摘要
提出了带约束多目标优化问题的一种新解法。首先定义了个体的序值和个体的约束度,利用这两个定义给出了一种新的适应度函数和开关选择算子,从而对种群中的个体进行评估或排序时无需特别关心个体是否可行,避免了罚函数选择参数的困难。用概率论有关理论证明了算法的收敛性。用标准的Benchmark函数进行了仿真实验,仿真结果表明,新算法对约束多目标优化问题的求解是有效的。
A new algorithm is proposed to solve the constrained multi-objective optimization problems. The rank and the scalar constraint violation of the individual are firstly defined. Then, based on the two definitions, a new fitness function and a switch selection operator are presented. Accordingly, when the individuals are evaluated or ranked, it doesn't need to care about the feasibility ot individuals; therefore it is a penalty-parameterless constraint-handling approach. Then the convergence ot this algorithm is proved using the theory ot probability. The computer simulations demonstrate the effectiveness ot the proposed algorithm.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2007年第2期277-280,共4页
Systems Engineering and Electronics
基金
国家自然科学基金(60374063)
陕西省自然科学基础研究计划项目(2006A12)
宝鸡文理学院院级重点科研项目资助课题(ZK0619)
关键词
多目标优化
进化算法
收敛性
multi-objective optimization
evolutionary algorithm
convergence