动态规划和粒子群算法在水电站厂内经济运行中的应用比较研究  

Comparison study between dynamic programming and particle swarm algorithm for the optimal operation of hydropower plant

作　　者：程春田[1] 唐子田[1] 李刚[1] 杨斌斌[1] 

机构地区：[1]大连理工大学水电水信息研究所,辽宁省大连市116024

出　　处：《水力发电学报》2008年第6期27-31,共5页Journal of Hydroelectric Engineering

基　　金：国家自然科学基金项目(50679011);高校博士点基金资助项目(20050141008)

摘　　要：随着电站装机容量和机组台数的不断增加,利用动态规划求解水电站厂内经济运行问题,将面临"维数灾"和实效性问题。近些年,粒子群算法作为一种新型的群体智能优化方法,由于能够弥补动态规划计算时间长、内存占用量大等诸多不足,在水电站厂内经济运行等方面得到了广泛重视。现有文献,大多数从方法的应用角度探讨较多,但从替代动态规划的必然性和潜力方面探讨较少,鲜有实例分析。本文以百万级装机千瓦的乌江渡水电站为实例,深入分析与比较了粒子群算法与动态规划的优劣,认为粒子群算法是代替动态规划、求解装机规模庞大的巨型水电站厂内经济运行的有效方法。With the appearance of huge hydropower plants with over 10 000MW of the installed capacity, dynamic programming （DP） will be faced with the curse of dimensionality for the optimal operation of hydropower plant. Recently, particle swarm optimization （PSO）, as a new evolutionary computational tool which can remedy the problem of calculation time and large amount of memory occupied by DP, has obtained wide attention in the hydroelectric power system. However, Most of existing literatures have only studied the application of PSO for hydroelectric power system but seldom give the comparison between PSO and DP from the inevitability and potential of replacement. For an example, Wujiangdu hydropower plant with 1250MW of installed capacity of 5 units is used to compare the performance of optimal operation between PSO and DP for hydropower plant operation. The results show that PSO has a great advantage over DP in the efficiency and it is one of effective methods for the optimal operation of the huge hydropower plant.

关 键 词：水电工程 动态规划 粒子群算法 巨型水电站 

分 类 号：TV697.4[水利工程—水利水电工程]