Application of BP Neural Networks in Inventory Dynamic Modeling

The authors establish an original inventory model based on BP neural networks in dynamic environment using the inventory history data, and overcome the net overfitting problem occurred with insufficient samples by using early stopping method. After a simple-relation inventory model training is attained, they get inventory model and analyze it dynamically by reconstructing the model after getting new sample data. An example is provided to illustrate the model. The theoretical evidence is provided for the inventory system to make management decision.     

BP Neural Networks and Its Application in Lung Cancer Intelligent Diagnosis System

The learning algorithm of networks is discussed. The programming example of 3 layer BP networks is given with Visual C++6.0 program langue. Based on this model, a lung cancer intelligent diagnosis system is successfully implemented. Furthermore, the paper introduces network's structure design, preferencesand the source of stylebookdatum in factual applications. The ameliorative arithmetic is applied to the study of networks and BP dynamic evolving process is designed. The experiments indicatecell images are recognized and classified by the trained neural network. The study illustrates the system has feasibility and clinical value in lung cancer diagnosis.     

Improved Ellipsoidal unit Neural Networks and Its Applications in CSTR

To overcome the limitations of the standard ellipsoidal unit neural networks, some new approaches used in ellipsoidal unit neural networks have been proposed. These new approaches address three main issues: firstly, to understand better and represent the nature of fault classification boundaries; secondly, to determine the network structure without the usual trial and error schemes; lastly, to avoid erroneous generalizations. The application in CSTR shows that the ellipsoidal unit networks can possess arbitrary nonlinear classifying ability, nonlinear interfacial describing ability, and obtain accurate and efficient diagnosis results.     

Forecast of the Horizontal Displacement of Cement-soil Gravity Retaining Wall Using BP Neural Networks

The authors are the first who succeeds in using BP neural networks to forecast the horizontal displacement of cement soil gravity retaining wall. This method is proved to be very useful and important. It has not only the ability to resolve many very complex problems, which cannot be solved by numerical methods, but also accords very well with the spot measurement values.     

Multiple Layer BP Neural Network for Fuzzy Calculating

The FCBP(Fuzzy calculating BP) algorithm which is proposed by this paper hasovercome the sensitivity for samples,reduced the number of input layer's samples,lightened the burden of input layer.it is suitable to fuzzy inference and pattern recognition.     

Using BP Neural Network to Determine the Membership Function of the Input of Fuzzy Control System

While design the fuzzy controller, it is very important to determine the membership function of fuzzy variables.The data can be broadly classified as fuzzy sets by using the classification property of the BP neural network. The author selects a BP neural network with one hide layer and uses S function to the input and hide layer, and linear function to the output layer.Advanced BP algorithm isused to train the BP neural network in the environment of MATLAB . The nearer to the target values is the better the last output is.With the trained BP network , the membership values of the inputs can be got ten. This method has high rate and low error.     

遗传算法优化人工神经网络模型在日光温室湿度预报中的应用

提出一种基于遗传算法优化BP神经网络的方法预测日光温室湿度环境因子。实测日光温室内影响空气湿度的环境因子组成数据样本作为神经网络的输入,采用基于实数编码的遗传算法替代随机设定神经网络的初始权阈值,然后通过改进的BP算法在由遗传算法确定的搜索空间中对网络进行精确训练。模型预报值和实测值基于1：1线的决定系数R2和预测平均相对误差MSE分别为0.9857和3．1％。结果表明,遗传算法优化BP神经网络预报模型收敛速度快、预测精度高。可为日光温室的湿度环境调控制提供理论依据和决策支持。     

基于BP神经网络的农业电子商务网站评价研究

提出了适合农业电子商务网站的评价指标体系。建立农业电子商务网站的BP神经网络评价模型,通过BP神经网络对数据进行训练、测试,对提出的评价指标体系进行了验证。通过对结果的分析发现,建立的基于BP神经网络的评价模型在农业电子商务网站评价中具有可行性,在网站建设方面具有较强的参考价值。     

The Method of Diagnosing Brain-dysfunction Based on Membership Function and BP Neural Network

Membership functions.formed with the characteristic values obtained by applying the continuous reaction time (CRT) theory from controls and heptic encephalopthy and proposedin this paper.whith these membership functions,the CRT data to be diagnosed are calculated in advance,and then discriminated by BP meural network to differentiate patients with or without braindysfunction.The troubles of low accuracy and efficiency,encountered in training BP network withfuzzy samples.scattered data and extended samples,are solved.     

Cost Estimation of Transmission Line Based on Artificial Neural Network

Based on the structure of cost factor in transmission lines project and the characteristic of neural networks, the paper shows up using artificial neural networks to estimate the cost of project and provide a method to investigate them. By analyzing the structure of project cost and influence factor, the paper builds both input and output neural cell for cost of transmission line project. The paper builds the model of artificial neural networks and exports the steps of the algorithmic. The project sample is used in history to train and test the model. At last, the model gets satisfactory result and meets the investigation requirement of engineering budgetary estimation. Therefore, the paper provides an impersonality and quick investigation method for budgetary estimation of power projects.     

