Next: Models Up: Models and Systems Previous: Models and Systems
The definition of system given the Institute of Electrical and Electronical Engineers professional group in System Science and Cybernetics (from [Rowe, 1965]): ``a large collection of interacting functional units that together achieve a defined purpose''. Or, in other words, a system is made up of three main components: a goal, a set of things and/or rules, and the way this things and/or rules are organized or connected in between them. There is a whole corpus of System Engineering but systems are also studied from a psychological or even philosophical point of view.
In [DeGreene, 1970], the author defends that systems are studied to: improve the system or its successor, determine general theories for new system development, and advance science. He also mentions that there are several scientific methods related to systems such as generalization across systems, analysis and synthesis, and modeling and simulation. All of these methods are of course interrelated. When analyzing a system, the basic steps are: (1) recognition that a problem exists and the solution may be related to systems analysis techniques; (2) definition of the problem in a appropriate form; (3) definition of the system itself (iterative process that starts with a gross approximation and results in a conceptual model); (4) definition of performance criteria; (5) definition of alternative configurations and their evaluation; (6) presentation of alternatives and tradeoff results to the user; (7) performance of iterative analysis during development; (9) analysis of operational systems to gather performance data [DeGreene, 1970].
But one of the most commonly accepted definitions for ``system'' is surprisingly related to the software engineering corpus. In the article entitled ``Definition of System'' [Hall and Fagen, 1956] the authors define: ``A system is a set of objects together with relationships between the objects and between their attributes.'' This definition, that is largely referenced and commented in the literature about system theory, was made in 1956, much before the term object orientation was even coined. So, according to Hall and Fagen, objects are simply the parts, or components of a system.
Attributes are the properties of objects. Relationships tie the system components together. A system cannot be considered as such if it does not have a purpose in itself. This assertion also implies that a system has properties, functions or purposes distinct from its constituent objects, relationships and attributes. Objects can be physical parts or abstract objects such as variables or equations. The relationships are those that tie the system together. The question of whether a relationship is important or trivial depends on the problem to be solved. And another interesting property of any system is that can be subdivided hierarchically into subsystems, sub-subsystems, components, units, parts and so forth. Any of this levels can be considered to be made up of objects as either a subsystem, component, unit or part is an object in itself.
If we accept Hall and Fagen's definition as approximately valid (note that the definition of system given by the IEEE is not conceptually very different) we can assert that the object-oriented paradigm is the best alternative when modeling a system. A system is in fact made of objects that relate in between them, we just have to identify them.
According to [DeGreene, 1970] different features characterize any system analysis: (1) emphasis on an uncertain future; (2) tendency to oversimplify; (3) evaluation of alternatives; (4) aid to conceptualization; (5) decline pressures resisting analysis; (6) use mutually reinforcing techniques; and (7) selection of relevant variables. System analysis have a qualitative beginning, quantification becomes possible as greater detail and precision is introduced. As a matter of fact, and according to [Boulding, 1969], one possible approach to general systems theory is through the arrangement of theoretical systems and constructs in a hierarchy of complexity. In any case, and as we will see in next section, this is usually accomplished through the use of appropriate models.
The state of a system is the collection of variables necessary to describe a system at a given time. An event is an instantaneous occurrence that may change the state of the system. In general, systems are classified as discrete or continuous (see [Law and Kelton, 2000]). A discrete system is one for which the state variables change instantaneously at separated points of time. In a continuous system these variables change continuously. Some systems are neither completely discrete nor completely continuous: combined discrete-continuous simulation. There are three types of interactions between discrete and continuous variables:
2004-10-18