Estimation of steady-state central moments by the regenerative method of simulation (Q1094801)

Let X(t) be a regenerative process taking values in a state space S. The probability measure \(\pi\) on S is said to be a steady-state distribution if \[ \lim_{t\to \infty}t^{- 1}\int^{t}_{0}f(X(s))ds=\int_{S}f(y)\pi (dy) \] for any f(.) such that \(\int_{S}| f(y)| \pi (dy)<\infty\). The authors propose estimators for some functionals of \(\pi\). The almost evident asymptotic properties of these statistics are established. The paper is supplied with examples of simulated regenerative processes.