The Almost Regenerative Method for Stochastic System Simulations

Abstract

The regenerative method for stochastic system simulation allows data collection each time the stochastic process enters a specific single state, r, called the regeneration state. The generated observations have the desireable property of being independent and identically distributed. Relative to a fixed run length, however, the mean time between entries into r may be excessively long for complicated stochastic systems, thus providing few observations and poor variance estimates. The almost regenerative method is an extension of the regenerative method designed to alleviate this problem for complicated stochastic systems (such as a network of queues). The almost regenerative method allows data collection each time the stochastic process enters a set of states. Simulations of simple queueing networks show that the almost regenerative method can provide an order to magnitude improvement over the regenerative method in terms of the mean-square-error of the estimator of total delay in queue, and this relative improvement increases with system complexity.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1975
Accession Number
ADA020129

Entities

People

  • Francis L. Gunther

Organizations

  • University of California, Berkeley

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Computational Science
  • Computer Programs
  • Computers
  • Data Science
  • Information Science
  • Normal Distribution
  • Operations Research
  • Queueing Theory
  • Random Variables
  • Simulations
  • Stationary Processes
  • Statistical Algorithms
  • Statistical Analysis
  • Statistics
  • Stochastic Processes
  • Theorems
  • Theses

Readers

  • Computational Modeling and Simulation
  • Mathematical Modeling and Probability Theory.
  • Statistical inference.