INTRODUCTION
Concepts of Operational Research
Pre-questions
What is the objective of OR?
What is the major approach of OR?
Why are the terms operational and research misleading?
What other names are given to OR?
What experts did the OR team consist of?
The management science of Operational Research (OR) is entirely practical
in intention and its objective is to propose solutions for complex problems
arising in the direction and management of large systems of people,
machines, materials and money in industry, business, government and
community services. The distinctive OR approach is to develop a scientific
model of the system under review, incorporating measurements of factors
such as chance and risk, with which to predict the outcomes of alternative
decisions, strategies or controls. OR should therefore enable management
to determine its policy and actions scientifically. The purpose of this
standard is to promote the use of standardized terminology for Operational
Research as an aid to effective communications, particularly in business
studies and where OR techniques have come into more general use.
There is a continuing debate within Operational Research circles as to
what OR actually is, because the words “operational” and “research” can
both be misleading. “Operational” suggests that OR ought to be concerned
with ongoing operations, the tactical as opposed to the strategic; whereas
“research” paradoxically suggests almost the opposite, namely that OR
should be concerned with possible future operations which cannot be put
into effect until certain necessary analysis, involving models,
experimentation and pilot studies has been satisfactorily completed. In
order to avoid these semantic difficulties, some organizations, while
making use of the OR approach and techniques, have replaced the term
Operational Research with other phrases such as Management Science,
Management Services, Systems Analysis, or Information Science. In many
occupations and professions, one or two particular techniques may be very
important, while the rest remain unused.
As originally conceived, an Operational Research team was composed of
experts drawn from various fields such as general management, finance,
marketing, logistics, manufacturing and quality assurance. Cross-
fertilization of ideas and pooling of expertise was encouraged, so that
complex and apparently intractable problems might be solved. Many of
the early OR successes resulted from the application of relatively simple
techniques such as linear regression and experimental design in situations
where these techniques had not previously been applied. In due course OR
became a subject in its own right and the OR expert emerged.
The OR approach could be described first in terms of the scientific method
where hypotheses are formulated and tested and second by the standard
approach to problem solving using the following process:
— define the problem and formulate it in such a way that a quantitative
approach is possible;
— develop an appropriate model;
— collect relevant data;
— validate the model by checking that it produces the same answer using
the same data found in the real world;
— evaluate the options by running hypothetical, forecast or what-if data
through the model and interpreting results;
— select the preferred option;
— prepare appropriate documentation and/or computer programs;
— implement.
A central concept within OR is that of a model, by which the OR
practitioner reaches a deeper understanding of the problem area. The
purpose of the model is to generate practical solutions with regard to
difficulties of implementation. As one of the pioneers of OR expressed it,
the third-best solution that can be produced tomorrow may be more useful
than the second-best in a month, or the best solution that could take years
or may be too difficult to put into operation. In OR terms, a model may be
a set of equations, a computer program, occasionally an analogue device,
or indeed any operating description, preferably quantitative, of the
dynamics of the situation. Where this cannot be represented
mathematically, it may be simulated. The relationships modelled may be
on a one-to-one basis, where each element in the real world corresponds
to an element in the model. More usually the relationships are on a one-to-
many basis, for example, where one number may represent all the
invoices passed in a month. Model is to OR as metaphor is to language.
The complexity of an OR problem frequently arises from the fact that there
is no evident structure in it and no obvious method of solution. Various
internal and external pressures, such as industrial relations and legal or
contractual constraints, may be involved and may require modifications to
an otherwise good solution. There may be no solution (in the mathematical
sense of the word) to an ill-defined problem, but enlightenment may be
gained by a structured approach. The first task of OR is therefore to
explore the problem situation itself to determine its structure. The factors
that affect the problem and are affected by it, should then be defined and
assessed quantitatively if possible. The aim should be to clarify the precise
nature of the problem and make it evident to all the parties involved.
When all parties involved in the problem situation agree that it is
adequately formulated, then the operational researcher can employ the
methods and techniques of OR. These are merely the tools that the
operational researcher uses on the way to solving problems and
implementing their solution; some of these tools are surprisingly powerful.
Although the bulk of this standard is devoted to terms used in these
techniques, it is essential to understand that OR is far more than a
collection of techniques.
The role of OR is not finished when proposals for resolving the original
problem are accepted; an important aspect is implementation, which
almost invariably raises difficulties of its own, for example the need to
adjust administrative procedures, to change attitudes, skills and
established practices.