Fully automated design of analog circuits using genetic operation and simplified transistor model

Summary form only given. In this talk, an automated design of analog circuits which can extremely reduce the design time is presented. Instead of using complicated transistor models, a simplified model is introduced at the first stage of topology generation. Firstly, the system randomly generates a set of circuits, and then evolves their topologies to fit an environment that is formed by the fitness function translated from the electrical specifications of the circuit. Thus expert knowledge about circuit topologies is not required. In order to create a circuit topology without help of analog circuit designer's knowledge, genetic algorithm (GA) is used. Together with the elite preservation, reuse of effective genetic operations (GO) is introduced to revise the circuit topologies to achieve higher performances. The reuse of GO together with using simplified transistor model, we can reduce the design time and improve the success rate to almost 100 % to obtain a satisfactory circuits.