Search

This paper is concerned with a multi-resolution tool for analysis of a time series aiming to detect abnormalities in various frequency regions. The task is treated as a kernel based novelty detection applied to a multi-level time series representation obtained from the discrete wavelet transform. Having a priori knowledge that the abnormalities manifest themselves in several frequency regions, a committee of detectors utilizing data dependent aggregation weights is build by combining outputs of detectors operating in those regions. [ABSTRACT FROM AUTHOR]

An all-optical neural network based on coupled ring lasers is proposed in this paper. Each laser in the network has a different wavelength, representing one neuron. The network status is determined by the wavelength of the network's light output. Inputs to the network are in the optical power domain. The nonlinear threshold function required for neural-network operation is achieved optically by interaction between the lasers. A simple laser model developed in the paper has illuminated the behavior of the coupled lasers. An experimental system is implemented using single mode fiber optic components at wavelengths near 1550 nm. A number of functions are implemented to demonstrate the practicality of the new network. From the experiment, a conclusion can be obtained that the neural network is particularly robust against input wavelength variations. [ABSTRACT FROM AUTHOR]

This paper proposes a reconfigurable architecture for VLSI implementation of BP neural networks with on-chip learning. Basing on systolic arrays, this architecture can flexibly adapt to neural networks with different scales, transfer functions or learning algorithms by reconfiguration of basic processing components,. Three kinds of reconfigurable processing units (RPU) are proposed firstly basing on the analysis of neural network's reconfiguration. Secondly, the paper proposes a reconfigurable systolic architecture and the method of mapping BP networks into this architecture. The implementation of an instance on FPGA is introduced in the last. The results show that this flexible architecture can also achieve a high learning speed of 432M CUPS (Connection Updated Per Second) at 100MHz using 22 multipliers. [ABSTRACT FROM AUTHOR]

In this paper there is a natural language interface presented, which consists of the intelligent mechanisms of human identification, speech recognition, word and command recognition, command syntax and result analysis, command safety assessment, technological process supervision as well as human reaction assessment. In this paper there is also a review of the selected issues on recognition of speech commands in natural language given by the operator of the technological device. A view is offered of the complexity of the recognition process of the operator's words and commands using neural networks made of a few layers of neurons. The paper presents research results of speech recognition and automatic recognition of commands in natural language using artificial neural networks. [ABSTRACT FROM AUTHOR]

New applications of artificial neural networks are capable of recognition and verification of effects and safety of commands given by the operator of the technological device. In this paper there is a review of the selected issues on estimation of results and safety of the operator's commands as well as supervision of the technological process. A view is offered of the complexity of effect analysis and safety assessment of commands given by the operator using neural networks. The first part of the paper introduces a new concept of modern supervising systems of the technological process using a natural language human-machine interface and discusses the general topics and issues. The second part is devoted to a discussion of more specific topics of the automatic command verification that have led to interesting new approaches and techniques. [ABSTRACT FROM AUTHOR]

It is well known, the Job-Shop Scheduling Problem (JSSP) is the most complicated and typical problem of all kinds of production scheduling problems, the allocation of resources over time to perform a collection of tasks. The current method has several shortcomings in solving the JSSP. In this paper, we correct these deficiencies by introducing a dynamic model that is based on an analysis of the run-time behavior of CSANN algorithm. At the same time, this paper proposes several new heuristics in order to improve the performance of CSANN. The computational simulations have shown that the proposed hybrid approach has good performance with respect to the quality of solution and the speed of computation. [ABSTRACT FROM AUTHOR]

As more and more industries experience the globalization of business activities, measuring productivity performance has become an area of concern for companies and policy makers in Europe, the United States, Japan and so on. A novel way about nonlinear regression modeling of high-technology manufacturing (HTM) productivity with the support vector machines (SVM) is presented in this paper. Optimization of labor productivity (LP) is also presented in this paper, which is based on chaos and uses the SVM regression model as the objective function. [ABSTRACT FROM AUTHOR]

