Sizing of Battery Considering Renewable Energy Bidding Strategy with Reinforcement Learning

This paper proposes a novel computationally efficient algorithm for optimal sizing of Battery Energy Storage Systems (BESS) considering renewable energy bidding strategies. Unlike existing two-stage methods, our algorithm enables the cooptimization of both by updating the BESS size during the training of the bidding policy, leveraging an extended reinforcement learning (RL) framework inspired by advancements in embodied cognition. By integrating the Deep Recurrent Q-Network (DRQN) with a distributed RL framework, the proposed algorithm effectively manages uncertainties in renewable generation and market prices while enabling parallel computation for efficiently handling long-term data.