A Game Theory Based Task Offloading Scheme for Maximizing Social Welfare in Edge Computing

Edge computing, as a computing paradigm that enables the decentralization of cloud computing services to the edge of the network, effectively addresses the issue of service unavailability caused by power constraints on end devices when handling user application requests. End users offload computational tasks and associated data to the infrastructures at the network edge. Even if executing tasks at the edge can reduce energy consumption and computational latency compared to local execution on end devices, offloading a large number of tasks consumes wireless channel resources and computational resources of the edge infrastructures, resulting in additional transmission costs and energy consumption. Moreover, competition among multiple users for limited resources at the edge nodes leads to a situation where it is challenging to balance the utilities of all the users and the Edge Service Provider (ESP). In this paper, we address the scenarios where both the battery capacity of end devices and the resource capacity at the edge are limited. We propose a computation offloading scheme based on a master-slave Stackelberg game. We provide theoretical proof of the existence of a unique Nash equilibrium in the proposed game and optimize the energy consumption and user benefits during the offloading process. Furthermore, ESP improves its revenue by servicing more user requests. The simulation results show that the proposed algorithm performs well in terms of energy consumption and user utility.