Spectral analysis of Sinai's walk for small eigenvalues (Q2468426)

One considers random walks on \(Z\) with a random potential \(V\) such that \(V\), under diffusive rescaling, converges weakly to a two-sided Brownian motion. This is the so called Sinai's walk. The authors study the spectrum of the generator of Sinai's walk on \([-N,N]\cap Z\) with Dirichlet conditions on \(-N,N\). One gives a refined and fully rigorous analysis of the bottom part of this spectrum, that is for eigenvalues of order \(o(\exp(-h\sqrt{N}))\), by using the distribution of the \(h\)-extrema of the rescaled potential \(V(Nx)/\sqrt{N}\) on \([-1,1]\) and the statistics of \(h\)-extrema of Brownian motion [cf. \textit{J. Neveu} and \textit{J. Pitman}, Lect. Notes Math. 1372, 239--247 (1989; Zbl 0741.60080)]. One applies the results to get an easy proof of a refined version of Sinai's localization theorem. One also express the fact that Sinai's random walk can be seen as a process that on an infinite sequence of random time scales appears to be approaching equilibrium exponentially.