A multi-algorithm framework for the evolvement of analog circuits
Mingguo Liu, Jingsong He · 2012
Analog circuits play an important part in modern electronic systems. Researchers have conducted a lot of investigations on the evolvement of analog circuits until now. Most works on the evolvement of analog circuits are simply the applications of evolutionary algorithms. As a result, it is significant to consider the following question: is it enough to use the existing algorithms for the analog circuit design problem without the consideration of the problem itself? Analog circuit design is a complex problem that is a mixture of combinatorial optimization and numerical optimization problems. The optimization of the component values can be considered as a numerical optimization problem. This paper propose a multi-algorithm framework for the evolvement of analog circuits. The multi-algorithm framework is proposed on the basis of study on the performances of different basic algorithms. Therefor, it is more suitable for analog circuit design. The multi-algorithm framework consists of several basic evolutionary algorithms: GA, DE, and ES. The multi-algorithm is a prototype and shows significant improvement on analog circuit design. We expect that it might have more inspirational effect on our fellow researchers to present more in-depth studies on analog circuit design problem and efficient evolutionary algorithms for analog circuit design.