A new multi-particle collision algorithm for optimization in a high performance environment

Abstract

A new meta-heuristics is introduced here: the Multi-Particle Collision Algorithm (M-PCA). The M-PCA is based on the implementation of a function optimization algorithm driven for a collision process of multiple particles. A parallel version for the M-PCA is also described. The complexity for PCA, M-PCA, and a parallel implementation for the MPCA is developed. The efficiency for optimization for PCA and M-PCA is evaluated for some test functions. The performance of the parallel implementation of the M-PCA is also presented. The results with M-PCA produced better optimized solutions for all test functions analyzed.