On some random walks on \(\mathbb{Z}\) in random medium (Q1872294)

The following model of random walks in random environment is considered. Let \((\Omega, {\mathcal F}, \mu, T)\) be an invertable dynamical system, i.e., \((\Omega, {\mathcal F}, \mu)\) is a probability space and \(T\) is a map on \(\Omega\), invertable and measurable together with its inverse. It is ergodic with respect to \(\mu\) and measure preserving. Given integers \(L, R \geq 1\), the set \(\Lambda\) is defined as \(\Lambda = \{ -L, -L+1 , \dots , 0 , \dots , R\}\). This is the set of possible jumps of the random walk on \(\mathbb Z\). The probabilities of such jumps are given by a collection \((p_z)_{z \in \Lambda}\) of positive random variables on \((\Omega, {\mathcal F}, \mu)\). They are supposed to obey the conditions: \(\sum_{z \in \Lambda }p_z (\omega) = 1\) \(\mu\)-almost surely; \(\exists \varepsilon >0\) such that \(\forall z \in \Lambda \setminus \{0\}\), one has \((p_z /p_R) \geq \varepsilon\) \(\mu\)-almost surely. The random walk \(\xi_n (\omega)\) is defined as a Markov chain by the transition probabilities \[ {\mathcal P}_\omega (\xi_{n+1} (\omega) = x + z \mid \xi_n (\omega) = x) = p_z (T^x \omega), \quad \;x \in {\mathbb Z}, \quad z \in \Lambda, \] and \(\xi_0 (\omega) = 0\). For this model with \(R=1\) the following results are obtained. A recurrence criterion expressed in terms of the sign of the maximal Lyapunov exponent of a certain random matrix is provided. An algorithm of calculating of this exponent is given. The existence of the absolutely continuous invariant measure for the Markov chain of ``environments viewed from the particle'' and of the nonzero drift (in the transient case) are characterized. A sufficient condition for the validity of the central limit theorem in the transient case is given.