摘要
由于遗传算法表现出良好的全局搜索性能 ,因此本文将其应用于高性能E面波导滤波器的设计中 .针对传统遗传算法所存在的局部搜索能力差的缺点 ,本文在并行遗传算法的基础上 ,通过将原有群体划分为多个子群体 ,利用多个子群进行局部极值点的搜索 ,并利用群间迁徙 ,使改进后的算法收敛速度平均提高了 2～ 3倍 .数值实验表明算法具有较快的收敛速度和较好的全局搜索性能 。
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.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2000年第3期121-124,共4页
Acta Electronica Sinica
