Information Age-Based Task Splitting Scheme for Edge Computing-Enabled Networks

Abstract

This letter focuses on data freshness-based task splitting in edge computing (EC)-enabled networks, where a user’s subtasks are offloaded to EC servers through radio access networks (RANs). A control tower (CT) receives information of available computing resources from EC servers and channel information. Considering the freshness of received information at CT, we define a successful offloading probability (SOP), then formulate the SOP maximization problem by optimizing the task splitting. We propose an alternative algorithm by decomposition with approximation and bisection methods. Numerical results demonstrate the proposed algorithm outperforms a baseline, emphasizing the importance of data freshness on the optimization.