English

Minimizing and Computing the Inverse Geodesic Length on Trees

Data Structures and Algorithms 2019-10-01 v2

Abstract

For any fixed measure HH that maps graphs to real numbers, the MinH problem is defined as follows: given a graph GG, an integer kk, and a target τ\tau, is there a set SS of kk vertices that can be deleted, so that H(GS)H(G - S) is at most τ\tau? In this paper, we consider the MinH problem on trees. We call HH "balanced on trees" if, whenever GG is a tree, there is an optimal choice of SS such that the components of GSG-S have sizes bounded by a polynomial in n/kn/k. We show that MinH on trees is FPT for parameter n/kn/k, and furthermore, can be solved in subexponential time, and polynomial space, if HH is additive, balanced on trees, and computable in polynomial time. A measure of interest is the Inverse Geodesic Length (IGL), which is used to gauge the connectedness of a graph. It is defined as the sum of inverse distances between every two vertices: IGL(G)={u,v}V1dG(u,v)IGL(G)=\sum_{\{u,v\} \subseteq V} \frac{1}{d_G(u,v)}. While MinIGL is W[1]-hard for parameter treewidth, and cannot be solved in 2o(k+n+m)2^{o(k+n+m)} time, even on bipartite graphs with nn vertices and mm edges, the complexity status of the problem remains open on trees. We show that IGL is balanced on trees, to give a 2O((nlogn)5/6)2^{O((n\log n)^{5/6})} time, polynomial space algorithm. The distance distribution of GG is the sequence {ai}\{a_i\} describing the number of vertex pairs distance ii apart in GG: ai={{u,v}:dG(u,v)=i}a_i=|\{\{u, v\}: d_G(u, v)=i\}|. We show that the distance distribution of a tree can be computed in O(nlog2n)O(n\log^2 n) time by reduction to polynomial multiplication. We extend our result to graphs with small treewidth by showing that the first pp values of the distance distribution can be computed in 2O(tw(G))n1+εp2^{O(tw(G))} n^{1+\varepsilon} \sqrt{p} time, and the entire distance distribution can be computed in 2O(tw(G))n1+ε2^{O(tw(G))} n^{1+\varepsilon} time, when the diameter of GG is O(nε)O(n^{\varepsilon'}) for every ε>0\varepsilon'>0.

Keywords

Cite

@article{arxiv.1811.03836,
  title  = {Minimizing and Computing the Inverse Geodesic Length on Trees},
  author = {Serge Gaspers and Joshua Lau},
  journal= {arXiv preprint arXiv:1811.03836},
  year   = {2019}
}

Comments

21 pages, 4 figures

R2 v1 2026-06-23T05:10:04.852Z