How to Leverage Mobile Vehicles to Balance the Workload in Multi-Access Edge Computing Systems

Abstract

In this correspondence, we investigate the joint relay and server selection problem for a novel vehicle-assisted multi-access edge computing (MEC) system, where vehicles on the road relay data from an end device to appropriate MEC servers for computing in a store-carry-forward manner. To harness the stochastic nature of task arrivals, vehicle availability, and resource availability at MEC servers, we design a relaying scheme to minimize the average task latency based on a Markov decision process. Our study shows that our proposed scheme is highly effective in balancing the communications and computing while minimizing the end-to-end latency for task completion.