Towards Efficient Computation Offloading in Multi-user MEC Environments: A Game-Theoretic Approach

Multi-access Edge Computing (MEC) is an emerging concept aimed at fulfilling the demands of low-latency and highly reliable applications in vehicular networks. Computation offloading from vehicles to MEC servers allows for the execution of computation-intensive and time-sensitive applications. However, the resources available in MEC servers are not limitless. To achieve optimal performance, it is necessary for vehicles to make real-time decisions regarding their task offloading strategies, considering the dynamic network environment. A new approach is presented in this paper by proposing a multi-user non-cooperative computational offloading game within vehicular MEC networks. The game adjusts the offloading probability for each vehicle, based on factors such as the distance between the MEC and vehicles, communication model, and the competition for MEC resources among multiple vehicles. Furthermore, a best-response learning algorithm is assigned based on the computational offloading game. It aims to maximize each vehicle's utility and guarantees convergence to a unique equilibrium under a set of conditions. To validate the effectiveness and performance of the proposed approach, a series of experiments are conducted.