Limit law for transition probabilities and moderate deviations for Sinai's random walk in random environment (Q1408500)

Let \(\xi^x=(\xi^x_t)_{t\geq 0}\) be the Markov process on \(Z\) starting at \(x\), such that, if currently at site \(y\), it jumps to \(y\pm 1\) with rate \(\omega^{\pm}_y\), where \(\omega^{\pm}_y\), \(y\in Z\), are given positive numbers. Suppose that \(\omega=(\omega^+_x,\omega^-_x)_{x\in Z}\) is a fixed realization of an i.i.d sequence of positive random variables. \(\omega\) is considered as the environment and \(\xi^x\) as the random walk in the random environment \(\omega\). The random walk in random environment is called ``Sinai's regime'' if \(E\ln(\omega^+_0/ \omega^-_0)=0\) and \(E\ln^2 (\omega^-_i/\omega^+_i)\in (0,\infty).\) A one-dimensional random walk in random environment in Sinai's regime is studied. The main result is that logarithms of the translation probabilites, after a suitable rescaling, converge in distribution as time tends to infinity, to some functional of the Brownian motion. The law of this functional when the initial and final points agree is computed. The probability of being at time \(t\) at distance at least \(z\) from the initial position is estimated, if \(z\) is greater than \(\ln^2t,\) but still of logarithmic order in time.