宏观进化多目标遗传算法在梯级水库调度中的应用  

Application of the macro-evolutionary multi-objective genetic algorithm to the operation of cascade reservoirs

作　　者：陈小兰[1] 熊立华[1] 万民[1] 盖永岗[1] 

机构地区：[1]武汉大学水资源与水电工程科学国家重点实验室,武汉430072

出　　处：《水力发电学报》2009年第3期5-9,68,共6页Journal of Hydroelectric Engineering

基　　金：霍英东青年教师基金(101077);国家自然科学基金(50409008);教育部新世纪优秀人才支持计划(NCET-05-0624)资助

摘　　要：宏观进化多目标遗传算法(macro-evolutionary multi-objective genetic algorithm,简称MMGA),是一种新的高等物种进化算法,它可以避免传统遗传算法(genetic algorithm,简称GA)在选择过程中出现的早熟收敛现象。MMGA是综合宏观进行化算法(macro-evolutionary algorithm,简称MA)与GA而形成的,该算法的特点是引进了MA算法中的种群间关联矩阵。利用种群间的适应度信息和个体间的距离信息,能够保持种群的多样性,为解决多目标规划问题提供了一条新的途径。本文将介绍MMGA算法的原理及步骤,并将其用到水库多目标优化调度中。Macro-evolutionary multi-objective genetic algorithm is a new kind of algorithm inspired by the high-level species evolution, which can avoid the premature convergence that arise during the selection process of conventional GA. MMGA is an integration of macro-evolutionary algorithm（MA） and genetic algorithm（GA）. By introducing it into the connectivity matrix W between species in MA, it can utilize the fitness information between species and the distance information between individuals. Consequently the diversity of the solutions can be maintained, thus provides a new alternative to solve the multi-objective optimization problem. The principle and solution step of MMGA is introduced and applied to the multi-objective optimization of reservoir operation.

关 键 词：水库调度 多目标优化 遗传算法 宏观进化 

分 类 号：TV697.12[水利工程—水利水电工程]