An Adaptive Learning Game Model for Interacting Electric Power Markets

Abstract

In the present paper, a simulation model of two interacting electric power markets is being introduced, with or with no restriction in the interconnection capacity, in order to study the behavior of the energy price under two different pricing methods: Uniform Pricing and Pay-As-Bid. The model simulates the operation of the two markets as a stochastic adaptive Nash game, where players use a learning algorithm to maximize their profit and counterbalance their lack of information. The comparison of the results between the independent operation of the markets and the one of the interacting operation shows that lower prices are recorded when both interconnected systems apply Uniform Pricing and markets are oligopolies, whereas higher prices arise when both markets apply the pay-as-bid rule and tend towards perfect competition. In the case where the two interacting markets apply different pricing methods the differences observed in the independent market operation are blunted and prices tend to converge in intermediary price levels. Finally, constrained interconnection capacity leads to slightly higher prices at all instances.