求解大规模机组组合问题的策略迭代近似动态规划.pdfVIP

第34 卷 第25 期 中 国 电 机 工 程 学 报 Vol.34 No.25 Sep.5, 2014 4420 2014 年9 月5 日 Proceedings of the CSEE ©2014 Chin.Soc.for Elec.Eng. DOI ：10.13334/j.0258-8013.pcsee.2014.25.026 文章编号：0258-8013 (2014) 25-4420-10 中图分类号：TM 73 求解大规模机组组合问题的策略迭代近似动态规划 韦化，龙丹丽，黎静华 (广西电力系统最优化与节能技术重点实验室(广西大学) ，广西壮族自治区 南宁市 530004) Policy Iteration-approximate Dynamic Programming for Large Scale Unit Commitment Problems WEI Hua, LONG Danli, LI Jinghua (Guangxi key Laboratory of Power System Optimization and Energy Technology(Guangxi University), Nanning 530004, Guangxi Zhuang Autonomous Region, China) ABSTRACT: To solve the ‘curses of dimensionality’ problem 解，从而为求解大规模电力系统UC 问题提供了参考。 of dynamic programming, this paper proposed a policy 关键词：电力系统；机组组合；近似动态规划；策略迭代； iteration-approximate dynamic programming (PI-ADP) method 近似值函数 for large scale power system unit commitment (UC) problems. Policy iteration was introduced to approximate the value 0 引言 functions in the dynamic programming procedure, replacing 机组组合(unit commitment，UC) 问题一直是现 the process of calculating the value functions accurately from 代电力系统日前调度主要的优化任务[1]，它是指在 feasible states, and the ‘curses of dimensionality’ was avoided. 满足系统安全运行前提下，制定总运行成本最小的 As approximating the value functions, the operation constraints of practical systems were used for compressing state space 机组启停机计划及出力安排[2] 。UC 问题的难点在 effectively and the alternative on/off actions were cut down, so 于包含大量的离散变量，是典型的大规模混合整数 that the calculation amount was reduced further. The results of 非线性规划问题[3]，至今尚无一种公认有效的求解 the systems ranging in size