Real-Time Scheduling Using Reinforcement Learning Technique for the Connected Vehicles

This paper proposes a real-time scheduling algorithm using Reinforcement Learning (RL) for the connected vehicles based on Software Defined Network (SDN) and fog computing. In the connected vehicles, there are various services that need to be processed in real time for the safety and entertainment of the driver. In such a situation, it is important for the driver to deliver the service within the deadline. Road Sid Units (RSUs) acting as fog server and SDN controller can make appropriate real time scheduling by utilizing current network situation and service request list. The proposed method finds a policy that minimizes the number of services that fail to meet deadlines for each scheduling period. Simulation results show that the proposed method has higher performance than the comparison method. The proposed method can guarantee effective scheduling in most situations by establishing adaptive policies in various environments through learning.