On Grünbaum's problem for symmetric configurations
Abstract
Let gn be the largest number of Euclidean balls of diameter $1$ which may be needed to cover a set of diameter $1$ in ℝn. We study this problem for finite sets invariant under all coordinate permutations. We prove that the exponential growth rate in this symmetric problem can be characterized exactly as a finite-alphabet squared-error rate-distortion supremum α0. Specialized to the two-point case, i.e., for subsets of Boolean cubes, this gives the explicit lower bound gn≥ (1.160235457…−o(1))n, improving the previous best bound (2/√3−o(1))n. Using Fix's Gaussian characterization of the rate-distortion problem, we give a numerical three-point construction with exponent base greater than $1.160497831$. Finally, we show that α0 is not attained by any finitely supported distribution.
Related
- Symmetric, rational and infinitely many (234) configurations, and the end of the existence problem for geometric (n4) configurations
- The HRT Conjecture for Symmetric Configurations and Real-Valued Functions
- Finiteness of Dziobek central configurations in the Newtonian five-body problem
- Finiteness of planar central configurations of five bodies for all positive masses: On Smale's sixth problem
- Uniqueness of convex central configurations in the planar (1 + 4)-body problem
BibTeX
@misc{arman2026grünbaums,
title = {{On Grünbaum's problem for symmetric configurations}},
author = {Andrii Arman and Andriy Bondarenko and Andriy Prymak and Danylo Radchenko},
howpublished = {HAL},
year = {2026},
month = jul,
url = {https://hal.science/hal-05851861v1},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - On Grünbaum's problem for symmetric configurations
AU - Andrii Arman
AU - Andriy Bondarenko
AU - Andriy Prymak
AU - Danylo Radchenko
PY - 2026
DA - 2026/07/24
PB - HAL
UR - https://hal.science/hal-05851861v1
AB - Let g_n be the largest number of Euclidean balls of diameter $1$ which may be needed to cover a set of diameter $1$ in ℝ^n. We study this problem for finite sets invariant under all coordinate permutations. We prove that the exponential growth rate in this symmetric problem can be characterized exactly as a finite-alphabet squared-error rate-distortion supremum α_0. Specialized to the two-point case, i.e., for subsets of Boolean cubes, this gives the explicit lower bound g_n≥ (1.160235457…−o(1))^n, improving the previous best bound (2/√3−o(1))^n. Using Fix's Gaussian characterization of the rate-distortion problem, we give a numerical three-point construction with exponent base greater than $1.160497831$. Finally, we show that α_0 is not attained by any finitely supported distribution.
ER -