Dynamic Bicycle Dispatching of Dockless Public Bicycle-sharing Systems Using Multi-objective Reinforcement Learning

As a new generation of Public Bicycle-sharing Systems (PBS), the Dockless PBS (DL-PBS) is an important application cyber-physical systems and intelligent transportation. How to use artificial intelligence provide efficient bicycle dispatching solutions based on dynamic rental demand essential issue for DL-PBS. In this article, we propose MORL-BD, algorithm multi-objective reinforcement learning optimal solution We model DL-PBS system from perspective deep predict layout parking spots dispatching. define multi-route problem as optimization by considering objectives costs, dispatch truck's initial load, workload balance among trucks, supply demand. On basis, collaborative multiple trucks modeled multi-agent model. All paths between are defined state spaces, reciprocal costs reward. Each truck equipped with agent learn path in network. create elite list store Pareto found each action, finally get frontier. Experimental results actual show that compared existing methods, MORL-BD can find higher quality frontier less execution time.