不完全需求信息下的寡头发电商报价学习模型

Bidding Learning Model in Oligopoly Power Market with Incomplete Demand Information

基于发电商主观推断得到的市场清除价与市场实际价格之差异,构建发电商报价学习模型,并对该模型的不动点及其稳定性进行了解析分析;在此基础上,针对存在2个发电商的报价市场,对学习模型的稳定域、分岔图和平均收益等复杂性现象进行了数值仿真分析。结果表明:学习模型唯一的不动点即为客观的市场需求曲线;在稳定域内发电商通过学习模型可发现客观的市场需求曲线,并得到Cournot均衡策略;在混沌状态下,有发电商获得超过Cournot均衡的平均收益,也有发电商获得少于Cournot均衡的平均收益,且可能形成一个新的Nash均衡。

The demand information is incomplete to power producers in oligopoly power bidding markets.In the paper,bidding learning model is presented based on the discrepancy between subjective power price deduced by power producer and objective market clear price,whose fixed point and stability is explored analytically.Based on the aforementioned analysis,the stable region,bifurcation of the fixed point,Cournot equilibrium,and the average profits of power producers are studied respectively through numerical simulation to the power markets consisting of two producers.The results indicate that（i） unique fixed point of the learning model coincides with objective demand function,（ii） the power producer can find the objective demand function and attain the Cournot-Nash equilibrium in stable regions and（iii） in chaos state,producers gain more or less profit than its Cournot profit and there is a possibility that a new Nash equilibrium deviated from the Cournot point will be appear.