291 results about "Penalty factor" patented technology

An MPC adaptation and tuning technique integrates feedback control performance better than methods commonly used today in MPC type controllers, resulting in an MPC adaptation / tuning technique that performs better than traditional MPC techniques in the presence of process model mismatch. The MPC controller performance is enhanced by adding a controller adaptation / tuning unit to an MPC controller, which adaptation / tuning unit implements an optimization routine to determine the best or most optimal set of controller design and / or tuning parameters to use within the MPC controller during on-line process control in the presence of a specific amount of model mismatch or a range of model mismatch. The adaptation / tuning unit determines one or more MPC controller tuning and design parameters, including for example, an MPC form, penalty factors for either or both of an MPC controller and an observer and a controller model for use in the MPC controller, based on a previously determined process model and either a known or an expected process model mismatch or process model mismatch range. A closed loop adaptation cycle may be implemented by performing an autocorrelation analysis on the prediction error or the control error to determine when significant process model mismatch exists or to determine an increase or a decrease in process model mismatch over time.

A microgrid including a renewable power source is optimized by treating the renewable source as operating in an isochronous mode, and by decomposing the optimization process. A renewable capability is defined for the renewable source to enable treatment of the renewable source as operating in an isochronous mode. The renewable capability may be based on a power rating of the renewable source. An objective function of the microgrid optimization problem may then be optimized while treating the renewable source as operating in an isochronous mode. A penalty factor may be applied to avoid power shortage.

A Deep Belief Network (DBN) feature extraction-based analogue circuit fault diagnosis method comprises the following steps: a time-domain response signal of a tested analogue circuit is acquired, where the acquired time-domain response signal is an output voltage signal of the tested analogue circuit; DBN-based feature extraction is performed on the acquired voltage signal, wherein learning rates of restricted Boltzmann machines in a DBN are optimized and acquired by virtue of a quantum-behaved particle swarm optimization (QPSO); a support vector machine (SVM)-based fault diagnosis model is constructed, wherein a penalty factor and a width factor of an SVM are optimized and acquired by virtue of the QPSO; and feature data of test data are input into the SVM-based fault diagnosis model, and a fault diagnosis result is output, where the feature data of the test data is generated by performing the DBN-based feature extraction on the test data.

The invention discloses a bearing fault diagnosis method based on improved EMD decomposition and sensitive characteristic selection. The method comprises steps of: performing wavelet noise reduction and EMD decomposition on the original vibration signals of a bearing in different fault states to obtain a plurality of IMF components; selecting, by quantitatively computing the correlation of each IMF component and the corresponding original vibration signal, the first h IMF components including the main fault information of the bearing as an object from which fault characteristic information is extracted, and extracting the characteristic parameters from the IMF components to form a original characteristic set; determining the sensitivity factor of each characteristic in the original characteristic set according to a distance evaluation method and constructing a sensitive characteristic set; inputting the sensitive characteristic vector of a training sample in the fault sample of the bearing into a SVM to be trained, optimizing the kernel function parameter g and the penalty factor c of the SVM according to a genetic algorithm, and identifying the fault of a tested sample. The method may reduce the dimensionality of the fault characteristic vector and the computational scale of a classifier, and increasing fault diagnosis accuracy of the antifriction bearing.

A system and method for determining a commanded engine and motor torque to minimize fuel consumption and manage battery state of charge in a hybrid electric vehicle. The method includes determining a penalty factor value that dictates the usage of the battery. A cost value is determined for every feasible engine torque for a selected or demanded vehicle torque and speed request. Each cost value is determined by a fuel consumption value, a change in the battery state of charge and the penalty factor value. The change in the battery state of charge is determined from the motor power and a nominal battery state of charge. For a specific penalty factor, a look-up table for optimal engine torque is generated for different requested vehicle torques at various vehicle speeds. A separate look-up table is provided for each penalty factor.