In this paper posture control of a human arm in the sagittal plane is investigated by means of model simulations. The arm is modeled by a nonlinear neuromusculoskeletal model with two degrees of freedom and six muscles. A multilayer perceptron network is used in this paper, and effectively adapted by Levenberg-Marquardt training algorithm. The duration of next movement is regulated according as current feedback states. Simulation Results indicate that this method can maintain two joints at different location in allowable bound. The control scheme provides novel insight into neural prosthesis control and robotic control. [ABSTRACT FROM AUTHOR]

In this paper, we aim to address the cell phase identification problem, and two important aspects, the feature extraction methods and the classifier design, are discussed. In our study, we first propose extracting high frequency information of different orientations using Steerable filters. Next, we employ a multi-layer neural network using the back-propagation algorithm to replace K-Nearest Neighbor (KNN) classifier which has been implemented in the Cellular Image Quantitator (CELLIQ) system [3]. Experimental results provide a comparison between the proposed steerable filter features and existing regular features which have been used in published papers [3, 5]. From the comparison, it can be concluded that Steerable filter features can effectively represent the cells in different phases and improve the classification accuracy. Neural network also has a better performance than KNN currently deployed in CELLIQ system [3]. [ABSTRACT FROM AUTHOR]

Many interaction data still exist only in the biomedical literature and they require much effort to construct well-structured data. Discovering useful knowledge from large collections of papers is becoming more important for efficient biological and biomedical researches as genomic research advances. In this paper, we present a relation kernel-based interaction extraction method to extract knowledge efficiently. We extract protein interactions of from text documents with relation kernel and Yeast was used as an example target organism. Kernel for relation extraction is constructed with predefined interaction corpus and set of interaction patterns. The proposed method only exploits shallow parsed documents. Experimental results show that the proposed kernel method achieves a recall rate of 79.0% and precision rate of 80.8% for protein interaction extraction from biomedical document without full parsing efforts. [ABSTRACT FROM AUTHOR]

In self and non-self discrimination (SNSD) model, it is very important to generate a desirable detector set since it decides the performance and scale of the SNSD model based task. By using the famous principle of negative selection in natural immune system, a novel generating algorithm of detector, multiple-point bit mutation method, is proposed in this paper. It utilizes random multiple-point mutation to look for non-self detectors in a large range in the whole space of detectors, such that we can obtain a required detector set in a reasonable computation time. This paper describes the work procedure of the proposed detector generating algorithm. We tested the algorithm by using many datasets and compared it with the Exhaustive Detector Generating Algorithm in details. The experimental results show that the proposed algorithm outperforms the Exhaustive Detector Generating Algorithm both in computational complexities and detection performance. [ABSTRACT FROM AUTHOR]

This paper addresses the important issues of error control for video transmission over 3G, considering the fact that wireless video delivery faces the huge challenge of the high error rate and time variability in wireless channel. This paper proposes a real world statistics based event-trigger bit error rate (BER) model, which can describe and handle the time-varying wireless channel error characteristics better. Moreover, a recurrent neural network is employed to decide the state transfer as a mechanism. Simulation results and comparisons demonstrate effectiveness and efficiency of the proposed method in term of visual performance and transmission efficiency over a variety of wireless channel conditions. [ABSTRACT FROM AUTHOR]

The recommendation G.728 depends on the Levinson-Durbin (L-D) algorithm to update its gain filter coefficients. In this paper, it is contrasted with BP neural network method. Because quantizer has not existed at optimizing gain filter, the quantization SNR can not be used to evaluate its performance. This paper proposes a scheme to estimate SNR so that gain predictor can be separately optimized with quantizer. Using BP neural network filter, the calculation quantity is only 6.7 percent of L-D method's and its average segment SNR is about 0.156dB higher than G.728. It is also used to evaluate the case that excitation vector is 16 or 20 samples, respectively, the BP neural network algorithm has similarly good result. [ABSTRACT FROM AUTHOR]

