High Leverage Interventions: Three Cases of Defensive Action and

OPERATIONS RESEARCHVol. 58, No. 6, November–December 2010, pp. 1535–1547issn 0030-364X �eissn 1526-5463 �10 �5806 �1535

informs ®

doi 10.1287/opre.1100.0887©2010 INFORMS

OR FORUM

High Leverage Interventions: Three Cases ofDefensive Action and Their Lessons for OR/MS Today

David C. LaneLondon School of Economics and Political Science, London WC2A 2AE, United Kingdom, [email protected]

This paper has two aims. First, to present cases in which scientists developed a defensive system for their homeland:Blackett and the air defense of Britain in WWII, Forrester and the SAGE system for North America in the Cold War, andArchimedes’ work defending Syracuse during the Second Punic War. In each case the historical context and the individual’sother achievements are outlined, and a description of the contribution’s relationship to OR/MS is given.

The second aim is to consider some of the features the cases share and examine them in terms of contemporary OR/MSmethodology. Particular reference is made to a recent analysis of the field’s strengths and weaknesses. This allows both acritical appraisal of the field and a set of potential responses for strengthening it. Although a mixed set of lessons arise,the overall conclusion is that the cases are examples to build on and that OR/MS retains the ability to do high stakes work.

Subject classifications : OR/MS implementation; applications; OR/MS philosophy.Area of review : OR Forum.History : Received December 2008; revision received September 2009; accepted July 2010.

Participate in the OR Forum discussion at http://orforum.blog.informs.org.

1. IntroductionThis paper has two aims. First, to present three cases inwhich scientists successfully developed a system to defendtheir homeland against attack. Second, to challenge readersto view the general OR/MS methodology today through thelens of those cases. The cases are of interest in themselves,but they benefit further from being considered together.Moreover, considering some of the features they shareprovides a critical comparison with contemporary OR/MSmethodology and offers potential responses for strengthen-ing the field.The cases involve one of the founders (Blackett), an

individual whose defense work had strong affinities withOR/MS (Forrester) and a scholar widely seen as a “pre-cursor” of the field (Archimedes). Key elements of OR/MScan be seen in the three cases.The nature of those key elements depends on the

understanding of OR used. A 1983 UK defence opera-tional research establishment booklet sees “The origins ofOperational Analysis” in Alexander the Great’s post-battlediscussions (Rowland 2006, p. 1). Alexander’s systemat-ically questioning the effectiveness of past operations toimprove future performance fits with the spirit of OR/MS.However, some additional scientific features are certainlyneeded to be comfortable using that particular label. Gassand Assad (2006), summarizing views widely expressedin the literature, propose two key elements: modeling andholism.

Modeling means generating abstract theories about anoperation being studied and then using those theories toexperiment with alternatives. Early accounts of OR/MSattach great importance to “imaginary models” (Goodeve1948, p. 379), mathematical analysis and analytical methods(Morse and Kimball 1951, pp. 6–7 and p. 9), and mathemat-ical models (Churchman et al. 1957, pp. 13–14 and Chap-ter 7). However, this element must be interpreted broadly.Gass and Assad (2006) acknowledge that it was not alwayspresent in early OR; statistical analysis and/or experimentalfield trials frequently substitute (see also Morse and Kimball1951, pp. 7–10a).1 “Holism”—“consideration of a problemin terms of its relationship to an entire operation”—is seenas a distinguishing feature of OR/MS by Trefethen (1954,p. 53). A similar idea is advanced by Waddington (1948)and Morse and Kimball (1951). It is therefore unsurpris-ing that OR/MS is seen as equivalent to “systems engineer-ing” (Morse and Kimball, p. 10b) and “systems analysis”(Churchman et al. 1957, p. 7; Miser 1963, p. 671). In thethree cases the key elements of modeling/data analysis andholism are used as tools for detecting work of an OR/MSnature.The paper is structured as follows. In the next three sec-

tions the cases are described in turn. Some backgroundis given, outlining the individual’s other achievements andthe historical context of their defense-related work. This isfollowed by a description of that defense contribution andthen its relationship to OR/MS. The paper’s final section

1535

Lane: High Leverage Interventions1536 Operations Research 58(6), pp. 1535–1547, © 2010 INFORMS

contains the comparison of some notable features of thesecases with contemporary methodological thinking and ideasfor potential responses.

2. Patrick BlackettThe first case concerns Patrick Blackett. One of the keyindividuals who developed the practice and theory of OR,his work related to the air defense of the United Kingdomduring the Second World War.

2.1. Background

P.M.S. Blackett (1897–1974) was one of a new technocraticbreed of British Navy officers and saw action during WWI(Hore 2002). He became a highly talented physicist, win-ning the 1948 Nobel Prize (Kirby 2003a). His defense sys-tem work occurred during the air conflict between Britainand Nazi Germany. To appreciate his contribution it isworth reprising attitudes toward aerial warfare in the mid-1930s (Kirby and Capey 1997). First, the immense casu-alties of WWI had led to a widespread wish to avoidlarge ground engagements. Second, the idea that this couldbe achieved using the destructive power of aerial bomb-ing. Third, the general belief that aerial attacks would beirresistible; “the bomber will always get through” (BritishPrime Minister Stanley Baldwin, 1932).The low quality of aircraft detection was at the heart

of this belief. In the mid-1930s detection methods wereacoustic and visual: parabolic devices sought the noise ofapproaching aircraft, while visual detection was attemptedby ground-based observers using binoculars, or by RoyalAir Force (RAF) crew flying fatigue-inducing patrols. Airexercises conducted in 1934 confirmed the inadequacy ofthe system—prompting Air Ministry civil servant A. P.Rowe to examine all 53 ministry reports on air defense andconclude that they lacked any useful ideas (Rowe 1948,Cunningham et al. 1984). However, his despairing conclu-sion resulted in the formation of the Committee for theScientific Survey of Air Defence (CSSAD) in 1935.

