"Systems Analysis for Business and Industrial Problem Solving" by S. Optner
Systems analysis is already being widely applied, though with varying degrees of success, to the solution of military and industrial problems. In the opinion of some, the success or failure of solving problems by means of systems analysis depends not on the skill with which the problem was formulated, but on the ability to experiment with the problem as a whole. The success of applying systems analysis and the viability of the solutions it produces are also influenced by the experimenter's ability to represent the real world of the problem in symbolic form.
One should not think that the generalization of problem-solving methods has created a universal method that eliminates the need for a systems analyst. To the extent that the repeated examination of alternatives constitutes the essence of the method, the method remains heuristic. Trial and error continue to exist, but within the framework of a formal procedure. The problem-solving method fixes the main elements of analysis in their proper relation to the problem. This a priori equips the person solving the problem with an understanding of what the structure of the problem's parts is and how a solution consistent with it can be obtained. For solving problems that inherently possess an ill-defined structure, the method has a set of techniques that facilitate the process of determining the structure.
Doubts may arise regarding the appropriateness of applying a general theory of problem solving to the problems of the business world. Some critics believe that the business world is not suited for the use of scientific methods. The assertion that the business world solves its tasks in an environment different from that in which scientific research is conducted is valid; that the problems of the business world are defined to a significantly lesser degree than scientific problems is possibly also valid; but that the problems of the business world are not suited for analysis by scientific methods is not valid.
The name "scientific method" in itself says nothing. The boundary between exact and inexact sciences is difficult to draw. For example, precision is only one of the tests by which a given activity can be classified as scientific. Perhaps the most essential test would be whether the analysis of a subject is conducted without purely descriptive terms. If abstraction and generalization are employed, the difference between sciences becomes mainly a difference in degree. On the other hand, if the arguments are illogical, concepts vague, decisions intuitive, and if the requirements are not precise but diffuse, then the difference between "sciences" becomes a difference in substance. However, such qualities are not characteristic of the business world. In any case, the logic, clarity, definiteness, and accuracy with which reality is represented by decision makers are such that they enable them to conduct their affairs.
Abstraction is a means, not an end in itself. It is a universal language and in both its more developed forms says little to the uninitiated. Under certain conditions, abstraction possesses the enormous advantage of its clarity. By moving from the real world to its various symbolic representations, the systems analyst gains the ability to analyze what he observes. The purpose of this transition is not to exclude the detail or completeness of the real world from its symbolic representation. The systems analyst will methodically move from a representation of the problem obtained by symbolic methods to the real world and back through a repetitive, loop-like process. This iterative process is the means by which the abstraction of the problem becomes similar to its counterpart in the real world.
Problem solving is defined as an activity that preserves or improves the characteristics of a system. A system is the means by which the problem-solving process is carried out. Preservation or improvement of systems is accomplished by introducing changes that increase the effectiveness of resource utilization. These resources are people, materials, equipment, devices, capital, and time.
Changes in the effectiveness of resource utilization are measured by:
- an increase or decrease in the need for resources without a corresponding change in the volume of value and profit;
- an increase or decrease in exposure to risk;
- a change in some relative value measured by criteria.
A problem-solving method more complex than the problem itself should not be applied. Clearly, information processing in solving business and military problems can become very complex. Therefore, when synthesizing an approach to problem solving, it is vitally important to recognize those difficulties of system improvement that may result in failure for the methodology. Such problems belong to a group that I call the group of qualitative or ill-structured problems, since they contain both the known and the unknown, with the unknown tending to dominate, thereby creating a need for particularly precise analytical methods.
Typical problems of this kind are those that:
- are scheduled for solution in the future;
- confront a wide range of alternatives;
- require large capital investments and contain elements of risk;
- depend on the current incompleteness of technological achievements;
- for which the requirements of cost or time are not fully defined;
- are inherently complex due to the combination of resources required for their solution.
A methodology that provides a systemic solution to a problem is primarily aimed at precisely such large-scale, complex problems. These problems are exceptionally difficult to solve and may consist of both quantitative and qualitative elements. The resolution of such problems, which are of a mixed and uncertain nature, is most critical for our world today and represents the most important area for the application of the abilities of executives and systems analysts.
Of particular interest are medium- and large-scale problems for the solution of which systems must be designed today, although their operation is planned for the future. The risk in solving such problems turns out to be great because, for their solution, the future must be predicted in sufficient detail. The further into the future, the greater the risk and the more the problem solver must venture. Great risk may manifest itself in irreparable losses or in unhealthy growth arising in various areas. It may take the form of large expenditure of funds, duplication of product development, doubling of personnel, devices and equipment, unjustified expenditure of time, or uncertainty about the degree of success.
The problem-solving methodology does not require that the desired level of success be precisely defined or that a system against which the solution could be compared exist from the very beginning. There is not even a need for the problem to be fully understood or for it to be clearly and completely formulated. The resolution of these questions is the task of the systems analyst, who must reconstruct all the missing elements and the structure of an incompletely defined problem, its alternatives, and solutions. The specialist solving the problem may identify the system (whose state gives rise to the problem) to be studied in a manner different from the way the problem was originally defined. If he finds that the original formulation of the problem contains something superfluous, contradictory, or unsatisfactory, he may eliminate these deficiencies and correctly pose the problem itself. Thus, the task of the problem-solving specialist also consists, in particular, in defining the problem.
One can say that in solving business problems, the goal is to conduct their analysis and solution with a precision inherent to the problem itself. Regardless of the magnitude or complexity of the problem, the goal is to improve the existing methods by which problems are evaluated, solutions are found, and their implementation is carried out. The problem-solving methodology provides additional means for introducing objectivity into the analysis of the situation. Although problems are solved by people and computers, and although computer programs for finding solutions are written by people, the most important points in finding solutions remain objectivity and logic. Objectivity is the primary requirement in observation. Rationality (logic) is defined as a thought process based on the use of logical inference. A body of knowledge, broadly confirmed by observations, becomes evidence. Observation is the process by which data are identified with a system for the subsequent explanation of that system. Explanation is defined as a logical inference of a statement from well-established facts. The process of explanation must be rational, that is, conducted logically.
Many identified industrial problems turn out to be quantitative-qualitative problems. Quantitative problems are those for which solutions are obtained through the use of predetermined methods of manipulating numbers. Qualitative problems are non-numerical and are concerned with the detailed enumeration of future or poorly defined resources and their properties or characteristics. As the understanding of problems having both quantitative and qualitative aspects improves, their quantitative sides are more easily captured and precise quantitative solutions become more feasible. Operations research has already made much useful progress by applying mathematics to solving the problems of business, defense, and government.
However, for solving those business problems that have not yet emerged from their qualitative state, quantitative methods have limited applicability. It is therefore necessary to bring into play other methods that allow qualitative problems to be solved rationally. Problems possessing both qualitative and quantitative aspects will be called mixed.
Systems analysis is the newest method that makes it possible to cope with such mixed problems. The most difficult to work with are qualitative problems, since they are incompletely structured. Moreover, qualitative problems cannot be easily expressed in their logical components. It is not surprising that in this broad area, the main role is played by judgment, intuition, experience, and sometimes simply caution or recklessness. The purpose of systems methodology is to create a workable structure for solving these difficult problems.
It follows from this that the methodology for solving business and industrial problems must make it possible to:
- prescribe a system that functionally organizes the overall problem-solving process;
- specify the system parameters that provide the structure necessary for solving the problem;
- describe models of the system and its capabilities, which allows for the iteration of output alternatives in the problem-solving process.