摘要： 利用B-P神经网络对作物生育动态进行模拟。在对B-P神经网络的结构和训练算法加以选择后,网络模型具有拟合精度高、收敛速度快和泛化能力强等特点。B-P网络可用于作物生育进程中单输入单输出、单输入多输出、多输入单输出和多输入多输出动态关系的模拟和作物生育动态的预测。
关键词:
作物生育动态,
B-P神经网络,
模拟,
预测
Abstract: B-P neural network was used for dynamic description of crop growth.After network structure was selected and training method was improved,the network model had characteristichs of high matching precision,high speed of convergence and powerful functionality, based on which dynamic description of single-output by single input and mufti-output by mufti-input as well as prediction of crop development dynamic could be achieved.
Key words:
Crop growth dynamic,
B-P neural network,
Simulation,
Prediction
中图分类号:
郭新宇, 张志鹏, 廖桂平, 郭建强. B-P神经网络在作物生育动态模拟中的应用[J]. 华北农学报, 2002, 17(1): 124-130. doi: 10.3321/j.issn:1000-7091.2002.01.025.
Guo Xinyu, Zhang Zhipeng, Liao Guiping, Guo Jianqiang. Applications of B-P Neural Network to Crop Growth Dynamic Simulation[J]. ACTA AGRICULTURAE BOREALI-SINICA, 2002, 17(1): 124-130. doi: 10.3321/j.issn:1000-7091.2002.01.025.