2.2. Blackett and the Air Defense of Britain

Blackett’s various roles during WWII were central to theestablishment of OR/MS (Kirby and Capey 1998, Haley2002). In August 1940 his group “Blackett’s Circus” stud-ied the use of radar to direct Army antiaircraft gunfire; inMarch 1941 he led the OR group at RAF Coastal Com-mand; and in January 1942 he became “Director of NavalOperational Research.” However, our focus is on his rolein the air defense story. It was as a member of CSSADfrom 1935 onward that Blackett contributed to the con-sistent championing and enabling of the innovations thattransformed Britain’s air defense.Prompted by Robert Watson-Watt, CSSAD recom-

mended radar as the basis of a new air defense system andfunded experiments into its use. From May 1935 in Orford-ness, Surrey and from spring 1936 at nearby Bawdsey

Manor, staff worked on detection: extracting altitude, bear-ing, and range information from radar signals. In the sum-mer of 1936 CSSAD authorized B. G. Dickens and a groupof RAF officers to begin working at Biggin Hill Airfield,Kent on interception: integrating radar and ground-basedobservations to direct fighter pilots to attack enemy aircraft.Similar work was also conducted at Bawdsey. Civilian sci-entists there included G. A. Roberts, who studied the opera-tional efficiency of the whole communications system, andE. C. Williams, who examined variations in performance ofstaff in early warning stations and ways of improving oper-ator technique. Members of the CSSAD also directly con-sulted with RAF officers on the operational requirementsof an early warning system (Air Ministry 1963).CSSAD recognized that both the technological and oper-

ational aspects of radar required development if its poten-tial was to be harnessed. Studies were commissioned inboth areas. Moreover, the CSSAD-sponsored work becamethe pattern for subsequent “operational research” studies—in 1938 producing the name for the new discipline(Kirby 2003b).These initiatives succeeded. Radar technology improved:

the “chain home” network of early warning radar stationsconstructed around the coast could detect aircraft 160 kmout, and different radar equipment detected low-altitude air-craft. The contribution of “OR” became particularly clear in1939, during the last pre-war exercise. The improvementsto the whole system were so apparent that the Commander-in-Chief of RAF Fighter Command, Air Chief MarshalDowding, asked for the permanent transfer of some of theBawdsey scientists (Rowe 1948). In June 1941 they becamethe “Operational Research Section, Fighter Command,” atestament to the effectiveness of having civilians work withmilitary staff to address operational problems using a sci-entific approach.Fighter response was only one element of air defense.

The Army’s Anti-Aircraft Command faced the problemof how to use radar to aim their guns. In August 1940a group formed around Blackett, subsequently becomingthe “Army Operational Research Group.” “Blackett’s Cir-cus,” as it was known, used a multidisciplinary, scientificapproach (McCloskey 1987a). Its research was distinctivelyoperational: radar equipment at gun sites did not performas under test conditions. The group researched the actualoperation of the equipment and improved performance. Thework here and in subsequent organizations is described intwo papers by Blackett, which are amongst the earliestOR/MS publications.2

In the summer of 1940, when Nazi Germany preparedto invade Britain by sending its bombers, radar technol-ogy was in place and operating effectively—a result of theactivities that flourished under the aegis of Blackett andhis CSSAD colleagues. However, OR/MS had made onefurther contribution. Information from radar and observerswas integrated in the “filter and operations room” at Fighter

Lane: High Leverage InterventionsOperations Research 58(6), pp. 1535–1547, © 2010 INFORMS 1537

Command. Here the positions of friendly and hostile air-craft were plotted on large maps. RAF controllers couldring Observer Corps sites to check sightings and could con-tact fighter squadrons and continuously monitor and adjusttheir interceptions (Price 2004). This development followeddirectly from the work on control room procedures byWilliams and others.In consequence, early in WWII the air defense of the

United Kingdom was achieved by a highly sophisticatedsystem of tightly integrated elements (see Figure 1). Radarreports of attacking aircraft arrived by telephone at FighterCommand for plotting. Sighting was supplemented and/or

Figure 1. Blackett and the air defense of Britain.

Radar detection stations

Intercepting fighters

Observerposts

Observercorps. center

Barrageballoons

Fighterbases

Direction finding

Sector airfieldoperations

room

Group HQoperations

room

HQ RAF fighter command:Filter room and operations room

Directionfinding

triangulation

Attacking bombers radar stationsAntiaircraft batteries

Notes. Top left: P. M. S. Blackett in 1942. Photograph courtesy of the National Portrait Gallery, London. Top right: the underground operations room atthe headquarters of RAF Fighter Command, Bentley Priory, Middlesex. Photograph courtesy of the Imperial War Museum, London. Bottom: the mainelements of the UK air defense system.

confirmed by visual data from Observer Corps posts. Thisinformation was coordinated in a plan position filter room.Information was exchanged with the relevant group andsector headquarters to direct the various responses. Whenattacking aircraft were far away, fighters were directed tointercept, their courses monitored by direction-finding radarstations. When attackers were closer, information was sentto direct antiaircraft guns. Even closer, barrage balloonswere launched, causing bombers to fly higher, reducingbomb-aiming accuracy.In the summer of 1940, the United Kingdom had “the

most efficient scheme of air defense in the world at the


time” (PRO AIR 41/15 1944, pp. 564–565). Radar multi-plied by 10 the effectiveness of the RAF, while OR/MStechniques increased that effectiveness further by a factorof two (Goodeve 1948, Kirby 2003b). The bomber offen-sive was beaten back.

2.3. “The Father of OR”

Consider now the extent of OR/MS in this work, usingthe key elements of modeling/data analysis and holism. Itsholistic nature is seen in the successful bringing togetherof innovative radar technology, visual observation, clearinformation presentation, radio and telephone communica-tion, and weapons technologies. Less apparent is the role ofmodeling. The analysis of data sets was a larger part of theCSSAD-sponsored work than abstract mathematical mod-eling. However, Blackett’s 1941 methodological reflectionson the work of CSSAD show that such data analysis hadmathematical foundations. For example, his “variationalanalysis” for calculating the marginal costs and benefits ofdifferent operational configurations was based on ideas ofpartial differentials. An example was the calculation of thebest number of antiaircraft guns to be located together in abattery. However, explicit modeling did have its place, fromthe use of simple geometric models to calculate interceptionvectors for fighters, to Blackett’s equation for comparingthe benefits of introducing new weapons versus improvingthe effectiveness of existing ones (Kirby 2003b).While many people were involved in developing the UK

air defense system, as a member of CSSAD and as head ofa number of groups, Blackett played a key role. He spreadOR/MS across all three British armed services and estab-lished its value to the war effort. His writings reflected onand advanced the new OR/MS methodology underlying thiswork. These activities are central to his status as “the fatherof operational research” (Zuckerman 1978, p. 266). OR/MSwas employed by the British in a range of theatres (AirMinistry 1963) and in other countries, notably the UnitedStates (Morse 1986), becoming a significant reason for theAllied victory (McCloskey 1987b, c; Kirby 2000). Blackettwas a central figure—and his achievements were groundedin his training in physics.

3. Jay ForresterThe second case concerns Jay Forrester, known to theOR/MS community for the system dynamics approach(Rand 2006, Lane 2006). His contribution to the air defenseof North America has affinities with OR/MS.

3.1. Background

J. W. Forrester (born 1918) was raised on a cattle ranchin Nebraska. After a first degree in electrical engineer-ing, he worked with MIT servo-mechanism pioneer GordonBrown. They developed devices to stabilize radar platformson ships and then ones that employed radar data to directantiaircraft guns.

The context of Forrester’s defense system contributionwas the Cold War. As WWII ended, the United Statesheld the monopoly on atomic weapons—until August1949, when the first Soviet bomb was tested. Attack byatomic-armed Soviet bombers seemed possible. This ledto an American equivalent of CSSAD, the US Air Force’sAir Defense System Engineering Committee, or ADSEC(Jacobs 1986). Committee chair George Valley and hiscolleagues were responsible for studying the nation’s airdefense. Exercises in June 1950 showed that the existingsystem offered little protection to attack, emphasizing theimportance of ADSEC’s task (Buderi 1996).

