Design of multilayered FSS and waveguide filters using genetic algorithms

In contrast to a neural network-based synthesis technique proposed by C. H. Chan et al. (1992), the present method does not search for the optimal design by optimizing the dimensions of a given element shape. Instead, such a design is obtained by cascading elementary building blocks of predefined element shapes and dielectric layers. A small database containing 16 different element types is used to illustrate the proposed technique.<<ETX>>