PreprintarXiv licencearXiv
The A-optimal design problem and the S-matrix conjecture
Abstract
The S-matrix conjecture was formulated by Sloane and Harwit in 1976, motivated by an A-optimal design problem arising in spectroscopy. It states that if A is a nonsingular real n× n matrix whose entries lie in [0,1], then |A−1|F≥2n/(n+1). Moreover, equality holds if and only if A is an S-matrix. In this paper, we prove the conjecture by combining a variational trace inequality in odd dimensions with a sharp range-constrained estimate for the Moore--Penrose inverse of a centered matrix in even dimensions.
Related
- A Walk From Free Probability to Matrix Discrepancy II: Weaver's Problem and the Kadison-Singer Conjecture
- On the Alzer-Berg problem: an optimal Bernstein boundary and a uniqueness conjecture
- Optimizing Flexibility and Commitments: Partially-Adaptive Multi-Stage Matrix Assembly Layout Design Problem
- On the matrix Ap conjecture
- A Complete Solution to the Conjecture on the Inverse Collatz–Sinogowitz Irregularity Problem
BibTeX
@misc{zhang2026aoptimal,
title = {{The $A$-optimal design problem and the $S$-matrix conjecture}},
author = {Teng Zhang},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.20306},
eprint = {2609.20306},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.20306},
}
RIS
TY - UNPB
TI - The A-optimal design problem and the S-matrix conjecture
AU - Teng Zhang
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2609.20306
UR - https://arxiv.org/abs/2609.20306
AB - The S-matrix conjecture was formulated by Sloane and Harwit in 1976, motivated by an A-optimal design problem arising in spectroscopy. It states that if A is a nonsingular real n× n matrix whose entries lie in [0,1], then |A^−1|_F≥2n/(n+1). Moreover, equality holds if and only if A is an S-matrix. In this paper, we prove the conjecture by combining a variational trace inequality in odd dimensions with a sharp range-constrained estimate for the Moore--Penrose inverse of a centered matrix in even dimensions.
ER -