Cost-Effective Task Scheduling in Mobile Cloud Computing under a Deadline Constraint.

Yalong Li,Linhui Wang,Jinyu Liu,Jin Sun

DOI: https://doi.org/10.1109/smartworld-uic-atc-scalcom-digitaltwin-pricomp-metaverse56740.2022.00049

2022-01-01

Abstract:Mobile cloud computing (MCC) is capable of delivering real-time and high-intensity computing services by offloading computation-intensive tasks from mobile devices to cloud datacenters. During this procedure, the performance of task scheduling strategies has a direct impact on the rental cost of a MCC system. Oriented toward a MCC environment, this paper studies the task scheduling problem that minimizes the rental cost considering both task offloading and result downloading process and the mobile device’s location movement under a deadline constraint. To solve this optimization problem, we introduce a novel evolutionary computation framework that employs metaheuristic algorithms to make decision on the offloading and downloading task sequences separately. The proposed framework includes a probabilistic mapping mechanism for enabling metaheuristic algorithms and a task assignment strategy for calculating the fitness value (i.e., the rental cost) of task sequences. With the aforementioned operators, a variety of meta-heuristic algorithms can be easily incorporated into the framework to solve the studied scheduling problem. Extensive simulations were performed to justify that the meta-heuristic framework proposed in this work can effectively reduce the rental cost of the MCC system. The evaluation results also demonstrate the easy-to-implement characteristic of the proposed framework in integrating various metaheuristic algorithms.