Hybrid Electric Vehicle Energy Management Strategy Based on Heterogeneous Multi-Agent Reinforcement Learning

Hybrid Electric Vehicle (HEV) plays a crucial role in the transition from traditional Internal Combustion Engine (ICE) vehicles to battery electric vehicles. However, the problem of power distribution of multiple energy sources during driving is still a bottleneck restricting the development of HEVs. This paper proposes an Energy Management Strategy (EMS) based on heterogeneous multi-agent reinforcement learning to optimize the energy distribution system for a HEV that includes an ICE, a lithium battery, and a supercapacitor which can effectively match the advantages of each power source. Simulation results show the proposed strategy can better distribute the energy with similar time consumption as the traditional optimization strategy.