Computers & Information Science

Kahn, Herman | Applications of Monte Carlo

  • First edition in the original wrappers. Though described as “revised”, this refers only to a shorter version presented at a meeting in 1954, and we can locate no copies of a 1954 edition in the record, only this 1956 edition. A significant and uncommon report by nuclear war theorist and futurist Herman Kahn (1922-1983) on using the Monte Carlo method in computer simulations.

    A significant and uncommon report by nuclear war theorist and futurist Herman Kahn (1922-1983) on using the Monte Carlo method in computer simulations.

    The Monte Carlo method was developed by John von Neumann, first as a way to simulate the physics of gases in the atmosphere, then to predict the behaviour of neutrons in nuclear chain reactions. Because the number of potential interactions in each case were so large, it wasn’t feasible to program each one into a computer simulation. Instead, von Neumann realised that a representative sample of the possible interactions could be produced. “A roulette wheel, later a table of “pseudo-random” numbers, could be used to generate the path any one particle would follow. What emerged was the shape — the upper and lower boundaries — of the probabilistic distribution of the event” (Ghamari-Tabrizi, The Worlds of Herman Kahn, p. 133).

    Kahn’s first years at the RAND Corporation “were devoted to simulating the activity of elementary particles and gamma rays that might penetrate various thicknesses of protective shields in nuclear reactors. He delivered papers on this topic at all three of the major unclassified conferences that introduced the Monte Carlo technique to the physics, engineering, mathematics, computing, and operations research communities in 1949 and 1954” (Ghamari-Tabrizi, The Worlds of Herman Kahn, pp. 132-133).

    Monte Carlo continued to influence Kahn’s methods and philosophy as he moved into more speculative pursuits, such as nuclear game theory. “He exulted in the ‘complete control’ that allowed him to tinker with probabilistic events. Like genies tumbling out of bottles, improbable events sprang forth and frolicked on command”, though he also warned against the dangers inherent in designing simulations based on statistical probabilities rather than real events. (Ghamari-Tabrizi, The Worlds of Herman Kahn, p. 136).

    In this document Kahn discusses “the general principle of doing Monte Carlo calculations with particular emphasis on reducing the amount of work involved”, but without discussing “important specific applications” as he hoped to later publish a more comprehensive book on the subject (foreword).

  • ...RM-1237-AEC. 19 April 1954. Revised 27 April 1956. Prepared under contract with the U. S. Atomic Energy Commission... Santa Monica, CA: The RAND Corporation, 1956.

    Perfect bound. Original grey, morocco-patterned paper wrappers printed in dark blue with a window for the title. Equations and charts within the text. Pencilled ownership signature to the upper wrapper. Wrappers partially tanned and rubbed and creased at the edges, light spotting to title and edges of text block. Very good condition.