Particle swarm optimization parallelism implementation to classify multiclass datasets

Abstract

This research concept deals with Parallelizes PSO-SA model which combines the particle swarm optimization (PSO) and Simulated Annealing (SA). Parallel PSO-SA operates by parallelizing the operation of each of the particles and Multistart PSO runs parallel several normal versions of PSO embedded with Simulated Annealing in parallel. The experimental results were conducted on benchmark data sets and the proposed approach can reduce the time complexity and also to increase classification accuracy.