Please use this identifier to cite or link to this item: https://research.matf.bg.ac.rs/handle/123456789/440
Title: Skewed Variable Neighborhood Search Method for the Weighted Generalized Regenerator Location Problem
Authors: Mrkela, Lazar
Stanimirović, Zorica 
Affiliations: Numerical Mathematics and Optimization 
Keywords: Optical networks;Skewed Variable Neighborhood Search;Telecommunication;Weighted Generalized Regenerator Location Problem
Issue Date: 1-Jan-2019
Related Publication(s): International Conference on Variable Neighborhood Search ICVNS 2018
Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Abstract: 
This paper deals with the Weighted Generalized Regenerator Location Problem (WGRLP) that arises in the design of optical telecommunication networks. During the transmission of optical signal, its quality deteriorates with the distance from the source, and therefore, it has to be regenerated by installing regenerators at some of the nodes in the network. The WGRLP involves weights assigned to potential regenerator locations, reflecting the costs of regenerator deployment. The objective of WGRLP is to minimize the sum of weights assigned to locations with installed regenerators, while ensuring a good quality communication among terminal nodes. As telecommunication networks usually involve large number of nodes, an efficient optimization method is required to deal with real-life problem dimensions. In this paper, a Skewed Variable Neighborhood Search method (SVNS) is proposed as solution approach for the WGRLP. The designed SVNS uses adequate data structures for solution representation and efficient procedures for objective function update, feasibility check, and solution repair. Computational results on the WGRLP data set from the literature show that the proposed SVNS reaches all known optimal solutions on small and medium size instances in short running times and outperforms existing heuristic approaches for the WGRLP. In addition, SVNS is tested on large scale WGRLP instances not considered in the literature so far. The presented computational results indicate the potential of SVNS as solution method for WGRLP and related network design problems.
URI: https://research.matf.bg.ac.rs/handle/123456789/440
ISBN: 9783030158422
ISSN: 03029743
DOI: 10.1007/978-3-030-15843-9_15
Appears in Collections:Research outputs

Show full item record

SCOPUSTM   
Citations

1
checked on Nov 15, 2024

Page view(s)

28
checked on Nov 15, 2024

Google ScholarTM

Check

Altmetric

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.