Automated design approach for analog circuit using genetic algorithm

The technology of electronic design automation (EDA) has improved the efficiency of design process, however, designer is still required much special knowledge of circuit. During the past decade, using genetic algorithm (CA) to design circuit had attracted many experts and scholars. However, too much more attention was focus on a circuit's function and many other factors had been neglected which caused the circuit had little applicability. This paper proposes an automated design approach for analog circuit based on a multi-objective adaptive CA. The multi-objective fitness evaluation method, which can dynamic adjust parameter, is selected. And a parallel evolution strategy which separates evolution of circuit structure and element value is adopted but also organically combined them by weight vectors. The experimental results indicate that this approach obviously be able to improve the evolution efficiency and could generate numbers of suitable circuits.