群延时均衡器的遗传算法优化设计

在已有文献报道中多采用Powell法设计均衡网络，但Powell法是确定性优化方法，结果的好坏与初始点的选择关系密切，并易陷入局部极值点。由于遗传算法是一种全局随机搜索方法，即使定义的适应性函数不可导或不连续，它也可能以很大的概率求得全局最优解，因而在工程领域中将有十分广泛的实际应用前景。将一种新的实数码遗传算法应用到群延时均衡器的设计中，并给出了巴特沃斯滤波器群延时均衡器的设计实例，结果表明提出的方法十分有效。

Optimal Design of the Group Delay Time Equalizer Using the Genetic Algorithm

It has been reported that the equalizer network was designed by Powell method in previous literatures. Since Powell method is a confirmed optimization method, the results, fine or bad, are closely related to the selection of initial points, and it is easy to fall into local optimal solutions. Because the genetic algorithm (GA) is a global stochastic searching method, and even if the defined fitness function is indifferential or uncontinuous, it is also possible to obtain the global optimal solution with a considerable probability. Thus there will be widely practical application in the engineering field. This paper attempts to present a new real coded GA for applying to design group delay time equalizers, and the examples about designing group delay time equalizers of the Butterworth filter are given This method is illustrated very effectively by results.