1: \begin{abstract}

2:    We study the convergence of random function iterations for finding

3:    an invariant measure of the corresponding Markov operator.  We call

4:    the problem of finding such an invariant measure the {\em

5:      stochastic fixed point problem}.  This generalizes earlier work

6:    studying the {\em stochastic feasibility problem}, namely, to find

7:    points that are, with probability 1, fixed points of the random

8:    functions \cite{HerLukStu19a}.  When no such points exist, the

9:    stochastic feasibility problem is called {\em inconsistent}, but

10:    still under certain assumptions, the more general stochastic fixed

11:    point problem has a solution and the random function iterations

12:    converge to an invariant measure for the corresponding Markov

13:    operator.  There are two major types of convergence: almost sure

14:    convergence of the iterates to a fixed point in the case of

15:    stochastic feasibility, and convergence in distribution more

16:    generally.  We show how common structures in deterministic fixed

17:    point theory can be exploited to establish existence of invariant

18:    measures and convergence of the Markov chain.  We show that weaker

19:    assumptions than are usually encountered in the analysis of Markov

20:    chains guarantee linear/geometric convergence.  This framework

21:    specializes to many applications of current interest including, for

22:    instance, stochastic algorithms for large-scale distributed

23:    computation, and deterministic iterative procedures with

24:    computational error.  The theory developed in 

25: this study provides a solid basis for describing the convergence

26: of simple computational methods without the assumption of infinite precision

27: arithmetic or vanishing computational errors.  

28: 

29: 

30: \end{abstract}

31: