Searcharxiv⌕ Search

arXiv subjects

Anh Thi Nguyen

Publications and source records attributed to Anh Thi Nguyen.

2 recordsLinked to original sources

Heron-Wasserstein majorization inequalities for spectral and Kubo-Ando geometric means

We prove sharp Heron-type majorization inequalities for two quadratic matrix expressions associated with the spectral and Kubo-Ando geometric means. For the spectral geometric mean cross term, we show that \[ λ\bigl(a^2A+b^2B+c(A\natural B)\bigr) \prec_w λ\bigl(W_{a,b}(A,B)\bigr), \qquad 0\le c\le 2ab, \] where $W_{a,b}(A,B)$ is the weighted Bures-Wasserstein expression. The coefficient $2ab$ is sharp, and at this endpoint the weak majorization becomes majorization. For the Kubo-Ando geometric mean, we prove the direct comparison \[ λ\bigl(a^2A+b^2B+2ab(A\#B)\bigr) \prec_w λ\bigl(W_{a,b}(A,B)\bigr). \] This settles, in the two-variable setting, Bhatia's question of whether the Heron-type norm inequality of Bhatia-Lim-Yamazaki admits a weak-majorization refinement. More precisely, we prove \[ λ\bigl(a^2A+b^2B+2ab(A\#B)\bigr) \prec_w λ\bigl((aA^{1/2}+bB^{1/2})^2\bigr), \] and consequently obtain the corresponding inequality for all unitarily invariant norms.

math.FA↗

Domino Tilings of Cruciform Regions

P. Di Francesco first introduced the "Aztec triangle" in his study of the relationship between the twenty-vertex model and domino tilings. He conjectured an exact formula for the number of tilings of the Aztec triangle, and it has since been proved by several authors. In an attempt to prove the conjecture, M. Ciucu showed that the tiling number of the Aztec triangle divides the tiling number of a new region called the "cruciform region," a superposition of two Aztec rectangles. Ciucu proved that the number of domino tilings of a cruciform region is given by a simple product formula. In this paper, we generalize Ciucu's tiling formula by providing a generating-function formula for the cruciform region.

math.CO↗