Showing papers on "Hybrid neural network published in 2004"

Hybrid neural network—prior knowledge model in temperature control of a semi-batch polymerization process

Abstract: Nonlinear process control is a challenging research topic at present. In recent years, neural network and hybrid neural networks have been much studied especially for modeling of nonlinear system. It has however been applied mainly as an estimator in parts of various control systems and the idea of utilizing it directly as a neural-controller has not been studied. Hence the contribution of this work is to use an inverse neural network in hybrid with a first principle model for the direct control of a nonlinear semi-batch polymerization process. These hybrid models were utilized in the direct inverse control strategy to track the set point of the temperature of the polymerization reactor under nominal condition and with various disturbances. For comparison purposes, the standard neural network and proportional-integral-derivative controller were also implemented in these control strategies. Adaptation mechanisms to improve the results have also been carried out to test the capability of these hybrid methods in control. The simulation results show the advantages and robustness of utilizing the neural network in this hybrid strategy especially when an adaptive algorithm is implemented.

Neural Networks for Hydrological Modeling

Abstract: 1. Why Use Neural Networks? Pauline E.Kneale, Linda M. See & Robert J.Abrahart 2. Neural Network Modelling: Basic Tools and Broader Issues Robert J.Abrahart 3. Single Network Modelling Solutions Christian W.Dawson & Robert L.Wilby 4. Hybrid Neural Network Modelling Solutions Asaad Y.Shamseldin 5. The Application of Time Delay Neural Networks to River Level Forecasting Linda M.See & Pauline E.Kneale 6. The Application of Cascade Correlation Neural Networks to River Flow Forecasting Claire E.Imrie 7. The Use of Partial Recurrent Neural Networks for Autoregressive Modelling of Dynamic Hydrological Systems Henk F.P.van den Boogaard 8. RLF1/ Flood Forecasting via the Internet Simon A.Corne & Stan Openshaw 9. Rainfall-Runoff Modelling Anthony W.Minns & Michael J.Hall 10. A Neural Network Approach to Rainfall Forecasting in Urban Environments James E.Ball & Kin Choi Luk 11. Water Quality and Ecological Management in Freshwaters Pauline E.Kneale 12. Neural Network Modelling of Sediment Supply and Transfer Susan M.White 13. Nowcasting products from Meteorological Satellite Imagery George S.Pankiewicz 14. Mapping Land Cover from Remotely Sensed Imagery for Input to Hydrological Models Giles M.Foody 15. Towards a Hydrological Research Agenda Robert J.Abrahart, Pauline E.Kneale & Linda M.See Index

Hybrid identification of nuclear power plant transients with artificial neural networks

Abstract: Proper and rapid identification of malfunctions (transients) is of premier importance for the safe operation of nuclear power plants. Feedforward neural networks trained with the backpropagation (BP) algorithm are frequently applied to model simulated nuclear power plant malfunctions. The correct identification of unlabeled transients-or transients of the "don't-know" type have proven to be especially challenging. A novel hybrid neural network methodology is presented which also correctly classifies the unlabeled transients. From this analysis the importance for properly accommodating practical aspects such as the drift of electronics elements of a simulator, the digitization of simulated and actual plant signals, and the accumulating errors during numerical integration became obvious. Beside the feedforward neural networks trained with the BP algorithm, many other types of networks and codes were used for finding the best (sensitive and robust) algorithms. Various neural network based models were successfully applied to identify labeled and unlabeled malfunctions of the Hungarian Paks nuclear power plant simulator. The BP and probabilistic methods have been proven as the most robust against the misleading recognition of unlabeled malfunctions.

