An Online Simulated Annealing-Based Task Offloading Strategy for a Mobile Edge Architecture
Ayeh Mahjoubi, Arunselvan Ramaswamy, Karl‐Johan Grinnemo
Abstract
This paper presents a new task-scheduling strategy, SATS, for Mobile Edge Computing (MEC). SATS uses a simple simulated annealing-based method for scheduling tasks and shows that it can be a promising solution for online task scheduling in a MEC architecture. The paper evaluates three types of predictors: neutral, conservative, and optimistic, and concludes that using a conservative predictor that overestimates the number of service requests leads to the best performance in terms of higher acceptance rates and shorter processing times. With its Simulated Annealing-based method, SATS offers an acceptance ratio that is only 5% lower than what it could have been if it had known the frequency of service request arrivals beforehand. The simplicity and efficiency of the SATS strategy are highlighted as it deviates less than 20% from this acceptance ratio in all conducted experiments.