Exponential decay of correlations for random Lasota-Yorke maps (Q1969096)

One considers random piecewise smooth, piecewise invertible maps on the (0,1) interval. One further assumes that the maps expand distances, do not have too many pieces, do not have too large distorsion and are strongly mixing. Markov properties are not required. It is shown (and this is the main result) that under some conditions (existence of good transfer operator on a space endowed with a probability), there is a normalized density which is globally invariant, this density can be approximated exponentially, and it exhibits the property of exponential mixing. As a result, one has exponential decay of random correlations. The mathematical asumptions are rather general, which causes that the results applies to a large class of stationary and ergodic processes.