Artificial Neural Networks (ANNs) which are derived from Biological Neural Networks (BNNs) are enhanced by many advanced mathematical techniques and have become powerful tools for solving complicated engineering problems. Integrating BNNs with mature ANNs is a very effective method of solving intricate biological problems and explaining neurophysiological data. In this paper we propose a neural network model that explains how the brain processes visual information about impending collisions with an object - in particular, how time-to-collision information is caculated in the brain. The model performs extremely well as a result of incorporating physiological data with the methods involved in the development of ANNs. By implementing this novel compuational neural network model, the results of the simulation demonstrate that this integrative approach is a very useful and efficient way to deal with complicated problems in neural computation. [ABSTRACT FROM AUTHOR]

Neural Networks are usually implemented in software on sequential machines but when implemented in hardware, they are extremely fast due to the massive parallelism inherent in the hardware devices. Implementation of Neural Networks in Programmable Logic Devices such as FPGAs (Field Programmable Gate Arrays) gives us more flexibility since these devices are reconfigurable and their design can be altered whenever needed. The design proposed in this paper shows the implementation of perceptron neural network in FPGAs for the character recognition problem. The characters here are the English language alphabets which are input to the network and after training; they are tested for recognition. Each alphabet is tested for three different fonts. After implementation, the simulations are done and performance issues of the design are analyzed.The post-layout simulation gives excellent results even if some noise is introduced to the input patterns. [ABSTRACT FROM AUTHOR]

In this paper, hybrid neural network processor (HANNP) is designed in VLSI. The HANNP has RISC based architecture leading to an effective general digital signal processing and artificial neural networks computation. The architecture of a HANNP including the general digital processing units such as 64-bit floating-point arithmetic unit (FPU), a control unit (CU) and neural network processing units such as artificial neural computing unit (NNPU), specialized neural data bus and interface unit, etc. The HANNP is modeled in Veilog HDL and implemented with FPGA. Character recognition problems and Kohonen self-organization problems are applied to the proposed HANNP to justify its applicability to real engineering problems. [ABSTRACT FROM AUTHOR]

In this paper we present the results of neural network hardware in-the-loop training for an analogue Local Cluster Neural Network (LCNN) chip. We use a Probabilistic Random Weight Change (PRWC) algorithm that is a combination of the random weight change and simulated annealing algorithms. We applied the PRWC algorithm to in-the-loop training for multi-dimensional function approximations and for predictions. We discuss the training strategy and the experiment results. [ABSTRACT FROM AUTHOR]

This paper proposes an efficient hardware architecture for a function generator suitable for an artificial neural network (ANN). A spline-based approximation function is designed that provides a good trade-off between accuracy and silicon area, whilst also being inherently scalable and adaptable for numerous activation functions. This has been achieved by using a minimax polynomial and through optimal placement of the approximating polynomials based on the results of a genetic algorithm. The approximation error of the proposed method compares favourably to all related research in this field. Efficient hardware multiplication circuitry is used in the implementation, which reduces the area overhead and increases the throughput. [ABSTRACT FROM AUTHOR]

A filtering method aided by neural network to improve the maneuvering target tracking performance is proposed in this paper. Based on unscented Kalman filter, the unscented particle filter (UPF) has more accurate proposal distribution and better approximation to non-linear tracking problem than other Sequential Monte-Carlo methods. The neural network is constructed and trained by the maneuvering features, and the outputs of NN are used as acceleration control parameters to correct model parameters. Simulation results show the performance of UPF aided by NN is much improved than extensive Kalman filter. [ABSTRACT FROM AUTHOR]

This paper presents the application of artificial neural network (ANN) in saturator. Phase Doppler Anemometry (PDA) is utilized to investigate the distribution of water droplets diameter and velocity in the saturator. The data obtained from experiment is used as input-output of ANN. Before using ANN method, some prerequisites have to be processed, including the selection of the number of input and output variables, hidden layer neurons, the network architecture and the normalization of data etc. The results indicate that the trained ANN can provide accurate prediction values which agree with real experimental data closely. [ABSTRACT FROM AUTHOR]

