Multiobjective Optimization of Microwave Circuits with Many Structural Parameters and Objectives

For many microwave circuit designs, there are multiple performance characteristics to be optimized simultaneously. It becomes more and more difficult to design these microwave circuits with the number of structural parameters and performance characteristics increasing. In this paper, a new multiobjective optimization technique, MOEA/D (multiobjective evolutionary algorithm based on decomposition) combined with FDTD (Finite Difference Time Domain) method, is proposed. As an example, a microstrip low-pass filter with 160 structural parameters and 4 performance characteristics is designed for obtaining low insertion loss, sharp out-of-band rejection, high stop-band attenuation, and good linearity. Comparisons on computational cost and searching speed are made to verify the high-efficiency of the proposed multiobjective optimization method.