A hybrid neural network model for noisy data regression

Abstract: A hybrid neural network model, based on the fusion of fuzzy adaptive resonance theory (FA ART) and the general regression neural network (GRNN), is proposed in this paper. Both FA and the GRNN are incremental learning systems and are very fast in network training. The proposed hybrid model, denoted as GRNNFA, is able to retain these advantages and, at the same time, to reduce the computational requirements in calculating and storing information of the kernels. A clustering version of the GRNN is designed with data compression by FA for noise removal. An adaptive gradient-based kernel width optimization algorithm has also been devised. Convergence of the gradient descent algorithm can be accelerated by the geometric incremental growth of the updating factor. A series of experiments with four benchmark datasets have been conducted to assess and compare effectiveness of GRNNFA with other approaches. The GRNNFA model is also employed in a novel application task for predicting the evacuation time of patrons at typical karaoke centers in Hong Kong in the event of fire. The results positively demonstrate the applicability of GRNNFA in noisy data regression problems.

Improving option pricing with the product constrained hybrid neural network

Abstract: In the past decade, many studies across various financial markets have shown conventional option pricing models to be inaccurate. To improve their accuracy, various researchers have turned to artificial neural networks (ANNs). In this work a neural network is constrained in such a way that pricing must be rational at the option-pricing boundaries. The constraints serve to change the regression surface of the ANN so that option pricing accuracy is improved in the locale of the boundaries. These constraints lead to statistically and economically significant out-performance, relative to both the most accurate conventional and nonconventional option pricing models.

Hybrid neural network and genetic algorithm based machining feature recognition

Abstract: In this research, neural networks (NNs) and genetic algorithms (GAs) are used together in a hybrid approach to reduce the computational complexity of feature recognition problem. The proposed approach combines the characteristics of evolutionary technique and NN to overcome the shortcomings of feature recognition problem. Consideration is given to reduce the computational complexity of network with specific interest to design the optimum network architecture using GA input selection approach. In order to evaluate the performance of the proposed system, experimental results are compared with previous NN based feature recognition research.

A hybrid neural network/genetic algorithm approach to optimizing feature extraction for signal classification

Abstract: In this paper, a hybrid neural network/genetic algorithm technique is presented, aiming at designing a feature extractor that leads to highly separable classes in the feature space. The application upon which the system is built, is the identification of the state of human peripheral vascular tissue (i.e., normal, fibrous and calcified). The system is further tested on the classification of spectra measured from the cell nuclei in blood samples in order to distinguish normal cells from those affected by Acute Lymphoblastic Leukemia. As advantages of the proposed technique we may encounter the algorithmic nature of the design procedure, the optimized classification results and the fact that the system performance is less dependent on the classifier type to be used.

A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model

Abstract: In this research, a new hybrid neural network is developed to model engagement behaviors of automotive transmission wet friction component. Utilizing known first principles on the physics of engagement, special modules are created to estimate viscous torque and asperity contact torque as preprocessors to a two-layer neural network. Inside these modules, all the physical parameters are represented by neurons with various activation functions derived from first principles. These new features contribute to the improved performance and trainability over a conventional two-layer network model. Both the hybrid and conventional neural net models are trained and tested with experimental data collected from an SAE#2 test stand. The results show that the performance of the hybrid model is much superior to that of the conventional model. It successfully captures detailed characteristics of the friction component engagement torque as a function of time over a wide operating range.

New approaches for license plate recognition system

Abstract: In this paper, we describe two new approaches for segmenting and recognizing license plate. The first method exploits the fact that edges are most densely found in the region that contains the license plate. We have proposed a weight allocation scheme which segments the region with the most dense edges. The second method uses a hybrid neural network which we propose. The efficiency of the license plate recognition systems using the proposed methods have been studied.

Seafloor classification using echo-waveforms: a method employing hybrid neural network architecture

Abstract: This letter presents seafloor classification study results of a hybrid artificial neural network architecture known as learning vector quantization. Single beam echo-sounding backscatter waveform data from three different seafloors of the western continental shelf of India are utilized. In this letter, an analysis is presented to establish the hybrid network as an efficient alternative for real-time seafloor classification of the acoustic backscatter data.

