An Efficient Variable Neighborhood Search for Generalized Regenerator Location Problems
-
Published:2022-08
Issue:05
Volume:31
Page:
-
ISSN:0218-2130
-
Container-title:International Journal on Artificial Intelligence Tools
-
language:en
-
Short-container-title:Int. J. Artif. Intell. Tools
Affiliation:
1. Faculty of Information Technology, Belgrade Metropolitan University, Tadeuša Košcuška 63, 11000 Belgrade, Serbia
Abstract
Optical networks are one of the fastest data communication networks, which use optical fiber cables for transmitting data in the form of light pulses between sender and receiver nodes. As the quality of optical signal deteriorates when increasing the distance among terminal nodes, expensive electronic devices (regenerators) must be deployed in the network to recover the signal quality. The objective of Generalized Regenerator Location Problem (GRLP) is to minimize the number of installed regenerators, while weighted variant (WGRLP) assumes different installation costs (weights) of potential regenerator locations and minimizes the sum of regenerator installation costs. In this study, an improved variant of Basic Variable Neighborhood Search (iBVNS) was developed as a metaheuristic solution approach to both GRLP and WGRLP. Computational results provided by the proposed iBVNS on all available data sets from the literature are compared with the results of existing solution methods for GRLP and WGRLP. The obtained results show that the proposed iBVNS quickly reaches optimal or near-optimal solutions and in many cases improves upper bounds provided by exact method. In addition, the conducted non-parametric statistical tests indicate that iBVNS outperforms all heuristic approaches proposed earlier in the literature for the considered problems.
Publisher
World Scientific Pub Co Pte Ltd
Subject
Artificial Intelligence,Artificial Intelligence