Title: Population-based variable neighbourhood search algorithm applied to unconstrained continuous optimisation

Authors: Wesklei Migliorini; Rafael Stubs Parpinelli

Addresses: Department of Computing Science, Santa Catarina State University, Joinville, Brazil ' Graduate Program in Applied Computing, Santa Catarina State University, Joinville, Brazil

Abstract: This work presents a population-based variable neighbourhood search approach for unconstrained continuous optimisation, called PRVNS. The main contributions of the proposed algorithm are to evolve a population of individuals (i.e., candidate solutions) and to allow each individual adapts its own neighbourhood search area accordingly to its performance. The adaptive amplitude control allows individuals to autonomously exploit and explore promising regions in the search space. Several unconstrained continuous benchmark functions with a high number of dimensions (d = 250) are used to evaluate the algorithm's performance. The PRVNS results are compared with the results obtained by some well known population-based approaches: differential evolution (DE), particle swarm optimisation (PSO) and artificial bee colony (ABC). Also, the standard VNS algorithm is considered in the experiments. The results and analyses suggest that the PRVNS approach is a promising and competitive algorithm for unconstrained continuous optimisation.

Keywords: unconstrained continuous optimisation; population-based algorithms; variable neighbourhood search; VNS; meta-heuristics; global search; adaptive behaviour.

DOI: 10.1504/IJBIC.2018.091232

International Journal of Bio-Inspired Computation, 2018 Vol.11 No.2, pp.73 - 80

Received: 27 Jan 2016
Accepted: 31 May 2016
Published online: 17 Apr 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article