电力市场模拟中基于信念的发电商智能个体决策模型  

A NEW BELIEF-BASED BID DECISION MODEL OF GENCO AGENT IN ELECTRICITY MARKET SIMULATION

作　　者：江健健[1] 康重庆[1] 夏清[1] 

机构地区：[1]清华大学电机工程与应用电子技术系,北京市海淀区100084

出　　处：《电网技术》2005年第14期4-9,共6页Power System Technology

基　　金：国家自然科学基金资助项目:(50377016)~~

摘　　要：为了模拟发电商在电力市场中的决策行为,基于报价中标概率信念函数,建立了一种新的发电商决策模型。该决策模型以追求发电商个体的收益期望最大为目标,能够为发电商决策最优报价曲线。作者针对模型的特点进行了深入的分析,给出了最优报价的理论判据,揭示了最优报价点的物理特征,并从报价曲线的几何形态上给出了直观的解释。利用该决策模型构建的Multi-Agent电力市场模拟方法,可以对电力市场进行定量和定性的分析。算例结果表明:市场成员的学习能力和发电成本分布是影响市场演化过程的关键因素。用本文所提出的思路同样可以建立购电者个体的决策模型。To simulate the bid decision-making behavior of GenCo, on the basis of Bid Acceptance Probability (BAP), the authors build up a new bid decision-making model of GenCo in electricity market. The new model can optimize the bidding curve of GenCo agent to maximize its expected profit. The features of the proposed model is analyzed in detail and theoretical criterion of optimized bidding price is given, therefore the physical characteristics of optimized bidding point are revealed and a directviewing explanation is given according to the bidding curve's geometrical shape. The proposed bid decision-making model can be applied to construct a Multi-Agent based simulation method to, qualitatively and quantitatively analyze electricity markets. Numerical examples show that the learning ability of market participants and the distribution of power generation costs are the key factors impacting the evolution process of electricity market. The buyer agent's decision-making model can also be built up by using the similar idea.

关 键 词：电力市场模拟 信念学习 中标概率 中标决策 最优判据 

分 类 号：TM734[电气工程—电力系统及自动化]