A new method for arbitrary 3d-object reconstruction in unknown environment is proposed in this paper. The implicit surface is reconstructed based on radial basis functions network from range scattered data. For the property of locality of radial basis function, the method is fast and robust with respect to large data. Furthermore, an adapted K-Means algorithm is used to reduce RBF centers for reconstruction. Experiment results show that the presented approach is helpful in speed improvement and is a good solution for large data reconstruction. [ABSTRACT FROM AUTHOR]

The issue of pre-splitting blasting design is investigated in this paper. According to the principle of expert system and neural networks, Visual C++ 6.0 and SQL Sever 2000 are employed to develop hybrid expert system of pre-splitting blasting design based on BP neural networks. The proposed expert system is a coupling system of engineering database and three-layered BP neural networks, which can be applied into cutting excavation of the expressway. The experiments show that the system can enhance the reliability of in-site pre-splitting blasting design scheme, efficiency and quality of construction. [ABSTRACT FROM AUTHOR]

Earthquakes will do great harms to the people, to estimate the future earthquake situation in Chinese mainland is still an open issue. There have been previous attempts to solve this problem by using artificial neural networks. In this paper, a novel algorithm named MIFEB is proposed to improve the estimation accuracy by combing bagging of neural networks with mutual information based feature selection for its individuals. MIFEB is compared with the general case of bagging on UCI data sets, then, MIFEB is used to forecast the seismicity of strong earthquakes in Chinese mainland, computation results show that MIFEB obtains higher accuracy than other several methods like bagging of neural networks and single neural networks do. [ABSTRACT FROM AUTHOR]

Electronic noses (E-nose) have gained popularity in various applications such as food inspection, cosmetics quality control [1], toxic vapor detection to counter terrorism, detection of Improvised Explosive Devices (IED), narcotics detection, etc. In the paper, we summarized our results on the application of Support Vector Machines (SVM) to gas detection and classification using E-nose. First, based on experimental data from Jet Propulsion Lab. (JPL), we created three different data sets based on different pre-processing techniques. Second, we used SVM to detect gas sample data from non-gas background data, and used three sensor selection methods to improve the detection rate. We were able to achieve 85% correct detection of gases. Third, SVM gas classifier was developed to classify 15 different single gases and mixtures. Different sensor selection methods were applied and FSS & BSS feature selection method yielded the best performance. [ABSTRACT FROM AUTHOR]

In this paper a novel optimization approach is presented. Network architecture and connection weights of neural networks (NN) are evolved by a particle swarm optimization (PSO) method, and then the appropriate network architecture and connection weights are fed into back-propagation (BP) networks. The ensemble strategy is carried out by simple averaging. The applied example is built with monthly mean rainfall of the whole area in Guangxi, China. The results show that the proposed approach can effectively improves convergence speed and generalization ability of NN. [ABSTRACT FROM AUTHOR]

Economic contribution rate of education is the key factor of education economy. In this paper, a soft computing method of economic contri-bution rate of education is proposed. The method is composed of four steps: The first step is doing fuzzy soft-clustering to object system based on levels of science technology and getting optimal number of clusters, which determines number of fuzzy rules. The second step is that the fuzzy neural networks FNN1 from human capital to economic growth is constructed and we obtain economic contribution rate of human capital αk. The third step is that the fuzzy neural networks FNN2 from education to human capital is constructed and we obtain human capital contribution rate of education $\alpha'_k$. The fourth step is calculating economic contribution rate of education $ECE_k = {\alpha_k} \times \alpha'_k$. At last, the economic contribution rate of education of China is obtained. [ABSTRACT FROM AUTHOR]

Conventional vibration signals processing techniques are most suitable for stationary processes. However, most mechanical faults in machinery reveal themselves through transient events in vibration signals. That is, the vibration generated by industrial machines always contains nonlinear and non-stationary signals. It is expected that a desired time-frequency analysis method should have good computation efficiency, and have good resolution in both time domain and frequency domain. In this paper, the auto-regressive model based pseudo-Wigner-Ville distribution for an integrated time-frequency signature extraction of the machine vibration is designed, the method offers the advantage of good localization of the vibration signal energy in the time-frequency domain. Kernel principal component analysis (KPCA) is used for the redundancy reduction and feature extraction in the time-frequency domain, and the self-organizing map (SOM) was employed to identify the faults of the rotating machinery. Experimental results show that the proposed method is very effective. [ABSTRACT FROM AUTHOR]

