摘要
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。
This paper discusses the thedamelltal elemenis of funtion transformation for Functional-Lind Neural Nforord (FLNN), and pointed out the characters of FLNN. Leaming rule and implementaion ofFLNN is discussed in detail. The paPer also presents several computational problems to which FLNN isapplied and high convereence speed is attained Finally compare of performance with the traditional BP algorithm is discussed.
出处
《计算机工程与设计》
CSCD
北大核心
1999年第5期38-41,共4页
Computer Engineering and Design
