摘要
针对遗传算法解决厂级负荷优化分配问题时易早熟,约束条件难满足的问题,提出一种多精英保留策略和基于可分配区间信息的解修补方法,结合厂级负荷优化分配问题的特点,在初始化、交叉与变异过程中,给出新的指导性规则。提高种群的多样性,通过对不可行解进行修补,提高初始化成功率,并将遗传操作过程中,维持种群中可行解比例在一个较高的水平,提高了随机解约束方法的生成效率,解决了遗传操作过程中,罚函数法难以保证不可行解比例的难题。最后通过数值仿真实验说明其正确性与有效性。
An improved multi-elitist preservation strategy and unit＇ s limitation based solution repair method is proposed, which focus on the solving of constraints handling and early convergence of genetic algorithm （ GA）. Combining with characteristics of plant load distribution, new guiding rules are formulated in solution initialization, crossover and mutation procedure, in order to improve the success rate of solution initialization, to preserve diversification and to sustain feasible solutions in whole population which avoids uncontrollable infeasible solution amounts of adaptive penalty function methods. Simulations on numerical examples show effectiveness of proposed methods.
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2012年第3期71-77,共7页
Journal of North China Electric Power University：Natural Science Edition
基金
国家自然科学基金重点项目(51036002)
北京市教委共建项目资助
关键词
遗传算法
精英保留
约束处理
解修补方法
genetic algorithms
elitist preservation
constraints handling
solution repairing method
