Automated Design and Synthesis of Analog Circuits with Practical Constraints

This paper develops a Genetic Algorithm (GA) based “growing” technique to design and synthesise analogue circuits with practical constraints such as the manufacturer’s preferred component values. Most existing problems with evolutionary search techniques applied to circuit design are addressed. The CA technique developed is then applied both to synthesise the topology of a network as well to perform value optimisation on the components based on a set of commonly used component values (E-12 series). Passive filter networks synthesise this way are realisable, effective and of novel topologies. It is anticipated that this technique can be extended to active networks.