To read this content please select one of the options below:

Hybrid differential evolution algorithms for the optimal camera placement problem

Mathieu Brévilliers (IRIMAS – Université de Haute-Alsace, Mulhouse, France)
Julien Lepagnot (IRIMAS – Université de Haute-Alsace, Mulhouse, France)
Lhassane Idoumghar (IRIMAS – Université de Haute-Alsace, Mulhouse, France)
Maher Rebai (Ecole Supérieure d’Ingénieurs Léonard de Vinci, Paris, France)
Julien Kritter (IRIMAS – Université de Haute-Alsace, Mulhouse, France)

Journal of Systems and Information Technology

ISSN: 1328-7265

Article publication date: 23 November 2018

Issue publication date: 30 November 2018

228

Abstract

Purpose

This paper aims to investigate to what extent hybrid differential evolution (DE) algorithms can be successful in solving the optimal camera placement problem.

Design/methodology/approach

This problem is stated as a unicost set covering problem (USCP) and 18 problem instances are defined according to practical operational needs. Three methods are selected from the literature to solve these instances: a CPLEX solver, greedy algorithm and row weighting local search (RWLS). Then, it is proposed to hybridize these algorithms with two hybrid DE approaches designed for combinatorial optimization problems. The first one is a set-based approach (DEset) from the literature. The second one is a new similarity-based approach (DEsim) that takes advantage of the geometric characteristics of a camera to find better solutions.

Findings

The experimental study highlights that RWLS and DEsim-CPLEX are the best proposed algorithms. Both easily outperform CPLEX, and it turns out that RWLS performs better on one class of problem instances, whereas DEsim-CPLEX performs better on another class, depending on the minimal resolution needed in practice.

Originality/value

Up to now, the efficiency of RWLS and the DEset approach has been investigated only for a few problems. Thus, the first contribution is to apply these methods for the first time in the context of camera placement. Moreover, new hybrid DE algorithms are proposed to solve the optimal camera placement problem when stated as a USCP. The second main contribution is the design of the DEsim approach that uses the distance between camera locations to fully benefit from the DE mutation scheme.

Keywords

Acknowledgements

This work was supported by the French Agence Nationale de la Recherche (ANR) as part of the OPMoPS project (ANR-16-SEBM-0004).

Citation

Brévilliers, M., Lepagnot, J., Idoumghar, L., Rebai, M. and Kritter, J. (2018), "Hybrid differential evolution algorithms for the optimal camera placement problem", Journal of Systems and Information Technology, Vol. 20 No. 4, pp. 446-467. https://doi.org/10.1108/JSIT-09-2017-0081

Publisher

:

Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited

Related articles