L-M算法在高层结构体系选型中的应用

Choice of structural styles for tall building based on L-M algorithm of artificial neural network

采用L-M算法人工神经网络进行高层建筑结构体系选择的研究。与传统的BP算法和改进的带动量自适应学习率BP算法比较表明,L-M算法可以解决高层建筑结构体系选择的问题,算法快,精度高,是运用人工神经网络解决高层建筑结构体系选择的较好网络。

In this paper,L-M(LevernbergMarquart)algorithm is discussed and compared with traditional BP(BackPropagation)algorithm and improved selfadaptive learning algorithm with momentum in the course of analysis,respectively.From the research,it is concluded that L-M algorithm of ANN can solve the problem on choice of the structural styles;whose runningspeed is 102～103 times faster than traditional BP algorithm's and is more efficient than traditional BP algorithm and improved selfadaptive learning algorithm with momentum.