技大学数学科学系,西安,710071 3. 华南理工大学自动控制工程系,广州,510640 |
| 基金项目： | 国家自然科学重点基金资助(编号:69934030) |
| 摘 要： | 文章首先构造了一个新的适应值函数,它以标准化后的目标函数的加权极大值作为个体的适应值,通过独特选取多组权重向量和正交试验方法,保持了种群的多样性,使得搜索方向在目标空间中均匀指向有效界面。该算法的显著特性是不管有效界面是否凸,都能找到足够多均匀分布的有效解。数值实验也表明了该算法非常有效。 |
| 关 键 词： | 遗传算法 多目标最优化 极大极小策略 正交试验设计法 |
| 文章编号： | 1002-8331-(2002)11-0027-03 |
A Novel Orthogonal Multiobjective Genetic Algorithms |
|
| Liu Hailin , Wang Yuping Liu Yongqing.A Novel Orthogonal Multiobjective Genetic Algorithms[J].Computer Engineering and Applications,2002,38(11):27-29,131. | |
| Authors: | Liu Hailin Wang Yuping Liu Yongqing |
| Affiliation: | Liu Hailin 1,2 Wang Yuping 3 Liu Yongqing 11 |
| Abstract: | A new fitness function is constructed by maximization of the weighted normalized- objectives,where the weight vectors are carefully and reasonably designed via generalized sphere coordinate transformation and orthogonal design.As a result ,the population can keep the diversity,and the search uniformly moves towards Pareto frontier.The most important characterization of the proposed algorithm is that it can always find enough solutions distributed uniformly on Pareto fron tier no mater whether the Pareto frontier is convex or not.The computer simulat ions also show the proposed algorithm is effective. |
| Keywords: | genetic algorithm Multiobjective optimization min-max strategy orthogonal design |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |