by Felisa J. Vázquez-Abad
and Yanick Champoux

Topics: Estimation
- Output Analysis
- Generation of RV's
- Efficiency Improvement

Topics: Optimization
- Discrete Optimization
- Functional Estimation
- Gradient-Based Methods
- Sensitivity Analysis

Case Studies:
- An Ice Cream Parlor
- Wine Provisioning in Air Lines

- The Outer Web
- References
- About SimSpiders

Simulation, from the latin simulare, means to "fake" or to "replicate". Today, computer simulation is used as a powefull tool to assess the impact of policies, to evaluate performance, to train professionals and more... All within the safety net of the computer, without actually having to experiment with or perturb the real system.

This page is dedicated to simulation of stochastic processes. There are two interelated scenarios when we use stochastic simulation:

  • If the system that we want to analyze and control is subject to uncertainty, then the performance (or outcome) cannot be accurately predicted in many complex problems of interest. We then perform simulations, where we program the computer to produce outcomes as if they resulted from actual experiments.
  • As a numerical tool for evaluating integrals. Known as Monte Carlo Simulation, this approach states that  definite integrals can be rewritten as expectations with respect to some probability. We then perform simulations that generate random variables according to these probabilities and evaluate their average.

There are different aspects to simulation,

  • Modeling and Programming
  • Statistical Analysis
  • Efficiency Improvement
  • Optimization

We shall explore in these pages the statistical principles of simulation as well as the mathematical basis of efficient estimation, all through case studies or model examples.

TO USE THIS PRIMER: Use the menu on the left of the window, and have fun while learning!

Don't forget to send us your comments.

© Copyright 1998 Felisa J. Vázquez-Abad and Yanick Champoux. All rights reserved.