A Strategic Game for Task Offloading among Capacitated UAV-Mounted Cloudlets

摘要

In this paper, we focus on the computational offloading problem among unmanned aerial vehicles (UAVs) acting as small flying cloudlets that receive compute-intensive tasks from Internet-of-Things (IoT) devices. Different from existing studies, we consider a network of capacitated UAV-mounted cloudlets (NUMC) covering a region, where each UAV is endowed with limited computational resources and a restricted capacity providing edge computing services to IoT users in that region. UAVs aim to optimize their energy consumption while satisfying quality of services of IoT tasks. We formulate the task offloading problem among UAVs as an integer program, and introduce a novel offloading game to model this problem. We prove the existence of pure-strategy Nash equilibrium of our game, where none of the UAVs has incentive to change its offloading decision. We propose a strategic offloading algorithm to solve our proposed game and find a Nash equilibrium. We evaluate the performance of our proposed algorithm by extensive experiments. The results show that a Nash equilibrium exists in NUMC and a desired system performance is achieved as well.