A Computation Offloading Strategy for LEO Satellite Mobile Edge Computing System

Abstract

Low Earth Orbit (LEO) satellite Mobile Edge Computing (MEC) system is expected to realize the vision of 6G. Considering the differences of computing task and computing capability among terminals as well as the resource limitation and mobility of LEO satellites. In this paper, a Joint Offloading Decision and Resource Allocation (JODRA) strategy is proposed to optimize system delay and energy consumption for multi-satellite coverage area. The offloading decision problem is modeled as a many-to-one match game with externality, and is tackled by improved Gale-Shapley(GS) algorithm and coalition game iterative algorithm. The resource allocation problem is solved by Rosen gradient projection method and Lagrange multiplier method. Simulation results indicate that the proposed strategy can effectively reduce system delay and energy consumption compared with reference strategies.