Meteorologist, oceanographer, Python solvem probler, experimenter, Raspberry Pi tinkerer, geek dad, manager, space enthusiast, ham radio lover, science columnist, ADHDer, synthetic intellect curious, linux admin cosplayer.

Nicholas Constantine Metropolis

Born: June 11, 1915, Chicago, Illinois Died: October 17, 1999, Los Alamos, New Mexico (aged 84)

Heritage and Family

Greek-American. Son of Greek immigrants. Stanislaw Ulam described him as “a Greek-American with a wonderful personality.” Had one son (Christopher) and two daughters (Penelope and Katharine).

Education

  • B.Sc., University of Chicago (1937)
  • Ph.D. in Physics (Chemical Physics), University of Chicago (1941)
  • Dissertation advisor: Robert Mulliken (Nobel laureate in Chemistry, 1966)

After his PhD, worked as an instructor at the University of Chicago under James Franck (Nobel laureate in Physics, 1925).

Los Alamos and the Manhattan Project (1943–1946)

Robert Oppenheimer recruited Metropolis in 1943 for the Manhattan Project. He arrived at Los Alamos in April 1943 as a member of the original staff of approximately fifty scientists – making him one of the founding members of the laboratory.

At Los Alamos, Metropolis worked in the Theoretical Division on implosion calculations and other computational problems essential to bomb design. After the war, he returned to the University of Chicago as an assistant professor.

ENIAC and the H-Bomb Calculations (1945–1946)

In August 1945, Metropolis traveled with Stan Frankel to the Moore School of Electrical Engineering at the University of Pennsylvania to learn to program ENIAC. That fall, they designed and ran the first nuclear physics calculations on an all-electronic computer: a mathematical model of a thermonuclear (hydrogen bomb) explosion.

The program ran in November 1945, with results available by December. The calculation required thousands of program steps and a million punch cards. The results revealed flaws in Edward Teller’s proposed “Super” design. This was one of the earliest demonstrations of electronic computers solving problems impossible by any other means.

In 1948, Metropolis returned to Los Alamos and rewired the ENIAC to perform Monte Carlo simulations of a nuclear fission core – an early example of the Monte Carlo method applied to computational physics.

The Monte Carlo Method (late 1940s)

The Monte Carlo method – using random sampling to obtain numerical results for problems that are deterministic in principle – was developed collaboratively at Los Alamos in the late 1940s. The key figures were Metropolis, Stanislaw Ulam, and John von Neumann.

The origin story belongs to Ulam (see Ulam entry), who conceived the idea while playing solitaire during recovery from encephalitis in 1946. Von Neumann immediately grasped its computational potential. Metropolis contributed the implementation: he had the hardware, the programming expertise, and the organizational ability to make the method work on actual computers.

The name “Monte Carlo” was suggested by Metropolis, inspired by Ulam’s uncle Michal, who “just had to go to Monte Carlo” to gamble. The first unclassified paper on the method was published by Metropolis and Ulam in 1949: “The Monte Carlo Method” in the Journal of the American Statistical Association.

The 1953 Paper and the Metropolis Algorithm

The landmark paper “Equation of State Calculations by Fast Computing Machines” was published in the Journal of Chemical Physics (Vol. 21, No. 6, pp. 1087-1092, June 1, 1953). The five co-authors were: Nicholas Metropolis, Arianna W. Rosenbluth, Marshall N. Rosenbluth, Augusta H. Teller, and Edward Teller.

The paper described a general method for computing the thermodynamic properties of interacting molecules using a modified Monte Carlo integration over configuration space. The calculations were performed on the Los Alamos MANIAC I.

The algorithm – generating a random walk through configuration space, accepting moves that lower energy and probabilistically accepting moves that raise it – became known as the Metropolis algorithm. It is the foundation of all Monte Carlo methods in statistical mechanics.

The relative contributions of the five authors have been debated. Marshall Rosenbluth later stated that he and his wife Arianna did most of the actual work. Regardless, the paper bears Metropolis’s name first.

Metropolis-Hastings Algorithm

In 1970, W. K. Hastings (Biometrika) generalized the Metropolis algorithm to allow asymmetric proposal distributions, creating the Metropolis-Hastings algorithm. This generalization is the foundation of modern Markov Chain Monte Carlo (MCMC) methods, now used across statistics, machine learning, computational biology, and finance. As of 2011, the original 1953 paper had been cited over 18000 times.

MANIAC I and II

MANIAC I (1952)

Metropolis returned to Los Alamos in 1948 and led the group in the Theoretical Division that designed and built the MANIAC I (Mathematical Analyzer, Numerical Integrator, and Computer). Operational in 1952, it was modeled on von Neumann’s IAS machine architecture and was used for the thermonuclear calculations that supported the development of the hydrogen bomb.

The Naming Anecdote

John von Neumann thought the acronym MANIAC was “too frivolous.” Metropolis claimed he chose the name deliberately “in the hope of stopping the rash of such acronyms for machine names” – but acknowledged it may have had the opposite effect. Astrophysicist George Gamow joked the name stood for “Metropolis And Neumann Invent Awful Contraption.”

MANIAC II (1957)

Metropolis led the design and construction of MANIAC II as well, a more powerful successor.

Academic Career

  • Assistant professor, University of Chicago (post-WWII)
  • Full professor of Physics, University of Chicago (1957–1965)
  • Founding director, Institute for Computer Research, University of Chicago
  • Returned to Los Alamos in 1965
  • Named Laboratory Senior Fellow at Los Alamos (1980)
  • First Los Alamos employee to receive “emeritus” status from the University of California (1987)

Personality and Anecdotes

  • Wonderful sense of humor and warm personality, by all accounts.
  • The poker anecdote: Metropolis once won $10 from John von Neumann at poker. He bought von Neumann’s book Theory of Games and Economic Behavior for $5 and pasted the remaining $5 bill inside the front cover as a trophy.
  • Appeared as a scientist in Woody Allen’s film Husbands and Wives (1992).
  • Avid skier and tennis player well into his mid-seventies.
  • Erdos number of 2; this gave Richard Feynman an Erdos number of 3 (through Metropolis).

Awards and Honors

  • IEEE Computer Pioneer Award (1984)
  • Member, American Academy of Arts and Sciences
  • Fellow, American Physical Society
  • Pioneer Award from the IEEE Aerospace and Electronic Systems Society
  • Nicholas Metropolis Award for Outstanding Doctoral Thesis Work in Computational Physics (established in his honor)

Connections to Others

  • Stan Frankel: Traveled together to learn ENIAC programming; co-designed H-bomb calculations; co-authored papers; two of the earliest electronic computer programmers
  • Stanislaw Ulam: Co-developer of the Monte Carlo method; co-authored the 1949 paper; Ulam provided the idea, Metropolis provided the name and the machines
  • John von Neumann: Collaborator on Monte Carlo method and MANIAC design; the poker anecdote; von Neumann objected to the MANIAC name
  • Edward Teller: Co-author of the 1953 paper; used results from Frankel-Metropolis ENIAC calculations
  • Robert Oppenheimer: Recruited Metropolis for Manhattan Project
  • Marshall and Arianna Rosenbluth: Co-authors of the 1953 paper who may have done most of the actual computational work
  • W. K. Hastings: Generalized the Metropolis algorithm in 1970

Sources