Community genetic algorithm design of symmetric E-plane microwave filters

The design of inductive E-plane microwave filters using a community genetic algorithm is discussed. The analysis of such devices necessary for their o ptimization is accomplished using an accurate equivalent circuit technique that speeds the electromagnetic characterization of the filters, making sto chastic optimization feasible. The community genetic algorithm is introduce d as a two-tiered simple genetic algorithm. The stratified nature of this n ovel algorithm allows it to reuse redundant computations in the filter char acterizations, speeding its convergence compared to a simple genetic algori thm. Moreover, the community structure of the algorithm is shown to enhance its robustness. The algorithm is applied to the design of bandpass filters with a prescribed maximum passband rejection, and numerical results are sh own for a four-post and a six-post filter. (C) 1999 John Wiley & Sons, Inc.