Introduction to Operations Research. Churchman C.W., Ackoff R., Arnoff L. 1957.
Operations research originally emerged during World War II in England, and its development was then quickly taken up in the United States. Thus, operations research was first applied to military problems, and only after the war — to problems of financial, industrial, and civil management. Initially, this process advanced more rapidly in England, until in 1951 operations research gained a firm foothold in American industry and began to develop rapidly in the United States.
Although work in the field that later came to be known as operations research began in military organizations, its development and formation as an independent discipline can be described by analogy with the well-known process of industrial development. Before the Industrial Revolution, production was mainly carried out by small enterprises where all management was conducted by a single owner, who made purchases, planned and directed work, marketed products, hired and fired workers, and so on.
Subsequently, a division of administrative functions took place. Heads of production departments, sales departments, financial departments, human resources departments, and others appeared. Increasing mechanization, accompanied by partial automation, led to even more rapid growth in industrial production, which caused a fragmentation of operations and further subdivision of administrative functions.
The emergence of numerous separate operations inevitably required the creation of administrative positions to manage these operations. At present, fewer and fewer heads of production departments have direct contact with the production workshops. They have effectively become employees performing purely organizational functions.
Simultaneously with the growing differentiation and subdivision of administrative functions, scientists' interest in problems arising in various production divisions of industrial organizations increased. In particular, scientists began to pay ever closer attention to questions of industrial production, and as a result of their joint efforts, a number of new branches of applied sciences emerged, with their specializations — mechanical engineers, chemical engineers, industrial construction engineers, statistical quality control engineers, and production management specialists. In other areas, specialists appeared in market research, industrial economics, econometrics, labor relations psychology, industrial sociology, and similar applied scientific disciplines.
During this period of differentiation and subdivision of administrative functions, a new class of management problems began to emerge and take shape — problems that may be called organizational-managerial in nature. These problems arose as a direct consequence of the functional division of labor within an enterprise, which gave rise to the need for organizational activity. In any organization, each specialized division performs a certain part of the overall work. Performing each part of the work is necessary to achieve the organization's overall objectives. However, as a result of such division of labor, each specialized division develops its own goals.
With the emergence of organizational-type problems, the position of management consultant appeared. The role of consultants is to help solve emerging organizational problems by drawing on their experience in solving similar problems in other areas. The method they use consists of identifying what is common to certain problems of a given class when analyzing proposed solutions. Naturally, attempts gradually began to be made to find a common structure (to build a "model") of these solutions and general principles for verifying the correctness of the structures found. These attempts led to the need for a scientific approach to the study of organizational-type problems, although it should be noted that scientific achievements had been used in this field from time to time, even before the advent of operations research.
During World War II, military authorities called upon a number of scientists to help solve strategic and tactical problems. Many of these problems belonged precisely to the type we have called organizational problems. Representatives of various branches of science were assembled into working groups tasked with optimizing the use of resources. These were the first operations research teams.
The purpose of operations research, a discipline that emerged alongside the growth of industrial organizations, is to provide management with a scientific basis for solving problems related to the interaction of various divisions of the organization in pursuit of the organization's overall objectives.
A solution that appears most advantageous for the organization as a whole is considered an optimal solution; a solution that is most advantageous for one or more parts of the organization is called a suboptimal solution.
In conducting operations research, efforts are made to find the best solutions for as large a part of the entire organization as possible. For example, when attempting to solve a maintenance problem at an enterprise using operations research methods, the aim is to account for the impact of various maintenance policies on production as a whole. One can go even further and attempt to determine how this impacts a given firm, which in turn affects the entire industry, and so on. Where possible, an attempt is also made to account for how the impact on the production department will affect other departments and the firm as a whole. In operations research, the aim is to account for interactions or cause-and-effect relationships to the extent that this appears essential. However, in any particular case, the scope of operations research is usually limited in practice, either because access to higher levels of the organization is restricted or because of limitations in time, money, and resources.
There is always a gap between ideal aspirations and real possibilities. Nevertheless, the fundamental goal of operations research comes down to finding optimal solutions to organizational-type problems with account taken of the functioning of the organization as a whole. This general character of the goal of operations research is an example of the systems approach, since by "system" we mean an interconnected complex of functionally related components.
A systems approach to problems does not mean that the problem formulated in the most general terms must be solved in a single research project. However desirable this might be, it is very rarely achievable in practice. In practice, the overall problem is usually solved in parts in a certain sequence. In many cases, the overall problem cannot be formulated in advance, but the solution at one stage makes it possible to determine the next step.
In short, although simultaneous optimization of all parts of a system is highly desirable, practical constraints generally require sequential optimization of individual parts of the system with parallel mutual adjustment of local optima to determine the optimal solution for the problem as a whole.
From the assertion that operations research strives to describe as large a part of the system as possible, it does not follow that one must begin by studying the system as a whole. Most operations research studies begin with known problems of limited scope. But subsequently, the boundaries of the study are expanded as far as circumstances permit. In essence, the scope of the study is a measure that determines one of the distinguishing features of operations research. Accordingly, an operations research study very often begins with the same problems with which a mechanical engineer, industrial construction engineer, chemical engineer, or market researcher might begin, but it rarely ends with problems of that type.
It is characteristic of operations research that in solving each problem, new problems arise. From this, it is clear that the application of operations research methods is ineffective when narrow, limited problems are being solved. The greatest benefit can be obtained from continuous, cumulative research — that is, from sequential transition from one problem to another. This follows from the very definition of the goals of operations research given above.
One of the most essential features of operations research is the aspiration to find an optimal solution, to determine the optimal strategy or design. The task is not to find a solution better than the existing one, but to find the very best solution among all possible ones. Such a solution is not always obtainable because of constraints imposed by the current state of science, insufficient time, resources, or capabilities. But efforts in operations research are continuously directed toward finding the optimal solution or a solution as close to it as possible.
The final decision rests with those who manage the operations, not with those who study them. The operations research team can only recommend solutions or provide a basis for choosing a particular solution. It can also assist in implementing the decision once it has been made.
To summarize, operations research in the most general sense can be characterized as the application of scientific principles, methods, and means to problems related to the functioning of organizational systems, with the aim of providing those who manage these systems with optimal solutions.