Premium
A multicriteria optimization model for cloud service provider selection in multicloud environments
Author(s) -
Mohamed Amany M.,
Abdelsalam Hisham M.
Publication year - 2020
Publication title -
software: practice and experience
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.437
H-Index - 70
eISSN - 1097-024X
pISSN - 0038-0644
DOI - 10.1002/spe.2803
Subject(s) - computer science , particle swarm optimization , taguchi methods , analytic hierarchy process , mathematical optimization , cloud computing , metaheuristic , service provider , selection (genetic algorithm) , genetic algorithm , simulated annealing , vendor , operations research , algorithm , service (business) , artificial intelligence , machine learning , engineering , mathematics , economy , economics , operating system , business , marketing
Summary Multicloud computing is a strategy that helps customers to reduce reliance on any single cloud provider (known as the vendor lock‐in problem). The value of such strategy increases with proper selection of qualified service providers. In this paper, a constrained multicriteria multicloud provider selection mathematical model is proposed. Three metaheuristics algorithms (simulated annealing [SA], genetic algorithm [GA], and particle swarm optimization algorithm [PSO]) were implemented to solve the model, and their performance was studied and compared using a hypothetical case study. For the sake of comparison, Taguchi's robust design method was used to select the algorithms' parameters values, an initial feasible solution was generated using analytic hierarchy process (AHP)—as the most used method to solve the cloud provider selection problem in the literature, all three algorithms used that solution and, in order to avoid AHP limitations, another initial solution was generated randomly and used by the three algorithm in a second set of performance experiments. Results showed that SA, GA, PSO improved the AHP solution by 53.75%, 60.41%, and 60.02%, respectively, SA and PSO are robust because of reaching the same best solution in spite of the initial solution.