An iterated "hyperplane exploration" approach for the quadratic knapsack problem - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Computers and Operations Research Année : 2017

An iterated "hyperplane exploration" approach for the quadratic knapsack problem

Résumé

The quadratic knapsack problem (QKP) is a well-known combinatorial optimization problem with numerous applications. Given its NP-hard nature, finding optimal solutions or even high quality suboptimal solutions to QKP in the general case is a highly challenging task. In this paper, we propose an iterated “hyperplane exploration” approach (IHEA) to solve QKP approximately. Instead of considering the whole solution space, the proposed approach adopts the idea of searching over a set of hyperplanes defined by a cardinality constraint to delimit the search to promising areas of the solution space. To explore these hyperplanes efficiently, IHEA employs a variable fixing strategy to reduce each hyperplane-constrained sub-problem and then applies a dedicated tabu search procedure to locate high quality solutions within the reduced solution space. Extensive experimental studies over three sets of 220 QKP instances indicate that IHEA competes very favorably with the state-of-the-art algorithms both in terms of solution quality and computing efficiency. We provide additional information to gain insight into the key components of the proposed approach.

Fichier principal
Vignette du fichier
chenhaocor2017.pdf (201.63 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01426415 , version 1 (04-01-2017)
hal-01426415 , version 2 (04-06-2017)

Identifiants

Citer

Yuning Chen, Jin-Kao Hao. An iterated "hyperplane exploration" approach for the quadratic knapsack problem. Computers and Operations Research, 2017, 77, pp.226-239. ⟨10.1016/j.cor.2016.08.006⟩. ⟨hal-01426415v2⟩

Collections

UNIV-ANGERS LERIA
74 Consultations
283 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More