P2P Trading of Multi-VPPs with Integrated PV Energy Storage Systems Based on Multi-Agent Rollout

Integrated PV energy storage system(IPESS) is one of the compact and integrated energy control systems. Energy storage can be combined with other renewable energy sources to achieve multi-energy complementarityac when integrated into the virtual power plant(VPP). However, the integration of IPESS will also bring in large trading action space. Therefore, this paper focuses on the scheduling of multi-VPPs with IPESS. The main contributions are as follows. Firstly, based on the Peer-toPeer(P2P) trading framework, a multi-energy trading model for multi-VPPs is proposed. This model considers the transporation delay of the stored energy. Secondly, a heuristic rule based on the Huff probability mechanism is proposed to address difficulty of large trading action space. This theory compares the energy trading probabilities of stored energy to reduce the decision space. Finally, the Multi-agent Rollout(MAR) algorithm is proposed to solve the energy optimization scheduling problem among VPPs. This algorithm improves policy optimization efficiency by policy improvement and local optimization. Numerical results demonstrate that the proposed method can optimize the trading policy of multi-VPPs more effectively.