2.1 Emergence and development of OR
2.1.2 Crisis of confidence in OR
The need for academic consideration of alternative paradigms to quantitative OR emerged in the 1970s when the ‘Crisis in OR’ was originally raised (Thunhurst,
1973). There was a growing dissatisfaction with the applicability of OR to strategic and social problems faced by industry at the time. The crisis reflected a dip in confidence of some practitioners and academics in the ability of OR to be useful in ‘solving’ the important problems of that time. This was the manifestation of several converging factors, which have been eloquently described by Kirby (2000, 2007). The emergent themes from the literature concerning the Crisis in OR seem to be characterised by an increasing chasm between practitioners’ use of OR and university-based teaching and research into OR, increasing importance of social problems faced by organisations, and the perception or realisation that the assumptions underpinning quantitative OR were ill-equipped to deal with these new problems.
The development of OR within universities had moved away from engaging with real world problems. “While we are training our students to use mathematical theory and techniques to solve well-formulated problems, we too often ignore the fact that problems of the real world do not present themselves in such well-formed structures” (Borsting, 1987 p.790). This led to these techniques being taught “without any recognition of the subsequent uses to which they might be put by OR workers” (Simpson, 1978 p.525). Morse (1976) commented that an OR textbook more closely resembles advanced mathematics than a physical science one. Alternative systems approaches were still not understood by academics, and despite proven analogues of success such as the RAND Corporation they were not integrated into educational programmes (Miser, 1987). This was despite calls made much earlier by Churchman (1970) to expand OR education. Pierskalla (1987) suggested that the reason a narrow focus of OR prevailed within universities was that in academia, tenure is based on producing many publications quickly; therefore, the temptation is to make small advances within the limits of the accepted paradigm—in this case, positivist OR focussing on quantitative modelling.
Quantitative OR was also easier to teach as mathematicians were able to codify mathematical techniques and reproduce them in educational programs. However, this quantitative approach had its drawbacks for practitioners, clients, and the reputation of OR. A survey of members of Operations Research Society of America reported by Balut & Armacost (1986) found that for every member conducting research, three others were working in applications/practice. However, the composition of the research published in OR at the time was mainly focussed on theoretical development. One of the findings of the survey was a call to increase the number of papers based on practical applications at the expense of theoretical work. It is evident that the type of research undertaken in the name of OR was moving away from the problems that practitioners were facing thereby leading to a growing chasm between OR development within universities and practitioner uses.
The perceived divorce between theory and practice led to a perception that OR was unable to deal with a new type of emerging problems, which were rooted in social interaction. Sadler (1978) suggested that in Britain there was a move away from an industrial society, where manufacturing efficiency was the most pressing question in OR. Therefore, for OR to survive, it needed “recognition that problems in the real world are ‘messy wholes’. They have quantitative and qualitative aspects, logical and irrational components, their technical, economic, social dimensions” (Sadler, 1978 p.6). These new types of problems were described as ‘messes’ (Ackoff, 1979), ‘swampy’ (Schon, 1987), ‘wicked’ (Rittel, 1972), and ‘turbulent’ (Emery & Trist, 1965), where factors change rapidly and unpredictably. Radford (1978) suggested that OR does not take into account these turbulent factors, as approaches like optimization are—for the benefits of a single participant in a static environment—developed for situations that are relatively well understood with reliable data on the characteristics of the situation and for situations much less complex than managers are dealing with. Moreover, OR was perceived to be unable
to answer these problems of increasing importance to senior management, so it was relegated down the organisational chain to deal with tactical problems (Eilon, 1977). For example, in organisational planning, OR lost its pre-eminent status due its misunderstandings of the multidimensional aspects of corporate planning (Harris, 1978). At the time, OR was being challenged with new types of social problems that many felt it was ill-equipped to answer; ignoring these problems threatened to stagnate OR, forcing abandonment of a much-needed decision making aid (Dando & Sharp, 1978). This led to the perception or realisation from some practitioners and academics that OR as researched and taught within universities had a number of limitations.
These limitations were outlined by McClelland (1975), who argues that the five assumptions made by quantitative OR imply that it is often not useful for managers. First, OR assumes that systems are discrete and operate in isolation, when in reality systems are interrelated. Second, OR assumes that the information available to managers and OR analysts is not contaminated when in reality it is; OR hides and occasionally amplifies these contaminations. Third, assumptions regarding managerial and organisational objectives have to be made, but these are rarely precise or simple or unitary. Fourth, traditional OR assumes that it is possible to represent the full range of alternatives open to managers; however, this is rarely the case as the mere construction of a model often eliminates certain alternatives. Fifth, OR assumes that outputs can be implemented; in reality, there may not be any political willpower to make the case for change and push changes through. As the types of problems OR was expected to address changed to the social/political problems outlined above, these assumptions started to disintegrate and certain practitioners and academics began to lose confidence in the existing OR approaches on offer.
The seriousness of this shift in perception of OR is highlighted by the action taken by the Operational Research Society in the UK, which commissioned a report on the future practice of OR: “The Commission was established out of a widespread apprehension of an impending, or even an actual, crisis confronting operational research. The focus of its investigation was Britain, and in particular the ways in which the Operational Research Society could most usefully respond to the situation.” (Operational Research Society, 1986 p.831).