English

Proximal Navigation Graphs and t-spanners

Computational Geometry 2022-03-02 v3

Abstract

Let (X,d)(X,\mathbf{d}) be a metric space, VXV\subseteq X a finite set, and EV×VE \subseteq V \times V. We call the graph G(E,V)G(E,V) a {\em metric} graph if each edge (u,v)E(u,v) \in E has weight d(u,v)d(u,v). In particular edge (u,u)(u,u) is in the graph and have distance 00. We call GG a {\em proximal navigation graph} or PNPN-graph if for each edge (u,v)E(u,v) \in E either u=vu=v or there is a node u1u_1 such that (u,u1)E(u,u_1) \in E and d(u,v)>d(u1,v)\mathbf{d}(u,v) > \mathbf{d}(u_1,v). In such graph it is possible to navigate greedily from an arbitrary source node to an arbitrary target node by reducing the distance between the current node and the target node in each step. The complete graph, the Delaunay triangulation and the Half Space Proximal (HSP) graph (defined below in the paper) are examples of PNPN-graphs. In this paper we study the relationship between PNPN-graphs and tt-spanners and prove that there are PNPN-graphs that are not tt-spanners for any tt. On the positive side we give sufficient conditions for a PNPN-graph to be a tt-spanner and prove that any PNPN-graph over Rn\mathbb{R}^n under the euclidean distance is a tt-spanner.

Keywords

Cite

@article{arxiv.1404.1646,
  title  = {Proximal Navigation Graphs and t-spanners},
  author = {Guillermo Ruiz and Edgar Chávez},
  journal= {arXiv preprint arXiv:1404.1646},
  year   = {2022}
}

Comments

There is an error in the main claim

R2 v1 2026-06-22T03:44:16.241Z