3.2. Forrester and the Air Defense ofNorth America

In 1944 Forrester was leading MIT’s Aircraft Stabilityand Control Analyzer (ASCA) project. The aim was todevelop a servo-mechanical aircraft flight simulator in orderto experiment with new aircraft designs. Forrester cameto believe that digital technology was required, and in1946 “Project Whirlwind” was founded to develop a high-speed digital computer to generate real time simulations(Redmond and Smith 1980). In 1947 Forrester becamethe director of MIT’s Digital Computing Laboratory, andthe “Whirlwind I” computer occupied nearly 250 m2. Yetby early 1950, delays, problems with memory reliabil-ity, and a tripling of the machine’s estimated developmentcosts meant that continued funding by the Office of NavalResearch seemed unlikely. Then ADSEC’s George Valleyvisited Forrester’s laboratory (See Figure 2).ADSEC had developed a bold plan for a new radar-based

air defense system using computer technology to processthe signal data. Valley’s visit was part of a survey of com-puter projects. He decided that Whirlwind I could be theplatform for the new system. Having already explored thecontribution that computing could make to military infor-mation systems (Forrester et al. 1948), Forrester readilyagreed. USAF funding started in March 1950.ADSEC’s October 1950 report recommended what came

to be called SAGE, for semi-automatic ground environ-ment (Jacobs 1986). SAGE would defend the airspace overCanada and the United States. It was to be a networkof radar stations and long-distance communication sys-tems based on telephone lines. Target tracking informationwould be relayed to digital computers managing the datain real time and automatically triangulating signals to cal-culate the position and velocity of targets. This would thenallow operators to deploy fighter aircraft and surface-to-airmissiles in response to perceived threats. At its heart wouldbe a development of the Whirlwind I computer (Astrahanand Jacobs 1983).In July 1951 what became MIT’s Lincoln Laboratory

was founded, with the aim of generating the technolo-gies required for air defense. The Whirlwind researchersjoined Division 6, Digital Computers. Led by Forrester,they began to develop Whirlwind II, the prototype for


Figure 2. Forrester and the air defense of North America.

STATUS

DIRECTIONCENTER

RADIO

GOC FILTERCENTERS

INTERCEPTORBASES

AMIS Air movements information service

GOC Ground observer corpsADC Adjacent direction centerAEW&C Airbome early warning and control

WEATHERTELETYPE

AIR ROUTETRAFFIC

CONTROL CENTER(AMIS)

TEXASTOWER

PICKET SHIP

AEW&C

LONG-RANGERADAR SITE GAP FILLER

RADIO SITE

AA GUNS

NIKE

AAOC

HEIGHTFINDER

CIVIL DEFENSE

ADJACENTDIRECTIONCENTER

ADC, FORCE, OTHER

COMBAT CENTERS

COMBATCENTER

AUTOMATIC AND

VOICE LINKS

ADJACENTDIRECTIONCENTERS

AAOC Antiaircraft operations center

Notes. Top left: magnetic core memory. Top right: Whirlwind I test control in 1950 with Forrester (standing centre left). Bottom: the various elements ofUS national air defense system. Pictures used with the permission of the MITRE Corporation. Copyright © the MITRE Corporation. All rights reserved.

the computer needed in SAGE. The computer was usedto explore designs for military information systems. Thisstarted in 1951 with “Project Charles” and developed intothe 1953 “Cape Cod System” experiments, which useddata relayed from radar stations on Cape Cod to show thatWhirlwind could be used to analyze digitalized radar datasupplied by telephone line to track approaching bombersand to direct interceptor aircraft (Forrester 2001).

While these experiments were a success, Whirlwind IIwas far from deployment. The need to increase the speedof the random access memory spurred Forrester to developcoincident-current magnetic core memory (Forrester 1951).Installed in 1953, access times dropped as core memory“doubled the operating speed [and] quadrupled the inputdata rate” (Jacobs 1983, p. 325). Such innovations producedthe machine at the heart of SAGE: the 275-ton AN/FSQ-7


(Army-Navy fixed special equipment), essentially the pro-duction version of Whirlwind II and the first computer builtin volume (Astrahan and Jacobs 1983).Inevitably, SAGE involved the tackling of numerous dif-

ferent problems. Many individuals helped develop the radarelements. C. Robert Wieser was responsible for SAGE’sreal-time software (then the world’s largest computer pro-gram). Enormous amounts of information were centralto the SAGE concept (Enticknap and Schuster 1958),and the problem of transmitting digital information overtelephone lines was solved under the direction of JackHarrington using technology that presaged today’s modem(Harrington 1983).In July 1958 the first direction centre became opera-

tional, and the whole SAGE system was completed in 1963.Attacking aircraft were identified by radar. This informa-tion was relayed to 24 direction centers, each operatingtwo AN/FSQ-7 computers in duplex mode. The infor-mation was processed and various responses coordinated.When the attacking aircraft were far away, fighters couldbe directed to intercept. When attackers were closer, loca-tional information was used to direct antiaircraft missiles.In the direction centers data were presented to operators asreal-time displays on cathode ray tubes, with “light pen”pointing devices used by operators to designate specific tar-gets on their screens and so get information on height andbearing (Everett et al. 1983). The display method and thepointing devices were key innovations in interactive com-puting (Palfreyman and Swade 1991).The last SAGE computer centre was decommissioned in

1984. Although never tested by a mass attack of Sovietbombers, the system is judged to have been very reliable(see, for example, Jenkins and Landis 2004), a source ofevident pride to Forrester (1992).

3.3. Affinities with OR/MS

As CSSAD was similar to ADSEC, so the fruits ofthose committees bear comparison. Although technologi-cally more advanced, SAGE was quite like the UK airdefense system. It faced many of the same problems andaddressed them using an approach that included key ele-ments of OR/MS. Holism is evident in the effective integra-tion of innovative computing technology, human-computerinterface methods, radio and telephone line communication,and weapons technologies.Elements of modeling/data analysis are seen in vari-

ous ways. One specific example of mathematical workis the development of the United Kingdom’s simple geo-metric modeling for fighters: “The next evolutionary step,which took place early in the era of Project Charles,was to program Whirlwind to compute collision-coursevectoring instructions for an interceptor aircraft automat-ically” (Wieser 1983, p. 365). A second example is theProject Charles and Cape Cod System explorations thatused Whirlwind as a defense information system. Thesewent beyond technical proof of concept and solving system

