Accelerating Science Discovery - Join the Discussion

The Unbelievable Accuracy of the Monte Carlo Method
by Kathy Chambers on Fri, 18 Jan, 2013
Stanislaw Ulam

The year was 1945, the year I was born. That in itself is of great significance to me.  However, it was a momentous year in history. World War II came to its merciful end and the development of the first electronic computer – the ENIAC—was nearing completion. At a post-war Los Alamos National Laboratory (LANL), mathematician Stanislaw Ulam envisioned the possibilities of reviving statistical techniques that would have a huge impact on science and technology research today. (Read the history of Stanislaw Ulam in the special edition of Los Alamos Science No. 15, 1987.)

Fifteen years earlier, a too-good-to-believe method to predict experimental results by statistical sampling techniques rather than using differential equations had been used by Enrico Fermi while studying neutron transport.  Fermi used his new method to mystify his colleagues with unbelievable accuracy of experimental results. This new prediction method was in its infancy.

By the late 1940s, Ulam was so intrigued with the ENIAC and the increased computing power it offered, he soon realized that Fermi’s computational methods were now appropriate. He began to use random statistical sampling to gain insight into phenomena for which there’s no obvious method of exact analysis. John von Neumann recognized the potential in Ulam’s techniques and championed his effort.  What we now know as the Monte Carlo method was introduced.  (Read more about the Monte Carlo method in OSTI’s January 2013 Science Showcase “Monte Carlo Methods” by Dr. William Watson, Physicist, OSTI staff.)

Glory days were to follow at LANL, across the DOE complex and throughout the scientific community because of this new method. Use of the Monte Carlo method applications in research exploded.  Techniques developed rapidly. Monte Carlo symposiums were held and proceedings were published. When other mathematical methods failed, the Monte Carlo method was there.

Today, researchers continue to apply Monte Carlo methods with unbelievable and unexpected results every day. These methods are used to determine whether Higgs bosons do exist and what properties they might have. They are used to analyze the behavior of fission reactors and improve modeling of fission chains. They are used to understand phenomena in computational mathematics, physics and biology, physical chemistry, engineering, computer graphics, applied statistics, games, design and visuals, telecommunications and finance and business.  

1945 was an unbelievably good year.

 

Kathleen Chambers

DOE/OSTI Staff

Comments

About the Author

Kathy Chambers's picture
Kathy Chambers
Senior STI Specialist, Information International Associates, Inc.