Extinction time for a random walk in a random environment (Q2515518)

The paper under review studies the upper bound of the survival probability up to time \(t\) of a random walk with death in \([-N, N]\) moving in a time-dependent environment. A random walk is alive if it is in the prescribed interval \([-N, N]\), and is dead otherwise. The walk on \(\{\pm N, \pm (N-1)\}\) besides moving may die, and this corresponds to the jump and death rates depending on the environment. The main result (Theorem 2.1) of this paper states that the survival probability of the random walk decreases exponentially in time, being bounded above by \(c\exp (-b N^{-2}t)\) for \(c>0\) and \(b>0\) independent of \(t\) and \(N\). The model and the evolution Markov process are given in Section 2 for all possible cases by giving the generators. Section 3 constructs a time-inhomogeneous Markov process whose jump intensities at time are obtained by averaging those of the original process over the environment. The auxiliary process reduces the proof of Theorem 2.1 to two estimates, one is the lower bound (3.8) of the exponent in the survival probability (3.7) and the other one is the upper bound in the exponential decay for the survival probability (3.9). Sections 4 and 5 are devoted to the proofs of (3.8) and (3.9). The techniques are related to the unit rate symmetric simple random walk on integers and the local central limit theorem is used for the proof of (3.9). The lower bound (3.8) is related to three bounds (5.4)--(5.6) by using probability under the stirring process and the law of the stirring process, the time reverting probability, the transition probabilities for the simple random walk in the Appendix as well as in [\textit{I. Karatzas} and \textit{S. E. Shreve}, Brownian motion and stochastic calculus. New York: Springer (1991; Zbl 0734.60060), Chapter 2], and the DPTV introduced in [\textit{A. De Masi} et al., J. Stat. Phys. 144, No. 6, 1151--1170 (2011; Zbl 1229.82128)]. Section 6 adds another estimate required in (3.8). It would be interesting to see how the symmetric simple random walk can be understood to be an asymmetric simple random walk in \([-M, N]\) for \(M\neq N\).