Malfunction Diagnosis of Three-phase Full-bridge Controlled Rectifier Based on BP Neural Network

The diagnosis of the power electronic circuit is very intricate. One of the main reasons is that the structure of the circuit will change if the power device is not working .The thyristor is the easiest to be mangled. So diagnosing the malfunction is the most important about the diagnosis of the power electronic circuit. The paper puts forward a malfunction diagnosis of the thyristors of three-phase full-bridge controlled rectifier with BP neural network. After analyzing the output waveforms of malfunctioning circuit and training a BP neural network with the sampling data of malfunctioning waveforms, a well training BP neural network is constructed and used to diagnose the malfunction. The simulation and experiment demonstrate that this method is valid.     

基于人工神经网络的植物油料实际压缩比与临界压榨压力预测

基于BP神经网络建立了菜籽、菜籽仁、花生、大豆、芝麻和亚麻籽的实际压缩比预测模型,实际压缩比神经网络预测值与试验实测值吻合。根据预测的实际压缩比曲线,确定出油料工程实际临界压榨压力分别为菜籽、菜籽仁、花生、亚麻籽80MPa,芝麻100MPa,大豆60MPa。     

基于逐步回归与BP神经网络的日光温室温湿度预测模型对比分析

为构建较准确的日光温室温湿度预测模型,于2011-2013年冬季(1月、2月、12月)天津市宝坻区开展温室内外环境监测试验,并建立3种天气类型(晴、多云、阴)下3个时段(0-8时、8-17时、17-23时)逐步回归与BP神经网络温室内温湿度预测模型。结果表明:1)温室内气温逐步回归模型9种情况下模拟值与实际值的绝对误差小于3℃的平均准确率Rate(≤3℃)为88%,平均均方根误差(RMSE)为2℃;BP神经网络模型9种情况下模拟值与实际值的绝对误差小于3℃的平均准确率Rate(≤3℃)为94%,平均均方根误差(RMSE)为1.6℃。应用BP神经网络建立的气温预测模型相对更为准确稳定。2) 相对湿度逐步回归模型9种情况下模拟值与实际值的绝对误差小于6%的平均准确率Rate(≤6%)为81%,平均均方根误差(RMSE)为5.7%;BP神经网络模型9种情况下模拟值与实际值的绝对误差小于6%的平均准确率Rate(≤6%)为80%,平均均方根误差(RMSE)为6.7%。两类模型均不适宜预测8-17时日光温室相对湿度,而17-23时与0-8时应用逐步回归建立的湿度预测模型相对更准确稳定。     

Improved BP Algorithm in Vibration Failure Diagosis of Steam Turbine-Generator Set

A failure diagnosis model of neural networks for the vibration failure feature of steam turbine-generator set is established on the basis of the improved BP algorithm ,and used for diagnosing practical generator set,the results of verification show that the method is effective.     

Calculation Model Based on Improved BP Network of Refrigerant State Parameters

BP network is improved by using Levenberg-Marquardt optimization method. Under large swatch input circumstance, more fast network constringency speed and more highly approach precision are obtained. Calculation model based on improved BP neural network of refrigerant state parameters is established. The main defect of refrigerant hot ties of matter model in existing refrigeration and air-conditioning appliances simulation system is overcome.     

A Kind of Fuzzy System Modelling and Control Based on Neural Networks

Combining artificial neural networks(ANN) with fuzzy system theory,a kind of modelling & control method of fuzzy system based on ANN is presented.The simulation researches have verified that the proposed approach can be applied effectively to a number of control systems which are defficult to build strict mathematical model.     

Predictive Control Approach for Active Power Filter Based on Neural Networks

In this paper, a predictive control approach using neural networks for active power filter is proposed. The control system of active power filter using this method make using of the internal model control technology based on neural networks, meanwhile, to solve its questions such as lag because of calculating using neural networks, a predictive model based on neural networks is introduced. Simulation analysis shows that this control approach can compensate the lag of system and take advantage of self-adaptive characteristic of neural networks. Good result can be obtained.     

Pre-bid Decision-making Process Based on Artificial Neural Networks Technology

There are complicated non-linear relations among influence factors, and between influence factors and decision - making results, which cannot be handled by common analytical method effectively. So the ANN technology was applied in the pre-bid decision-making processes. An indicator system of influence factors for pre-bid decision-making process was built based on extensive investigations, comparisons and analyses, with due considerations to the ANN analysis requirements. And then the criterion and method for quantifying the indicators were put forward by quantificational methods and fuzzy theories. An ANN model for pre-bid decision- making process was constructed and performed in Matlab, which has powerful learning ability and fast learning speed. The model was verified by a few examples from engineering practice.     

Construction of Wavelet Neural Networks and its Robustness Analysis of Learning Algorithm

In this paper,a method of constructing the wavelet neural networks for nonlin-ear fun ctional approximation is discussed.The expon ential convergence of the training process andits robust stability to the noise perturbances and the network design errors are also proved.