基于离散Markov决策过程的发电公司多阶段决策  

Discrete Time Markov Decision Process Based Multi-Period Decision Problem for Generation Companies

作　　者：张宏刚[1] 宋依群[1] 

机构地区：[1]上海交通大学电气工程系,上海200030

出　　处：《上海交通大学学报》2004年第8期1238-1240,1245,共4页Journal of Shanghai Jiaotong University

摘　　要：采用离散时间Markov决策过程(DTMDP)对以多阶段总利润最优为目标的发电公司决策问题进行研究.市场环境下,发电公司根据自身条件,其竞争策略可以是价格的接受者,也可以是价格的制定者.考虑了发电公司不同策略情况下市场均衡状态间的转换概率,分别给出了发电公司作为价格接受者和价格制定者时的多阶段决策模型.通过算例验证了所提模型的有效性和可行性.A novel strategy model based on discrete time Markov decision process (DTMDP) was proposed for a generation company to maximize its total multi-period profit in deregulation environment. In terms of the company's scale and market share in the market, it will compete as a price taker or a price maker. Different decision models with corresponding transition probabilities between market equilibria were developed for strategies of price-taker and price-maker respectively. Finally, it is shown via the numerical results that the proposed model of multi-period profit optimization for generation companies is effective.

关 键 词：电力市场 离散时间Markov决策过程 决策问题 

分 类 号：TM732[电气工程—电力系统及自动化]