Table of Contents
Fetching ...

Minimum 0-Extension Problems on Directed Metrics

Hiroshi Hirai, Ryuhei Mizutani

TL;DR

This work extends the minimum 0-extension problem to directed metrics, removing symmetry in both the metric and the edge costs. It provides a dichotomy-like classification: tractable cases occur when the directed metric arises from a modular lattice with orbit-invariant edge-length, while hardness arises if the metric is non-modular, the underlying graph is non-orientable, or the directed orbit-invariance fails; it also introduces a bias-based NP-hardness criterion via biased non-collinear triples and completes a star-structure dichotomy. The results connect the tractability of directed valued constraint satisfaction problems to rich geometric and lattice-theoretic structures, yielding efficient verification of tractability and concrete hardness gadgets. Overall, the paper advances a finer understanding of when directed variants of classic network optimization problems admit polynomial-time solutions or are intractable, with implications for directed facility location and related VCSPs.

Abstract

For a metric $μ$ on a finite set $T$, the minimum 0-extension problem 0-Ext$[μ]$ is defined as follows: Given $V\supseteq T$ and $\ c:{V \choose 2}\rightarrow \mathbf{Q_+}$, minimize $\sum c(xy)μ(γ(x),γ(y))$ subject to $γ:V\rightarrow T,\ γ(t)=t\ (\forall t\in T)$, where the sum is taken over all unordered pairs in $V$. This problem generalizes several classical combinatorial optimization problems such as the minimum cut problem or the multiterminal cut problem. Karzanov and Hirai established a complete classification of metrics $μ$ for which 0-Ext$[μ]$ is polynomial time solvable or NP-hard. This result can also be viewed as a sharpening of the general dichotomy theorem for finite-valued CSPs (Thapper and Živný 2016) specialized to 0-Ext$[μ]$. In this paper, we consider a directed version $\overrightarrow{0}$-Ext$[μ]$ of the minimum 0-extension problem, where $μ$ and $c$ are not assumed to be symmetric. We extend the NP-hardness condition of 0-Ext$[μ]$ to $\overrightarrow{0}$-Ext$[μ]$: If $μ$ cannot be represented as the shortest path metric of an orientable modular graph with an orbit-invariant ``directed'' edge-length, then $\overrightarrow{0}$-Ext$[μ]$ is NP-hard. We also show a partial converse: If $μ$ is a directed metric of a modular lattice with an orbit-invariant directed edge-length, then $\overrightarrow{0}$-Ext$[μ]$ is tractable. We further provide a new NP-hardness condition characteristic of $\overrightarrow{0}$-Ext$[μ]$, and establish a dichotomy for the case where $μ$ is a directed metric of a star.

Minimum 0-Extension Problems on Directed Metrics

TL;DR

This work extends the minimum 0-extension problem to directed metrics, removing symmetry in both the metric and the edge costs. It provides a dichotomy-like classification: tractable cases occur when the directed metric arises from a modular lattice with orbit-invariant edge-length, while hardness arises if the metric is non-modular, the underlying graph is non-orientable, or the directed orbit-invariance fails; it also introduces a bias-based NP-hardness criterion via biased non-collinear triples and completes a star-structure dichotomy. The results connect the tractability of directed valued constraint satisfaction problems to rich geometric and lattice-theoretic structures, yielding efficient verification of tractability and concrete hardness gadgets. Overall, the paper advances a finer understanding of when directed variants of classic network optimization problems admit polynomial-time solutions or are intractable, with implications for directed facility location and related VCSPs.

Abstract

For a metric on a finite set , the minimum 0-extension problem 0-Ext is defined as follows: Given and , minimize subject to , where the sum is taken over all unordered pairs in . This problem generalizes several classical combinatorial optimization problems such as the minimum cut problem or the multiterminal cut problem. Karzanov and Hirai established a complete classification of metrics for which 0-Ext is polynomial time solvable or NP-hard. This result can also be viewed as a sharpening of the general dichotomy theorem for finite-valued CSPs (Thapper and Živný 2016) specialized to 0-Ext. In this paper, we consider a directed version -Ext of the minimum 0-extension problem, where and are not assumed to be symmetric. We extend the NP-hardness condition of 0-Ext to -Ext: If cannot be represented as the shortest path metric of an orientable modular graph with an orbit-invariant ``directed'' edge-length, then -Ext is NP-hard. We also show a partial converse: If is a directed metric of a modular lattice with an orbit-invariant directed edge-length, then -Ext is tractable. We further provide a new NP-hardness condition characteristic of -Ext, and establish a dichotomy for the case where is a directed metric of a star.

Paper Structure

This paper contains 18 sections, 23 theorems, 62 equations, 7 figures.

Key Result

Theorem 1.1

Let$\mu$be a rational-valued metric. 0-Ext$[\mu]$is strongly NP-hardA problem $\mathcal{P}$ is called strongly NP-hard if $\mathcal{P}$ is still NP-hard when all numbers of the instance are bounded by some polynomial in the length of the instance. if

Figures (7)

  • Figure 1: (a) NP-hard case (b) tractable case (c) tractable case
  • Figure 2: (a) a modular graph (b) a nonmodular graph
  • Figure 3: (a) a modular directed metric (b) the underlying graph of (a)
  • Figure 4: (a) the case (i) (b) the case (ii) ($k=4$)
  • Figure 5: the function $c'$
  • ...and 2 more figures

Theorems & Definitions (33)

  • Theorem 1.1: karzanov2004
  • Theorem 1.2: hirai2016
  • Theorem 1.3
  • Theorem 1.4
  • Theorem 1.5
  • Theorem 1.6
  • Theorem 2.1: kolmogorov2015
  • Remark 2.2
  • Proposition 2.3: cohen2006; see thapper2016
  • Remark 2.4
  • ...and 23 more