integration problems: they were experimental field trials inthe broadest sense. “The Cape Cod System was � � � a modelof the SAGE system, scaled down in size but realisticallyembodying all the SAGE air-defense functions” (Wieser1983, p. 366). As such “[it] was intended to demonstratethe operations that were to be executed for field use”(Astrahan and Jacobs 1983, p. 342). The 1950 Whirlwindcould only display the solution to a simple mechanics prob-lem on a cathode ray screen; the Cape Cod System work—activities comparable to the CSSAD-initiated OR work atBawdsey on improving radar operator technique—led tothe sophisticated real-time displays and “light pen” pointingdevices. A final example is the project management mod-eling used to propose a detailed 15-year plan for develop-ing military and other information systems based on digitalcomputing (Forrester et al. 1948). This document includesmeasures of project scope, proposed activities, and keymilestones. Fine-grained calculations for resource usageproduce graphs over time of spending and staff employed,while a set of cost-vs.-duration curves present an analy-sis of project acceleration options. This plan, formulatedbefore ADSEC ever met, indicates the forward thinking ofForrester and his group and provides a context for SAGE’sdevelopment.In 1956 Forrester moved to MIT’s new Sloan School

of Management (Forrester 1992). He argued that servo-mechanistic concepts were relevant to management(Forrester 1956, 1958); soon “system dynamics” was born(Forrester 1961).Returning to Forrester’s defense system work, naturally

he was one of many people involved. In particular, Valleyis known as “the father of SAGE.” Nevertheless, Forrester’sachievements were considerable. He “guided the planningand design of the SAGE system” (Everett 1983, p. 321),this broad contribution having affinities with OR. As hereflected on joining Sloan, “moving to a managementschool was not a break from a purely technical background.I was already in management. We had been running avast operation � � � from basic research to military operationalplanning” (Forrester 1992, p. 343). Forrester was a centralfigure in SAGE—and his achievements were grounded inhis knowledge of servo-mechanism theory and computing.

4. ArchimedesThe final case concerns the system Archimedes developedto defend his home city of Syracuse.

4.1. Background

Archimedes (287-212 BCE) ranks amongst the greatestmathematicians and physicists (Drachmann 1968). He stud-ied natural phenomena to understand their underlying prin-ciples and then experimented to test his theories—a keyshift into a scientific way of understanding the world. Hemade major contributions to astronomy and geometry. Inmechanics he formulated the principle of the lever and is


credited with inventing compound pulleys. He also discov-ered that a floating body displaces a mass of liquid equalto its own mass and studied the stability of floating bodies.The focus here is his defensive preparation for the siege

of Syracuse. This work is often cited as an example of sci-entists and military personnel cooperating (Flood 1962, AirMinistry 1963), and others refer to Archimedes as a “pre-cursor” or “Vorläufer” of OR/MS (respectively, Trefethen1954, Brusberg 1965). Larnder, who worked in FighterCommand’s OR section, refers to Archimedes as someonewho “studied warlike operations and carried out analysisthat might well qualify as ‘operational research’ ” (Larnder1984, p. 466).The context was the Second Punic War (218-202 BCE),

in which Carthage and Rome contested control of theMediterranean. While allied with Rome, Syracuse’s KingHieron II asked Archimedes to improve the city’s defenses.When Hieron died in 215, the new rulers sided withCarthage—and a Roman attack became a certainty.To develop a critical examination of Archimedes’ contri-

bution, the primary source is The Histories (Polybius 1923).Polybius (c. 200—after 118 BCE) had a good understand-ing of military matters and might have interviewed eye wit-nesses to the siege (Burrow 2007). Other sources are Livy(1965) and Plutarch (1912). In addition, work on Greekengineering (Landels 1978), military operations (Connolly1998), and the machinery and tactics of sieges (Kern 1999)have been drawn on to assemble the account below.

4.2. Archimedes and the Defense of Syracuse

Archimedes’ response to Hieron’s request was a systematicexamination of the city’s defense requirements, followedby a program of action. This included raising, strength-ening, and adding to the city’s walls and designing bothspecific “engines” and an appropriate system for operat-ing them effectively: “Archimedes now made � � � extensivepreparations, both within the city and also to guard againstan attack from the sea” (Polybius 8.3.5).3

Knowing, from lookouts, the position of an attacker,there might be a variety of responses. At a distance, largerocks or spears could be thrown by catapult artillery. Closerto, smaller arrows could be employed. Very close, objectscould be dropped directly onto an enemy. Syracuse hadall these responses available—and more. The point uponwhich the three historians agree is the layered nature of thesystem. For example, “Archimedes � � �had prepared enginesconstructed to carry to any distance” (Polybius 8.5.2).In 213 BCE the Roman 22nd and 23rd legions arrived

with 60 ships. Appius Pulcher led the army while ConsulMarcus Marcellus commanded the fleet. Plutarch tells us“But the apparatus was, in most opportune time, ready athand for the Syracusans, and with it also the engineer him-self” (Plutarch Marcellus, 14.9). The Romans planned totake Syracuse in five days; “but in this they did not reckonwith the ability of Archimedes, or foresee that in some

cases the genius of one man accomplishes much more thanany number of hands” (Polybius 8.3.3).A joint land and sea attack followed. The defenders

watched Marcellus’ ships approach and aimed the appro-priate weapon. Polybius describes how Archimedes “sodamaged the assailants at long range � � �with his more pow-erful [stone-throwers] and heavier missiles as to throw theminto much difficulty and distress” (Polybius 8.5.2, withcorrection). Such catapults utilized both lever-based andpulley-based mechanisms. Then, “as soon as these enginesshot too high he continued using smaller and smaller onesas the range became shorter, and � � �put a complete stop totheir advance” (Polybius 8.5.3).Later, attempting a night assault, Marcellus encoun-

tered the close-range defenses as archers and small “scor-pion” catapults shot through slits cut low in the sea wall.Nevertheless, close engagement was needed to bring up“sambucae,” ship-mounted walkways across which Romanmarines could storm the walls. These encountered the nextlayer of Archimedes’ defense system: large cranes thatswung out over the walls and dropped objects weighing upto 250 kg onto the vessels. More levers and pulleys were inuse. A final type of “engine” used is known as Archimedes’“shipshaker” (James and Thorpe 1995), or his “iron hand,”a crane:

� � �with which the man who piloted the beam would clutch atthe ship, and when he had got hold of her by the prow, wouldpress down the opposite end of the machine � � �Then whenhe had thus by lifting up the ship’s prow made her standupright on her stern, he � � �by means of a rope and pulleylet the chain and hand suddenly drop from it � � � [S]ome ofthe vessels fell on their sides, some entirely capsized, whilethe greater number � � �went under water and filled. (Polybius8.6.2-4)

Pulcher’s land assault had faced a similarly layereddefense and suffered a similar fate. Plutarch comments,“the rest of the Syracusans were but the body ofArchimedes’ designs, one soul moving and governing all”(Plutarch Marcellus, 17.2). The Romans abandoned directassault. After a prolonged siege, drunk and inattentiveguards allowed the Romans to storm the city. Archimedeswas killed.

