摘要
针对有等式约束的优化问题,提出了一种新的遗传算法。该算法是在种群初始化、交叉、变异操作过程中使用求解参数方程的方法处理等式约束,违反不等式约束的个体用死亡罚函数进行惩罚设计出的实数编码遗传算法。数值实验结果表明,新算法性能优于现有其它算法;它不仅可以处理线性等式约束,而且还可以处理非线性等式约束,同时提高了收敛速度和解的精度,是一种通用强、高效稳健的智能算法。
A new genetic algorithm is presented to solve equality-constrained optimization problems. Parametric equation method is taken to keep particles satisfying with equality constraints, and death penalty method is used to handle inequality constraints during the process of population initiation, crossover and mutation, and a new real-code genetic algorithm is proposed. The experiment results demonstrate that the new genetic algorithm is superior to some other techniques; the proposed algorithm is a general, effective and robust method, it can handle not only linear equality constraints, but also nonlinear equality constraints, furthermore, the speed ofconvergence and the precision are improved.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第13期3184-3185,3194,共3页
Computer Engineering and Design
基金
广东工业大学青年基金项目(052039)
关键词
参数方程
等式约束
遗传算法
死亡罚函数
约束优化
parametric equation
equality constraints
genetic algorithm
death penalty method
constrained optimization
