Research Of The Power Plant's Price System Based On Multi-Agent Technology And Game Theory

Electricity market is an inevitable trend of the electric power industry reformation. With "Separation of power plants and power net, bidding for generation" in China and a number of pilot provinces and cities in the power generation side of the electricity market putting into use, China's electricity market gradually enters a substantive stage that the power generation side has been opening up.Under this background, power plants must be in an open and fair competitive environment to participate in bidding and obtain their own survival and development through market competition.Based on the pricing demand of the power generation side, this paper offers an in-depth research of the power plants' pricing system combined with Multi-Agent technology and game theory.Firstly, this paper analyzes pricing mechanism in the electricity market, Multi-Agent technology, game theory and the Cournot model. A research topic that the realization of the power plants' pricing system used of multi-agent technology and game theory is proposed.Secondly, the three-tier architecture of the power plants' pricing system is submitted only under the market's environment of the power generation side. Application server architecture of the power plants' pricing system is built to use Multi-Agent technology. All Agents in the system's application server are classified and deployed. And the consultation and communication mechanism for price are also established. The mechanism, which makes the system have the characteristics of interactive reasoning, consultation and cooperation, of many power plants' bidding simulation in this system is achieved.And then, according to the analysis of pricing factors, the bidding mathematical model of power plant, which is based on game theory and belongs to the incomplete information Cournot model, is builded up. At the same time analysis of the example's data is done under different circumstances and the realization of corresponding algorithm is described.Finally, prototype of the price system is implemented based on JADE platform. The advantages of JADE facilitate the agent's management and communications, and Java programming language is used to finish the internal mechanism of the Agent including the game algorithm of information processing and data computing, information receiving and feedback among the Agents, and some Agents' bidding simulation and so on.In short, this pricing system used coordination and cooperation mechanism of Agents to make the pricing system have good intelligence and cooperation. And game theory methods are used to guide pricing and make the price is more in line with the characteristics of the competitive market. As the same time, the bidding simulation in the system avoides the blindness of the price and increased the possibility of bidding. In addition, the three-tier architecture of the system exhibits strong scalability, easy maintainability and high safety.