Provided is an urban traffic jam intelligent combination prediction method based on the track of a floating vehicle. The method includes: firstly, performing data processing and extracting parameters: connecting sampling points of the same floating vehicle ID to a movable track according to time relevance, matching a map with corresponding roads of an urban road network, and extracting the parameters capable of representing traffic jam; secondly, performing combined prediction of traffic flow parameters: predicting the traffic and the road average speed in the future by employing an SVM algorithm, bringing forward a penalty factor of SVM by PSO, optimizing corresponding kernel parameters by employing a plurality of kernel functions, and taking the prediction error minimum as the final prediction value; and finally making fuzzy divisions according to the jam state: processing the predicted traffic flow parameters, and converting the parameters to a jam state known by travelers via a fuzzy comprehensive evaluation method of multi-index weight adaptation change. According to the prediction method, compared with the classic optimization algorithm and the prediction method of an SVM single kernel function via the PSO, the accuracy, the timeliness, and the stability are all excellent.

The invention discloses a wind turbine generator gear case fault diagnosis method based on VMD and FA_PNN. Firstly, gear case vibration signals acquired by a sensor are subjected to de-trending processing, then, the processed gear case vibration signals are subjected to VMD variation modal decomposition under the condition of different decomposition numbers and penalty factors, k modal componentsare obtained with a Pearson's correlation coefficient method, singular value entropy, power spectral entropy, marginal spectral entropy and instantaneous energy spectral entropy of the k modal components are extracted from three angles of time domain, frequency domain and time-frequency domain, a feature vector matrix capable of describing operating states of a wind turbine generator gear case ina quantization manner is formed, and finally, test sample data are tested with well-trained firefly optimized probabilistic neural network FA_PNN, so that fault diagnosis of the wind turbine generatorgear case is completed. Classified recognition of faults of the wind turbine generator gear case is realized.

A method and computer program product for optimization of large scale resource scheduling problems. Large scale resource scheduling problems are computationally very hard and extremely time consuming to solve. This invention provides a Lagrangian relaxation based solution method. The method has two distinct characteristics. First, the method is formal. It is completely structure-based and does not use any problem domain specific knowledge in the solution process, either in the dual optimization or the primal feasibility enforcement process. Second, updating the Lagrangian multipliers after solution of every sub-problem without using penalty factors results in fast and smooth convergence in the dual optimization. The combination of high quality dual solution and the structure-based primal feasibility enforcement produces a high quality primal solution with very small solution gap. An optimal solution is first found to the dual of the resource scheduling problem by sequentially finding a solution to a plurality of sub-problems and updating a set of values used in the dual problem formulation after each sub-problem solution is obtained. Coupling constraint violations are systematically reduced and the set of values are updated until a feasible solution to the primal resource scheduling problem is obtained. An initial set of multiplier values is further determined by solving a relaxed version of the primal problem where most of the local constraints except the variable bounds are relaxed.

The invention discloses a microgrid multi-time scale self-adaption energy dispatching method under a source-grid-load interaction mode. The invention starts from the aspect of microgrid multi-time scale optimal management, and provides a self-adaption energy management method comprising basic day conventional dispatching and control day rolling dispatching. According to the method, labor is contributed for an energy storage device and a diesel generator according to a global object function and a global constraint condition, active power consumption of power and controllable and respectable load interactive with a large power grid is optimized, and the magnitude of the contributed labor can be predicted according to wind power and photovoltaic power generation systems, self-adaption adjustment for energy storage SOC penalty factors and control day rolling optimization time duration is realized, the influence of the error of the contributed labor predicted by the wind power and photovoltaic power generation systems on microgrid energy management is reduced, and economic and environmentally-friendly coordinated operation of a microgrid is realized.

An expression recognition method based on AVR and enhanced LBP in the technical field of model recognition comprises the following steps: collecting an original image; expanding virtual samples; carrying out wavelet decomposition on human image; extracting local binary pattern characteristic LBP; calculating an enhanced variance ratio AVR characteristic value and adding with a penalty factor, then extracting a plurality of groups of characteristic values of different dimensionality discriminated by AVR value, carrying out support vector machine classification accuracy test, and taking the characteristic dimensionality with the highest accuracy and corresponding characteristic value as the LBP characteristic. The method of the invention integrates image acquisition, human face test and human eye test, enhances the LBP characteristic by wavelet decomposition, and effectively improves the accuracy by adopting AVR method to extract effective characteristic.

