摘要
设计一种求解约束优化问题的粒子进化变异遗传算法(IGA_PSE).首先,分析候选解约束条件离差统计信息与约束违反函数之间的关系及其性质,基于约束条件离差统计信息提出一种改进约束处理方法;其次,基于粒子进化策略提出3种新变异算子;然后,讨论该算法早熟收敛的3种情况,并提出相应的种群多样化维持策略;最后,通过数值实验表明所提出的算法能够有效求解约束优化问题.
An improved genetic algorithm(GA) with particle swarm's evolutionary(IGA_PSE) strategy is proposed to solve constrained optimization problems(COP), t^irstly, the relation and its characters between the statistics information of the degree of constraint deviation and the constraint violation functions of candidate solutions are analyzed, and an improved constraint handling method is proposed by using statistics information of the degree of constraint condition deviation. Secondly, three novel mutation operators with particle swarm's evolutionary strategy are applied to IGA..PSE. Then, three situations of premature convergence are argued, and the corresponding strategy of diversity maintenance is proposed. Finally, numerical experiments of standard test functions show that the proposed method can solve the constraint optimization problems effectively.
出处
《控制与决策》
EI
CSCD
北大核心
2012年第10期1441-1446,共6页
Control and Decision
基金
国家自然科学基金项目(70901074
70971131
71031007)
关键词
约束优化问题
遗传算法
粒子进化变异算子
早熟收敛
constrained optimization problems: genetic algorithm: particle swarm's evolutionary mutation operator
premature convergence