Abstract
Currently, constant innovations in service-oriented architectures lead to extending the Cloud services with promising solutions such as Fog computing. As the Cloud-Fog environment still remains in its infancy stage, several issues remain among the considerable challenges to be handled. One of the key issues in a Cloud-Fog environment is the scheduling of business processes tasks, i. e. selecting the suitable Cloud-Fog resources to support the execution of the business processes tasks while considering budget and temporal constraints. Indeed, these constraints are generally contradictory. Indeed, the use of cheaper resources increases the execution time and vice versa. Furthermore, minimizing the energy consumption is among the prominent considerations when dealing with Cloud-Fog environment. Hence, finding out the trade-off set of optimal solutions is required considering minimizing cost, time and energy consumption. To address such an issue, we propose, in this paper, a Multi-Objectives Particle Swarm Optimization (MOPSO) algorithm based on a non-dominance sort to handle the scheduling problem of time-aware business processes with many conflicting objective functions. Our algorithm aims to optimize three conflicting objectives namely, the makespan (total execution time), the monetary cost and the energy consumption while taking into account budget and temporal constraints of the business process. The output of our MOPSO algorithm represents a set of Pareto optimal solutions from which the user can select the best one. The elaborated experimentation illustrates the good performance of the proposed algorithm.