演化算法求解约束多目标优化问题  

Evolution algorithm for constrained multi-objective optimization problem

作　　者：杨林根[1] 李红梅[1] 

机构地区：[1]广东白云学院计算机系

出　　处：《计算机工程与设计》2008年第24期6243-6245,共3页Computer Engineering and Design

摘　　要：求解多目标优化问题的演化算法主要考虑如何处理相互冲突的多个目标间的优化,很少考虑对约束条件的处理。通过引入约束主导原理,提出一种无需采用罚函数,完全是基于个体排序的求解约束多目标优化问题的演化算法。对测试函数进行了实验,实验结果表明了该算法的可行性和有效性。Evolution algorithm for constrained multi-objective optimization problems mainly focus on optimizing the conflicting multiple objectives without considering the constraint conditions. A genetic algorithm is described which uses constraint domain principle and bases on a individual ranking approach without using penalty function methods. Test results show that the new approach is feasible and effective.

关 键 词：约束多目标优化 演化算法 适应度赋值 约束主导原理 个体排序 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]