基于多目标遗传算法的综合利用水库优化调度图求解

分析了常规方法求解综合利用水库调度图存在的问题.提出了一个基于遗传算法的调度图求解模型.模型同时以保证率和缺水量为目标,直接以调度图的基本调度线为决策变量,通过水库的模拟运行结果评价可行解,然后利用遗传算子不断改进调度线,能搜索得出调度图的非劣解集.实例计算结果显示多目标方法较传统仅以保证率为目标推求调度图的做法合理

Deriving multipurpose reservoir operating rule curves using multi-objective genetic algorithms

Analyzing some problems of traditional methods in deriving multipurpose reservoir operating rule curves,this paper puts forward a GA-based model for multipurpose reservoir operation optimization with two objectives to maximize water supply reliability and minimize total water deficiency.In this model,basic parameters of operating rule curves are used as its decision variables;and feasible decision variable sets are evaluated through reservoir simulation;and their fitness are determined by corresponding operation performances;then three genetic operators are used to improve operating rule curves.This process is continued until certain convergence condition is met.Finally,the non-inferior solutions of operating rule curves are obtained.The model is applied to a multipurpose reservoir in North China where the operating results demonstrate its applicability.