# The complexity of the reliable connectivity problem

• Lefteris M. Kirousis
• Paul Spirakis
Contributions
Part of the Lecture Notes in Computer Science book series (LNCS, volume 520)

## Abstract

Let G=(V, E) be a graph together with two distinguished nodes s and t, and suppose that to every node vV, a nonnegative integer f(v)≤degree(v) is assigned. Suppose, moreover, that each node v can cause at most f(v) of its incident edges to “fail” (these f(v) edges can be arbitrarily chosen). The Reliable Connectivity Problem is to test whether node s remains connected with t with a path of non-failed edges for all possible choices of the failed edges. We first show that the complement of the Reliable Connectivity Problem is NP-complete and that this remains true even if G is restricted to the class of directed and acyclic graphs. However, we show that the problem is in P for directed and acyclic graphs if we assume that the edges caused to fail by each node v are chosen only from the edges incoming to v. Concerning the parallel complexity of this version of the problem, it turns out that it is P-complete. Moreover, approximating the maximum d such that for any choice of failed edges there is a directed path of non-failed edges that starts from s and has length d turns out to be P-complete as well, for any given degree of relative accuracy of the approximation. On the contrary, given that every node v will cause at least f(v) incoming edges to fail, the question whether there is a choice of failed edges such that s remains connected with t via non-failed edges turns out to be in NC, even for general graphs.

## Keywords

Directed Path Directed Acyclic Graph Undirected Graph Incident Edge Failure Pattern
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

## References

1. R. Anderson and E.W. Mayr [1984] A P-Complete Problem and Approximations to It, Technical Report, Computer Science Dept., Stanford University, California.Google Scholar
2. M.O. Ball [1977], Network Reliability and Analysis: Algorithms and Complexity, Doctoral Thesis, Operations Research Dept., Cornell University, Ithaca, New York.Google Scholar
3. L.M. Kirousis, M. Serna and P. Spirakis [1989], The parallel complexity of the subgraph connectivity problem, Proceedings of the 30th Annual Symposium on Foundations of Computer Science, (IEEE Computer Society Press), 294–299.Google Scholar
4. A. Rosenthal [1974], Computing Reliability of Complex Systems, Doctoral Thesis, Dept. of Electrical Engineering and Computer Science, University of California, Berkeley, California.Google Scholar
5. M. Serna and P. Spirakis [1991], “Tight RNC approximations to max flow,” to appear in: Proceedings of the 8th Symposium on Theoretical Aspects of Computer Science (Lecture Notes in Computer Science, Springer-Verlag).Google Scholar
6. M. Yannakakis [1981], Edge-deletion problems, SIAM J. Computing 10, 297–309.Google Scholar

## Authors and Affiliations

• 2
• Lefteris M. Kirousis
• 1
• 2
• Paul Spirakis
• 1
• 2
• 3
1. 1.Department of Computer Science and EngineeringUniversity of PatrasPatrasGreece
2. 2.Computer Technology InstitutePatrasGreece
3. 3.Courant Institute of Mathematical Sciences, NYUU.S.A.