License plate recognition system using hybrid neural networks

Abstract: License plate recognition is one of the machine vision problem which has a wide range of practical applications. Various methodologies have been suggested with varying efficiencies. In this paper we propose an ASIC implementation of a new methodology which uses a hybrid neural network in conjunction with the weight based density map technique for the license plate localization. The combination of the two techniques, that have already been used individually, has produced higher recognition rates. Further the ASIC implementation means that the time complexity of the methodology can be completely ignored. The system design and the experimental results of the system have been discussed briefly.

A Hybrid Intelligent Active Force Controller for Articulated Robot Arms Using Dynamic Structure Neural Network

Abstract: The key feature of this paper is the application of a robotic control concept – Active Force Control (AFC) In this type of control, the unknown friction effect of the robotic arm may be compensated by the AFC method AFC involves the direct measurement of the acceleration and force quantities and therefore, the process of estimating the system ‘disturbance’ due to friction becomes instantaneous and purely algebraic However, the AFC strategy is very practical provided a good estimation of the inertia matrix of articulated robot arm is acquired A dynamic structure neural network – Growing Multi-experts Network (GMN) is developed to estimate the robot inertia matrix The growing and pruning mechanism of GMN ensures the optimum size of the network that results in an excellent generalization capability of the network Active Force Control (AFC) in conjunction with GMN successfully reduces the velocity and position tracking errors in spite of robot joint friction The embedded GMN is capable of coupling the inertia matrix estimation on-line that clearly enhances the performance of AFC controller The robustness and effectiveness of the new hybrid neural network-based AFC scheme are demonstrated clearly with regard to two link articulated robot and a simulated two-degree of freedom Puma 560 robot

Advanced hybrid neural network automotive friction component model for powertrain system dynamic analysis. Part 1: Model development

Abstract: An accurate and easy-to-implement dynamic friction component model is necessary for powertrain system design and performance studies. A neural network approach developed by Cao et al. [1] for friction component modelling has illustrated some very promising results. However, this model has complex architecture that may lead to reduced training efficiency; also, owing to the lack of time information, the network cannot adapt to time step variations. Therefore, it cannot be easily integrated with powertrain system models, which in general require variable time steps for superior numerical integration performance. In this paper, a new first-principle-based hybrid network friction component model, the advanced hybrid neural network (AHNN), is derived for dynamic engagement analysis with variable time steps. With improvement over the previous work by Cao et al. [1], the time pattern information is added to the inputs and a simpler architecture is developed through more explicit utilization of the physic...

Innovations in intelligent systems

Abstract: Use of multi-category proximal SVM for data set reduction.- Bayesian control of dynamic systems.- AppART: a hybrid neural network based on adaptive resonance theory for universal function approximation.- An algorithmic approach to the main concepts of rough set theory.- Automated case selection from databases using similarity-based rough approximation.- An induction algorithm with selection significance based on a fuzzy derivative.- Model and fixpoint semantics for fuzzy disjunctive programs with weak similarity.- An automated report generation tool for the data understanding phase.- Finding trigonometric identities with tree adjunct grammar guided genetic Modeling a distributed knowledge management for autonomous cooperative agents with knowledge migration programming.- Intelligent information systems based on paraconsistent logic programs.- Neuro-fuzzy paradigms for intelligent energy management.- Information space optimization for inductive learning.- Detecting, tracking, and classifying human movement using active contour Fuzzy sets in investigation of human cognition processes models and neural networks.- A full explanation facility for an MLP network that classifies .- Automatic translation to controlled medical vocabularies low-back-pain patients and for predicting MLP reliability.- A genetic programming for the induction of natural language parser.

