Abstract: The immune algorithm(IA) selects thickness and affinity as the optimality criterions of individuals, and accordingly reproduces the individuals that have small thickness and large fitness. The IA increases the ability to fight premature by calculating individuals' multiplicaties. A numerical example demonstrates its capability to solve the multimodal function. Finally, broadband stepped impedance transformers are designed by four methods respectively. The result shows that microwave circuit optimization based on the IA is feasible and high efficient.

Key words: immune algorithm, genetic algorithm, premature convergence, stepped impedance transformer