Peer-to-Peer Electricity-Carbon Coupling Trading Market Based on Multi-Agent Reinforcement Learning

With the the vigorous development of new power system, energy and carbon coupling trading has emerged as an effective way to promote the renewable generation and reduce carbon footprint. It is a challenge to design an effective market framework for joint electricity and carbon trading to achieve better strategies. This paper introduces a peer-to-peer electricity-carbon coupling trading (ECCT) market and develops a carbon-electricity virtual battery (CEVB) model to represent the relationship between dynamic carbon credits and energy consumption. The trading optimization problem is formulated as a partially observable Markov decision process, then multi-agent double soft actor-critic (MADSAC) algorithm is introduced to solve the problem. The algorithm makes the method more exploratory, protects personal privacy and solves the overestimation problems effectively. Result studies demonstrate the effectiveness of the proposed market and algorithm in optimizing total cost and reducing carbon emissions.