Reliable prediction of melt index (MI) is crucial in practical propylene polymerization processes. In this paper, a least squares support vector machines (LS-SVM) soft-sensor model is developed first to infer the MI of polypropylene from other process variables. A weighted least squares support vector machines (weighted LS-SVM) approach is further proposed to obtain rather robust estimate. Detailed comparative researches are carried out among standard SVM, LS-SVM, and weighted LS-SVM. The research results confirm the effectiveness of the presented methods. [ABSTRACT FROM AUTHOR]

This paper describes the design and implementation of a wood defect classifier. The defects are four different types of knots found in wood surfaces. Classification is based on features obtained from Gabor filters and supervised and non supervised artificial neural networks are used as classifiers. A Self-organizing neural network and a fuzzy Self-organizing neural network were designed as classifiers. The fuzzy SONN shows a reduction on the training time and had a better performance. A final classifier, a feedforward perceptron using the weights of the fuzzy SONN as initial weights turn to be the best classifier with a performance of 97.22% in training and 91.17% in testing. The perceptron classifier surpasses a human inspector task which has a maximum performance of 85%. [ABSTRACT FROM AUTHOR]

The principal object of this paper is to develop a neural network model, which can simulate the plasma enhanced chemical vapor deposition (PECVD) process in TFT-Array procedure. Then the Boolean logic rules are extracted from the trained neural network in order to establish a knowledge base of expert system. The input data of neural network was collected form the process parameters of PECVD machines in the TFT-Array department, included the flow rate of all gases, pressure and temperature of the chamber, etc. After checking, explaining and integrating the extraction rules into knowledge base, the rules can be the basics of membrane thickness prediction and alarm diagnosis in PECVD system. [ABSTRACT FROM AUTHOR]

For the on-line estimation of some directly immeasurable crucial variables in erythromycin fermentation process, this paper presents an Artificial Neural Network (ANN) left-inversion based on the "assumed inherent sensor" and its left-inversion concepts. The ANN left-inversion is composed of two relatively independent parts ( a static ANN used to approximate the complex nonlinear function and several differentiators used to represent its dynamic behaviors, so that the ANN left-inversion is a special kind of dynamic ANN in essence. Different from common dynamic ANNs, such a separate structure makes the ANN left-inversion easier to use, hence facilitating its application. The ANN left-inversion has been used to estimate such immeasurable variables as mycelia concentration, sugar concentration and chemical potency in erythromycin fermentation process. The experimental results show its validity. [ABSTRACT FROM AUTHOR]

Wafer yield estimation is a very complicated nonlinear problem due to many variations in fabrication processes at different silicon foundries. The purpose of this paper is to use Support Vector Machines (SVMs) to analyze and predict electrical test data, which are traditionally captured by probing each chip on the wafer. The predicted data produced by the support vector machines is then compared with the known measured data to determine the accuracy. Once the SVM has captured nonlinear relationship between fabrication processes and wafer yields, it can be used to predict wafer yield in other lots fabricated by the same silicon foundry. The advantage of using this approach is to save time due to probing hardware constraints, predict wafer yield across the same fabrication process and give an alternative method of device simulation. Our experiments show that the SVMs predict more accurate than classical device physics equations and in some cases SPICE simulation software in comparison with the actual measured electrical data. Electrical data used for this research include threshold voltages, saturation currents, and leakage currents. [ABSTRACT FROM AUTHOR]