Dynamic On‐Line Reoptimization Control of a Batch MMA Polymerization Reactor Using Hybrid Neural Network Models

Abstract: A hybrid neural network model based on-line reoptimization control strategy is developed for a batch polymerization reactor. To address the difficulties in batch polymerization reactor modeling, the hybrid neural network model contains a simplified mechanistic model covering material balance assuming perfect temperature control, and recurrent neural networks modeling the residuals of the simplified mechanistic model due to imperfect temperature control. This hybrid neural network model is used to calculate the optimal control policy. A difficulty in the optimal control of batch polymerization reactors is that the optimization effort can be seriously hampered by unknown disturbances such as reactive impurities and reactor fouling. With the presence of an unknown amount of reactive impurities, the off-line calculated optimal control profile will be no longer optimal. To address this issue, a strategy combining on-line reactive impurity estimation and on-line reoptimization is proposed in this paper. The amount of reactive impurities is estimated on-line during the early stage of a batch by using a neural network based inverse model. Based on the estimated amount of reactive impurities, on-line reoptimization is then applied to calculate the optimal reactor temperature profile for the remaining time period of the batch reactor operation. This approach is illustrated on the optimization control of a simulated batch methyl methacrylate polymerization process.

Vehicle detection with projection histogram and type recognition using hybrid neural networks

Abstract: Most vehicle head faces have one window, two illuminative lamps and a license plate, a method is proposed using features about them to detect vehicle head face. Edges of window are straight lines commonly, in the projection histogram whose direction is not parallel to window edge, there is an abrupt change at the edge location. This property can be used to locate window candidate position. If lamps and license plate have all been located, the candidate is verified surely and truly. In order to recognize vehicle type, a hybrid neural network is designed to analyze the extracted features. The network contains two parts, the 1/sup st/ part uses support vector machines, the 2/sup nd/ part contains four perceptrons which verifies the output of 1/sup st/ and produces recognized result. Synthetic experiment of this method is given and some discussions have also been made.

Advanced hybrid neural network automotive friction component model for powertrain system dynamic analysis. Part 2: System simulation

Abstract: In this research, the advanced hybrid neural network (AHNN) friction-component model, presented in Part 1 of this two-part paper, is integrated with an automotive drivetrain model for system simulations. The AHNN model accurately predicts the dynamic behaviours of transmission friction components over a broad operating range. It also allows variable sampling time steps in a numerical integration process. In this investigation, the AHNN model is trained using experimental data obtained from a powertrain dynamometer test stand. Since typical dynamometer measurements are acquired at locations away from friction components, a backtracking algorithm is developed to evaluate friction component torque during engagement. The trained AHNN model, together with a comprehensive drivetrain model, is implemented to simulate the shifting process of an automatic transmission system under various operating conditions, including different oil-temperature and engine-throttle levels. Simulation results demonstrate that the AHNN friction component model can be effectively utilized as a part of the drivetrain model to accurately predict transmission shift dynamics.

A hybrid neural network/rule based system for bilingual text-to-phoneme mapping

Abstract: Text-to-phoneme (TTP) mapping is a preliminary step in text-to-speech synthesis and it affects the naturalness and understandability of synthetic speech In this paper, we propose a hybrid neural network/rule based system for bilingual text-to-phoneme mapping Our system uses three neural networks and a simple rule to perform the phoneme transcription The first network is trained to convert the letters from the first language into their corresponding phonemes, the second one is used to obtain the phonemes for the second language whereas the third neural network together with a simple rule is responsible of the language recognition The proposed approach can be easily extended for multilingual applications when more neural networks are introduced Simulations performed on a bilingual dictionary (English+French) show the improvements in terms of phoneme accuracy of our method against the approach that uses a single neural network for multilingual TTP

A hybrid neural network for input that is both categorical and quantitative

Abstract: The data on which a MLP (multilayer perceptron) is normally trained to approximate a continuous function may include inputs that are categorical in addition to the numeric or quantitative inputs. Examples of categorical variables are gender, race, and so on. An approach examined in this article is to train a hybrid network consisting of a MLP and an encoder with multiple output units; that is, a separate output unit for each of the various combinations of values of the categorical variables. Input to the feed forward subnetwork of the hybrid network is then restricted to truly numerical quantities. A MLP with connection matrices that multiply input values and sigmoid functions that further transform values represents a continuous mapping in all input variables. A MLP therefore requires that all inputs correspond to numeric, continuously valued variables and represents a continuous function in all input variables. A categorical variable, on the other hand, produces a discontinuous relationship between an input variable and the output. The way that this problem is often dealt with is to replace the categorical values by numeric ones and treat them as if they were continuously valued. However there is no meaningful correspondence between the continuous quantities generated this way and the original categorical values. The basic difficulty with using these variables is that they define a metric for the categories that may not be reasonable. This suggests that the categorical inputs should be segregated from the continuous inputs as explained above. Results show that the method utilizing a hybrid network and separating numerical from quantitative input, as discussed here, is quite effective. © 2004 Wiley Periodicals, Inc. Int J Int Syst 19: 979–1001, 2004.

