A multilayered shielded microwave circuit design method based on genetic algorithms and neural networks

In this work we propose the use of genetic algorithms combined with neural networks to design shielded printed microwave filters. Optimization based on genetic algorithm techniques (GA) is usually very time consuming. By using neural networks, the GA can be applied to the design of multilayered shielded circuits in reasonable time. The neural networks approximates the Green´s functions employed in the integral equation (IE) approach. Fast analysis of the printed circuits inside the shielded structure can be done, therefore limiting the total execution time of the GA. A new specific fitness function specially well suited for the design of microwave filters has also been derived. Results show good performance of the CAD developed for the design of practical printed shielded microwave filters