4.3. Precursor to OR/MS

Archimedes’ contributions lay in three areas (see Figure 3).First, the devices operating at the city walls. The cranes and“iron hands” derived from existing technology, but the lat-ter was clearly tailored to the specific needs of the situation,surely drawing inspiration from Archimedes’ work on thestability of floating bodies (Rorres and Harris 2001). Sec-ond, the deployment of catapults. This technology was welldeveloped; Archimedes’ contribution was the design of thelayered artillery defense. The Romans would have expectedto face projectile weapons; not being able to find gapsin their coverage is what surprised them. Last, his devel-opment of the city’s fortifications: the walls, ditches, and


Figure 3. Archimedes and the defense of Syracuse.

a

b

d

c

50 m.

D

1 m.

Notes. Top left: a large torsion catapult, showing the diameter of the spring. Top right: an “iron hand” capsizing a quinquireme; note the slits at the baseof the wall. Bottom: reconstruction of the Euryalos fort, whose artillery tower (a), protective barriers (b), and outer ditch (c) defended the city’s gate (d).

outworks. Most striking is the Euryalos fort that defendedthe city’s gate, but all the 27 km of walls made Syra-cuse one of “two sites � � �whose defensive systems displaya complexity almost unrivalled in the Hellenistic world”(Winter 1967, p. 363).Elements of OR/MS can be discerned here. Model-

ing/data analysis was certainly present. Consider the cranesand iron hands, and recall that Archimedes formulatedthe mathematical principles underlying levers and pulleys.Another usage concerned catapults. By the end of thethird century BCE, catapult technology was widely used(Campbell and Delf 2003, Rihll 2007), so Archimedesstrengthened the city walls and built protective barriers forthe city gate to resist catapult fire. As reflected in the tech-nical manuals of the period, analysis of field experimentshad led to standard designs (Marsden 1969). Constructing astandard machine required use of a formula for the diameter

of the torsion spring, D (in Athenian dactyls). In the caseof a stone-thrower

D = 1�1× 3√100× M�

where M = Weight (in Athenian minae).

The size of every component followed directly: breadthof the spring frame (6 1

2D�, thickness of the stanchions( 58D�, etc. Although these standard catapults started withthe same range, they could subsequently be used toachieve different distances of shot. This is how the lay-ered defense might have been achieved: Archimedes coulddesign his defenses at a systemic level, imagining wherehe might place catapults and what range was required, andthen use the formula to specify the dimensions of eachartillery piece.The consequences of such calculations can still be seen

in the stones of the Euryalos fort. For example, the protec-tive barriers meant that any catapults directed against the


gate needed to move within range of the catapults on theartillery tower. Similarly, the positioning of the outermostditch meant that assaulting catapults would be ∼180 m.from the tower and “outside the effective � � � range of themost powerful [catapult], the type which could throw a50-pound stone” (Lawrence 1946, p. 105). In contrast, thetower’s elevation meant that attackers at the outer ditchcould be struck by Syracusan artillery. In this sense, mod-eling was central to the work in Syracuse, playing a keyrole in the grand design.The holistic nature of this work is clear. From strength-

ening walls to positioning catapults and cranes, how theseelements would work together had been thought throughwith great care. Hence, “It is [Archimedes’] comprehen-sive, systematic approach to the art of warfare that isunique” (Simms 1995, p. 68).Archimedes was involved in the design of this system in

a central way—and his achievements were grounded in hisknowledge of geometry, physics, and mechanics.

5. Methodological DiscussionThe three cases benefit from being seen together becausetheir parallels become apparent. This section goes furtherby discussing whether they have general methodologicallessons for OR/MS today.Because the cases are success stories, the following ques-

tions arise. In terms of OR/MS methodology, what notablefeatures do they share? How do those features comparewith contemporary OR/MS? These questions are addressedas follows. First, some of the features that contributed tothe success of the cases, and that are common to the three,are identified. Second, to make a comparison with mod-ern OR/MS, a recent analysis of the “OR/MS ecosystem”is used (Sodhi and Tang 2008). This analysis is chosenbecause it emerged from discussions at the 2006 INFORMS

Table 1. Notable features relating to the personalities involved in the three cases, comparisons with OR/MS today, andpotential responses.

Feature relating to personalities Comparison with contemporary OR/MS Potential response

Training and achievements in otherdisciplines

� OR/MS continues to draw on multipledisciplines

� Take interdisciplinarity seriously andsustain it

Interest in putting ideas into practice ? Disengagement of research andpractice; retreat from practicalapplications

� Strengthen end-user links:—More practice-driven research—More tailored publications

Concerned with methods of operatingwhole systems

? Tends to solve narrow parts ofcomplex problems; less impact atstrategic level

� Rebalance strategic/operational levelfocus:

—Use success stories as levers to getaccess

—Deliver with “strategic approaches”

Access to, and credibility with, high-leveldecision makers

? Low profile and unclear identity ofOR/MS

� Incentivize OR/MS workers to befamiliar with “sharp end” practicalities� Reflect this in branding

Note. See text for detail and references.

meeting. Finally, using that comparison, an additional ques-tion is asked: if current OR/MS methodology is foundwanting, then what response is needed?

5.1. Personalities

Four features relate to the personalities of those involvedin the cases (see Table 1).First, the individuals had training and achievements

in various disciplines not obviously related to the tasks.All brought an element of interdisciplinarity; Blackett’semphasis on interdisciplinary teams became a characteris-tic of OR/MS (see Churchman et al. 1957, Chapter 22).Sodhi and Tang’s analysis finds modern OR/MS contin-uing to draw on multiple disciplines, viewing this as amajor strength. This chimes with an earlier recommenda-tion that “interdisciplinarity be taken seriously” (Ackoff1979, p. 101). The appropriate response, surely, is to takeit as a feature to be valued and sustained.A second notable feature is that all were interested

in putting scientific ideas into practice to develop realimprovements. In contrast, Sodhi and Tang see the dis-engagement of research and practice as a key weaknessof OR/MS today. They attribute this to an “imbalance”in journal publications (practical papers declining in rela-tive number) and an orientation towards tools rather thanproblems. Sodhi and Tang recommend connecting morestrongly with end-users, suggesting more practice-drivenresearch and new publications tailored to appeal to end-users. But how to persuade young colleagues to allocatetime implementing—rather than constructing—their ideas?Clearly the incentive structure today is significantly differ-ent from that in the three cases.A third feature is the organizational level of the work

done. It is the holistic perspective that is most apparent:the concern with methods of operating whole systems ofinteracting components. Sodhi and Tang report that modern


Table 2. Notable features relating to the context of the three cases, comparisons with OR/MS today, and potentialresponses.

Feature relating to context Comparison with contemporary OR/MS Potential response

Modeling/data analysis beneficial � Power and generalizability of theorybuilding is a key strength

� Retain as hallmark of field andcontinue use and development

Predominantly technocraticproblem

� OR/MS very effective here � Continue such work but also � � �

? Other problems exist � Continue using and developingOR/MS approaches for moresociotechnical problems

High stakes, conflict situation � Problems exist in these domains but inother domains too.

