A Diferent Approach in the Parameters' Identification of a Jfet Using Genetic Algorithms

Abstract

The genetic algorithms are developing in three directions: the genetic algorithms theory, the genetic algorithms programming and the study of the problems that can be solved with genetic algorithms. In this paper it is presented a study on the identification of the parameters of a JFET (Junction Field Effect Transistor). The problem is very exciting because the JFET has two mathematical models: an empirical one, and an analytic one, both of the models being nonlinear in parameters. In a parametric identification problem, it is minimized the distance between an experimental data set and an analytical function, which represent the mathematical model of the studied phenomenon. Basically, a genetic algorithm can maximize a fitness function, which is a positive defined function whose maximum is searched. However, genetic algorithms can also solve minimum problems, on condition that to the minimum problem can be applied an algebraic transform or a rank based transform in a maximum problem.