A hybrid ACO approach to the Matrix Bandwidth Minimization Problem

TL;DR

This paper presents a hybrid Ant Colony Optimization algorithm that combines local search techniques to effectively minimize matrix bandwidth, addressing a complex P-complete problem with practical industrial applications.

Contribution

The paper introduces a novel hybrid ACO approach with local search for the Matrix Bandwidth Minimization Problem, improving solution quality and computational efficiency.

Findings

The proposed algorithm performs well on benchmark MBMP instances.

Hybridization with local search enhances solution quality.

Experimental results demonstrate competitive performance.

Abstract

The evolution of the human society raises more and more difficult endeavors. For some of the real-life problems, the computing time-restriction enhances their complexity. The Matrix Bandwidth Minimization Problem (MBMP) seeks for a simultaneous permutation of the rows and the columns of a square matrix in order to keep its nonzero entries close to the main diagonal. The MBMP is a highly investigated P-complete problem, as it has broad applications in industry, logistics, artificial intelligence or information recovery. This paper describes a new attempt to use the Ant Colony Optimization framework in tackling MBMP. The introduced model is based on the hybridization of the Ant Colony System technique with new local search mechanisms. Computational experiments confirm a good performance of the proposed algorithm for the considered set of MBMP instances.