改进差分进化算法在水库优化调度中的应用研究

Application of Improved Differential Evolution Algorithm to Optimal Operation of Reservoirs

差分进化算法在求解水库优化调度时,进化后期种群多样性急剧下降,导致算法无法跳出局部最优解而出现"早熟"收敛。针对该问题,该文对算法的贪婪选择策略进行改进,使其以一定的概率动态接受稍差解作为子代个体,从而提高算法的种群多样性;同时,提出种群基因重生策略,进一步改善种群进化的基因信息结构。将改进的差分进化算法应用于清江梯级发电调度问题,并与差分进化算法、模拟退火算法求解结果进行对比。模拟结果表明,改进算法具有更强的全局搜索能力,求解梯级水库优化调度问题更具有优势。

When differential evolution algorithm is used to solve the problem of optimal operation of reservoirs , a sharp decline in population diversity leads to premature convergence in the later stage of evolution .To solve this problem , the improved greedy selection strategy is discussed in this paper .Poor solution is accepted as offspring with a dynamic probability .Meanwhile, population genetic rebirth strategy is proposed to improve the structure of genetic information in population evolution .The improved differential evolution algorithm is applied to optimal operation of cascade reservoirs in Qingjiang .Comparison is also made between the results of the improved algorithm , differential evolution algorithm and simulated annealing algorithm .It is shown that improved algorithm has stronger global search ability and more suitable for solving the problem of optimal scheduling of cascade reservoirs .