基于泛函分析思想的动态规划算法及其在水库调度中的应用研究

发布时间：2018-01-11 02:04

本文关键词：基于泛函分析思想的动态规划算法及其在水库调度中的应用研究　出处：《水利学报》2016年01期 　论文类型：期刊论文

【摘要】：泛函分析具有高度抽象性和概括性,因此可用来对某些复杂的问题进行建模,进而采用更加简便的数学方法进行求解。本文针对动态规划算法在水库优化调度中计算规模大和时间长的问题,引入泛函分析思想,构建了时段平均出力的泛函计算模型,并基于此提出了一种改进的动态规划算法,该算法省去了传统动态规划算法中大量重复的计算过程,减小了计算规模,从而提高求解效率。分别以我国某流域中M水电站和M、N组成的梯级水电站为例进行实例研究,从多年平均发电量和计算用时等方面将改进动态规划算法与传统动态规划算法以及逐步优化算法进行了详细的对比分析。结果表明,该算法能在保证全局收敛的基础上减少动态规划计算量,缩短计算时间。
[Abstract]:Functional analysis is highly abstract and general, so it can be used to model some complex problems. Furthermore, a more convenient mathematical method is used to solve the problem. In this paper, the functional analysis is introduced to solve the problem of large scale and long time of dynamic programming algorithm in reservoir optimal operation. The functional model of time-averaged force is constructed, and an improved dynamic programming algorithm is proposed based on it, which eliminates a lot of repeated calculation process in traditional dynamic programming algorithm and reduces the scale of calculation. In order to improve the efficiency of solution, the examples of M hydropower station and Mon N cascade hydropower station in a river basin in China are taken as an example. The improved dynamic programming algorithm is compared with the traditional dynamic programming algorithm and the stepwise optimization algorithm from the aspects of the average power generation and calculation time. The results show that the improved dynamic programming algorithm is better than the traditional dynamic programming algorithm and the stepwise optimization algorithm. The algorithm can reduce the computational load of dynamic programming and shorten the computing time on the basis of global convergence.
【作者单位】：华北电力大学可再生能源学院;雅砻江流域水电开发有限公司;
【基金】：国家自然科学基金项目(51279062) 中央高校基本科研业务费专项资金(13QN22;2014ZD12;JB2015161) 科技部“十二五”科技支撑计划(2012BAB05B05) 雅砻江流域水电开发有限公司资助(JKZX-201416-01)
【分类号】：TV697.11
【正文快照】： 1研究背景在水库优化调度模型求解的众多方法中,动态规划以其对阶段性、非线性问题的有效处理而获得了广泛的应用[1]。动态规划是一种全局搜索法,它把原问题化成一系列结构相似且相对简单的子问题,再对所有子问题进行组合遍历寻优,其最大的优点在于可求出给定离散程度下的全局

【参考文献】