Convergence of the Mean of a Regenerative Process.
Abstract
Conditions guaranteeing the convergence of EX(t) for regenerative stochastic processes X are studied. It is shown that the continuous time theory is more complicated than discrete time, in that an extra condition with no discrete time counterpart is required for convergence of the mean in continuous time. However, an easily verified sufficient condition for semi-Markov processes is obtained and applied to the M/G/1 queue and continuous time Markov chains.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1986
- Accession Number
- ADA180947
Entities
People
- Peter W. Glynn
Organizations
- University of Wisconsin–Madison