72 results about "Linear number" patented technology

The present invention applies a novel iterative algorithm to the problem of multidimensional optimization by supplying a strict, nonlinear mathematical solution to what has traditionally been treated as a linear multidimensional problem. The process consists of randomly selecting a statistically significant sample of a prospect list, calculating the value of the utility function for each pair of an offer and selected prospects, reducing the original linear multidimensional problem to a non-linear problem with a feasible number of dimensions, solving the non-linear problem for the selected sample numerically with the desired tolerance using an iterative algorithm, and using the results to calculate an optimal set of offers in one pass for the full prospect list.

A computer-implemented method and system are provided for authenticating the identity of a user registered with a computer system. The authentication method comprises generating a multi-dimensional array of elements that are addressable by respective sets of indices, generating a challenge code comprising a linear array of elements for addressing a first set of indices of the array of elements, transmitting the multi-dimensional array of elements and challenge code to at least one computing device associated with the user, receiving a response code from the user, and verifying the user's identity when the received response code matches a derived code obtained by retrieving elements from the multi-dimensional array at locations addressed by elements taken from the challenge code and a personal code stored at the computer system, wherein the personal code comprises a linear array of elements for addressing a different set of indices of the array of elements.

Aspects of the present disclosure include a quick and easy user interface system for allowing a user to format mathematical expression on a math program as they would be formatted in longhand. The user interface is configured to switch between a linear math expression entry scheme and a structured math expression entry scheme based on a user's math expression entries in order to predict the user's desired entry format. The user interface contains additional shortcuts and display features which enable a user to enter mathematical expressions quickly and easily, with less confusion.

A method and system for optimizing a data structure for manipulation of matrices in a parallel environment limits computational branching. The data structure further is further optimized for linear data storage and synchronization among multiple processing threads.

The invention discloses a fault diagnosis method based on Stacked LSTM (Long Short Term Memory). The method comprises the following steps: acquiring measurement data of each measurement device of a chemical system in a normal state and K fault states and acquiring training samples by construction , then, creating a fault diagnosis model based on Stacked LSTM; wherein the model comprises a StackedLSTM network, a full connection layer and a softmax layer, wherein the Stacked LSTM network is obtained by superposing D layers of LSTM networks; and training the fault diagnosis model by adopting thetraining sample, collecting actual operation data in the operation process of the chemical engineering system, constructing an input sequence, and inputting the input sequence into the fault diagnosis model based on Stacked LSTM to obtain a fault identification result. The Stacked LSTM network is formed by stacking multiple layers of LSTM networks, the dynamic time sequence information of the original data can be automatically extracted under different time scales, and the method has relatively high expression capability for complex nonlinear data, so that the accuracy and robustness of faultdiagnosis are improved.

Provided are a dispatch device and a method of a batch processing machine capable of sequencing facing a product family. The dispatch device of the batch processing machine sequencing facing the product family comprises two sub-modules including a real-time scheduling simulation platform module based on a heuristic algorithm and a sequencing solving module based on a slack mixed integer linear mathematical model. The two modules conduct loop execution under a variable time window based on a rolling time domain strategy. The sequencing solving module does not conduct lot grouping and batch sequencing on workpieces of different product families in a buffer period before the batch processing machine and directly conducts optimized sequencing on the different product families. The real-time scheduling simulation platform conducts scheduling on workpieces with the highest priority level in the product families according to the processing capability of a free and available batch processing machine. The dispatch device and the method of the batch processing machine sequencing facing the product family greatly reduces time complexity of the algorithm, well achieves real-time optimal scheduling of a reentrant low batch processing machine and meets just-in-time production requirements of the semiconductor industry on the premise of guaranteeing high accuracy. The dispatch device and the method of the batch processing machine capable of sequencing facing the product family is simple in concept, convenient to achieve and in favor of popularization and application to reentrant industries such as the semiconductor industry.

