Using Genetic Algorithms for Device Modeling

In this paper, genetic algorithms are used to find the optimal values for the lumped elements of a given discrete model such that its response best match the modeled device's response. A novel objective function is used by the algorithm which is suited for matching responses with pairs of conjugate poles or zeroes. By applying the method for different values of the device's parameters, the relationship between the model's and the device's parameters can be found. An application of the method to a bandpass filter shows that it can give quite precise results.