A robust deep reinforcement learning approach to driverless taxi dispatching under uncertain demand

Xiaoting Zhou, Lubin Wu, Yu Zhang, Zhen-Song Chen, Shancheng Jiang

Abstract: Highlights•A novel RL model integrated with robust optimization is proposed to control driverless taxi fleets under uncertainty.•Robust optimization is exploited to guide the convergence of the critic network and improve the learning efficiency.•Experiment results show that the proposed model outperforms the general TD3 and other state-of-the-art RL approaches.•The model can provide a reliable dispatching strategy in different ratios between driverless taxis and passenger demand.