首页 | 本学科首页   官方微博 | 高级检索  
     

基于可行域搜索映射的并行动态规划
引用本文:纪昌明,马皓宇,李传刚,李宁宁,俞洪杰.基于可行域搜索映射的并行动态规划[J].水利学报,2018,49(6):649-661.
作者姓名:纪昌明  马皓宇  李传刚  李宁宁  俞洪杰
作者单位:华北电力大学可再生能源学院
基金项目:“十三五”国家重点研发计划项目(2016YFC0402208);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项目(2016XS53)
摘    要:为解决传统动态规划在处理水库群联合优化调度时面临的约束处理机制选择和计算时间长的问题,引入映射思想,基于映射和集合论知识构建可行域搜索映射模型,并结合动态规划的并行性,提出基于可行域搜索映射的并行动态规划。该算法通过构建时段可行搜索空间和并行模式,以规避无效状态组合计算并充分发挥计算机多核优势,提高计算效率。以李仙江流域三库联合调度为实例进行研究,从年发电量、计算耗时等方面将改进算法与传统动态规划以及逐步优化算法(POA)进行详细的对比分析。结果表明,该算法能在保证解全局收敛性的前提下减少计算耗时,制定梯级水库最优调度策略。

关 键 词:映射  可行域  并行  动态规划  联合调度
收稿时间:2018/1/12 0:00:00

Research on parallel dynamic programming based on feasible region search mapping
JI Changming,MA Haoyu,LI Chuangang,LI Ningning and YU Hongjie.Research on parallel dynamic programming based on feasible region search mapping[J].Journal of Hydraulic Engineering,2018,49(6):649-661.
Authors:JI Changming  MA Haoyu  LI Chuangang  LI Ningning and YU Hongjie
Affiliation:School of Renewable Energy, North China Electric Power University, Beijing 102206, China,School of Renewable Energy, North China Electric Power University, Beijing 102206, China,School of Renewable Energy, North China Electric Power University, Beijing 102206, China,School of Renewable Energy, North China Electric Power University, Beijing 102206, China and School of Renewable Energy, North China Electric Power University, Beijing 102206, China
Abstract:To overcome the difficulties in selection of the constraint processing mechanism and long calcnlation time by using the traditional dynamic programming for optimizing reservoir group operation,a mapping model for feasible region search is developed based on knowledge of the set and mapping theory.A parallel dynamic programming algorithm is put forward on the basis of feasible region search mapping.Through construction of the feasible solution search space and a parallel computing mode,this algorithm can avoid the calculation of invalid state combinations and give full play to the power of multi-core computers,thus improving computational efficiency.As an example,the joint scheduling of three reservoirs in Lixianjiang river basin is taken for case study.In terms of annual generated energy,calculation time and other criteria,detailed comparative analysis is conducted over the improved algorithm,the traditional dynamic programming and also progressive optimization algorithm.The results show that the proposed algorithm can reduce computation time while ensuring global convergence of the solutions,which provides scientific reference for making the optimal scheduling strategy of cascade reservoirs.
Keywords:mapping  feasible region  parallel computing  dynamic programming  joint operation
本文献已被 CNKI 等数据库收录!
点击此处可从《水利学报》浏览原始摘要信息
点击此处可从《水利学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号