Operations Research Office/RAC

Military OR at the Operations Research Office/Research Analysis Corporation in the 1950s and 1960s

While other branches of the US Armed Forces considered and funded the futures of their respective civilian research organizations after World War II, Army leadership (excluding those within the Air Force) was not committed to such a program. Army scientists returned to academia and industry, and all non-hardware relationships were formally severed. By the late 1940s, however, then-Army Chief of Staff Dwight Eisenhower and Lt. General A. C. McAuliffe felt an Army specific research organization was necessary and established the General Research Office in June 1948. The name was changed to the “Operations Research Office” (ORO) by the time business commenced in December to prevent confusion with generalities. ORO was created under contract with Baltimore’s Johns Hopkins University but was initially headquartered in Washington and operated from Chevy Chase, Maryland. Neither location was on university grounds and staffers did not have academic appointments. The decision to associate with Hopkins was driven by a desire to create an atmosphere of intellectual independence, conductive to university-quality research and a means to draw academically minded individuals to government work (US Congress Office of Technology Assessment 1995). This trend would continue with other research contractors to the United States Department of Defense, such as with IDA CCR’s relationship with Princeton and SORO’s relationship with American University.

The ORO faced its first field tests in the Korean War (1950-1953). Ellis Johnson, the organization’s director and an OR veteran of the Navy’s Mine Warfare Operational Research Group, considered the East Asian conflict “an excellent laboratory for operations research,” and personally led a team into the theater of combat. This observation proved correct as numerous OR theories and models crafted and implemented in the field were eventually published in leading operations research publications. One of their successful implementations during the war, for instance, was a systematic data collection on the causes of enemy tank losses. Other projects included preliminary evaluations of air support operations, the utilization of indigenous manpower, the effectiveness of North Korean propaganda, and a study on Chinese soldiers’ attitudes toward the war (Research Analysis Corporation 1965). ORO’s research infrastructure in the peninsula laid the groundwork for the establishment of the Human Resources Research Officer (HumRRO) and the Special Operations Research Officer (SORO) to study psychological warfare and counterinsurgency warfare, respectively. Over 100 ORO personnel across eight different teams worked in Japan or Korea during the 1950s and 113 of them received the Korean Service Medal from the United Nations.

Each group and project within ORO were assigned an “Army Advisory Groups” to ensure proper implementation of methods and best utilization of civilian and military persons. ORO research in the early 1950s was not exclusive to field operations in Asia. While Johnson handled the teams in Korea, George Shortley, an eventual Editor-in-Chief of Operations Research and director of Booz Allen’s Applied Research in Bethesda, MD, served as acting director of stateside operations and worked as a liaison between Army leadership and domestic ORO projects. Among other collaborative opportunities, ORO personnel were invited to attend DOD-hosted lectures at the National War College. Enthusiasm for these talks reached a tipping point, however, as ORO administrators had to enforce limitations on who and how many were able to attend. The comparatively casual habits of the academically minded ORO staff conflicted with the rigid expectations of the Army. One such example was an office-wide reprimand to late arrivals at invited talks, as that behavior was deemed “embarrassing to the organization supporting the lecture, Operations Research Office, and the individuals concerned.” (Operations Research Office 1950b)

ORO researchers also developed early computational systems and models. Paul Dunn, Charles Flagle, and Philip A. Hick developed the Queuiac, an electromechanical analog for the simulation of waiting line problems. This project stemmed from a key goal of ORO – to improve the efficiency of Army communications systems. The Queuiac was assembled simulation equipment to address these problems. The two basic unit functions were to serve as a source of arrival-time or service-time information (“Arrival and Service Units”) and to accept said information, store it, and send it to the next node (“Register Unit”). The circuity design was made by Sherman Rigby of Engineering Products and Specialties Company in Boston Massachusetts. Programming on the system was prepared on five-hole teletype paper tape. While created for military operations, the Queuiac was built with civilian industry implementation also in mind (Dunn et. al 1956). Dunn and company’s paper on the model was presented at the Fourth Annual Meeting of the Operations Research Society of America (ORSA) in Washington, D.C.

