Maximum Likelihood Estimation of the Distribution of the Sum of Three Independent Exponential Random Variables
Abstract
This paper describes a procedure for computing the maximum likelihood estimates of the parameters of the distribution of the sum of three independent exponential random variables. By fitting sample interevent time data from a real system to this distribution, one can create a simulation of the system that exploits the regenerative representation of queueing systems to analyze the simulation's output by relatively elementary statistical methods. The paper also describes computation of the sample asymptotic covariance matrix and an implementation of the likelihood ratio for testing six hypotheses that are special cases of interest. A set of FORTRAN subroutines for executing these procedures appears in the Appendix.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1976
- Accession Number
- ADA026851
Entities
People
- George S. Fishman
Organizations
- University of North Carolina at Chapel Hill