On a class of stable random dynamical systems: Theory and applications (Q5931249): Difference between revisions

Products of independent identically distributed maps from a (not necessarily bounded) interval to itself are considered. The maps are assumed to be monotone, either non-decreasing or non-increasing. It is shown that there exists a unique invariant measure for the induced Markov operator if there exists a point in the interval and an \(N\in\mathbb{N}\) such that both the probability that all points from the interval are moved below that point by the \(N\)th iterate, and the probability that all points are moved above that point by the \(N\)th iterate, are positive. This condition implies furthermore that the Kolmogorov distance between the distributions of the \(n\)th iterate and the invariant measure converges to zero exponentially fast, for every initial condition. In case of a compact interval also a central limit theorem is derived. In a section on applications the case of an evolution of the capital stock of an economy given by an interaction of growth and cyclic forces is discussed.