Equilibrium Analysis for Electricity Market Considering Carbon Emission Trading Based on Multi-Agent Deep Reinforcement Learning

With the proposal of China's carbon peaking and carbon neutrality goals, carbon emission trading (CET) is gradually participating in the electricity market to accelerate carbon emission reduction and the improvement of power supply structure. In this study, we analyze the impacts of CET on the electricity market based on the electricity market equilibrium model and multi-agent deep reinforcement learning (MADL) method. We firstly establish the electricity market clearing process involved CET and develop the bi-level problem to model the electricity market equilibrium with strategic generation company (GENCO) bidders. Then, a multi-agent Twin Delayed Deep Deterministic Policy Gradient (MATD3) algorithm is applied to solve the market equilibrium described above. Finally, we simulate multiple cases based on a modified IEEE 30-bus system. The result shows that an excessive carbon price can raise the nodal electricity price and have a negative influence on reducing carbon emission, and an appropriately low carbon emission quota setting can help for carbon emission reduction.