The invention discloses a manifold learning and Hilbert-Huang transformation combined structural modal parameter identification method. The method comprises the following steps of: 1, acquiring time domain response data of a measure point in a structure; 2, processing the time domain response data acquired in the step 1 by adoption of a manifold learning algorithm so as to obtain a vibration modeand a fixed frequency of the structure; and 3, processing the time domain response data acquired in the step 1 by adoption of a Hilbert-Huang transformation method so as to obtain a damping ratio of the structure. Compared with the prior art, the method has the beneficial effects as follows: 1, when modal parameter extraction is carried out by utilizing the manifold learning and Hilbert-Huang transformation combined method, vibration modes, fixed frequencies and damping ratios with relatively high precision can be obtained through response data when material parameters and experiment conditions of structures are unknown; and 2, the method can be used for processing nonlinear data and retaining nonlinear manifolds of the structures.

The invention discloses a fault-tolerant control method for a nonlinear actuator fault system based on an observer, which belongs to the field of intelligent control. Mainly under a data-driven framework, aiming at the fault-tolerant control problem of a repeatable nonlinear system with an actuator fault, the control scheme is as follows: establishing the repeatable nonlinear system with the actuator fault; iteratively linearizing the nonlinear system into an equivalent linear data model; designing an iterative learning observer to estimate system output; designing an iterative updating algorithm to estimate an unknown gradient matrix in the linear data model; and designing an observer-based controller to realize fault-tolerant control of the nonlinear actuator fault system. According to the observer-based fault-tolerant control method for the nonlinear actuator fault system, the inhibiting effect of the observer on uncertainty is utilized, good tracking performance and high robustnessare achieved, and in addition, due to the fact that a data driving technology is used, the fault-tolerant control method does not depend on model information and has good adaptability.

The invention discloses a learning-oriented disturbance observer design method, and belongs to the field of intelligent control. Mainly under a data-driven framework, aiming at the total disturbance estimation problem of a repeatable system with non-repetitive external disturbance, the disturbance observer scheme is as follows: establishing a system with non-repetitive disturbance; iteratively linearizing the system into a linear data model, and converting non-repetitive disturbance into total disturbance; designing an iterative updating algorithm to estimate a gradient matrix in the linear data model; designing a state-based learning disturbance observer for a state-measurable system; and for a system with an unmeasurable state, designing an output-based learning disturbance observer. According to the learning-oriented disturbance observer design method disclosed by the invention, the total disturbance in the system can be estimated under a data-driven framework, only a small amount of model information is utilized, the method has good adaptability, learning can be carried out from the previous batch of operation data, and the estimation of the total disturbance is more accurate.

The invention provides a character emotion identification system and method. The system comprises a character receiving device, a neural network layer, a linear layer and an emotion analysis device, wherein the character receiving device is used for receiving tested characters and converting the tested characters into character vectors; the neural network layer is used for processing the character vectors to obtain character abstract representation data; the linear layer is used for carrying out linearization on the character abstract representation data to obtain linear data; and the emotion analysis device is used for calculating emotion label data corresponding to the tested characters according to the linear data, and carrying out emotion identification on the tested characters through the emotion label data. By use of the character emotion identification system and method, the tested character still can be subjected to the emotion identification when the tested characters do not contain emotion keywords.

The invention discloses an intelligent welding method, an intelligent welding system and a computer storage medium, belongs to the technical field of welding, and aims to solve the technical problem of low welding precision in the prior art. The intelligent welding method comprises the following steps of predicting linear data through a linear model to obtain a first prediction result; and / or predicting nonlinear data through a nonlinear model to obtain a second prediction result; and obtaining target operation parameters based on the first prediction result and / or the second prediction result, and emitting target laser through a laser device body according to the target operation parameters to weld a current to-be-welded object. By means of the intelligent welding method, the welding precision can be improved.

