An Efficient Simulated Annealing-based Task Scheduling Technique for Task Offloading in a Mobile Edge Architecture

The Internet of Things (IoT) has emerged as a fundamental cornerstone in the digitalization of industry and society. Still, IoT devices’ limited processing and memory capacities pose a problem for conducting complex and time-sensitive computations such as AI-based shop floor monitoring or personalized health tracking on these devices, and offloading to the cloud is not an option due to excessive delays. Edge computing has recently appeared to address the requirements of these IoT applications. This paper formulates the scheduling of tasks between IoT devices, edge servers, and the cloud in a three-layer Mobile Edge Computing (MEC) architecture as a Mixed-Integer Linear Programming (MILP) problem. The paper proposes a simulated annealing-based task scheduling technique and demonstrates that it schedules tasks almost as time-efficient as if the MILP problem had been solved with a mixed integer programming optimization package; however, at a fraction of the cost in terms of CPU, memory, and network resources. Also, the paper demonstrates that the proposed task scheduling technique compares favorably in terms of efficiency, resource consumption, and timeliness with previously proposed techniques based on heuristics, including genetic programming.