Neural network approach for modification and fitting of digitized data in reverse engineering

Reverse engineering in the manufacturing field is a process in which the digitized data are obtained from an existing object model or a part of it, and then the CAD model is reconstructed. This paper presents an RBF neural network approach to modify and fit the digitized data. The centers for the RBF are selected by using the orthogonal least squares learning algorithm. A mathematically known surface is used for generating a number of samples for training the networks. The trained networks then generated a number of new points which were compared with the calculating points from the equations. Moreover, a series of practice digitizing curves are used to test the approach. The results showed that this approach is effective in modifying and fitting digitized data and generating data points to reconstruct the surface model.     

基于混合优化的RBF神经网络集成的降水预报模型

针对传统的单个RBF神经网络集成中个体的隐节点个数和初始参数难以客观确定的不足,为了提高泛化能力,提出一种以高斯核函数的混合优化的RBF神经网络的方法,首先引入正交最小二乘法动态客观的获取数据中心的个数、数据中心及权值;然后通过计算隐层中心点间最小距离作为扩展常数;最后使用剃度法调节权值、中心及扩展常数使网络参数和结构达到最优.该方法结合了正交最小二乘法和剃度算法的优点,通过从结构和算法两方面的调整提升了单个的传统的RBF网络的性能.并将上述优化混合的RBF神经网络与主成分分析方法相结合建立模型.本文以广西5月逐日降水事先初选的众多预报因子进行主成分分析算法提取有效的几个综合因子,然后使用混合算法优化的径向基网络建立降水预测模型.结果表明,该模型具有较好的收敛效果和泛化能力,在预报性能上明显优于同期的T213降水预报,具有一定的普遍适用性.     

Nonlinear modelling of a SOFC stack by improved neural networks identification

The solid oxide fuel cell (SOFC) is a nonlinear system that is hard to model by conventional methods. So far,most existing models are based on conversion laws,which are too complicated to be applied to design a control system. To facilitate a valid control strategy design,this paper tries to avoid the internal complexities and presents a modelling study of SOFC per-formance by using a radial basis function (RBF) neural network based on a genetic algorithm (GA). During the process of mod-elling,the GA aims to optimize the parameters of RBF neural networks and the optimum values are regarded as the initial values of the RBF neural network parameters. The validity and accuracy of modelling are tested by simulations,whose results reveal that it is feasible to establish the model of SOFC stack by using RBF neural networks identification based on the GA. Furthermore,it is possible to design an online controller of a SOFC stack based on this GA-RBF neural network identification model.     

基于正交矩的纹理分割

在识别一幅图像中的界面或物体时，一般先要进行纹理分割．本提出了基于勒让得矩的纹理分割方法．首先在图像的小窗口中计算矩值，然后用一个非线性转换器把它转化成纹理特征．再用这些特征组成特征向量作为输入数据．接着采用RBF人工神经网络对提取的特征进行分割．用k均值算法训练RBF人工神经网络的隐层．输出层的训练是采用基于LMS的监督式数学模型．该算法成功地分割了许多灰度级图像．和基于几何矩的纹理分割相比，用正交矩可以降低分割错误率．     

具有饱和非线性输入的一类混沌系统自适应神经网络控制

针对一类具有饱和非线性输入的混沌系统,基于RBF神经网络的逼近能力提出一种控制方案。该方法利用自适应控制和鲁棒控制,使系统可在模型函数和外扰未知下,设计出结构简单有效的控制器,有效消除了现实中由于饱和非线性输入的存在而引起的控制器抖动的不良控制效果。仿真结果表明了所提控制方法的可行性。     

LS-SVM model based nonlinear predictive control for MCFC system

This paper describes a nonlinear model predictive controller for regulating a molten carbonate fuel cell (MCFC). In order to improve MCFC’s generating performance, prolong its life and guarantee safety, it must be controlled efficiently. First, the output voltage of an MCFC stack is identified by a least squares support vector machine (LS-SVM) method with radial basis function (RBF) kernel so as to implement nonlinear predictive control. And then, the optimal control sequences are obtained by applying genetic algorithm (GA). The model and controller have been realized in the MATLAB environment. Simulation results indicated that the proposed controller exhibits satisfying control effect.     

