English

Dense Steiner problems: Approximation algorithms and inapproximability

Data Structures and Algorithms 2020-04-30 v1

Abstract

The Steiner Tree problem is a classical problem in combinatorial optimization: the goal is to connect a set TT of terminals in a graph GG by a tree of minimum size. Karpinski and Zelikovsky (1996) studied the δ\delta-dense version of {\sc Steiner Tree}, where each terminal has at least δV(G)T\delta |V(G)\setminus T| neighbours outside TT, for a fixed δ>0\delta > 0. They gave a PTAS for this problem. We study a generalization of pairwise δ\delta-dense {\sc Steiner Forest}, which asks for a minimum-size forest in GG in which the nodes in each terminal set T1,,TkT_1,\dots,T_k are connected, and every terminal in TiT_i has at least δTj\delta |T_j| neighbours in TjT_j, and at least δS\delta|S| nodes in S=V(G)(T1Tk)S = V(G)\setminus (T_1\cup\dots\cup T_k), for each i,ji, j in {1,,k}\{1,\dots, k\} with iji\neq j. Our first result is a polynomial-time approximation scheme for all δ>1/2\delta > 1/2. Then, we show a (1312+ε)(\frac{13}{12}+\varepsilon)-approximation algorithm for δ=1/2\delta = 1/2 and any ε>0\varepsilon > 0. We also consider the δ\delta-dense Group Steiner Tree problem as defined by Hauptmann and show that the problem is APX\mathsf{APX}-hard.

Keywords

Cite

@article{arxiv.2004.14102,
  title  = {Dense Steiner problems: Approximation algorithms and inapproximability},
  author = {Marek Karpinski and Mateusz Lewandowski and Syed Mohammad Meesum and Matthias Mnich},
  journal= {arXiv preprint arXiv:2004.14102},
  year   = {2020}
}