NIPS 2018 Expo Demo
Dec. 2, 2018
Deep RL for Online Order Dispatching and Driver Repositioning
Sponsor: Didi Chuxing
Tony Qin (DiDi), Yan Jiao (DiDi), Xiaocheng Tang (DiDi), Jieping Ye (DiDi), Chenxi Wang (DiDi)
Tony Qin (DiDi), Yan Jiao (DiDi), Xiaocheng Tang (DiDi), Jieping Ye (DiDi)
In this demo, we will show a simulation-based presentation of deep RL in action on order dispatching and driver repositioning. Specifically, we will demonstrate through several specially designed domains how we use deep RL to train agents (drivers) to have longer optimization horizon and to cooperate to achieve higher objective values collectively.