SigmaPedia - The Free Online Lean Six Sigma Encyclopedia

English |  Español |  Français |  Português |  Deutsch |  中文

Monte Carlo Analysis

Go Back

Definition

A technique which simulates (mimics) the behavior of a process in order to identify solutions to improve or optimize the process. It Models the process by sampling from the probability distributions of individual elements/inputs of the process. This type of analysis is often used in problems involving a large number of variables, or variables that have a lot of variability associated with them, or problems that are hard to physically replicate.

Examples

In order to optimize the staffing levels at a call center, a model may be constructed to incorporate the distributions of staffing levels, call volumes and other operating characteristics, and then the simulation is used to track how changes in these variables affect the call wait times.

Application

The advantage of this method is that effects of changes to the simulated process can be studied virtually, allowing the evaluation of ‘best’ and ‘worst’-case scenarios without breaking the actual process. The method takes advantage of the computing speed of modern computers to generate thousands of random samples from the specified distributions and uses the relationships among them to paint a detailed picture of the process that brings them together.

External Links

Statistics and Simulation by MoreSteam.com - http://www.moresteam.com/whitepapers/stats-and-sim.pdf