A Hybrid Neural Network and Genetic Algorithm Approach for Multicast QoS Routing

Abstract: Computing the Multicast QoS routing is an NP-complete problem. Generally, it was solved by heuristic algorithms, which include tabu search, simulated annealing, genetic algorithms (GA), neural networks (NN), etc. In this paper, a hybrid neural network and genetic algorithm approach is described to compute the multicast QoS routing tree. The integration of neural network and genetic algorithm can overcome the premature and increase the convergence speed. The simulation results show that the proposed approach outperforms the traditional GA and NN algorithm in terms of both solution accuracy and convergence speed.

Short-term traffic flow prediction methods comparison of road networks

Abstract: A large number of techniques have been applied into short-term traffic flow prediction, which can be classified into two groups: statistical models and artificial neural network model The models and their application were discussed and compared Several models, including historical average, ARIMA(auto regressive integrated moving average) model, nonparametric regression, RBF(radial basis function) neural network and Bayesian combined neural network model were applied into a numerical example of short-term traffic volume prediction in a field network, their prediction results and performances were compared It was found that the error of hybrid neural network model is littlest, its prediction reliability is highest, it is the most effective method to predicte short-term traffic flow 2 tabs, 2 figs, 17 refs

A fully automated offline handwriting recognition system incorporating rule based neural network validated segmentation and hybrid neural network classifier

Abstract: In this paper we propose a fully automated offline handwriting recognition system that incorporates rule based segmentation, contour based feature extraction, neural network validation, a hybrid neural network classifier and a hamming neural network lexicon. The work is based on our earlier promising results in this area using heuristic segmentation and contour based feature extraction. The segmentation is done using many heuristic based set of rules in an iterative manner and finally followed by a neural network validation system. The extraction of feature is performed using both contour and structure based feature extraction algorithm. The classification is performed by a hybrid neural network that incorporates a hybrid combination of evolutionary algorithm and matrix based solution method. Finally a hamming neural network is used as a lexicon. A benchmark dataset from CEDAR has been used for training and testing.

Modeling and Optimization of a Styrene Monomer Reactor System Using a Hybrid Neural Network Model

Abstract: The operation of a current styrene monomer plant requires large quantities of energy to heat and cool its processing streams. Especially, operating a styrene monomer reactor system under proper conditions is very important because this reactor system occupies a large portion of the total operating cost using a large amount of expensive high-pressure steam. The optimization of the operating conditions of this reactor system can therefore be used to significantly reduce the total cost. To predict the dehydrogenation reactor conditions and take into account the effects of catalyst deactivation, we propose in this paper an alternative hybrid model of the reactor that is composed of a mathematical model and a neural network model. The mathematical model is a first principle model that predicts the compositions and the temperature and pressure profiles from the reaction mechanism and reactor geometry. The catalyst deactivation factor used in the mathematical model is calculated with the neural network model. Actual plant data were used in this study to test the hybrid model. Using this reactor model, we were able to solve the optimization problem for this plant. The objective of the optimization was to maximize the performance of the dehydrogenation reactor. A trajectory optimization method is proposed in this study that reduces the calculation required for the optimization. In this method, the trajectory of each operating variable is optimized while the other operating variables are held constant at their average values. Empirical equations are then obtained from the optimal trajectories, and the parameters of the empirical equations for all operation trajectories are optimized simultaneously. We found that the optimal profit was greater than that currently obtained by the plant.

