Title:Fully Decentralized Computation Offloading in Priority-Driven Edge Computing Systems

Abstract:We develop a novel framework for fully decentralized offloading policy design in multi-access edge computing (MEC) systems. The system comprises $N$ power-constrained user equipments (UEs) assisted by an edge server (ES) to process incoming tasks. Tasks are labeled with urgency flags, and in this paper, we classify them under three urgency levels, namely, high, moderate, and low urgency. We formulate the problem of designing computation decisions for the UEs within a large population noncooperative game framework, where each UE selfishly decides on how to split task execution between its local onboard processor and the ES. We employ the weighted average age of information (AoI) metric to quantify information freshness at the UEs. Increased onboard processing consumes more local power, while increased offloading may potentially incur a higher average AoI due to other UEs' packets being offloaded to the same ES. Thus, we use the mean-field game (MFG) formulation to compute approximate decentralized Nash equilibrium offloading and local computation policies for the UEs to balance between the information freshness and local power consumption. Finally, we provide a projected gradient descent-based algorithm to numerically assess the merits of our approach.