A More Accurate Delay Model based Task Scheduling in Cellular Edge Computing Systems

Abstract

This paper studies task scheduling for delay minimization in cellular edge computing systems. Existing works depend on the delay model implied by Little’s Law. However, that model could be inaccurate. This work designs scheduling algorithm without using Little’s Law. Firstly, a more accurate delay model for cellular edge computing systems is established. Then, a Lyapunov function of delay is defined based on which a novel task scheduling algorithm is proposed. Simulation results show that the proposed scheduling algorithm can further reduce the delay compared with the traditional scheduling algorithm. Specifically, the proposed scheduling algorithm can achieve maximum delay reduction of 55% for light-traffic users.