TS-SMOSA: A Multi-Objective Optimization Method for Task Scheduling in Mobile Edge Computing

Abstract

Deploying cloud service at base stations (Cloudlets, considered as Mobile Edge Computing) or sharing resources between mobile devices (Mobile Cloudlets) are expected to be enablers of Edge Computing or Fog computing. Different Mobile Edge Computing systems have been proposed based on hierarchical cloud architecture. However, the dynamics and diversity of mobile cloudlets makes it challenging to offload tasks to optimal nodes (process locally, or offload to a device in Mobile Cloudlets or Cloudlets). In this paper, offloading availability is formulated by considering user mobility similarity and the residual energy of mobile devices. To guarantee offloading availability and minimize energy consumption as well as data size transmitted through the cellular access links, offloading decisions for multi-tasks (task scheduling problem) are formulated to be a multi-objective optimization problem and we develop a centralized algorithm TS-SMOSA such that efficient-energy offloading decisions are made and tasks can be allocated to optimal nodes. Finally, experimental results are promising and show near optimal solutions for all of our studied scenario.