� Retain confidence in OR/MS ability tosolve important problems

Note. See text for detail and references.

OR/MS tends not to be involved in high level issues, theinterest in tools and in solving “a narrow part of a complexproblem” (p. 270) tending to locate work at the operationallevel. This chimes with previous concerns about the field’s“lack of impact on strategic issues” (Zahedi 1984, p. 58).Sodhi and Tang acknowledge the need to “rebalance thefocus between strategic and operational levels” (p. 273).They recommend using success stories as levers to cre-ate management interest further up organizations. Whetherany specifically “strategic” approaches are also needed todeliver effective work is considered in §5.2 below.A final feature is that the individuals had access to people

at high levels and had credibility with them. Archimedesdemonstrated his pulley and lever ideas to King Hieronby moving a beached ship back into the water (Simms1995). The king’s response was that “from that day forthArchimedes was to be believed in everything that he maysay” (Heath 2002, p. xix). A plainer statement of client con-fidence is hard to find. In sharp contrast, Sodhi and Tangreport as an important weakness the unclear identity ofOR/MS, relating this to poor levels of access at high levels.This clearly overlaps with the previous feature: the respectcommanded by the three individuals can be contrasted withthe low profile of OR/MS amongst top managers identifiedby Bell and Anderson (2002). For Blackett and Forrester,military experience might have been a factor. Blackettserved at the Battle of the Falkland Islands and at Jutland(Lovell 1988). Forrester’s decision in 1943 to stay with theUSS Lexington to repair a stabilizing device meant that hewas present when the carrier participated in the retaking ofthe Marshall Islands and was torpedoed. Exposure to thesharp end of the task cannot have harmed their credibilityin the eyes of those they sought to influence—and Morseand Kimball (1951) stress the need for familiarity withactual practice and first-hand experience. What is today’sequivalent? For individuals, experience in organizations viaprevious jobs, placements, and consulting assignments. Forthe whole profession, incentives that acknowledge the valueof such experience. Responding to the concerns about thefield’s identity, these changes also need to be reflectedin the clear branding of OR/MS as a discipline with an

understanding of the “sharp” end. This would increase thechances of obtaining high-level access and offering recom-mendations with credibility.

5.2. Context

Three notable features emerge when considering the con-texts of the cases (see Table 2).First, modeling/data analysis tools were relevant and

effective. Sodhi and Tang see the ability to create new the-ories offering both generality and elegance as one of thefield’s greatest strengths. Considering the advances madesince the experiences of the three cases, it is clear that theuse and continued development of modeling is a hallmarkof OR/MS and is central to its success.Second, these were predominantly “technocratic” prob-

lems. The usefulness of modeling tools is one aspect ofthis, but there are others: the problems were clear, therewas consensus about the aims of the work, and imple-mentation was via hierarchical chains of command. Suchproblems exist today, and OR/MS retains its ability totackle them. However, in this regard the cases offer slightlytoo narrow a lesson; other types of problems exist, andOR/MS can contribute to these also. With more “sociotech-nical” problems the individuals concerned and how theywork together are as much elements of the situation as anyanalytical problem (Pasmore and Sherwood 1978). Theremight be no consensus about the problem definition; aimsmight be hard to agree on, and implementing change mightrequire support from a range of stakeholders. The earli-est OR/MS publications certainly emphasized the need toaddress the human relations aspect of problems—hencethe discussion above on credibility. However, as the lim-itations of a purely analytical approach were recognized(Hoos 1972, Lee 1973), critiques of OR/MS appeared inboth the United States (Caywood 1970, Zahedi 1984) andBritain (Rosenhead and Thunhurst 1982, Checkland 1972,Jackson 1982, Eden 1982, Rosenhead 1989a). Today, Sodhiand Tang report that OR/MS might be neglected by topmanagers exactly because it is primarily about “problemsolving” not “problem defining.”4 They also see the rela-tive absence of discussions about how to generate actual


change in organizations as a weakness of OR/MS, causedby its “tool-orientation.” In fact, recent decades have seenthe development of modeling tools and approaches appro-priate for more sociotechnical situations (Ackoff 1981;Rosenhead 1989b, 1996). Effective work in organizationshas resulted (Ormerod 1996, Eden and Ackermann 2004,Mingers and Rosenhead 2004). Note that “technocratic”and “sociotechnical” are not a dichotomy; rather, two pointson a spectrum. A sense of balance is key: “The technocraticview is faulty, not because it is incorrect, but because it isincomplete” (Tinker and Lowe 1984, p. 45). The opportu-nity here is for OR/MS to be used in exciting areas dif-ferent from those in the three cases, ones in which humanrelations play an increasing role, and yet still applying thefield’s hallmark of rigorous modeling. These sociotechnicalapproaches might also be useful for performing the systemslevel, or “strategic” work, mentioned in §5.1.The final notable feature is that these were high stakes,

direct conflict situations. Sadly, the defense and secu-rity applications community faces not dissimilar situationstoday. However, for most OR/MS people, lives are not inthe balance. Does this mean that OR/MS cannot normallyexpect the scale of accomplishment seen in these threecases? That does not follow. War is not the only circum-stance where there is value in getting things right. Althoughbusiness is not war, competition between companies shouldexpose inefficiencies and errors. Sodhi and Tang see theextra risk produced by the globalization of supply/servicechains as one opportunity for OR/MS, and it is easy toadd examples: from revenue management and optimizedmaintenance schedules by airlines to the inventory controlunderwritten by OR/MS algorithms in large supermarketchains, sound analysis continues to provide benefit to mil-lions of people around the world. No one would describethese as low-stakes situations, and there are many more towhich OR/MS can contribute boldly.

5.3. Closing Remarks

This discussion offers a mixed set of lessons: contemporaryOR/MS strongly retains many of the success factors in thethree cases—but not always all. Not surprisingly, debateabout ways of improving the methodology of OR/MS iswidespread. Important as such methodological considera-tions are, it is worth retaining confidence in the field’s abil-ity to improve an extraordinarily wide range of situations.Additionally, it would be wrong to be so impressed by theachievements of Blackett, Forrester, and Archimedes thatthey are seen as past glories, never to be repeated. Theyare exemplars of what OR/MS is capable of, successes tobe built upon boldly when looking for the next high-stakesproblem.Archimedes showed such boldness. Having moved the

boat off the beach, Archimedes’ reply to King Hieron tellsus what OR/MS is capable of today: “There is no limit.Just give me somewhere to stand, and I shall move the

earth” (Drachmann 1958, p. 281). This is the most impor-tant methodological lesson of these three cases: our field isstill capable of high leverage interventions.

AcknowledgmentsThe author is grateful to three referees and the area editorfor their detailed and constructive comments on an earlierdraft. Any errors remaining are the author’s responsibility.

