Modeling and optimization of planar microwave filters

Abstract

In this thesis, the FEM-GA and FEM-PSO approaches are proposed for innovative design of new irregular planar filter structures with higher performance or smaller size. The traditional design restricts filter structures to regular shapes. The proposed approaches allow arbitrary structures and make the optimal filter design possible. The approaches combine powerful and flexible methods — finite element method (FEM) as the field simulator and the genetic algorithms (GA) or particle swarm optimization (PSO) as the optimization engine. FEM is a very powerful and efficient numerical technique for problems with arbitrary structures and material profiles. GA and PSO are stochastic search and optimization techniques which can optimize large and complicated problems. The combinations greatly increase the design capability, allowing innovative filter design with improved performance.