Uniform asymptotic expansion conjecture for eigenvalues of higher-order Toeplitz symbols

From papers

Let gm(x)=(2sinx2)2mg_m(x)=\left(2\sin\frac{x}{2}\right)^{2m} for an integer m3m\geq 3, and let λn,j\lambda_{n,j} be the eigenvalues of the Toeplitz matrices generated by gmg_m. Let dkd_k denote the coefficient functions in the proposed regular expansion. Higher-order Toeplitz expansion conjecture. If p2m1p\leq 2m-1, there exist NpNN_p\in\mathbb{N} and Dp>0D_p>0 such that

λn,jk=0pdk(jπn+2)(n+2)kDp(n+2)p+1\left|\lambda_{n,j}-\sum_{k=0}^p \frac{d_k\left(\frac{j\pi}{n+2}\right)}{(n+2)^k}\right|\leq\frac{D_p}{(n+2)^{p+1}}

for all nNpn\geq N_p and all j{1,,n}j\in\{1,\ldots,n\}. For p=2mp=2m, this inequality does not hold for all sufficiently large nn and all 1jn1\leq j\leq n, but it does hold for all sufficiently large nn and (log(n+2))2jn(\log(n+2))^2\leq j\leq n. This predicts uniform regular asymptotics through order 2m12m-1, while identifying the boundary layer near j=1j=1 as the obstruction at order 2m2m.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

Mauricio Barrera, Albrecht Boettcher, Sergei M. Grudsky and Egor A. Maximenko, “Eigenvalues of even very nice Toeplitz matrices can be unexpectedly erratic”, arXiv:1710.05243 (2017).

Solutions 0

No solutions have been posted yet.