Endnotes1. The two aspects are seen in Lanchester’s (1916) equa-tions for forces in combat and Edison’s analysis of sta-tistical data relating to antisubmarine warfare when hewas president of the Naval Consulting Board (Scott 1920,Whitmore 1953).2. “Scientists at the Operational Level” (1941) and “A Noteon Certain Aspects of the Methodology of OperationalResearch” (1943); available in Blackett (1948, 1962, 1995).3. The convention of citing ancient sources by book andchapter number of the original texts is used here. However,the specific translations are cited in the text and listed inthe references.4. On this point also, practice has diverged from the soundrecommendations of Morse and Kimball (1951). See “Theproblem of finding the problem” (pp. 5–6).

ReferencesAckoff, R. L. 1979. Resurrecting the future of operational research.

J. Oper. Res. Soc. 30(3) 189–199.

Ackoff, R. L. 1981. The art and science of mess management. Interfaces11(1) 20–26.

Air Ministry. 1963. The Origins and Development of OperationalResearch in the Royal Air Force. Her Majesty’s Stationery Office,London.

Astrahan, M. M., J. F. Jacobs. 1983. History of the design of the SAGEcomputer—The AN/FSQ-7. Ann. Hist. Comput. 5(4) 340–349.

Bell, P. C., C. K. Anderson. 2002. In search of strategic operationsresearch/management science. Interfaces 32(2) 28–40.

Blackett, P. M. S. 1948. Operational research. Advancement Sci. V(17)26–38.

Blackett, P. M. S. 1962. Studies of War. Oliver and Boyd, Edinburgh.

Blackett, P. M. S. 1995. Operational research. P. Keys, ed. Understand-ing the Process of Operational Research: Collected Readings. Wiley,Chichester, 7–30.

Brusberg, H. 1965. Der Entwicklungsstand Der UnternehmensforschungMit Besonderer Berucksichtigung Der Bundesrepublik Deutschland.Franz Steiner, Wiesbaden, Germany.

Buderi, R. 1996. The Invention That Changed the World: How a SmallGroup of Radar Pioneers Won the Second World War and Launcheda Technological Revolution. Simon and Schuster, New York.

Burrow, J. 2007. A History of Histories: Epics, Chronicles, Romances, andEnquiries from Herodotus and Thucydides to the Twentieth Century.Allen Lane, London.

Campbell, D. B., B. Delf. 2003. Greek and Roman Artillery 399 BC–AD 363. Osprey Publishing, Oxford, UK.

Caywood, T. E. 1970. How can we improve operations research? Oper.Res. 18(4) 569–576.


Checkland, P. B. 1972. Towards a systems-based methodology for real-world problem solving. J. Systems Engrg. 3 87–116.

Churchman, C. W., R. L. Ackoff, E. L. Arnoff. 1957. Introduction toOperations Research (8th printing, 1964). Wiley, London.

Connolly, P. 1998. Greece and Rome at War. Greenhill Books, London.

Cunningham, W. P., D. Freeman, J. F. McCloskey. 1984. Of radar andoperations research: An appreciation of A. P. Rowe (1898–1976).Oper. Res. 32(4) 958–967.

Drachmann, A. G. 1958. How Archimedes expected to move theearth. Centaurus—Internat. J. Hist. Sci. Its Cultural Aspects 5(3–4)278–282.

Drachmann, A. G. 1968. Archimedes and the science of physics.Centaurus—Internat. J. Hist. Sci. Its Cultural Aspects 12(1) 1–11.

Eden, C. 1982. Problem construction and the influence of O.R. Interfaces12(2) 50–60.

Eden, C., F. Ackermann. 2004. Cognitive mapping expert views for policyanalysis in the public sector. Eur. J. Oper. Res. 152(3) 615–630.

Enticknap, R. G., E. F. Schuster. 1958. SAGE data system considerations.AIEE Trans. (January 1959 section) 77(I) 824–832.

Everett, R. R. 1983. About this issue and contributors. Ann. Hist. Comput.5(4) 319–322.

Everett, R. R., C. A. Zraket, H. D. Benington. 1983. SAGE—A data-processing system for air defense. Ann. Hist. Comput. 5(4) 330–339.

Flood, M. M. 1962. New operations research potentials (address of theretiring president of ORSA). Oper. Res. 10(4) 423–436.

Forrester, J. W. 1951. Digital information storage in three dimensionsusing magnetic cores. J. Appl. Phys. 22(1) 44–48.

Forrester, J. W. 1956. Dynamic models of economic systems and industrialorganizations (2003 republication of “Note to the faculty researchseminar. November 5, 1956” and MIT “D-memo” zero). System Dyn.Rev. 19(4) 331–345.

Forrester, J. W. 1958. Industrial dynamics: A major breakthrough for deci-sion makers. Harvard Bus. Rev. 36(4) 37–66.

Forrester, J. W. 1961. Industrial Dynamics. MIT Press, Cambridge, MA.

Forrester, J. W. 1992. From the ranch to system dynamics. A. G. Bedeian,ed. Management Laureates: A Collection of Autobiographical Essays,Vol. 1. JAI Press, Greenwich, CT, 335–370.

Forrester, J. W. 2001. Lincoln Laboratory, MIT: Historical comments(Lecture Given on the 50th Anniversary of the Lincoln Laboratory,26th November). MIT D-Memo. 4878, 1–7.

Forrester, J. W., H. R. Boyd, R. R. Everett, H. Fahnestock, R. A. Nelson.1948. Forecast for military systems using electronic digital comput-ers. Memorandum L-3, Project Whirlwind, Servomechanisms Labo-ratory, Massachusetts Institute of Technology, Cambridge, MA.

Gass, S. I., A. A. Assad. 2006. An Annotated Timeline of Opera-tions Research: An Informal History. Kluwer Academic Publishers,New York.

Goodeve, C. 1948. Operational research. Nature 161(4089, March 13)377–384.

Haley, B. 2002. War and peace: The first 25 years of OR in Great Britain.Oper. Res. 50(1) 82–88.

Harrington, J. V. 1983. Radar data transmission. Ann. Hist. Comput. 5(4)370–374.

Heath, T. L. 2002. The Works of Archimedes (republished edition of 1897work of the same name and of “The Method of Archimedes, RecentlyDiscovered by Heiberg” from 1912). Dover, Mineola, NY.

Hoos, I. R. 1972. Systems Analysis in Public Policy: A Critique. Universityof California Press, Los Angeles.

Hore, P., ed. 2002. Patrick Blackett: Sailor, Scientist, Socialist. Routledge,London.

Jackson, M. C. 1982. The nature of “soft” systems thinking: The workof Churchman, Ackoff and Checkland. J. Appl. Systems Anal. 9(1)17–29.

Jacobs, J. F. 1983. SAGE Overview. Ann. Hist. Comput. 5(4) 323–329.

Jacobs, J. F. 1986. The SAGE Air Defense System: A Personal History.MITRE Corporation, Cambridge, MA.

