Technical Articles

Using Multiple Processors for Monte Carlo Analysis of System Models

By Amory Wakefield, MathWorks


Monte Carlo analysis is a standard method of simulating variability that occurs in real physical parameters. Automotive companies use Monte Carlo testing to ensure high quality and robust designs. Due to time and resource constraints, engineers often examine only a limited number of key parameters rather than an entire set. This leaves the design vulnerable to problems caused by missing the full potential impact of parameters that were unvaried during testing. New high-performance computing tools and multiprocessor machines have eliminated the time and resource limitations in many cases by providing the processing power needed to vary large numbers of parameters in complex dynamic models. This paper presents new methods for distributing Monte Carlo analyses of system models across multiple machines. These methods reduce testing time and enable more complete analyses, ensuring better quality when designs go into production.

Copyright © 2008 by The MathWorks, Inc. Published by SAE International, with permission.

This paper was presented at SAE World Congress.

Read full paper.

Published 2008

View Articles for Related Industries