A supply-demand interaction optimization method based on blockchain and deep reinforcement learning

With the proposal of the global carbon neutrality goal and the evolution of industrial park energy towards intelligence, traditional power dispatching and supply-demand management approaches have fallen short in meeting the demands for flexibility and precise regulation in modern industrial parks. This paper introduces an optimization method for supply-demand interaction, grounded in blockchain technology and deep reinforcement learning (DRL). By incorporating smart contracts and a multi-level virtual power plant (VPP) architecture, coordinated regulation of power resources and efficient peer-to-peer energy trading both within and beyond the park's confines are achieved. Specifically, the paper delineates the design of a blockchain-based distributed energy management system (DEMS) that integrates multi-agent DRL mechanisms, enhancing the adaptability and response speed of the industrial park. Case studies demonstrate that the proposed method markedly outperforms traditional approaches in terms of flexibility, energy efficiency, and economic benefits, underscoring its extens ive application potential in the future energy management of intelligent industrial parks.