Estimation for Discrete-time Semi-Markov Reward Processes: Analysis and Inference
- 92 Downloads
In this paper, the moments of a class of reward processes defined on a discrete-time semi-Markov process and the asymptotic behaviors of the corresponding empirical estimators have been investigated. Some known results concerning the asymptotic distribution and properties of semi-Markov kernel have been obtained by a different approach. By using the empirical estimator of the semi-Markov kernel and the mentioned approach, the estimators for the moments of the reward process have been introduced and their asymptotic properties have been established. As a consequence of the strong consistency and asymptotic normality, the confidence intervals have also been obtained. A numerical example illustrates the results.
KeywordsDiscrete-time semi-Markov processes Reward process Empirical estimator Delta method Asymptotic properties
Mathematics Subject Classification (2010)60K15 62K20 60J0
Unable to display preview. Download preview PDF.
- Van Der Vart AW (2000). Asymptotic statistics, Cambridge series in statistical and probability mathematics, 3, Cambridge University PressGoogle Scholar