Optimal Joint Bidding and Pricing of Electricity Retailers Using Multi-agent Deep Reinforcement Learning

In real electricity markets, it is more valuable and difficult to study the optimal decision-making of a price-maker entity than those of a price-taker entity. This paper concerns with the optimal joint bidding and pricing problem of a price-maker electricity retailer. After formulating the original optimization problem into a Markov decision process (MDP), a novel multi-agent deep reinforcement learning (MADRL) approach is proposed to solve for the optimal joint bidding and pricing strategy. Two agents are responsible for the bidding strategy and pricing srategy respectively. Numerical results prove the effectiveness of the proposed method. Insightful comparative analysis is also conducted for price-maker and price-taker cases.