English

Worst-case time decremental connectivity and k-edge witness

Data Structures and Algorithms 2008-10-31 v1

Abstract

We give a simple algorithm for decremental graph connectivity that handles edge deletions in worst-case time O(klogn)O(k \log n) and connectivity queries in O(logk)O(\log k), where kk is the number of edges deleted so far, and uses worst-case space O(m2)O(m^2). We use this to give an algorithm for kk-edge witness (``does the removal of a given set of kk edges disconnect two vertices u,vu,v?'') with worst-case time O(k2logn)O(k^2 \log n) and space O(k2n2)O(k^2 n^2). For k=o(n)k = o(\sqrt{n}) these improve the worst-case O(n)O(\sqrt{n}) bound for deletion due to Eppstein et al. We also give a decremental connectivity algorithm using O(n2logn/loglogn)O(n^2 \log n / \log \log n) space, whose time complexity depends on the toughness and independence number of the input graph. Finally, we show how to construct a distributed data structure for \kvw by giving a labeling scheme. This is the first data structure for \kvw that can efficiently distributed without just giving each vertex a copy of the whole structure. Its complexity depends on being able to construct a linear layout with good properties.

Keywords

Cite

@article{arxiv.0810.5477,
  title  = {Worst-case time decremental connectivity and k-edge witness},
  author = {Andrew Twigg},
  journal= {arXiv preprint arXiv:0810.5477},
  year   = {2008}
}
R2 v1 2026-06-21T11:36:34.147Z