Counterexamples for directed and node capacitated cut-trees

by András A. Benczúr

Siam Journal on Computing 24(3):505-510 (1995)

ps.gz ps pdf

Document in Citeseer database

Citations from the paper (the infamous incorrect Schnorr paper) as in the Hypertext Bibliography Project (apparently not maintained since 1998)

Abstract

We show that there is no cut-tree for various connectivity concepts, hence pointing out to errors in the papers of Schnorr and Gusfield and Naor. Gomory and Hu constructed a cut-tree for undirected graphs which compactly represents a minimum cut for each pair of vertices. This has a straightforward generalization to directed Eulerian graphs, cf. Gupta. To arbitrary directed graphs a generalization was given by Schnorr. There is a well-known transformation of vetex connectivity to directed edge connectivity; directed edge cuts correspond to vertex cuts in some weak sense. The result of Schnorr was later applied by Gusfield and Naor for such a cut-tree construction. In this paper counterexamples are described to show that for directed graphs there is no cut-tree and therefore the cut-tree results of Schnorr and Gusfield and Naor are incorrect. Our final example shows that, without weakening the notion of vertex connectivity, it is impossible to construct vertex cut-trees for undirected graphs in general.

Related articles: these results are incorrect

Schnorr, C.P.
Bottlenecks and edge connectivity in unsymmetrical networks
SIAM. J. Comput. 8 (1979) 265--274.

Gusfield, D. and D. Naor
Efficient Algorithms for Generalized Cut-Trees
Networks 21 (1991) 505--520.

Citations

  1. Romeo Rizzi
    Excluding a simple good pair approach to directed cuts
    Graphs and Combinatorics, to appear