神经网络与符号推理集成系统在桩基检测中的初步应用

An Integrated System for Neural Network and Symbolic Inference and Its Application to Integrity Detection of Foundation Piles

本文提出一种神经网络与符号推理集成系统ＩＮＮＳＩ（ＩｎｔｅｇｒａｔｉｏｎｏｆＮｅｕｒａｌＮｅｔｗｏｒｋａｎｄＳｙｍｂｏｌｉｃＩｎｆｅｒｅｎｃｅ）。该系统的特点是领域规则和学习样本能够共同对系统做出贡献。通过桩基完整性检测仿真试验，验证了应用该方法检测桩基的完整性是有效的，并且优于单独使用符号推理方法或神经网络方法。

An Integrated system for neural network and symbolic inference is presented. In the system the two intelligent functions, neural network and symbolic inference, may work together to make contribution. By applying it to the computer simulation for integrity detection of foundation piles, the system is proved to be effective.