Abstract
M.Ing. (Electrical & Electronic Engineering)
A dynamic process is considered modelled and identified when the model can predict its future behaviour as a result of a known stimulus. However, practical reality is complex and it is quite difficult to totally encompass a model representing a physical phenomenon in a mathematical formulation. Besides, to keep such formulations tractable, certain restrictive assumptions such as, for example, linearity, are often required. The common feature of general control-theoretic methods used for modelling is that they presuppose the valid and accurate knowledge of the processes to be controlled. If, however, one does not understand the inner workings of a complex process that one wishes to model, traditional techniques rarely yield satisfactory results. As systems become more complex it becomes increasingly difficult to make mathematical statements about them which are both meaningful and precise. Thus one is compelled to concede that imprecision and inexactness must be accepted in any real system application. The theory of fuzzy sets is a methodology for the handling of qualitative, inexact, imprecise, information in a systematic and rigorous way. This approach provides an excellent tool for the modelling of human-centered systems, especially because fuzziness seems to be an important facet of the human thinking process. Instead of using a precisely defined or
measured value of a variable, a human being tends to summarize available information by classifying into vague and imprecise categories such as, for example, low, medium, high. In this way, the information received from the outside world is reduced to just what is needed to perform the task on hand with the required precision. Thus there is no need for precise mathematical models and thereby the human (i.e. fuzzy) decision-making mechanism has considerably less computational overhead and is thus faster and more conducive to biological survival than an equivalent precise mathematical model...