Abstract：Conventional optimization methods have ‘dimension disaster’ problems and poor search efficiencies in solving the long-term scheduling optimization problem of large-scale cascade hydropower stations. In order to achieve better results, the genetic algorithm has been improved in this study. The adaptive control theory is applied to the crossover and mutation operators, such that they can be automatically changed according to the fitness value. Chaos theory is used to generate the initial population, and the simulated annealing method is also introduced. An improved simulated annealing genetic algorithm is proposed by combining the advantages of these two algorithms, which can enhance global optimization capabilities and local search capabilities, and reduce the probability of stuck on local optima. The improved algorithm is then applied to an established model. Through compared with the conventional genetic algorithms, the results show that the improved simulated annealing genetic algorithm has strong global search ability and good solution effect, which can be served as a reference for solving the long-term scheduling optimization problem of large-scale cascade hydropower stations.
标题：基于改进模拟退火遗传算法的梯级水电站长期优化调度
title:Long-Term Optimal Scheduling of Cascade Hydropower Stations Based on Improved Simulated Annealing Genetic Algorithm
