Multi-Resonator Generation by Genetic Optimization for Application to Planar-Circuit Bandpass Filters

Abstract

This paper proposes a design method of arbitrarily-shaped bandpass filters with multiple resonators in the limited space. Although the proposed method is based on the genetic-algorithm (GA) optimization, we newly introduce the fitness function which estimates not only the magnitude of S-parameters but also the phase characteristics in a passband. Conventional GAs without phase evaluation are difficult to design a filter having the specified number of resonators efficiently, whereas the present method can easily construct such a filter. As an example, arbitrarily-shaped planar-circuit filters with 3 and 4 resonators are optimized in the limited space and fabricated. The effectiveness of the present technique is verified by comparison of frequency responses between the calculated and the measured results.