Sparse Bounded Hop-Spanners for Geometric Intersection Graphs

πŸ“… 2025-04-08
πŸ“ˆ Citations: 0
✨ Influential: 0
πŸ“„ PDF
πŸ€– AI Summary
This paper investigates tight asymptotic bounds on the size of 2-hop and 3-hop sparse bounded-hop spanners in *d*-dimensional geometric intersection graphsβ€”such as unit-ball graphs, fat axis-aligned boxes, and semi-algebraic objects. Leveraging a synthesis of combinatorial geometry, divide-and-conquer with grid-based partitioning, VC-dimension and Ξ΅-net theory, and semi-algebraic methods, it establishes near-optimal upper and lower bounds for classical geometric graph families in high dimensions. Key contributions include: (i) an upper bound of $O^*(n^{3/2 - 1/(2lfloor d/2 floor + 1)})$ on the size of 2-hop spanners for *d*-dimensional ball graphs; and (ii) a lower bound of $Omega(n^{4/3})$ for 3-hop spanners in 3D tetrahedron intersection graphs. Collectively, these results characterize the fundamental trade-offs between dimensionality, geometric complexity, and sparsity in bounded-hop spanner construction.

Technology Category

Application Category

πŸ“ Abstract
We present new results on $2$- and $3$-hop spanners for geometric intersection graphs. These include improved upper and lower bounds for $2$- and $3$-hop spanners for many geometric intersection graphs in $mathbb{R}^d$. For example, we show that the intersection graph of $n$ balls in $mathbb{R}^d$ admits a $2$-hop spanner of size $O^*left(n^{frac{3}{2}-frac{1}{2(2lfloor d/2 floor +1)}} ight)$ and the intersection graph of $n$ fat axis-parallel boxes in $mathbb{R}^d$ admits a $2$-hop spanner of size $O(n log^{d+1}n)$. Furthermore, we show that the intersection graph of general semi-algebraic objects in $mathbb{R}^d$ admits a $3$-hop spanner of size $O^*left(n^{frac{3}{2}-frac{1}{2(2D-1)}} ight)$, where $D$ is a parameter associated with the description complexity of the objects. For such families (or more specifically, for tetrahedra in $mathbb{R}^3$), we provide a lower bound of $Omega(n^{frac{4}{3}})$. For $3$-hop and axis-parallel boxes in $mathbb{R}^d$, we provide the upper bound $O(n log ^{d-1}n)$ and lower bound $Omegaleft(n (frac{log n}{log log n})^{d-2} ight)$.
Problem

Research questions and friction points this paper is trying to address.

Study 2- and 3-hop spanners for geometric intersection graphs
Improve bounds for spanners in various geometric graphs
Analyze size complexity for different object families
Innovation

Methods, ideas, or system contributions that make the work stand out.

Improved 2-hop spanners for geometric graphs
Enhanced 3-hop spanners for semi-algebraic objects
Optimized bounds for axis-parallel box spanners
πŸ”Ž Similar Papers
No similar papers found.