为了解决曲线相似性的模式识别问题，提出了一种 基于遗传学习算法的模式识别方法，即层次逐级分解法。并应用该方法解决了配线法确定含 水层参数问题。结果表明，应用该方法确定含水层参数不仅是可行的，而且提高了参数的计 算精度。文中提出的方法可以较好地解决曲线相似性的模式识别问题。

In order to solve the problem of pattern recognition of similarity of curves,a k ind of pattern recognition technique based on genetic algorithms is put forward in this article,namely,progressive hierarchy decomposite technique,and this tech nique is applied to determine the parameters of aquifer using type curve method. The results indicate that this technique is not only feasible,but also can impro ve the accuracy of aquifer's parameters.The technique presented in this art icle is a preferable one for solving the problem of pattern recognition of simil arity of curves.