James, P., N. Thorpe. 1995. Ancient Inventions. Michael O’Mara Books,London.

Jenkins, D. R., T. R. Landis. 2004. Valkyrie: North American’s Mach 3Superbomber. Speciality Press, North Branch, MN.

Kern, P. B. 1999. Ancient Siege Warfare. Souvenir Press, London.Kirby, M. 2000. Operations research and the defeat of Nazi Germany.

Military Oper. Res. 5 57–70.Kirby, M. 2003a. IFORS’ Operational Research Hall of Fame: Patrick

Maynard Stuart Blackett. Internat. Trans. Oper. Res. 10(4) 405–407.Kirby, M. 2003b. Operational Research in War and Peace: The British

Experience from the 1930s to 1970. Imperial College Press, London.Kirby, M., R. Capey. 1997. The air defence of Great Britain, 1920–

1940: An operational research perspective. J. Oper. Res. Soc. 48(6)555–568.

Kirby, M., R. Capey. 1998. The origins and diffusion of operationalresearch in the UK. J. Oper. Res. Soc. 49(4) 307–326.

Lanchester, F. W. 1916. Aircraft in War: The Dawn of the Fourth Arm.Constable, London.

Landels, J. G. 1978. Engineering in the Ancient World. Chatto and Win-dus, London.

Lane, D. C. 2006. IFORS’ Operational Research Hall of Fame—JayWright Forrester. Internat. Trans. Oper. Res. 13(5) 483–492.

Larnder, H. 1984. The origin of operational research. Oper. Res. 32(2)465–475.

Lawrence, A. W. 1946. Archimedes and the design of the Euryalus fort.J. Hellenic Studies 66(1) 99–107.

Lee, D. B. 1973. Requiem for large-scale models. J. Amer. Inst. Planners39(3) 163–178.

Livy. 1965. The War with Hannibal: Books XXI–XXX of the History ofRome from Its Foundation (translated by A. de Sélincourt). Penguin,London.

Lovell, B. 1988. Blackett in war and peace. J. Oper. Res. Soc. 39(3)221—233.

Marsden, E. W. 1969. Greek and Roman Artillery: Historical Develop-ment. Oxford University Press, London.

McCloskey, J. F. 1987a. The beginnings of operations research:1934–1941. Oper. Res. 35(1) 143–152.

McCloskey, J. F. 1987b. British operational research in World War II.Oper. Res. 35(3) 453–470.

McCloskey, J. F. 1987c. U.S. operations research in World War II. Oper.Res. 35(6) 910–925.

Mingers, J., J. Rosenhead. 2004. Problem structuring methods in action.Eur. J. Oper. Res. 152(3) 530–554.

Miser, H. J. 1963. Operations research in perspective. Oper. Res. 11(5)669–677.

Morse, P. M. 1986. The beginnings of operations research in the UnitedStates. Oper. Res. 34(1) 10–17.

Morse, P. M., G. E. Kimball. 1951. Methods of Operations Research. MITPress, Cambridge, MA.

Ormerod, R. J. 1996. Information systems strategy development at Sains-bury’s supermarkets using “soft” OR. Interfaces 26(1) 102–130.

Palfreyman, J., D. Swade. 1991. The Dream Machine: Exploring the Com-puter Age. BBC Books, London.

Pasmore, W. A., J. J. Sherwood. 1978. Sociotechnical Systems: A Source-book. Pfeiffer, San Diego.

Plutarch. 1912. Plutarch’s Lives: The “Dryden Plutarch” (revised by A. H.Clough), Marcellus, Vol. 1. J. M. Dent, London.

Polybius. 1923. The Histories Vol. III, Loeb Classical Library Edition(translated by W. R. Paton). Harvard University Press, Cambridge,MA.

Price, A. 2004. Britain’s Air Defences 1939–45. Osprey Publishing,Oxford, UK.

PRO AIR 41/15. 1944. The air defence of Great Britain, vol. 2, The Battleof Britain.


Rand, G. 2006. IFORS’ Operational Research Hall of Fame. Internat.Trans. Oper. Res. 13(6) 583–584.

Redmond, K. C., T. M. Smith. 1980. Project Whirlwind. Digital Press,Bedford, MA.

Rihll, T. 2007. The Catapult: A History. Westholme, Yardley, PA.Rorres, C., H. G. Harris. 2001. A formidable war machine: Construc-

tion and operation of Archimedes’ iron hand. Sympos. ExtraordinaryMachines Structures in Antiquity, August 19–24, 2001—Olympia,Greece, 1–18.

Rosenhead, J. 1989a. Introduction: Old and new paradigms of analysis.J. Rosenhead, ed. Rational Analysis for a Problematic World: Prob-lem Structuring Methods for Complexity, Uncertainty and Conflict.Wiley, Chichester, UK, 1–20.

Rosenhead, J., ed. 1989b. Rational Analysis for a Problematic World:Problem Structuring Methods for Complexity, Uncertainty and Con-flict. Wiley, Chichester, UK.

Rosenhead, J. 1996. What’s the problem? An introduction to problemstructuring methods. Interfaces 26(6) 1–24.

Rosenhead, J., C. Thunhurst. 1982. A materialist analysis of operationalresearch. J. Oper. Res. Soc. 33(2) 111–122.

Rowe, A. P. 1948. One Story of Radar. Cambridge University Press, Cam-bridge, UK.

Rowland, D. 2006. The Stress of Battle: Quantifying Human Performancein Combat. TSO, London.

Scott, L. N. 1920. Naval Consulting Board of the United States. Govern-ment Printing Office, Washington, DC.

Simms, D. L. 1995. Archimedes the engineer. Hist. Tech. 17(1)45–111.

Sodhi, M. S., C. S. Tang. 2008. The OR/MS ecosystem: Strengths, weak-nesses, opportunities, and threats. Oper. Res. 56(2) 267–277.

Tinker, A., A. Lowe. 1984. One-dimensional management science:The making of a technocratic consciousness. Interfaces 14(2)40–49.

Trefethen, F. N. 1954. A history of operations research (1995 repub-lication). P. Keys, ed. Understanding the Process of OperationalResearch: Collected Readings. Wiley, Chichester, UK, 47–73.

Waddington, C. H. 1948. Operational research. Nature 161(4089,March 13) 404.

Whitmore, W. F. 1953. Edison and operations research. J. Oper. Res. Soc.America (Oper. Res.) 1(2) 83–85.

Wieser, C. R. 1983. The Cape Cod System. Ann. Hist. Comput. 5(4)362–369.

Winter, F. E. 1967. The chronology of the Euryalos fortress at Syracuse.Amer. J. Archaeology 67(4) 363–387.

Zahedi, F. 1984. A survey of issues in the MS/OR field. Interfaces 14(2)57–68.

Zuckerman, S. 1978. From Apes to Warlords. Hamish Hamilton, London.

Documents

High Leverage Interventions: Three Cases of Defensive Action and