Multi-objective fuzzy approach to scheduling and offloading workflow tasks in Fog–Cloud computing

• Published in: Simulation modelling practice and theory Vol. 123; p. 102687
• Main Authors: Mokni, Marwa, Yassa, Sonia, Hajlaoui, Jalel Eddine, Omri, Mohamed Nazih, Chelouah, Rachid
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.02.2023; Elsevier
• Subjects: Cloud computing; Computer Science; Fog computing; Fuzzy inference system; MAS; Negotiation; Optimization; QoS; Scheduling; Workflow; Fog computing; Cloud computing; Negotiation; Fuzzy inference system; QoS; Workflow; Scheduling; Optimization; MAS
• ISSN: 1569-190X; 1878-1462
• DOI: 10.1016/j.simpat.2022.102687

Fog–Cloud computing environments have become attractive focus areas for scheduling workflow automation of different processes in different domains. The planning of workflow containing tasks having different characteristics, in Fog-Cloud computing, can affect the gain in terms of resources. The main objective of this article is to efficiently schedule a workflow in such heterogeneous environments under conflicting constraints between the user and the resource provider. In this paper, we propose a new approach to designing an environment that combines partitioning, sequencing, and scheduling algorithms while guaranteeing a multi-objective optimization of users’ and providers’ conflicting requirements. Our approach also integrates a distributed resolution with Multi-Agent System (MAS), in order to reduce the complexity of the problem, and a fuzzy inference system to deal with the uncertainty problem. To study and confirm the effectiveness of our approach, we conducted an experimental study on different data collections. We compared our approach with the main related approaches that we studied. The analysis of the obtained results clearly shows the limits of these studied approaches and confirms the performance of our approach, as well as its effectiveness in planning workflows on Fog–Cloud computing while taking into account the utilization of resources. In terms of Makespan, our solution recorded a reduction of 48% compared to the Energy Makespan Multi-Objective Optimization (EM-MOO) approach, and 41% compared to the Multi-Agent System based Genetic Algorithm (MAS-GA), while respecting the time constraint.