The analysis of the effects of surface roughness of shafts on journal bearings using recurrent hybrid neural network

Abstract: This paper presents an investigation for analysing the load carrying capacity of journal bearing in a variety of conditions using a proposed neural network (NN). The NN structure is very suitable for this kind of system. The network is capable of predicting the pressures of the experimental system. The network has parallel structure and fast learning capacity. It can be outlined from the results for both approaches, NN could be used to model journal bearing systems in real time applications.

An application of adaptive genetic-neural algorithm to Sinter's BTP process

Abstract: This work presents adaptive genetic-neural network for sinter's burning through point (BTP), since BTP control is most important, which is tightly coupled with sinter ore quality. In off-line, the adaptive genetic algorithm (AGA) is used to optimize the connection weights and thresholds, and during on-line hybrid neural network (HNN) inherited from the principle of back propagation is used to train the map parameters and improve the system precision in each sampling period. The results obtained from the actual process demonstrate that the performance and capability of the proposed system are superior.

A new hybrid neural architecture (MLP+RPE) for hetero association: multi layer perceptron and coupled recursive processing elements neural networks

Abstract: This paper addresses a new hybrid neural network, which joins the classical multi layer perceptron (MLP) with a neural network composed of coupled recursive processing elements (RPEs). The individual characteristics of each one of these architectures, once combined, permitted the implementation of input-output mappings where the input patterns can be either discrete or analog and the output patterns can be discrete. Experiments for the performance evaluation of this hybrid neural architecture employing nodes that exhibit bifurcation and chaotic dynamics are described and the results addressing the operation under analog noise added to the input patterns are presented and analyzed. The performance of the MLP+RPE architecture was contrasted with the performance of other hybrid arrangements that have the same functionality (arrangements joining the MLP and Hopfield architectures), and the obtained results indicate that the MLP+RPE architecture presents significant superiority.

A building of the genetic-neural network for sinter's burning through point

Abstract: This paper presents the genetic-neural network for sinter's burning through point since BTP control is the most important, which is tightly coupled with sinter ore quality. In offline, advanced genetic algorithm (GA) is used to optimize the original connection weights and thresholds, and during online, hybrid neural network (HNN) inherited from the principle of backpropagation is used to train the map parameters and improve the system precision in each sampling period. The results obtained from the actual process demonstrate that the performance and capability of the proposed system are superior.

Hybrid BP/CNN neural network for intrusion detection

Abstract: In order to improve the intrusion detection rates and reduce false positives, a hybrid BP/CNN neural network is constructed, which has both the capability of real-time classification which BP has and the functionality of time-delay, collection and judgment which chaotic neuron has. Because this intrusion detection approach has flexible time-delay characteristic, it corresponds to the requirement of intrusion detection nowadays. The simulation tests to FTP brute-force attacks are conducted using samples captured from data traffic in local computer network. The test results are drawn by ROC curves. The intrusion criterion with high rate intrusion detection and low rates of false alarms can be found. The intrusion detection approach in this paper may be generalized to other intrusion detection systems.

A hybrid neural network model for the dynamics of the Kuramoto-Sivashinsky equation

Abstract: A hybrid approach consisting of two neural networks is used to model the oscillatory dynamical behavior of the Kuramoto-Sivashinsky (KS) equation at a bifurcation parameter α=84.25. This oscillatory behavior results from a fixed point that occurs at α=72 having a shape of two-humped curve that becomes unstable and undergoes a Hopf bifurcation at α=83.75. First, Karhunen-Loeve (KL) decomposition was used to extract five coherent structures of the oscillatory behavior capturing almost 100% of the energy. Based on the five coherent structures, a system offive ordinary differential equations (ODEs) whose dynamics is similar to the original dynamics of the KS equation was derived via KL Galerkin projection. Then, an autoassociative neural network was utilized on the amplitudes of the ODEs system with the task of reducing the dimension of the dynamical behavior to its intrinsic dimension, and a feedforward neural network was usedto model the dynamics at a future time. We show that by combining KL decomposition and neural networks, a reduced dynamical model of the KS equation is obtained.