ORO had a significant presence in the professional operations research community at large. In 1956, fifty-four ORO persons were  listed as members of ORSA (ORSA 1956). This number marked a significant portion of their growing number of employees. By 1960, one year prior to their severance from Johns Hopkins, there were one hundred ninety-seven persons on staff at ORO. These employees, advisors, and contractors held degrees from over one hundred twenty different universities in over 41 different fields, ranging from Applied Mathematics to Zoology. The plurality held undergraduate degrees in mathematics (>19%), engineering (>13%), physics (>10%), and economics (> 6%). Over a quarter of analysts held graduate degrees in either mathematics or physics; and thirty-five ORO persons were PhDs. A small handful of analysts and research assistants were simultaneously completing their doctorate and master’s degrees while on staff, many from nearby American University and Johns Hopkins. Of the one hundred ninety-seven named ORO staffers in 1960, only sixteen were women. Margaret Emerson, the Chief of ORO’s Library Division, was the only holding a director-level position at the time (Operations Research Office 1961).  

One problem that plagued ORO from its inception was the issue of security clearances. As early as 1950, ORO leadership recognized the time for clearing new personnel could take anywhere from six months to a year. The organization was unable to effectively use researchers in limbo as classified problem sets remained untouched in the interim. To solve this, Ellis Johnson took on “the personal risk of clearing individuals of ORO for access to highly classified materials at his discretion.” This move was initially supported by a joint letter from the US Armed Forces and was commended by the Army in March 1950 (Operations Research Office 1950a). As Johnson expanded the office, however, his Army superiors began scaling back their excitement and support for the organization. One point of contention was ORO’s summer high school program. While ORO benefited from the influx of young talent, their classified work eventually leaked to the public. The Washington Post famously published an article capturing the nature of this research headlined “A-Attack to Kill 80%, 11 Boy Scientists Predict”, instigating a sense of confusion and alarm. Johnson’s further openness to the press about ORO projects further upset the military brass (Thomas 2015). Ultimately, the Army decided not to renew their partnership with Johns Hopkins, leading to Johnson’s resignation in 1961. The organization’s facilities, staff, and obligations were taken on by the newly formed Research Analysis Corporation (RAC) in McLean, Virginia.

Structurally, a significant difference between ORO and RAC was governance by an independent Board of Trustees rather than a partnership with a university. As the name “corporation” would suggest, the resulting organization was formed more as a company than a university research group. Early on, the board faced a slew of immediate problems with the transition, ranging from selecting a new president to minimizing the loss of departing personnel. Though there were some resignations and interruptions, work was largely unimpaired (Thomson 1975). RAC staff continued to apply operations research to pressing military problems including the deployment of atomic weaponry (such as minimizing the zone of detonation) and simulations of tank combat duels (Research Analysis Corporation 1965).

Though ORO had been reprimanded for operating outside its direct purview, RAC researchers continued to work with external OR persons. P.D. Robers co-authored a solution to linear economic models with Abraham Charnes of the University of Texas and Northwestern University’s Arthur P. Hurter and Adi Ben-Israel (Ben-Israel et al. 1970). Participation in the greater OR community continued as well, with many RAC staffers holding societal positions. From 1961 through 1966, Joseph O. Harrison, served as an Associate Editor of Operations Reserarch. In 1968-1969, Charles P. Chadsey served as Managing Editor and Advertising Manager of Operations Research. During that same window, Bernard B. Watson was as alternate ORSA representative to the American Association for the Advancement of Science (ORSA 1969).

In 1972, RAC assets, leases, and contracts were sold to the General Research Corporation (GRC), closing the chapter on ORO’s twenty-four-year legacy. The decision to close was not made lightly, as the Board of Trustees spent the later part of the 1960s and early 1970s exploring alternative paths for the organization’s long-term survival. The Army was not allowing RAC to engage in fundamental, long-range problems, which limited its ability to compete for substantial projects and fund its staff and their research. After the sale, RAC staff was dispersed among GRC’s Operations Analysis Division, Washington Operations Office, and the American Technical Assistance Corporation, a GRC affiliated entity (Thomson 1975). Though ORO/RAC did not last organizationally beyond the 1970s, many former personnel went on to accomplish significant advancements in operations research across academia, industry, and the federal sector.  

Compiled by Reed Devany

Edited by Linus Schrage



The author would like to thank the Special Collections archivists at John Hopkins University’s Sheridan Libraries for digitizing and sharing their Operations Research Office primary source materials (Record Group 08.060, Special Collections, Sheridan Libraries, Johns Hopkins University).

Image Gallery and Slideshow