Task Offloading for Large-Scale Asynchronous Mobile Edge Computing: An Index Policy Approach

Mobile-edge computing (MEC) offloads computational tasks from wireless devices to network edge, and enables real-time information transmission computing. Most existing work concerns a small-scale synchronous MEC system. In this paper, we focus on large-scale asynchronous system with random task arrivals, distinct workloads, diverse deadlines. We formulate the offloading policy design as restless multi-armed bandit (RMAB) maximize total discounted reward over time horizon. However, formulated RMAB is related PSPACE-hard sequential decision-making problem, which intractable. To address issue, by exploiting Whittle index (WI) theory, rigorously establish WI indexability derive scalable closed-form solution. Consequently, in our policy, each user only needs calculate its report it BS, users highest indices are selected for offloading. Furthermore, when completion ratio becomes focus, shorter slack less remaining workload (STLW) priority rule introduced into performance improvement. When knowledge of energy consumption not available prior offloading, develop Bayesian learning-enabled policies, including maximum likelihood estimation, learning conjugate prior, prior-swapping techniques. Simulation results show that proposed policies significantly outperform other policies.