Design Of Broadband Matching Network And Gain Equalizer Based On Genetic Algorithm

Traditional design methods for microwave circuit needs formula derivation or circuit parameters calculation, and then, draw the prototype circuit in simulation software to optimize. In the absence of a mature design theory, such as the gain equalizer, only professional designers can propose the original circuit based on experience, and then, optimize it in simulation software. This design method relies heavily on professional knowledge and experience. It is difficult to give satisfactory results when facing the complex and difficult design requirements.In this paper, we propose an automatic RF and microwave circuit design method based on genetic algorithm. This method overcomes the disadvantages of the traditional genetic algorithm, such as the slow convergence speed and premature convergence, and make the proposed algorithm suitable for the automatic design of RF and microwave circuit. The method can search the circuit topology and the corresponding electrical parameters simultaneously by global search, the whole process need not human intervention or the given circuit topology, what we need to do is to input related data, and then, we will obtain the optimal circuit. We have successfully applied the method to the design of broadband matching circuit and gain equalizer in this paper. Comparing with real frequency technique, the broadband matching networks designed by the proposed method achieve better performance under the circumstance of using less lumped element. Moreover, an broadband matching network is designed to meet the need of actual situation. The designed gain equalizers, which have various attenuation curve from simple to complex, have a very good balance effect. The two circuit design process with a high degree of automation, greatly improve the efficiency and accuracy of design, and are of great significance in engineering practice and theoretical research.