Synthesis of active-RC filters using genetic algorithms

Abstract

Traditionally, analogue filter designers synthesise their circuit topologies using pencil-and-paper method. Based on their previous experience, they generally use their intuition to reach proper network configuration for the desired filter function. However, this approach might be very time-consuming and cumbersome. With the advent of computer technology, the generation of analogue filters (or any other types of circuit for that matter) can be done using software. Taking advantage of this high computing power, some evolutionary methods, like genetic algorithms (GAs), can be employed to obtain the desired circuits sophistically, which generally leads reaching the goal faster. In this paper, automatic synthesis of filters utilizing GAs is presented. Although the approach can be applied to all analogue filters, this paper specifically focuses on active-RC filter topologies, which compose of an active element, e.g. operational amplifier, and resistors and capacitors. Some of the filter topologies that are generated by the constructed software are presented. Experimental results regarding to GA performance are also included.