Short-Term Electricity Market Simulation for Pool-Based Multi-Period Auctions

This paper presents a simulation model based on the diagonalization algorithm for the study of the short-term operation of a pool-based oligopolistic electricity market with multi-period auctions. A detailed producer's profit maximization model as well as a detailed non-convex independent system operator (ISO) market clearing algorithm are used. A multi-period auction is assumed, where generators, under conditions of perfect information (regarding system load and competitor offers), submit stepwise energy and reserves offer curves and, subsequently, the ISO clears the market. An innovative correction of the residual demand curve (RDC) is introduced to account for the effect of some of the non-convexities of the ISO market clearing. Convergence of the diagonalization algorithm is enhanced using reinforcement learning (RL), through which the knowledge on the success of the producers' past behavior is used to adjust the supply offers.