市场环境下发电公司的动态学习与市场猜测均衡

Dynamic Learning of Generation Companies and Conjectural Equilibrium in Electricity Spot Market

提出了基于猜测供给函数模型的发电公司动态学习模型,用于分析和预测重复博弈的不完全信息电力市场环境下发电公司潜在的市场力.各发电公司根据公布的前阶段市场运行数据动态地调整其微增响应猜测,使之能够准确反映所有竞争对手相应于市场结清电价单位变化时所愿意引起的出力变化,从而动态调整提交的供给函数以获得短期最大利润.算例分析表明,发电公司动态学习有利可图;经过各发电公司动态学习并不断调整优化猜测供给函数将引导市场最终达到一猜测均衡结果.

A dynamic learning model of generation companies under deregulation was proposed based on conjectured supply function (CSF) equilibrium. It can be used to forecast and evaluate potential market power generation companies possessed from repeatedly gaming in electricity market with incomplete information. On the basis of published historical data of market operation, individual companies can learn dynamically and regulate their conjectures to fit accurate strategic behaviors of their rivals in order to maximize their own profits. It is shown via case studies that companies have incentive to learn dynamically since learning will bring them more profit, and that the market will get to a conjectural equilibrium when all companies learn and regulate corresponding conjectures.