An Enhanced Binary Particle Swarm Optimization (E-BPSO) algorithm for service placement in hybrid cloud platforms

• Published in: Neural computing & applications 2023, Vol.35 (2), p.1343-1361
• Main Authors: Abbes, Wissem, Kechaou, Zied, Hussain, Amir, Qahtani, Abdulrahman M., Almutiry, Omar, Dhahri, Habib, Alimi, Adel M.
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial Intelligence; Cloud computing; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Evolutionary algorithms; Image Processing and Computer Vision; Original Article; Particle swarm optimization; Placement; Probability and Statistics in Computer Science; Profitability; Solution space
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-022-07839-5

Hybrid cloud platforms offer an attractive solution to organizations interested in implementing integrated private and public cloud applications to meet their profitability requirements. However, this can only be achieved by utilizing available resources while speeding up execution processes. Accordingly, deploying new applications entails dedicating some of these processes to a private cloud while allocating others to the public cloud. In this context, the current work aims to minimize relevant costs and deliver effective choices for an optimal service placement solution within minimal execution time. To date, several evolutionary algorithms have been applied to solve the challenging service placement problem by dealing with complex solution spaces to provide an optimal placement with relatively short execution times. In particular, the standard BPSO algorithm has been found to display a significant disadvantage, namely getting trapped in local optima and demonstrating a noticeable lack of robustness in dealing with service placement problems. Hence, to overcome the critical shortcomings associated with the standard BPSO, an enhanced binary particle swarm optimization (E-BPSO) algorithm is proposed, comprising a modification inspired by the continuous PSO for the particle position updating equation. Our proposed E-BPSO algorithm is shown to outperform state-of-the-art approaches using a real benchmark task in terms of both cost and execution time.