Analysis of electricity spot market equilibrium with forward contract based on Stackelberg game and reinforcement learning

Forward market is an effective means of curbing electricity market risk by mitigating the use of market power by generation companies (GenCos) in the spot market. This paper establishes a Stackelberg game model for the electricity spot market to study the strategic bidding behavior of GenCos in the spot market under different forward market settings and its impact on the market clearing results. The GenCos act as the leaders to maximize revenues from the market, and the spot market clearing process acts as the follower to maximize total social welfare. A multi-agent deep reinforcement learning (MADRL) algorithm is applied to solve the Stackelberg game problem and obtain the market equilibrium solution. Finally, multiple cases are set up in an IEEE 30-bus system including six generators for simulation and analysis. The simulation results show that raising the proportion of forward market transactions can effectively compress the bidding space in the spot market, reduce the strategic bidding behavior of GenCos, and alleviate the spike and volatility of spot market prices.