Bayesian student modeling and the problem of parameter specification

In this paper, the application of Bayesian networks to student modeling is discussed. A review of related work is made, and then the structural model is defined. Two of the most commonly cited reasons for not using Bayesian networks in student modeling are the computational complexity of the algorithms and the difficulty of the knowledge acquisition process . We propose an approach to simplify knowledge acquisition. Our approach applies causal independence to factor the conditional probabilities and decrease the parameters required for each question to a number linear in the number of concepts. This also provides the new parameters with an intuitive meaning that makes their specification easier. Finally, we present an example to illustrate the use of our approach.     

基于神经网络预报的动态矩阵预测控制

提出了一种基于径向基神经网络预报的动态矩阵预测控制新算法，在该算法中，先用RBF神经网络辨识对象模型，同时预测对象的未来输出，然后用动态矩阵控制算法进行滚动优化和反馈校正。该方法解决了非线性、时变对象难以建模的问题，仿真结果验证了这一新型算法的可行性。     

Local and global approaches of affinity propagation clustering for large scale data

Recently a new clustering algorithm called ＇affinity propagation＇ （AP） has been proposed, which efficiently clustered sparsely related data by passing messages between data points. However, we want to cluster large scale data where the similarities are not sparse in many cases. This paper presents two variants of AP for grouping large scale data with a dense similarity matrix. The local approach is partition affinity propagation （PAP） and the global method is landmark affinity propagation （LAP）. PAP passes messages in the subsets of data first and then merges them as the number of initial step of iterations; it can effectively reduce the number of iterations of clustering. LAP passes messages between the landmark data points first and then clusters non-landmark data points; it is a large global approximation method to speed up clustering. Experiments are conducted on many datasets, such as random data points, manifold subspaces, images of faces and Chinese calligraphy, and the results demonstrate that the two approaches are feasible and practicable.     

基于粒子群算法优化的城市供水量预测模型研究

鉴于BP神经网络、RBF神经网络在城市供水量预测精度上的不足,利用粒子群算法优化两者相关参数,实现更高预测精度,并通过建立BP神经网络、RBF神经网络、PSO-BP神经网络、PSO-RBF神经网络分别对城市供水量数据进行仿真预测。最终测试样本统计结果显示：RBF神经网络比BP神经网络平均相对误差（MRE）低约1%,在拟合度（R2）上高约0.014;PSO-BP神经网络比BP神经网络在MRE上降低约1.25%,在R2上提高约0.05;PSO-RBF神经网络比RBF神经网络在MRE上降低约0.3%,在R2上提高约0.072。由此说明RBF神经网络比BP神经网络在城市供水量预测方面更有优势,并且利用粒子群算法优化神经网络模型参数可有效提升神经网络预测精度。     

Estimation of vegetation biophysical parameters by remote sensing using radial basis function neural network

Hyperspectral reflectance (350~2500 nm) data were recorded at two different sites of rice in two experiment fields including two cultivars, and three levels of nitrogen (N) application. Twenty-five Vegetation Indices (VIs) were used to predict the rice agronomic parameters including Leaf Area Index (LAI, m2 green leaf/m2 soil) and Green Leaf Chlorophyll Density (GLCD, mg chlorophyll/m2 soil) by the traditional regression models and Radial Basis Function Neural Network (RBF). RBF emerged as a variant of Artificial Neural Networks (ANNs) in the late 1980’s. A large variety of training algorithms has been tested for training RBF networks. In this study, Original RBF (ORBF), Gradient Descent RBF (GDRBF), and Generalized Regression Neural Network (GRNN) were employed. Results showed that green waveband Normalized Difference Vegetation Index (NDVIgreen) and TCARI/OSAVI have the best prediction power for LAI by exponent model and ORBF respectively, and that TCARI/OSAVI has the best prediction power for GLCD by exponent model and GDRBF. The best performances of RBF are compared with the traditional models, showing that the relationship between VIs and agronomic variables are further improved when RBF is used. Compared with the best traditional models, ORBF using TCARI/OSAVI improves the prediction power for LAI by lowering the Root Mean Square Error (RMSE) for 0.1119, and GDRBF using TCARI/OSAVI improves the prediction power for GLCD by lowering the RMSE for 26.7853. It is concluded that RBF provides a useful exploratory and predictive tool when applied to the sensitive VIs.     

