Genetic Algorithm as Discrete-Time Filter Solver

Abstract

Since the first investigations into Genetic Algorithms (GAs), they have been proposed as control-related systems. A characteristic of GAs is their versatility when used in optimizing problems. In this work a GA is used to adjust a discrete-time transfer function in order to get it working as a filter. Data used to do this work are given input-output sequences. The results of the GA are the values of the parameters of a digital filter. GA is checked in order to work as a high pass and a low pass filter. Five parameters are adjusted from a discrete time second-order digital filter to get the desired behavior of the system. Several possible output sequences are considered and related with the way of working of the GA. Some ideas about the translation of this scheme in hardware are also mentioned.