基于多目标差分进化算法的水库多目标防洪调度研究

为求解水库多目标防洪优化调度问题，提出一种基于自适应柯西变异的多目标差分进化算法。算法采用一种自适应柯西变异策略克服早熟收敛问题，提高了收敛精度；根据多目标优化的特点对差分算子进行修正，并引入外部档案技术，提高了算法的收敛速度。通过对典型历史洪水的多目标调度研究，结果表明基于自适应柯西变异的多目标差分进化算法在可接受的时间内，可生成大量在各目标分布均匀、分布范围广的非劣调度方案供决策者评价优选，为水库多目标防洪调度决策提供了一种新的调度方案生成方法。

Multi-objective optimization in reservoir flood control based on multi-objective differential evolution

A multi-objective differential evolution based on adaptive Cauchy mutation (MODE-ACM) is present to solve reservoir multi-objective flood control problem. MODE-ACM adopts an adaptive Cauchy mutation (ACM) strategy to avoid premature convergence and improve its convergence precision. The operators of differential evolution are modified to suit multi-objective optimization. By using archiving technique, the convergence rate of MODE-ACM is increased. MODE-ACM was applied in a practical reservoir multi-objective flood control operation; the results showed that the algorithm could generate a lot of non-dominated scheduling schemes for decision-maker in a short time, and these schemes distributed widely and uniformly on every optimized objective. MODE-ACM provides a novel way to generate feasible scheduling schemes for reservoir multi-objective flood control operation.