"Systems Analysis and Long-Range Planning." Conference proceedings. G.S. Pospelov, 1972 (in Russian).
Systems analysis, as an applied branch of general systems theory that historically arose from the methods of designing and developing large-scale technical systems, turns out to be closely connected with the entire range of problems in long-range national economic planning. By its very nature, systems analysis involves the formulation of a comprehensive plan for an operation or undertaking.
Plans of operations implemented by hierarchical organizational systems include instructions on who must do what, when, and which resources are allocated for this purpose. In other words, an operational plan decomposes the operation's overarching goal into a hierarchy of goals and tasks to be performed by individual executors. This means, among other things, that an operational plan constitutes a system of decisions made in advance (before the operation begins) regarding who is to do what and when.
In long-range planning, the interval between a decision made at the moment of planning and its execution may span 10–15 years or more, necessitating planning for the development of the organization's executors, their education, and so forth.
A plan, as a system of goals and tasks, can be represented by a so-called hierarchical graph whose vertices represent goals and tasks at various levels, and whose arcs represent relations of subordination, interaction, and significance among them. Such a graph well illustrates two principles of systems analysis:
- The means of achieving a goal follow from the goal itself and its analysis.
- Goals and tasks at any lower level are treated as means of achieving the goals of the upper level.
From the definition of a plan follows the concept of planning as a process of making decisions before the start of an operation, or somewhat differently: planning is a hierarchical process of forming preliminary decisions in a system that determines the order in which a sequence of individual measures, operations, and actions must be carried out. The result of planning (its goal) is the construction of a model of the future operation in the shared understanding of the organization.
Let us dwell on the relationship between the methods of operations research (economic-mathematical optimization methods) and systems analysis.
Operations research, as a theory of mathematical models for decision making, presupposes that the goal is given or the task is posed, that alternative means (strategies, controls) for its achievement are defined as countable or uncountable sets, and that criteria for selecting among alternatives are specified. Next, a mathematical decision-making model is constructed; mathematical expressions are found for the optimality criterion and the constraining factors (conditions); data and norms are collected for solving the problem. At the next stage, an algorithm for finding the optimal solution is developed and the problem is programmed; the computer then automatically finds it.
It should be noted that the optimal solution obtained in this way considers only quantitative factors and, even from this standpoint, focuses on only one aspect of the problem. Thus, the class of operations research problems known as allocation and assignment problems, also called optimal planning problems, is among the most common (economic-mathematical methods) problems solved by mathematical programming algorithms. This class includes transportation problems, locating new enterprises in a given industry, and assigning them to consumers. In all these cases, the goal is set, and the optimal plan, automatically generated by a computer, is by no means a plan of operations for transportation or industry development. The optimal plan or optimal solution indicates how consumers should be distributed among suppliers or where enterprises should be located and what the production volume of each enterprise should be. Since such plans are not operational, they cannot be comprehensive and, in this sense, are one-sided.
Systems analysis always involves comprehensive or operational plans that are hierarchical in nature. In this context, the determination of the goal of activity itself, of the alternatives for its achievement, and of the optimality criteria is the subject of systems analysis. Whereas in operations research the goal of the operation is given, in systems analysis what should be considered as given is a change in the situation or the realization of a desired scenario, where a situation (scenario) is understood as the aggregate state of the system and its environment. The situation in a system can be changed by performing an operation from a set of alternatives and, accordingly, achieving one of the overarching goals.
Thus, the task of systems analysis includes determining alternative overarching goals, decomposing each overarching goal into its own hierarchy of goals and tasks — sometimes called a course of action; calculating and assessing the consequences of each alternative course of action; and selecting, on this basis, the optimal course of action with its corresponding overarching goal.
As can be seen, the problems of systems analysis and operations research differ substantially, despite the fact that in the process of constructing comprehensive plans (hierarchies of goals and tasks), operations research methods are widely used.
It should also be noted that operations research, as a mathematical theory of decision making, was developed for application to single-goal, single-level systems. Due to the difficulty and complexity of the problem, the mathematical theory of decision making for the case of multi-goal, single-level systems (various branches of game theory) proved less fully developed and less widely used in practice.
As for multi-goal and multi-level (hierarchical) systems, it is still too early to speak of their mathematical optimization theory. However, in the practice of planning and decision making, it is precisely this last type of system that one has to deal with. Systems analysis also pertains precisely to this type of system, and the level of mathematical formalization is such that we are compelled to agree that systems analysis is "enlightened common sense, in whose service mathematical methods have been placed."