Frogs in random environment (Q5931544)

The problem of recurrence/transience for random motions in random media is studied. Apparently, \textit{F. Solomon} [Ann. Probab. 3, No. 1, 1-31 (1975; Zbl 0305.60029)] was the first to treat the problems of such kind: he studied one-dimensional random walk (r.w.) in random environment (r.e.). Some attention was given to r.w. in r.e. in more complex spaces, e.g., trees, words from a finite alphabet, \(\mathbb Z^{d}\). Branching r.w. (b.r.w.) in r.e. was also studied in one-dimensional r.e. and many-dimensional r.e. In the present paper the random motion in question is, in some sense, something in the frontier between r.w. and b.r.w. Namely, it is studied the so-called frog model: initially there are some ``sleeping'' particles and one ``active'' particle. A sleeping particle is activated when an active particle hits it, after that the activated particle starts to walk independently of everything and can activate other sleeping particles as well. The initial configuration of sleeping particles is random with density \(p(x).\) There is no interaction between active particles. The folkloric name for this kind of model is ``frog model''. The critical rate of decay of \(p(x)\) separating transience from recurrence is identified, and some other properties of the model are studied.