Markov Chain Moment Formulas for Regenerative Simulation
STANFORD UNIV CA DEPT OF OPERATIONS RESEARCH
Pagination or Media Count:
Let X sub n n or be a regenerative Markov chain on a general state space, and f a real-valued bounded function. Let tau and Z be random variables that have the distribution of a regeneration cycle length and the sum of fXk over a cycle, respectively. This paper derives expressions for moments of the form E tau superscript j Z superscript k, which are then used to gain insight into the qualities of regenerative estimators based on different regeneration points.
- Statistics and Probability