大变异遗传退火法及其在水库优化调度中的应用

The Genetic Algorithm Simulated Annealing of Large Probability of Mutation and its Application in Reservoir Optimization

根据水电站优化调度特点,建立水库调度优化模型,并针对遗传算法易早熟收敛和进化后期搜索效率较低的问题,将遗传操作与模拟退火(SA)思想相结合,并加入大变异(LPM)思想,最后将大变异遗传退火算法(LPM GASA)用于隔河岩水库。通过LPM GASA在隔河岩水库优化调度中的模拟,表明改进策略在解决水库优化调度问题方面的有效性。最后与动态规划(DP)法以及几种策略下的浮点编码遗传算法(FGA)作比较,体现出改进GA的优点及优越性。

According to the characteristics of hydropower optimization, the establishment of reservoir operation optimization model, and be likely to premature convergence of genetic algorithms and evolutionary efficiency of anaphase lower demand issues, genetic manipulation and simulated annealing （SA） a combination of thinking, and to add thought of LPM, the final will be large probability mutation of genetic-annealing algorithm （LPM GASA） for Geheyan reservoir simulation. GA Geheyan reservoir in optimal scheduling of the simulation shows that the reservoir to improve the strategy in addressing the issue of optimizing the effectiveness of scheduling. Finally, with the dynamic programming （DP） method and several strategies of genetic algorithm for comparison shows the advantages and the superiority of improved GA.