Approximations for the Disjoint Paths Problem in High-Diameter Planar Networks

39Citations
Citations of this article
27Readers
Mendeley users who have this article in their library.

This article is free to access.

Abstract

We consider the problem of connecting distinguished terminal pairs in a graph via edge-disjoint paths. This is a classical NP-complete problem for which no general approximation techniques are known; it has recently been brought into focus in papers discussing applications to admission control in high-speed networks and to routing in all-optical networks. In this paper we provide O(log n)-approximation algorithms for two natural optimization versions of this problem for the class of nearly Eulerian, uniformly high-diameter planar graphs, which includes two-dimensional meshes and other common planar interconnection networks. We give an O(log n)-approximation to the maximum number of terminal pairs that can be simultaneously connected via edge-disjoint paths, and an O(log n)-approximation to the minimum number of wavelengths needed to route a collection of terminal pairs in the "optical routing" model considered by Raghavan, Upfal, and others. The latter result improves on an O(log2n)-approximation for the special case of the mesh obtained independently by Aumann and Rabani. For both problems the O(log n)-approximation is a consequence of an O(1)-approximation for the special case when all terminal pairs are roughly the same distance apart. Our algorithms make use of a number of new techniques, including the construction of a "crossbar" structure in any nearly Eulerian planar graph, and develops some connections with classical matroid algorithms. © 1998 Academic Press.

Cite

CITATION STYLE

APA

Kleinberg, J., & Tardos, É. (1998). Approximations for the Disjoint Paths Problem in High-Diameter Planar Networks. Journal of Computer and System Sciences, 57(1), 61–73. https://doi.org/10.1006/jcss.1998.1579

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free