基于改进PSO算法的跨流域水库群联合调度图优化

Optimization of joint operating rule curves for inter-basin multi-reservoir system based on modified PSO algorithm

针对跨流域水库群优化调度具有高维非线性和动态性的特点,提出一种改进粒子群优化算法用于跨流域水库群引(调)水与供水联合调度图的求解.该算法针对基本粒子群优化算法易陷入局部最优且进化后期收敛速度慢的缺陷,引入交叉和变异策略、模拟退火策略和反射边界策略,以增加种群的多样性并提高算法的进化速度和全局搜索能力.数值测试证明了该算法用于求解高维复杂优化问题的有效性,实例研究也表明该算法提高了整个系统的供水与引水效益,是一种优化复杂水库群联合调度高效实用的方法.

Aiming at the inter-basin multi-reservoir optimal operation with high-dimensional nonlinearity and dynamic characteristics,a modified particle swarm optimization （MPSO）algorithm is proposed for deriving the water supply-transfer operating rule curves.Several strategies,such as crossover and mutation,simulated annealing and reflective boundary,are introduced in the basic PSO algorithm.These strategies can keep the diversity of population,improve the convergence speed and enhance the global searching ability,so that it can avoid the prematurity and slow convergence in later evolution.Numerical tests results manifest the effectiveness of the proposed algorithm for solving high-dimensional complex optimization problems.A case study indicates that the benefit of water transfer and supply has been improved considerably by the proposed algorithm. The proposed algorithm can be an effective method for optimizing the joint operation of such complex multi-reservoir system.