基于 DenseNet 的人脸脸型分类研究

脸型在三维人脸重建、人脸识别与检索等领域有着重要应用。针对脸型分类问题提出一种基于傅里叶描述符、三角形半径描述符和几何特征描述符的 DenseNet 网络脸型分类方法。通过主动形状模型方法定位得到人脸边缘轮廓点,分别使用傅里叶描述符、三角形半径描述符和几何特征描述符将轮廓点数值化为一维向量,并将其输入 DenseNet 网络进行训练,实现脸型分类。通过实验对比分析,该方法能够充分挖掘人脸形状信息,有效提高脸型分类准确率,同时避免旋转、尺度等影响。     

Numerical simulation of interaction between laminar flow and elastic sheet

A numerical simulation of the interaction between laminar flow with low Reynolds number and a highly flexible elastic sheet is presented. The mathematical model for the simulation includes a three-dimensional finitevolume based fluid solver for incompressible viscous flow and a combined finite-discrete element method for the three-dimensional deformation of solid. An immersed boundary method is used to couple the simulation of fluid and solid. It is implemented through a set of immersed boundary points scattered on the solid surface. These points provide a deformable solid wall boundary for the fluid by adding body force to Navier-Stokes equations. The force from the fluid is also obtained for each point and then applied on the boundary nodes of the solid. The vortex-induced vibration of the highly flexible elastic sheet is simulated with the established mathematical model. The simulated results for both swing pattern and oscillation frequency of the elastic sheet in low Reynolds number flow agree well with experimental data.     

一种利用等效模型与遗传算法的动态有限元模型修正方法

为了解决现有动态有限元模型修正方法计算效率不高或者可能获得局部最优解的问题,提出了一种利用等效模型和遗传算法的动态有限元模型修正新方法.首先,在设计参数的取值范围内,根据预设的多项式模型的阶次以及自变量的个数,利用试验设计方法获得拟合响应面模型所需要的最优样本点;通过有限元分析获得样本数据,并利用回归分析获得响应面模型,从而以响应面模型逼近结构特征与设计参数之间的函数关系.然后,在遗传算法的适应度评估环节,利用响应面模型替代有限元模型计算对应于一组设计参数的结构特征,并计算遗传个体的适应度,最终通过进化获得最优解,即为修正后的设计参数.以汽车车架模型为例,对其进行有限元分析与模态试验,并利用所提出的方法进行模型修正.修正后,模态频率误差的均方值小于2%.用修改后结构的动态特性的测试结果,对修正后有限元模型的预测能力进行检验,模态频率预测误差的均方值小于2%.     

Artificial neural network modeling of water quality of the Yangtze River system: a case study in reaches crossing the city of Chongqing

An effective approach for describing complicated water quality processes is very important for river water quality management. We built two artificial neural network （ANN） models, a feed-forward back-propagation （BP） model and a radial basis function （RBF） model, to simulate the water quality of the Yangtze and Jialing Rivers in reaches crossing the city of Chongqing, P. R. China. Our models used the historical monitoring data of biological oxygen demand, dissolved oxygen, ammonia, oil and volatile phenolic compounds. Comparison with the one-dimensional traditional water quality model suggest that both BP and RBF models are superior; their higher accuracy and better goodness-of-fit indicate that the ANN calculation of water quality agrees better with measurement. It is demonstrated that ANN modeling can be a tool for estimating the water quality of the Yangtze River. Of the two ANN models, the RBF model calculates with a smaller mean error, but a larger root mean square error. More effort to identify out the causes of these differences would help optimize the structures of neural network water-quality models.     

Application of BP NN and RBF NN in Modeling Activated Sludge System

Based on the operation data from a certain wastewater treatment plant(WWTP) in northeast China,the models of back propagation neural network ( BP NN ) and radial basis function neural network ( RBF NN ) have been designed respectively and the ability of convergence and generalization has been analyzed separately.As for BP NN, the effects of numbers of layers and nodes have been studied ; as for RBF NN, the influences of the number of nodes and the RBF‘s width have been studied. It is concluded that BP NN has converged much slowly in comparison with RBF NN. The conclusion that the RBF NN is suitable for modeling activated sludge system has been drawn. An automatically optimum design program for RBF NN has been developed, through which the RBF NN model of traditional activated sludge system has been established.     

