I. Introduction
Autonomous ground vehicles (AGVs) have been widely used for industrial automation. Despite the advancement that has been achieved, there are still many challenges associated with this domain. For instance, in many industrial applications such as manufacturing assembly lines, body in white processes, and intelligent warehouses, an AGV could be used to travel to different spots in a facility to complete tasks that are stochastically assigned to the AGV. Each task could have different priority level, which will influence the routing of the AGV. The online controller needs to adjust the AGV routing in real time in order to achieve an optimal tradeoff between completing the tasks in their assigned priority order and completing the tasks within shortest time.