基于L-M优化算法的水稻螟虫预测模型及其初步应用  

Prediction model for rice stem borer based on L-M optimized algorithm and its preliminary application

作　　者：高艳萍[1] 于红[1] 崔新忠[1] 姜国兴[1] 王美妮[1] 

机构地区：[1]大连水产学院信息工程学院,大连116023

出　　处：《农业工程学报》2007年第7期162-165,共4页Transactions of the Chinese Society of Agricultural Engineering

基　　金：辽宁省教育厅攻关项目(05L090)

摘　　要：农村稻区水稻螟虫发生量与多种气候因素相关,各因素之间存在相互作用,是非线性系统。神经网络能有效地描述非线性模型多输入和不确定的特性。传统的BP网络在训练时易陷入局部极小点从而导致训练时间长、收敛速度慢,采用Levenberg-Marquardt优化算法(简称L-M算法)能克服其缺点。在MATLAB中应用L-M算法对辽宁盘锦田间稻区进行水稻螟虫发生量的仿真预测,试验结果表明L-M优化算法的预测精度和收敛速度明显提高,为稻区防控虫害和精确喷药提供参考,具有实用价值。Rice stem borer is a pest, pest emergence volume is related to different climatic factors. The prediction of pest emergence volume is a problem of multi-variable and non-linearity, The artificial neural network is a nonlinear optimization tool. However, the training of neural network by conventional back-propagation （BP） method has intrinsic vulnerable weakness in slow convergence and local minima, L-M optimized algorithm not only possesses the advantages of artificial neural network, but also offsets disadvantages caused by the BP- neural network. On the basis of this, the project selected L-M optimized algorithm as modeling method for prediction of pest emergence volume in rice planting area. The experimental results by MATLAB show that this method is effective and more accurate than BP neural network, and it can be applied for prediction of pest emergence.

关 键 词：神经网络 L-M优化算法 预测模型 水稻螟虫 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]