The embodiment of the invention discloses an image iteration reconstruction method and a device. The method comprises steps of acquiring projection data subjected to water hardening correction after a target object is subjected to ray scanning, and according to the projection data, generating an initial reconstruction image and a noise weight matrix; according to the initial reconstruction image, acquiring an initial regularization penalty factor and an initial bone image segmentation matrix; according to the projection data, the initial reconstruction image, the initial bone image segmentation matrix, the noise weight matrix and the initial regularization penalty factor, using a pre-established bone hardening iteration reconstruction model to carry out image iteration reconstruction and acquiring a reconstruction image subjected to the bone hardening correction. In this way, it is achieved that bone hardening artifacts are gradually eliminated in an image iteration reconstruction process; bone hardening artifacts in the iteration reconstruction image can be effectively reduced; and quality of the iteration reconstruction image is ensured.

Disclosed is a protein-ligand affinity predicting method based on molecule descriptors. The protein-ligand affinity is reflected through construction of perfect and systematic molecule descriptors, and the relation between the descriptors and the affinity is constructed through a supporting vector regression (SVR) mode. The method includes the steps of training set preparation: preparing a large amount of data containing the crystal structure and the affinity of a protein-ligand complex; construction and calculation of the molecule descriptors: constructing 50 kinds of molecule descriptors which belong to nine categories, and calculating concrete values of all the complex descriptors in the training set; regression model construction: fitting the relation between the descriptors and the affinity through the SVR mode, and introducing a conjugate gradient method to optimize a penalty factor C and a kernel function parameter; and novel scoring function building which is used for predicting the affinity of the complex. The method has the advantages of being high in prediction capacity, small in target dependence, high in homolog sensitivity and the like.

Disclosed is a spatial gravity model based fuzzy c-means remote sensing image automatic classification method which is suitable for automatic segmentation and classification of remote sensing images, medical images and other image. The spatial gravity model based fuzzy c-means remote sensing image automatic classification method comprises the steps of determining the number of pixels of a remote sensing digital image and performing clustering on the image through a standard FCM (Fuzzy C-Means) model to obtain an initial fuzzy membership matrix and clustering center; solving the spatial gravity and spatial constraint penalty factors between every pixel and other pixels in neighborhood windows of the pixel in turn to obtain a fuzzy factor as the formula finally; adding the fuzzy factor as the formula to the standard FCM model to obtain a new clustering objective function; solving the fuzzy matrix as the formula and the clustering center as the formula in a circulating mode until the clustering center does not continue to be changed any longer or the operation reaches the maximum number of iterations; performing class marking on every pixel point of the remote sensing image through the maximum membership matrix criterion according to the finally solved fuzzy membership matrix as the formula to determine the class of every pixel to form a remote sensing image classification thematic map so as to implement the automatic classification of remote sensing digital images. The spatial gravity model based fuzzy c-means remote sensing image automatic classification method is simple, high in automation degree, small affected by image noise and high in image segmentation classification accuracy.

An augmented Lagrangian genetic algorithm that may be used to generate solutions for optimization problems subject to linear, bound, and non-linear constraints is discussed. The augmented Lagrangian genetic algorithm uses an adaptive mutation operator to separately handle the linear, and bound constraints, and uses an augmented Lagrangian framework to handle non-linear constraints. The non-linear constraints are handled by creating a sub-problem without the linear and bound constraints and solving the sub-problem using Lagrange parameter estimates and a penalty factor. The exclusion of the linear constraints and boundary constraints from the sub-problem allows the sub-problem to be resolved in a more effective manner than is possible using conventional techniques.

The invention discloses an intelligent storage route identification device and method used for multiple AGVs. The method is characterized by S100, determining the shortest path of offline phase AGV vehicles, improving real-time performance of a system, and reducing a burden of online operation of a scheduling system; S200, generating an optimal scheduling scheme; S300, obtaining a real-time running state of the system, adding different penalty factor processing when a fitness value of a scheduling scheme is calculated, and extracting real-time information of the system; S400, implementing a dynamic scheduling strategy of the multiple AGVs, and using a time window to grasp the real-time state of the entire system; and S500, realizing a communication interaction system, establishing a reliable communication connection between a AGV scheduling system and the AGV vehicles so as to transmit a command sent by an upper computer, an AGV operation state collected by a lower computer and environmental information. The whole system is beneficial to balancing time consumed by re-generation of an offline shortest path library, and at the same time, the system can rapidly respond when a road section has a fault, a fitness value of a scheduling scheme is increased, and efficiency and stability of the system are effectively improved.

