In this paper

an improved Genetic Algorithm (GA) was applied to the design of high performance E-plane waveguide filters at the first time.For dealing with the shortcoming of GA as low efficiency in local optimum searching

the idea based on Parallel Genetic Algorithm (PGA) known as separate local optimization was presented.By separating the whole population into sub-populations

and adding a new operator known as crossover between two sub-population

the new method can provide as 2~3 times quickly as traditional GA.All numerical examinations illustrated the high convergence speed and good global searching of the new algorithm

and proved the validity of it.