论文部分内容阅读
对函数型神经网络的函数变换原理进行论述，指出函数型神经网络的特点。对函数型网络的学习规则及其实现进行了详细的描述，将该网络应用于土木工程的若干计算问题，得到了较高的收敛速度。给出了与传统BP算法的性能比较。
The functional transformation principle of functional neural network is discussed, and the characteristics of functional neural network are pointed out. A detailed description is given of the learning rules of functional networks and their implementation. The network is applied to several computational problems of civil engineering and a high convergence rate is obtained. The performance comparison with the traditional BP algorithm is given.