The present invention discloses an improved deep Boltzmann machine-based pulmonary nodule feature extraction and benign and malignant classification method. The method includes the following steps that: step A, pulmonary nodules are segmented from CT images through using a threshold probability image graph method, so that regions of interest (ROI) are obtained, and the regions of interest are cut into nodule images of the same size; and step B, a supervised deep learning algorithm Pnd-EBM is designed to realize the diagnosis of a pulmonary nodule, wherein the diagnosis of the pulmonary nodule further includes three major steps: B1, a deep Boltzmann machine (DBM) is adopted to extract the features of the ROI of the pulmonary nodule which have deep expression abilities; B2, a sparse cross-entropy penalty factor is adopted to improve a cost function, so that the phenomenon of feature homogenization in a training process can be avoided; and B3, an extreme learning machine (ELM) is adopted to perform benign and malignant classification on the extracted features of the pulmonary nodule. The improved deep Boltzmann machine-based pulmonary nodule feature extraction method is superior to a traditional feature extraction method. With the method adopted, the complexity of manual extraction and the difference of feature selection can be avoided, and references can be provided for clinical diagnosis.

The invention discloses a photoacoustic spectroscopy-based transformer fault diagnosis method employing a parameter optimization SVM (support vector machine). According to the photoacoustic spectroscopy, contents of five characteristic gases of transformer oil are detected and calculated; 5 SVMs and 4 kernel functions are subjected to cross combination, and 20 SVM models are formed; the heuristic algorithm is adopted to perform parameter optimization on values of penalty factors c and g, and accordingly the SVM model highest in transformer fault diagnosis accuracy and operation speed is established. The experimental result shows that the SVM model composed of a C-SVC (C-support vector classification) model, an RBF (radial basis function) and the genetic algorithm optimization is highest in transformer fault diagnosis accuracy, with test set of 97.5% and training set of 98.3333%, and the optimizing speed of the genetic algorithm is faster than the particle swarm optimization by double. The photoacoustic spectroscopy-based transformer fault diagnosis method has the advantages such as simple operation, non-contact measurement, no consumption of carrier gas, short detection period, stability and high sensitivity.

The invention discloses a milling flutter recognition method based on variation modal decomposition and energy entropy, and belongs to the technical field of machine tool machining flutter recognition. The milling flutter recognition method comprises the following steps of S1, establishing a VMD mathematical model; S2, establishing a mathematical model of the energy entropy; S3, conducting three sets of milling machining experiments representing three cutting states, including stable cutting, a weak flutter and a severe flutter, and obtaining three sets of milling force signals through a dynamometer; S4, conducting FFT analysis on the three sets of milling force signals, and proving that the three sets of milling force signals represent the states of stable cutting, the weak flutter and the severe flutter respectively; S5, determining the number K of optimal modals of VMD decomposition and a penalty factor alpha through a VMD parameter automatic selection method based on a kurtosis value; S6, solving instantaneous frequencies of multiple IMF, and determining a milling flutter characteristic frequency band; S7, adopting a hammering experiment to obtain the modal of a knife; S8, extracting a flutter characteristic vector of each IMF based on the energy entropy. According to the milling flutter recognition method based on the variation modal decomposition and the energy entropy, the VDM decomposition effect is improved, and automatic flutter recognition is achieved.

The invention discloses a prediction and assessment method for stability of surrounding rocks of a mining roadway. The prediction and assessment method comprises the following steps: inputting measured data of influence factors of surrounding rocks of the mining roadway into a computer; then, retrieving sample data to set up a prediction model of a single kernel function support vector machine; combining selected global kernel functions with local kernel functions to form mixed kernel functions to set up a prediction model of a mixed kernel function support vector machine; utilizing a particle swarm algorithm to optimize penalty factors, kernel parameters and single kernel function coefficients of the predication model to set up a prediction model of a particle swarm algorithm-mixed kernel function support vector machine for stability of surrounding rocks of the mining roadway; and calculating prediction precision and reliability of to-be-predicted samples and assessing the maturity of the model. The prediction and assessment method for stability of surrounding rocks of the mining roadway is easily adopted and scientific. The grade prediction result of stability of surrounding rocks of the mining roadway and the maturity assessment result of the model are accurate and reliable and are highly consistent to the reality.

