QQ客服 官方微博 反馈留言

	电力市场模拟中的报价中标概率函数与发电商个体学习模型
	江健健 ; 康重庆 ; 夏清 ; JIANG Jian-jian ; KANG Chong-qing ; XIA Qing
	2010-06-10 ; 2010-06-10
关键词	电力市场模拟 多智能体 信念学习 报价中标概率 学习模型 Electricity market simulation Multi-agent Belief learning Bid acceptance probability Learning model F407.61
其他题名	BID ACCEPTANCE PROBABILITY AND LEARNING MODEL OF GENCO AGENT IN ELECTRICITY MARKET SIMULATION
中文摘要	为了在基于Multi-Agent结构的电力市场模拟中建立高效、合理的智能个体行为模型,作者引入信念学习的思想,提出市场交易者报价中标概率函数的概念。报价中标概率函数中隐性地包含了大量市场交易相关信息,是市场交易者对电力市场交易的宏观认识,能够为其进行报价决策提供相关信息。基于报价中标概率函数,建立了发电商的学习模型,每个发电商都从交易数据中不断学习和调整各自报价的中标概率信念。算例结果表明:报价中标概率函数能够反映市场供求关系、电网阻塞以及发电商投机倾向对市场价格的影响,同时使发电商个体具有记忆能力,并能减少决策问题约束条件、简化发电商学习与决策模型。所提出的学习思路同样适用于建立购电者个体的学习模型。; To build an efficient and rational learning behavioral model in Multi-Agent based electricity market simulation, the authors induct an idea of belief learning and propose a concept of Bid Acceptance Probability (BAP). A lot of information related to transaction in the market is implicitly involved in BAP, so BAP is market participants’ macroscopic knowledge on transactions in the market and can provide associated information for their bid strategy. Based on BAP the learning model of GenCo agent is built up, each GenGo can unceasingly learn and modify its own belief of BAP from simulation iteration. Numerical examples show that BAP can reflect the impacts of demands-supply relation, grid congestion and the trend of GenCo’s market manipulation on market price, in addition it can also endow the GenGo agent with memory capacity, reduce the constraints in decision-making, simplify GenCo’s learning model and decision-making model. The proposed learning model can also be applied in building the learning model for buyer agents.; 国家自然科学基金资助项目(50377016)。~~
语种	中文 ; 中文
内容类型	期刊论文
源URL	[http://hdl.handle.net/123456789/61895]
专题	清华大学
推荐引用方式 GB/T 7714	江健健,康重庆,夏清,等. 电力市场模拟中的报价中标概率函数与发电商个体学习模型[J],2010, 2010.
APA	江健健,康重庆,夏清,JIANG Jian-jian,KANG Chong-qing,&XIA Qing.(2010).电力市场模拟中的报价中标概率函数与发电商个体学习模型..
MLA	江健健,et al."电力市场模拟中的报价中标概率函数与发电商个体学习模型".(2010).

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。