Simulation of synergistic trading in medium- and long-term power and carbon markets based on a multiagent model and reinforcement learning algorithm

Abstract

Abstract This paper simulates the transactions of thermal power plants participating in medium- and long-term power markets and carbon markets simultaneously. The bidding adjustment function in bilateral contracts can realistically reflect the adjustment of participants’ bids and ask prices. This paper introduces the carbon assets management module to improve the flexibility of transactions. The thermal power plants’ bidding strategies affect the market clearing prices in the centralized trading markets, which in turn will affect their profits. We implement a reinforcement learning algorithm to investigate the bidding adjustment of thermal power plants in central bidding markets. The simulation results show that an increase in the carbon quota auction ratio can increase the prices of electricity and carbon quotas. Our simulations also reveal that excessive emission penalties and carbon incentive mechanisms can decrease the total carbon emissions of the thermal power industry, which is beneficial for the green power transition.