The invention discloses a method for extracting bearing fault characteristic frequency through information-entropy optimized VMD and application thereof. The method for extracting the bearing fault characteristic frequency through the information-entropy optimized VMD includes the steps that firstly, according to the information entropy minimum principle, the modal number of VMD is optimized, andthe optimized modal number is adopted to optimize the penalty factor of the VMD according to the information entropy minimum principle; then, the optimized modal number and penalty factor are adoptedto conduct VMD on bearing original vibration signals to obtain IMF components of the vested modal number, and through comparison, the IMF component where the information-entropy minimum is located canbe obtained and serve as a sensitive IMF component; finally, the selected IMF component is subjected to envelope demodulation analysis to extract the bearing fault characteristic frequency. By the adoption of the method for extracting the bearing fault characteristic frequency through the information-entropy optimized VMD, the bearing fault characteristic frequency can be effectively extracted; the method for extracting the bearing fault characteristic frequency through the information-entropy optimized VMD is applied to analyzing bearing simulation signals and actual bearing signals and hasrelatively broad practicability.

This application discloses a method implemented of recognizing a keyword in a speech that includes a sequence of audio frames further including a current frame and a subsequent frame. A candidate keyword is determined for the current frame using a decoding network that includes keywords and filler words of multiple languages, and used to determine a confidence score for the audio frame sequence. A word option is also determined for the subsequent frame based on the decoding network, and when the candidate keyword and the word option are associated with two distinct types of languages, the confidence score of the audio frame sequence is updated at least based on a penalty factor associated with the two distinct types of languages. The audio frame sequence is then determined to include both the candidate keyword and the word option by evaluating the updated confidence score according to a keyword determination criterion.

The invention provides a hyperspectrum classification method based on a composite kernel function. The hyperspectrum classification method comprises the steps of inputting a set of hyperspectrum images in N classes, taking a support vector machine as a base classifier, and meanwhile, randomly selecting S samples from every classes of the hyperspectrum images to form a training set and forming a test set with the left samples, determining the change range of each parameter, next, determining the optimal performance parameters, including a penalty factor and a kernel of the support vector machine by virtue of cross validation for K times, constructing a composite kernel function by use of a composite kernel construction policy and training the support vector machine, and circulating for N times by use of the parameters of a support vector machine decision function obtained in the training process to obtain decision function values of which the test set belongs to every classes and to form a matrix as shown in the specification, and then determining multiple classifier policies, namely finding the maximum values of every columns of the matrix as shown in the specification. The hyperspectrum classification method based on the composite kernel function has the characteristics of better description of distribution features of a data set, relatively high classification accuracy and the like. The time taken by parameter optimization of the hyperspectrum classification method is also relatively short in contrast with a traditional multi-kernel learning method.

The invention provides a method for predicting defects of object-oriented software. The method comprises the steps of processing a training data set, acquiring valid feature attributes and building a new training data set according to the valid feature attributes; training a support vector machine (SVM) according to the new training data set and performing parameter optimization via particle swarm optimization (PSO), wherein parameters include a penalty factor and a Gaussian kernel bandwidth; using an SVM model to perform defect prediction on prediction data according to the optimized parameters and obtaining a prediction result. A training data set is processed to acquire valid feature attributes and a new training data set is built according to the valid feature attributes, so that curse of dimensionality is avoided effectively, processing cost is reduced and data processing speed is increased; PSO is employed for parameter optimization, so that optimal parameters can be selected and defect prediction accuracy is improved.

