We survey the history of simulation up to 1981, with special emphasis on some of the critical advances in the field and some of the individuals who played leading roles in those advances. INTRODUCTIONA history of simulation can be written from many perspectives-for example, uses of simulation (analysis, training, research); types of simulation models (discrete-event, continuous, combined discrete-continuous); simulation programming languages or environments (GPSS, SIMSCRIPT, SIMULA, SLAM, Arena, AutoMod, Simio); and application domains or communities of interest (communications, manufacturing, military, transportation). Examples of the various perspectives and combinations can be readily found in the published histories; see Nance (1996), Nance andSargent (2002), and Hollocks (2006). We offer this brief treatment from a very informal perspective; the objective is to highlight people, places, and events that have marked the development of discrete-event and Monte Carlo simulation. Within this informal perspective, and sometimes anecdotal description, lies a secondary objective: to motivate others to document their historical contributions or knowledge so that a comprehensive history can be captured for posterity in places like the Simulation Archive at North Carolina State University . THE PRECOMPUTER ERA: FROM BUFFON TO WORLD WAR II (1777-1945)The Monte Carlo method is generally considered to have originated with the Buffon "needle experiment" in 1777. The experiment is to "throw" needles onto a plane with equally spaced parallel lines in order to estimate the value of π. The reader interested in the details may consult . Since Buffon's published solution contained an error that was corrected by Laplace in 1812, the terminology Buffon-Laplace needle problem is also used; see . About a century after Laplace's contribution is the perhaps surprising role played by simulation in one of the most important applied statistics developments. William Sealy Gosset, trained in mathematics and chemistry, became a brewer with Arthur Guiness, Son & Co. Ltd., in 1899 at the age of 23. Guiness allowed Gosset to publish certain major statistical results, provided he used a pseudonym and no proprietary data was used. These results were published under the pseudonym "Student" beginning in 1908 with a paper formulating what is now known as Student's t-distribution. Because Gosset had incomplete analytical results, he used a crude form of manual simulation to verify his conjecture about the exact form of the probability density function for Student's t-distribution. This inaugural application of simulation to the field of industrial process control is a remarkable example of the synergy of simulation-based experimentation and analytic techniques in the discovery of the exact solution of what is arguably a classical industrial-engineering problem.

In response to a request from the WSC Foundation and the WSC 2010 Program Committee, we review and slightly revise our survey of the history of simulation up to 1982, with special emphasis on some of the critical advances in the field and some of the individuals who played leading roles in those advances. Documenting the history of simulation remains a work in progress on our part, and we encourage individuals and organizations in the simulation community to bring significant historical data to our attention.

No abstract

Output analysis methods that provide reliable point and confidence-interval estimators for system performance characteristics are critical elements of any modern simulation project. Remarkable advances in simulation output analysis have been achieved over the last thirty years, in part owing to the application of data-reuse techniques designed to improve estimator accuracy and efficiency. Many of the key insights regarding data reuse are given in the seminal 1984 Winter Simulation Conference paper by Meketon and Schmeiser that is titled "Overlapping Batch Means: Something for Nothing?" and that introduced the method of overlapping batch means (OBM). We trace the development of OBM from the original work of Meketon and Schmeiser, and we discuss some recent extensions of the method.

No abstract