[关键词]
[摘要]
采用标准差分进化算法在求解梯级水库优化问题时,随着解链长度的增加,算法求解性能下降,进化后期种群多样性降低,算法极易陷入局部最优解。为此,定义了个体差异参数来动态控制差分进化算法的缩放因子,即定义算法可进化度参数来动态控制算法的选择机制。通过对比标准差分进化算法、逐步优化算法和动态差分进化算法求解2个标准测试函数和某梯级水库优化调度的模拟仿真结果,发现后者较前者的全局搜索能力有了显著提高。
[Key word]
[Abstract]
The standard difference evolution algorithm is often used to analyze the optimal operation of cascade reservoirs, but with the increasing of melting length, the performance of algorithm solution decreases, the population diversity at late evolution decrease, and thus the algorithm may only determine the local optimal solution. In this study, the parameter of individual difference was defined to perform dynamic control on the scaling factor of difference evolution algorithm, and the parameter of evolution possibility was defined to perform dynamic control on the selection mechanism of the algorithm. Two standard testing functions and optimal operation of the cascade reservoirs were solved using the standard difference evolution algorithm, progressive optimization algorithm, and dynamic difference evolution algorithm. The simulation results indicated that the global searching ability of dynamic difference evolution algorithm increases significantly compared to that of standard difference evolution algorithm
[中图分类号]
[基金项目]