The invention provides a multi-area interconnected power system wind power consumption optimization method considering connection line electricity transaction scheme. The method comprises the following steps: determining time period and step length of problem research, and collecting calculation data of each province; summarizing data of system thermal power units, grouping the units having same operation parameters into same class and summarizing the number of units in each class; establishing a multi-province interconnected power system wind power consumption optimization module; solving the optimization module by utilizing a solver CPLEX to obtain an optimal solution; and calculating wind power consumption quantity and wind power abandon quantity of each province. The method greatly reduces the number of variables and improves problem solving speed; by introducing wind power abandon section penalty factors and setting value of each section of penalty factor, wind power abandon sequence can be controlled; and with the connection line electricity transaction scheme being taken into consideration, by optimizing connection line power, resource optimized allocation is realized in multiple provinces, and whole-grid wind power priority consumption is realized.

The present invention provides a blasting vibration predicting method based on a particle swarm algorithm optimization support vector machine. According to the method, firstly, blasting vibration influence factors are subjected to feature extraction. Secondly, the kernel function, the penalty factor, the slack variable and the kernel parameters of the support vector machine are subjected to combined optimization thorough the improved PSO algorithm, and then an optimal support vector regression machine and an optimal support vector classifier model are respectively obtained. In this way, the classified prediction of blasting vibration data is realized. Compared with the traditional blasting vibration velocity predicting method for support vector machines, optimized combined parameters are obtained, so that the performances of models can be better improved. The prediction accuracy of models is improved, and the prediction accuracy of the blasting vibration strength is greatly improved.

The invention relates to an undersampled CT image reconstruction method. At first, values of regularization parameters in the object functions of an iteration algorithm are determined by a trained and constructed self-adaptive parameter selection model, and then an iteration reconstruction model based on a dictionary learning theory is built, weight penalty factors are added to the regularization bound items of the object functions to form a weight dictionary learning iteration reconstruction model, every iteration process of the model alternatively updates the sparse representation of the weight penalty factors, the self-adaptive dictionary, and image blocks and reconstructs the images, and the iteration stopping condition is reached at the end. The undersampled CT image reconstruction method can be better adapted to the undersampled condition with insufficient projection data to meet the need of reducing the radiation dosege, and when the data sampling rate is reduced to a certain degree, a CT image can be reconstructed in high quality, and good robustness is provided for the projection noises caused by low dosage.

The invention discloses a multivariate quality process out-of-control signal diagnostic method based on a support vector machine and genetic optimization. The multivariate quality process out-of-control signal diagnostic method based on the support vector machine and the genetic optimization is characterized in that first, the types of signals likely to lead to abnormality of a multivariate process are determined according to the mean value dimensions of the multivariate process, namely the structure of a classifier model is determined; second, radial basis function parameters and penalty factors of the support vector machine are optimized with the genetic algorithm; third, the optimal support vector machine classifier model is obtained through the acquired optimal parameters, and the multivariate process out-of-control signals are diagnosed on the basis of the optimal support vector machine classifier model. The parameters of the SVM are selected dynamically through global searching ability of the genetic algorithm, and thus automatic optimization selection of the parameters of the SVM classifier is achieved, and quality diagnosis effects of the multivariate process are also promoted. The multivariate quality process out-of-control signal diagnostic method based on the support vector machine and the genetic optimization integrates the GA global searching ability and the classifying ability of the SVM, and meanwhile avoids complex calculation, simplifies the network structure of the classifier and promotes generalization ability and identification efficiency of the classifier.

The invention provides a CCPDI-IMPM method for large deformation analysis pf a saturated porous medium based on a u-p control equation of a generalized Biot theory and a convective particle domain interpolation technology, and the method comprises saturated porous medium large deformation consolidation process analysis and saturated porous medium large deformation dynamic process analysis. According to the method, an implicit time integral strategy is adopted to overcome the defect that a traditional explicit material point method is limited by a critical time step length, and particularly, the calculation efficiency is remarkably improved when a quasi-static problem is solved; and the grid boundary smooth interpolation advantage of the CPDI and the Euler-Lagrangian description advantage of the implicit substance point method are combined, so that compared with the traditional FEM, the calculation performance is more stable when large deformation and extremely large deformation are processed. The invention further provides a method for expanding the substance penalty factor, and the method is used for processing the boundary condition of the fluid-solid coupling system more simplyand accurately.