Managers have long attempted to solve management problems scientifically, but operations researchers have supplied an element of novelty in the orderliness and completeness of their approach. They have emphasized defining the problem and goals, carefully collecting and evaluating data, developing and testing hypotheses, determining relationships among data, developing and checking predictions based on hypotheses, and devising measures to evaluate the effectiveness of a course of action.
Thus, the essential characteristics of operations research as applied to decision making can be summarized as follows:
It emphasizes models – the logical physical representation of a reality or problem. Models can, of course, be simple or complex. For example, the accounting formula “Assets minus liabilities equals proprietorship” is a model, since it represents an idea and, within the limits of the terms used, symbolizes the relationship among the variables involved.
It emphasizes goals in a problem area and the development of measures of effectiveness in determining whether a given solution shows promise of achieving these goals. For example, if the goal is profit, the measure of effectiveness may be the rate of return on investment, and every proposed solution will arrange the variables so that the end result can be weighted against this measure.
It incorporates in a model the variables in a problem, or at least those that appear to be important to its solution. Managers can control some variables; others may be uncontrollable factors in the problem.
It puts the model, and its variables, constraints and goals, in mathematical terms so that they may be clearly identified, subjected to mathematical simplification and readily utilized for calculation by substitution of quantities for symbols.
It quantifies the variables in a problem to the extent possible, since only quantifiable data can be inserted into a model to yield a measurable result.
It supplements much unavailable data with such usable mathematical and statistical devices as the probabilities in a situation, thus often making the mathematical and computing problem workable within a small margin of error despite gaps in accurate quantifiable data.
Of all these characteristics, perhaps the basic tool – and the major contribution – of operations research is the construction and use of conceptual models. There are many types of models. Some assert logical relationships among variables. These may be referred to as “simulative” or “descriptive” if they are designed only to describe the relationship of elements in a situation. The models useful for planning are referred to as “decision” or “optimizing” models, designed to lead to the selection of a best course of action among available alternatives.