Optimization Letters

, Volume 6, Issue 3, pp 415–420

On the maximum TSP with γ-parameterized triangle inequality

Original Paper

DOI: 10.1007/s11590-010-0266-y

Cite this article as:
Li, W. & Shi, Y. Optim Lett (2012) 6: 415. doi:10.1007/s11590-010-0266-y
  • 71 Downloads

Abstract

The maximum TSP with γ-parameterized triangle inequality is defined as follows. Given a complete graph G = (V, E, w) in which the edge weights satisfy w(uv) ≤ γ · (w(ux) + w(xv)) for all distinct nodes \({u,x,v \in V}\), find a tour with maximum weight that visits each node exactly once. Recently, Zhang et al. (Theor Comput Sci 411(26–28):2537–2541, 2010) proposed a \({\frac{\gamma+1}{3\gamma}}\)-approximation algorithm for \({\gamma\in\left[\frac{1}{2},1\right)}\). In this paper, we show that the approximation ratio of Kostochka and Serdyukov’s algorithm (Upravlyaemye Sistemy 26:55–59, 1985) is \({\frac{4\gamma+1}{6\gamma}}\), and the expected approximation ratio of Hassin and Rubinstein’s randomized algorithm (Inf Process Lett 81(5):247–251, 2002) is \({\frac{3\gamma+\frac{1}{2}}{4\gamma}-O\left(\frac{1}{\sqrt{n}}\right)}\), for \({\gamma\in\left[\frac{1}{2},+\infty\right)}\). These improve the result in Zhang et al. (Theor Comput Sci 411(26–28):2537–2541, 2010) and generalize the results in Hassin and Rubinstein and Kostochka and Serdyukov (Inf Process Lett 81(5):247–251, 2002; Upravlyaemye Sistemy 26:55–59, 1985).

Keywords

Maximum traveling salesman problemParameterized triangle inequalityApproximation algorithm

Copyright information

© Springer-Verlag 2010

Authors and Affiliations

  1. 1.Department of Atmospheric ScienceYunnan UniversityKunmingPeople’s Republic of China
  2. 2.Department of Architectural EngineeringChongQing Technology and Business InstituteChongqingPeople’s Republic of China