arXiv subjects
Publications and source records attributed to Evgenij Kritchevski.
We revisit the classical problem of alignment in products of positive matrices. A direct analysis yields a sharp product misalignment bound, providing an elementary route to alignment and Birkhoff's contraction formula.
We consider two models of one-dimensional discrete random Schrodinger operators (H_n ψ)_l =ψ_{l-1}+ψ_{l +1}+v_l ψ_l, ψ_0=ψ_{n+1}=0 in the cases v_k=σω_k/\sqrt{n} and v_k=σω_k/ \sqrt{k}. Here ω_k are independent random variables with mean 0 and variance 1. We show that the eigenvectors are delocalized and the transfer matrix evolution has a scaling limit given by a stochastic differential equation. In both cases, eigenvalues near a fixed bulk energy E have a point process limit. We give bounds on the eigenvalue repulsion, large gap probability, identify the limiting intensity and provide a central limit theorem. In the second model, the limiting processes are the same as the point processes obtained as the bulk scaling limits of the beta-ensembles of random matrix theory. In the first model, the eigenvalue repulsion is much stronger.
We study the eigenvalue statistics for the hieracharchial Anderson model of Molchanov. We prove Poisson fluctuations at arbitrary disorder, when the the model has spectral dimension d<1. The proof is based on Minami's technique and we give an elementary exposition of the probabilistic arguments.
We prove that a large class of hierarchical Anderson models with spectral dimension ${\rm d}\leq 2$ has only pure point spectrum.