OPTIMIZATION TECHNIQUE FOR MODELING OF STRUCTURE AND PARAMETERS OF SYSTEMS.

To design a control loop for a system and to study the stability of that loop, it is important to know a model which represents that system. Mathematical models with different parameters has been suggested in the literature. The response of the original system to a certain signal has been measured in the time domain or in the frequency domain. An error objective function has been built between the response of the original system and the response of the mathematical model to the same input signal. The parameters are changed according to a certain strategy, in order to make the objective function minimum. The number of parameters; their form i. e. the structure; and an a priori value for each must be known beforehand. In applying this method a new technique for the optimization is suggested here to optimize the structure and the parameters. This has resulted in more exact models compared to those obtained using previous methods.