The construction of models for prediction and control of initially unknown, potentially nonlinear system is a difficult, fundamental problem in machine learning and engineering control. Since the neural network as a tool for machine learning was introduced, significant progress has been made on data handling and learning algorithms. Currently, the most popular learning algorithm in neural network training is feed forward error back-propagation (FFEBP) algorithm. Aside from the success of the FFEBP algorithm, a polynomial neural networks (PNN) learning has been proposed as a new learning method. The PNN learning is a self-organizing process designed to determine an appropriate set of Ivakhnenko polynomials that allow the activation of many neurons to achieve a desired state of activation that mimics a given set of sampled patterns. These neurons are interconnected in such a way that the knowledge is stored in Ivakhnenko coefficients. In this paper, PNN model has been developed using the nonlinear reactive ion etching (RIE) experimental data utilizing Group Method of Data Handling (GMDH). To characterize the RIE process using PNN, a low-k dielectric polymer benzocyclobutene (BCB) is etched in an SF6 and O2 plasma in parallel plate system. Data from 24 factorial experimental design to characterize etch process variation with controllable input factors consisting of the two gas flows, RF power and chamber pressure are used to build PNN models of etch rate, uniformity, selectivity and anisotropy. The modeling and prediction performance of PNN is compared with those of FFEBP. The results show that the prediction capability of the PNN models is at least 16.9% better than that of the conventional neural network models. [ABSTRACT FROM AUTHOR]

In semiconductor manufacturing, complex and nonlinear fabrication processes are ubiquitous. Plasma processing such as plasma enhanced chemical vapor deposition (PECVD) and reactive ion etching (RIE) are workhorses in semiconductor fabrication, but also play as yield limiters due the nature of complexity of plasma process. In this paper, modular neural network (MNN) is applied for the purpose of plasma process modeling and characterization in the area of semiconductor manufacturing. MNN consists of a number of local expert networks (LENs) and one gating network. LENs compete using supervised learning to learn different regions of the data space under the supervision of gating network. Once proper MNNs for various responses of interest are established, response surfaces are generated to visually assist the characterization of the processes. As either an alternative or an augmentation to existing methods, this can provide more reliable and flexible flat form of process modeling and characterization in semiconductor manufacturing environment. [ABSTRACT FROM AUTHOR]

In this paper, an artificial intelligent system to predict the temperature distribution of the semiconductor chip with multi-heat sources is presented by integrating the back-propagation neural network (BNN) and the computational fluid dynamics (CFD) techniques. Six randomly generated coordinates of three power sections on the chip die are the inputs and sixty-four temperature monitoring points on the top of the chip die are the outputs. In the present methodology, one hundred sets of training data obtained from the CFD simulations results were sent to the BNN for the intelligent training. There are other sixteen generated input sets to be the test data and compared the results between CFD simulation and BNN, it shows that the BNN model is able to accurately estimate the corresponding temperature distribution as well as the maximum temperature values under different power distribution after well trained. [ABSTRACT FROM AUTHOR]

The Conventional methods of color separation in computer-based machine vision offer only weak performance because of environmental factors such as light source, camera sensitivity, and others. In this paper, we propose an improved color separation method using fuzzy membership for feature implementation and a neural network for feature classification. In addition, we choose HLS color coordination. The HLS includes hue, light, and saturation. There are the most human-like color recognition elements. A proposed color recognition algorithm is applied to a line order detection system of harness. The detection system was designed and implemented as a testbed to evaluate the physical performance. The proposed color separation algorithm is tested with different kinds of harness line. [ABSTRACT FROM AUTHOR]

For design of micro spring based vertical type probe card, accurate micro spring tension force modeling is essential to guarantee the probe testing performance and reliability. In this paper, neural network based micro spring model was developed to find optimal spring height and shift value for appropriate tension force. Modeling results are applied to design and fabrication of vertical type probe card using 80(m and 100(m tungsten wires for micro spring type probing on silicon substrate. Compare to conventional statistical modeling scheme, neural network based model showed superior modeling accuracy with limited a priori information. Proposed high pad density probe card can be applied to high-density multi-die testing as well as advanced bumping type chip test. [ABSTRACT FROM AUTHOR]