基于核主成分降维的RBF网络降水预测

针对径向基函数神经网络（RBF网络）的隐层节点数、中心和宽度难以确定的问题,为提高网络性能,首先采用模糊聚类分析法对样本数据进行初始聚类,以初始分类间的最小距离作为初始宽度;然后引入正交最小二乘法训练出新的数据中心、个数及权值,修改宽度为当前数据中心间的最小距离;最后采用梯度下降法训练并调整中心、宽度及权值;几种算法进行的组合优化改进,使网络泛化性能更优.由于降水影响因子众多,采用了核主成分分析法（KPCA）对样本数据进行特征提取降维预处理.对广西5月3区的日降水量使用上述模型进行预报实验,结果表明,该模型具有较好的泛化性能,预报准确率高于T213降水预报模式,具有一定的推广价值.     

Generating ethnographic research questions: an anthropological contribution to the study of higher education

As part of recent complex transformations, it seems that higher educational organisations are being forced to reorganise, standardise and streamline in order to survive in the new political and economic context. How are ethnographers in general going to approach these contemporary phenomena? By drawing on the conceptual history of anthropology, the aim of this article is to generate ethnographic-oriented research questions concerned with higher education. The first part of the article provides an ethnographic background, while the second part focuses on Paul Willis's reasoning on ethnographic imagination, as a prerequisite for generating alternative research questions. The third part makes explicit anthropologist Maurice Godelier's theoretical imagination, carving out some specific theoretical parts which may be used in the generating process. The conclusion then suggests a number of questions to be asked by future ethnographers of higher education. The questions are followed by a reflection upon the consequences of doing ethnography within contemporary higher education settings, which are increasingly dominated by policy-makers; ethnography is thus to be seen as an intervening instrument.     

Study of CNG/diesel dual fuel engine's emissions by means of RBF neural network

Great efforts have been made to resolve the serious environmental pollution and inevitable declining of energy resources. A review of Chinese fuel reserves and engine technology showed that compressed natural gas (CNG)/diesel dual fuel engine (DFE) was one of the best solutions for the above problems at present. In order to study and improve the emission performance of CNG/diesel DFE, an emission model for DFE based on radial basis function (RBF) neural network was developed which was a black-box input-output training data model not require priori knowledge. The RBF centers and the connected weights could be selected automatically according to the distribution of the training data in input-output space and the given approximating error. Studies showed that the predicted results accorded well with the experimental data over a large range of operating conditions from low load to high load. The developed emissions model based on the RBF neural network could be used to successfully predict and optimize the emissions performance of DFE. And the effect of the DFE main performance parameters, such as rotation speed, load, pilot quantity and injection timing, were also predicted by means of this model. In resume, an emission prediction model for CNG/diesel DFE based on RBF neural network was built for analyzing the effect of the main performance parameters on the CO, NOx emissions of DFE. The predicted results agreed quite well with the traditional emissions model, which indicated that the model had certain application value, although it still has some limitations, because of its high dependence on the quantity of the experimental sample data.     

Beginning to Lecture at University: A Complex Web of Socialisation Patterns

This article reports on a longitudinal study which investigated the process of becoming and being a first-year junior lecturer in a South African linguistics department. The lecturer arrived as a recent doctoral graduate from England at the beginning of the academic year. Qualitative data collection and analysis revealed that there were a number of areas in which the lecturer experienced conflicts between his own expectations of the job and what he actually encountered: for example, teaching practice, departmental politics, linguistic paradigms, affirmative action. In this article, the lecturer's experiences of learning to lecture are examined in terms of how he found himself conforming, coping, generating and resisting. These socialisation patterns provide a framework for considering the early experiences of all beginning lecturers and are used, in this article, as the basis for an approach to induction for academic staff in universities. After eighteen months he left the department to lecture linguistics and then teach English as a foreign language in Japan.