SAW transducer optimization by the evolutionary computation

A novel method is presented which applies the principles of biological evolution to the optimization of interdigital transducer (IDT) layouts based on the impulse model. The main point is a random variation of the finger overlapping lengths within the IDT by an algorithm simulating the biological mutation. The responses of both the original and mutated IDT are calculated and compared. The better one is selected for a further mutation. The cycle of mutation and selection is repeated until the created layout meets the requirements on the transfer function or it is terminated by limiting the computing time. The method has been used to meet standard requirements presented in literature. The results show that IDTs designed by mutation and selection yield the same height level of performance as those optimized by well known methods. An advantage of our method should be the applicability of secondary conditions on some layout parameters