For on-line batch process monitoring, multiway principal component analysis (MPCA) is a useful tool. But the MPCA-based methods suffer two disadvantages: (i) it restricts itself to a linear setting, where high-order statistical information is discarded; (ii) all the measurement variables must follow Gaussian distribution and the objective of MPCA is only to decorrelate variables, but not to make them independent. To improve the ability of batch process monitoring, this paper proposes a monitoring method named multiway kernel independent component analysis (MKICA). By using kernel trick, the new monitoring indices are investigated, which have been mapped into high-dimensional feature space. On the benchmark simulator of fed-batch penicillin production, the presented method has been validated. [ABSTRACT FROM AUTHOR]

Conventional PCA-based monitoring method relies on the assumption that process data is normally distributed, which the actual industrial processes often don't satisfy. Instead, mixture probabilistic principal component analysis (MPPCA) models are suitable to process with any probability density function. But, it suffers a drawback that the needed charts are too many to be watched in practice while the number of sub-models in MPPCA is large. Different from existing MPPCA, this paper proposes a novel method, which integrates every monitoring chart of MPPCA models into only one chart via probability and field process monitoring can rely on just one chart. The application in real chemical separation process shows validity of the proposed method. [ABSTRACT FROM AUTHOR]

As a valid statistical tool, principal component analysis (PCA) has been widely used in industrial process monitoring. But due to its intrinsic linear character, it performs badly in nonlinear process monitoring. Kernel PCA (KPCA) can extract useful information in nonlinear data. However KPCA-based monitoring is not suitable for on-line monitoring because of large calculation and much memory occupation. The paper introduces an on-line monitoring method based on KPCA and neural network (NN), where KPCA is used to extract nonlinear principal components (PCs) and then NN approximates the relationship between process data and nonlinear PCs. We can obtain nonlinear PCs by NN to compute the monitoring indices and then achieve the on-line monitoring. The case study shows the validity of the method. [ABSTRACT FROM AUTHOR]

Scheduling of re-entrant lines is very important for manufacturing systems. For some dynamic scheduling methodologies, it is necessary to model a production system with finite-state discrete-time Markov process. However, proper state cannot be found as absorbing state of Markov process when general Mean Output Rate is employed as an evaluating criterion. Mean-Output-parts Number Before First Block is presented to be a new evaluating criterion in this paper to evaluate scheduling policies for Closed Re-entrant Lines(CRL). Simulations of four static scheduling policies verify the new criterion. In order to apply a Neuro-Dynamic Programming (NDP) method to scheduling of a CRL, cost-to-go value function and transition cost function are presented as new forms under the new criterion. In addition, the policy obtained in a less-number parts system by the NDP is applied in a more-number parts system directly, whose results are satisfactory. [ABSTRACT FROM AUTHOR]

Based on the features of production plan and control in core enterprise under supply chain, according to extension matter-element theory, the paper presents an automatic on-line measuring method of distributed production plan track using the multi-sensor and a new extension measurement method which has two BP networks for realizing the compensation of the measurable matter-element and matter-element transform respectively and uses the D-S Evidence theory for matter-element focus to realize the machining accuracy data of a working procedure that has finished justly and the faults of the machine tools and cutting-tools etc. By the data and the faults, it will reschedule Job-Shop production plan, so as to realize the right time to finish the production plan and to supply data guarantee for the production plan and control in core enterprise under supply chain. The result presented shows that the method is feasible and efficient. [ABSTRACT FROM AUTHOR]

This paper proposes the application of Meta-Learning Evolutionary Artificial Neural Network (MLEANN) in selecting flexible manufacturing systems (FMS) from a group of candidate FMS's. First, multi-criteria decisionmaking (MCDM) methodology using an improved S-shaped membership function has been developed for finding out the ‘best candidate FMS alternative' from a set of candidate-FMSs. The MCDM model trade-offs among various parameters, namely, design parameters, economic considerations, etc., affecting the FMS selection process in multi-criteria decision-making environment. Genetic algorithm is used to evolve the architecture and weights of the proposed neural network method. Further, a back-propagation (BP) algorithm is used as the local search algorithm. The selection of FMS is made according to the error output of the results found from the MCDM model. [ABSTRACT FROM AUTHOR]

