混沌粒子群微分进化算法及其在水库发电优化调度中的应用

Chaotic Particle Swarm Optimization and Differential Evolution Algorithm and its application to Reservoir Optimal Scheduling of Generation in Hydropower Systems

粒子群优化算法(PSO)与微分进化算法(DE)都是有效的基于群体智能的全局优化算法,但它们都容易过早收敛,陷入局部最优。针对以上问题,提出了混沌粒子群微分进化算法(CPSO-DE),该算法引入可变的惯性权重和学习因子,以基于logical映射的混沌序列代替标准PSO中的随机序列来对粒子群进行初始化,同时将微分进化算法(DE)中的变异、交叉和选择思想引入标准PSO算法中,改变标准PSO算法单一的进化策略,在全局范围内搜索最优解。作为实证的需要,通过对水库优化调度所存在问题的分析,建立了基于CPSO-DE算法的水库优化调度数学模型与求解算法,并以某水库实际运行数据进行计算,结果表明CPSO-DE算法具有较好的全局最优解,验证了CPSO-DE算法的可行性与健壮性。

Particle warm optimization and differential evolution algorithm belong to effective global optimization based on swarm intelligence algorithms. However, they have premature convergence and enter into partial optimum. To solve this problem, this paper proposes chaotic particle swarm optimization-differential evolution algorithm. It introduces a variable inertia weight and learning factor. The particle swarm is initiated by using the chaotic sequence based on logical instead of random sequence in standard PSO. At the same time, by introducing the idea of mutation, crossover and selection in differential evolution into the criteria PSO algorithm, the single evolution strategy in standard PSO algorithm is changed in order to search the global optimal solution. As an empirical need, the optimal operation model of reservoir and solution algorithm is established by analyzing the optimal operation problems of reservoir, based on CPSO-DE algorithm. A reservoir of data to calculate the actual operation, the results show that the CPSO-DE algorithm good global optimal solution, verify the CPSO-DE feasibility and robustness.