A UCB1-Based Online Job Dispatcher for Heterogeneous Mobile Edge Computing System

Abstract

Mobile edge computing is widely believed to be a key technology in 5G networks. It promises dramatic reduction in latency and mobile energy consumption by offloading computation intensive and latency critical jobs to edge servers for execution. In this paper, we consider the job dispatching problem at the access proxies in a heterogeneous mobile edge computing system. The problem is modeled as a multi-armed bandit problem, and an online job dispatcher algorithm based on UCBI strategy is proposed. The performance of the proposed dispatcher and the edge servers are theoretically modeled and analyzed. Simulation experiments are conducted to evaluate the effectiveness of the proposed algorithm.