Game Theory Research On Power Market And It's Application Of Decision-making System

With the current of power industry from monopoly to competition, electricity market has been established in our country step by step. In the electricity market environment, generation companies as the providers of power must bid based on the regulation of electricity market. Profits of generation companies depend, to a large extent, on bidding strategies. Hence, how to develop optimal bidding strategies has become a major concern of generation companies. This dissertation focuses on the study of bidding strategies of generation companies in the electricity market environment and some significant results are obtained.This paper firstly discusses the issue of bidding strategy based on marginal price forecasting, its difficulty lies in prediction of system marginal price. Here adopts BP neural network to predict SMP, historical data are used for training and forecasting results shows that the proposed method is effective. Meanwhile, it provides the bidding model and method based on SMP, and works out the strategy optimization under some conditions.Supposing the cost information of generation companies is incomplete, the Cournot model is introduced and static game theory of incomplete information is used to study the bidding strategy of generation companies with the incomplete information being asymmetric or symmetric.We design an assistant decision-making system for generation bidding based on theoretical research. This system mainly includes the modules, such as inquiring of generator parameter, inquiring of historical trade, cost analysis, bidding strategy etc. the system is made to help the decision makers of generation company to make reasonable and efficient bidding decision in the electricity market.