Optimization Design and Analysis of a Multisection impedance Transforming Branch-Line Hybrid Using Genetic Algorithm

• Main Authors: Chien, Che-Hung, 簡哲弘
• Other Authors: Rainfield Y.Yen
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/90026753688927596758

碩士 === 淡江大學 === 電機工程學系 === 85 === In this thesis, we apply genetic algorithm (GA) to optimally design a broadband multisection impedance transforming branch-line hybrid, which includes two section and three section hybrids. For a given practical realizable impedance range, the return loss, isolation loss and OPIB (output power imbalance) of the branch-line hybrid is optimized (minimized) within a specified frequency band (or bandwidth) when the power division ratio k2, impedance transforming ratio r are specified. Theoretically, the even-odd mode technique is used to analyze the multisection branch-line hybrid, so that we can obtain the scattering parameters from which the objective function is defined. Genetic algorithm is then applied to optimize the objective function. Generally, when GA is used to serach through the parameter space, it tends to give the global optimum solution, and thus suitable for those problem with lots of parameters, such as multisection branch-line hybrid.