In this paper, a novel multiple Neural Network (NN) models including forecasting model, presetting model, adjusting model and judgment model for Basic Oxygen Furnace (BOF) steelmaking dynamic process is introduced. The control system is composed of the preset model of the dynamic requirement for oxygen blowing and coolant adding, bath [C] and temperature prediction model, and judgment model for blowing-stop. In this method, NN technology is used to construct these models above; Fuzzy Inference (FI) is adopted to derive the control law. The control method of BOF steelmaking process has been successfully applied in some steelmaking plants to improve the bath Hit Ratio (HR) significantly. [ABSTRACT FROM AUTHOR]

This paper presents two optimization procedures-single and multi objective optimization for 1370mm tandem cold rolling schedules, in which back propagation (BP) neural network is adopted to predict the rolling force instead of traditional models. Analysis and comparison with existing schedules are offered. The results show that the proposed schedules are more promising. [ABSTRACT FROM AUTHOR]

Performance of controlled laminar cooling is usually poor because of the difficulty in continuous online temperature measurement and the complex nature of the laminar cooling process (e.g., highly nonlinear, time varying). This paper developed a hybrid control strategy for the laminar cooling process that integrates Radial Basis Function (RBF) networks and Case-Based Reasoning (CBR). The spraying pattern and the first activated headers are found by a case-based reasoner, while the number of activated headers is calculated in real time by RBF networks. Experimental studies using production data from a hot strip mill show the superior performance of the proposed control strategy. [ABSTRACT FROM AUTHOR]

Due to the complex nature (e.g., highly nonlinear, time varying, and spatially varying) of the laminar cooling process, accurate mathematical modeling of the process is difficult. This paper developed a hybrid model of the laminar cooling process by integrating Radial Basis Function (RBF) networks into the first principles dynamical model. The heat transfer coefficients of water cooling in the dynamical model were found by RBF networks. The developed model is capable of predicting the through-thickness temperature evolutions of the moving strip during the laminar cooling process. Experimental studies using real data from a hot strip mill show the superiority of the proposed model. [ABSTRACT FROM AUTHOR]

Via two-stage back-propagation neural network (BNN) learning algorithm, this paper establishes the relationship between different heatsink design parameters and performance evaluation, and induces 5 corresponding performance outputs from 6 different heatsink design and operating condition parameters (inlet airflow velocity, heatsink length or width, fin thickness, fin gap, fin height and heatsink base height) by using Computation Fluid Dynamics (CFD). After two stages well-trained, the BNN model with error compensator is able to accurate estimate the output values under different heatsink design and operation conditions. [ABSTRACT FROM AUTHOR]

This paper presents a number of neural network (NN) models for examining how a given product form affects product images perceived by customers. An experimental study on mobile phones is conducted. The concept of consumer oriented design is used to extract the experimental samples as a design database for the numerical analysis. The result of the experiment demonstrates the advantages of using NN models for the product form design. NN models can help product designers understand consumers' perception and translate consumers' feeling of a product into design elements. [ABSTRACT FROM AUTHOR]

Differentiation of syndromes is the kernel theory of Traditional Chinese Medicine (TCM). How to diagnose syndromes correctly with scientific means according to symptoms is the first problem in TCM. Several modern approaches have been applied, but no satisfied results have been obtained because of the complexity of diagnosis procedure. Support Vector Machine (SVM) is a new classification technique and has drawn much attention on this topic in recent years. In this paper, we combine non-linear Principle Component Analysis (PCA) neural network with multi-class SVM to realize differentiation of syndromes. Non-linear PCA is used to preprocess clinical data to save computational cost and reduce noise. The multi-class SVM takes the non-linear principle components as its inputs and determines a corresponding syndrome. Analyzing of a TCM example shows its effectiveness. [ABSTRACT FROM AUTHOR]