基于约束强度的演化算法  

Evolutionary algorithm for constrained single-objective optimization

作　　者：李红梅[1,2] 

机构地区：[1]中山大学计算机学院,广东广州510641 [2]广东白云学院计算机系,广东广州510450

出　　处：《计算机工程与设计》2009年第7期1719-1721,共3页Computer Engineering and Design

摘　　要：为了克服惩罚函数法存在的罚参数难以选择和控制的主要缺陷,利用个体违反约束条件的程度函数,定义了约束强度指标,并设计了一种新的具有较强全局搜索能力的多父体杂交算子,从而提出一种基于约束强度的有效的演化算法。通过数值验证比较其性能优于现有的一些约束单目标优化演化算法。In order to overcome the major shortcomings of penalty function method that is difficult in choosing the penalty parameters and control, the degree of constraint of individual violation is used completly, the binding strength indicators are defined and a new strong global search ability of the father of many body crossover operator is designed, which presents a constraint based on the strength of effective evolutionary algorithm, through the validation to compare their performance is better than some of the existing constrained single-objective optimization evolutionary algorithms.

关 键 词：演化算法 约束单目标优化 约束强度 多父体杂交 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]