Near-optimal energy management for plug-in hybrid fuel cell and battery propulsion using deep reinforcement learning

Plug-in hybrid fuel cell and battery propulsion systems appear promising for decarbonising transportation applications such as road vehicles coastal ships. However, it is challenging to develop optimal or near-optimal energy management these without exact knowledge of future load profiles. Although efforts have been made strategies in a stochastic environment with discrete state space using Q-learning Double Q-learning, tabular reinforcement learning agents’ effectiveness limited due the resolution. This article aims an improved system deep achieve enhanced cost-saving by extending parameters be continuous. The based upon Deep Q-Network. Real-world collected profiles are applied train Q-Network ferry. results suggest that acquired strategy has achieved further 5.5% cost reduction 93.8% decrease training time, compared produced agent function approximations. In addition, this also proposes adaptive scheme practical hybrid-electric operating changing environments.