Broadcast Distributed PSO for Large Scale Problems

By | October 20, 2018

Farid Bourennani

Department of Computer Science, Faculty of Computing and IT, University of Jeddah, Saudi Arabia.

Citation: Farid Bourennani , “Broadcast Distributed PSO for Large Scale Problems “, The World of Computer Science and Information Technology Journal (WSCIT). 2018 Volume 8, Issue 5, pp.43.49.

Published on: 20 October 2018

Abstract—Nowadays, we have access to unprecedented high-performance computing (HPC) resources that can be utilized to solve complex and computationally expensive optimization problem. However, one of the problems with existing metaheuristics algorithms is that they do not scale well. For example, particle swarm optimization (PSO) which is one of the most known metaheuristics performs poorly in terms of accuracy and convergence speed with large dimensional problems. In this paper, we propose a broadcast and distributed PSO using message passing interface (MPI) that showed to be faster and more accurate than the commonly utilized distributed master-slave version of PSO for the studied large-scale optimization problems.

Keywords-Particle swarm optimization; distributed computing; large scale optimization; big optimization.

Leave a Reply

Your email address will not be published. Required fields are marked *