CAJ | 학술논문

阻抗匹配问题在射频技术中有着极其重要的意义，良好的阻抗匹配不仅有利于电路的低压低耗实现，而且较小的反射系数为系统的可靠运行提供了保证。针对遗传算法在自动阻抗匹配过程中收敛速度慢且易陷入早熟（即陷入局部最优），提出利用分层机制来预防早熟，并对遗传算法中的交叉操作算子进行了改进来加快收敛。通过MATLAB对上述算法进行实现，并采用不同频率及不同负载阻抗对上述算法及遗传算法进行测试从而进行比较。结果表明分层遗传算法很好地预防了早熟，收敛速度比原算法得到较大的提升。
조항필배문제재사빈기술중유착겁기중요적의의，량호적조항필배불부유리우전로적저압저모실현，이차교소적반사계수위계통적가고운행제공료보증。침대유전산법재자동조항필배과정중수렴속도만차역함입조숙（즉함입국부최우），제출이용분층궤제래예방조숙，병대유전산법중적교차조작산자진행료개진래가쾌수렴。통과MATLAB대상술산법진행실현，병채용불동빈솔급불동부재조항대상술산법급유전산법진행측시종이진행비교。결과표명분층유전산법흔호지예방료조숙，수렴속도비원산법득도교대적제승。
Matching impedance is very important to design RF antenna circuits. Optimized impedance is beneficial to realize low voltage and low power antenna circuits, and lower reflection coefficient will ensure that the antenna system can be running smoothly and reliably too. To avoid slow and premature convergence, it proposes a layered process technique and performs refor-mation including improving crossover operation. It runs the program on the MATLAB simulation platform with different loads at different frequencies and analyzes results in different conditions. These results indicate that the reformative GA can escape from premature convergence and it gets solutions whose errors are lower than predefined errors more speedy than the conven-tional GA.