自适应混沌差分进化算法在梯级水库优化调度中的应用

介绍了一种基于自适应混沌映射的差分进化算法,该算法采用混沌映射的方式产生初始种群,并综合考虑算法迭代进度和个体进化程度两个因素,对缩放因子进行动态调整以促进算法全局搜索和局部寻优的平衡。同时,在算法进化的不同阶段采取不同尺度的扰动策略,进一步提高算法的再搜索能力。将该算法应用于某梯级水库发电调度的研究中,通过实例计算,并与基本差分进化算法、模拟退火算法相比,得到了更优的全局最优解,验证了该算法的可靠性和实用性。

Application of adaptive chaotic differential evolution algorithm to optimal dispatching of cascade reservoirs

An adaptive chaotic mapping-based differential evolution algorithm is introduced herein. The algorithm generates initial population in the chaotic mapping mode，and then dynamically adjusts scaling factor under the comprehensive considerartion of two factors，i.e. both the iterative progress and individual evolution degree of the algorithm，so as to promote the balance between the global search and the local search. Meanwhile，the disturbance strategies with various scales are taken in various stages for further enhancing the re-search performance of the algorithm. This algorithm is applied to the study of the power generation dispatching of a cascade of reservoirs，and then is compared with the basic differential evolution algorithm and the simulated annealing algorithm through the calculation on an actual case，from which a more optimized global optimal solution is obtained，while the reliability and practicability of it are verified as well.