中文摘要：

根据混联梯级水电站优化调度特点,建立并行粒子群算法求解多阶段最优化问题数学模型,重点研究了粗粒度并行粒子群算法。在基于单向环结构交流局部最优解的并行粒子群算法（PPSO）研究的基础上,提出了基于处理机上局部最优解间距离自适应选择信息交流对象策略的PPSO。应用开发的分布式水库群优化调度并行计算系统,将上述两种策略的PPSO和串行粒子群算法（PSO）应用于金沙江与雅砻江混联水库群优化调度中。通过对其优化结果的分析表明,PPSO有利于提高运算速度和求解精度以及改善算法的收敛性能。

英文摘要：

Based on the characteristic of optimal operation of series-parallel hybrid cascade hydropower stations,a mathematical model was established to solve the multi-stage optimization problem by using particle swarm optimization（PSO） algorithm and the coarse grain parallel particle swarm optimization algorithm（PPSO） was mainly studied.Before this study,a PPSO algorithm which communicates the local optimum solutions based on one direction loop had been discussed.In this study,an improved PPSO algorithm which chooses information communicate object adaptively based on the local optimum solution interval in processor was developed.By using the developed parallel computation software system for distributed reservoir groups optimization dispatching,the aforementioned two PPSOs strategies and PSO were applied to the parallel-serial reservoir groups optimization dispatching of Yalong River and the Jinsha River.The final analysis of the optimization results revealed that PPSO can better speed up the computation and improve the convergence of algorithm.