Abstract: Recent years have witnessed the new challenges that have emerged in power grid. A special challenge is represented by peaks in the power demand of customers. The most promising solution to tackle the peak demand challenge is smart grid. Since power grids have little capacity to store energy, power demand and supply must balance at all times. As a consequence, demand response has become a powerful tool to solve the power needs of different users. Price response mechanism is the main research areas of demand response in smart grid, which mainly reflects the price and real-time power demand situation through dynamic pricing. The fixed power pricing is no longer suitable for modern electricity market. Thus, real-time pricing becomes the most promising demand response method, and the large users will also face real-time pricing bargaining with electricity retailers for reflecting the response behavior of users to real-time price.In fact, under the smart grid environment,the electricity generation, transmission and distribution are separated in electricity market, power price changes according to the real-time changes of the user power demand. The real-time short-term contracts is signed for power trading between large users and power retailers to ensure short-term electricity price is relatively stable. Therefore, retailers and large user need to bargain the contract price and the contract power amount to ensure that market participants can hedge the real-time price risk in the next time. In this paper, a incomplete information real-time bargaining game model between power retailer and large user is formulated under the smart grid environment by means of Rubinstein' ideas. Both sides of the bargaining determine their own strategies by estimating bargaining costs and strategies for each other according to the last bargaining result. Eventually, the next contract power price is determined. The results of the study show that successful bargaining between the two sides is mainly related to the conversion coefficient and prediction interval of real-time power price. We know also the bargaining is successful when the conversion coefficient is in the range between 0.8 and 1. At the same time, the payoff of retailers and large users increase with the increasing of conversion coefficient. Meanwhile, it is also found that the probabilities of success will increase when parameter breduces. So an real-time power market bargaining mechanism designed effectively is not only improve the efficiency of real-time power market, but also reach more long-term contract by choosing appropriate conversion coefficient and interval parameters, so as to optimize smart grid.

Key words: smart grid, real-time pricing, game, bargaining