Approximation result toward nearest neighbor heuristic

In this paper, we revisit the famous heuristic called nearest neighbor (N N) for the traveling salesman problem under maximization and minimization goal. We deal with variants where the edge costs belong to interval Ša;taĆ for a>0 and t>1, which certainly corresponds to practical cases...

Full description

Bibliographic Details
Main Author: Monnot Jér"me
Format: Article
Language:English
Published: University of Belgrade 2002-01-01
Series:Yugoslav Journal of Operations Research
Subjects:
Online Access:http://www.doiserbia.nb.rs/img/doi/0354-0243/2002/0354-02430201011M.pdf
Description
Summary:In this paper, we revisit the famous heuristic called nearest neighbor (N N) for the traveling salesman problem under maximization and minimization goal. We deal with variants where the edge costs belong to interval Ša;taĆ for a>0 and t>1, which certainly corresponds to practical cases of these problems. We prove that NN is a (t+1)/2t-approximation for maxTSPŠa;taĆ and a 2/(t+1)-approximation for minTSPŠa;taĆ under the standard performance ratio. Moreover, we show that these ratios are tight for some instances.
ISSN:0354-0243
1820-743X