The comparisons between pricing methods on pool-based electricity market using agent-based simulation

Because of the existences of market power and economies of scale, there have been various pricing methods proposed for pool-based electricity market, for example, uniform clearing pricing method (UCP), pay as bid pricing (PAB) and the electricity value equivalent (EVE) pricing method. An agent-based simulation model is developed in this paper to compare the market characteristics under different pricing methods. In this model the generators learn bidding strategy using reinforced learning algorithm in repeated bidding game to seek for their maximum profits. Simulation result is presented based on the data from IEEE Reliability Test System, showing that the EVE pricing method has many market characteristics better than other pricing methods. For example, when EVE is used in market pricing, there exists little room for a power supplier to raise the market price by his strategic bidding and the market becomes robust in some sense. And also EVE provides an intrinsic and reasonable mechanism to compensate the capacity investment automatically.