The invention discloses a method for calculating the ultimate load of an independent foundation in a karst foundation. The method comprises the following steps of taking a covering layer and caving rock mass at the top of a karst cave as a study object; carrying out stress analysis on the caved rock mass at the top of the karst cave; establishing an additional stress condition on a rock-soil interface of a collapsed rock mass at the top of the karst cave; establishing a linear mathematical programming model for solving the limit load of the independent foundation in the karst foundation; and solving the linear mathematical programming model of the independent foundation limit load in the karst foundation to obtain the limit load of the independent foundation, so that the result of the independent foundation limit load obtained by the method is more accurate than that obtained by the traditional method.

The invention provides a memory coupling compiling method and system for a reconfigurable chip. The memory coupling compiling method comprises the following steps: acquiring the periodicity of a DFG data flow diagram; obtaining a linear conversion vector of the periodicity through the mapping time difference; and judging whether a linear array of the linear conversion vector can be obtained through a heuristic algorithm or not, obtaining a memory mapping result according to a judgment result or through the current DFG data flow diagram, and obtaining and adjusting the current DFG data flow diagram until the linear array is obtained. According to the method and the device, each node in the DFG data flow diagram is adjusted, so that the requirement of obtaining the linear array is met, the compiling time is greatly shortened, and the compiling experience of a user is very strong. The reconfigurable acceleration chip is high in practicability, can be used, and has extremely high reusability for programmable devices.

The invention discloses a novel chemical process quality prediction method, comprising the following steps: step 1, collecting sensor data in the chemical process, processing and establishing a novelprocess prediction model; Step 2: Collecting the newly obtained data during the chemical process operation, and using the prediction model obtained in step 1 to carry out on-line prediction. First, the orthogonal signal correction algorithm is used to deal with the historical data collected from the process, and then the prediction model is established by using the processed data. In order to improve the ability of the model to deal with nonlinear data, support vector machine is introduced into the prediction model. This method improves the disadvantage of the traditional prediction model thatthe modeling process is complex and can not deal with the nonlinear data, and improves the ability of the model to track the real data.

The invention discloses an intelligent control algorithm for analyzing ecological oxygenation of a river channel based on parameter uncertainty. The invention discloses a riverway oxygenation algorithm based on a deep learning algorithm. The riverway oxygenation algorithm comprises the following steps: 1) collecting parameters; 2) estimating a dissolved oxygen balance model; 3) determining n objective functions for performing model parameter uncertainty analysis; 4) substituting into a dissolved oxygen balance model for calculation; 5) comparing to obtain a dominant parameter set corresponding to the Pareto set under the target function; 6) drawing a joint probability density function, and performing two-parameter uncertainty analysis; 7) drawing probability density distribution of each parameter in the dominant set; 8) calculating the oxygen demand of the river channel; and 9) determining the oxygenation amount of the equipment according to the oxygen demand of the river and the oxygenation efficiency of the equipment. The method overcomes the defects that in an existing technical scheme, an oxygenation device cannot be accurately controlled according to multi-parameter data, and a simple linear mathematical model can only be judged through artificial experience or applied mechanically, so that the aim of accurately controlling oxygenation according to the parameters such as the river water flow speed, the river water quality and the environmental influence factors is achieved.

The invention discloses a batch prediction compensation method for data packet loss in a network system, and belongs to the field of intelligent control. Mainly aiming at the problem that data on an input side and an output side in a network control system is lost, a control scheme provided by the invention comprises the following steps of: establishing a virtual linear data model, and describingthe input and output dynamic states of a multiple-input multiple-output linear repetitive system in an iterative domain; establishing a batch iterative prediction model based on the lifting technologyby utilizing the linear data model, wherein the batch iterative prediction model actually exists in a computer and is used for predicting and compensating for lost data. The lifting model predictioncompensation method for data packet loss in a network system is popularized to an unknown nonlinear non-affine system. The lifting model prediction compensation method for data packet loss in a network system mainly solves the problem of data packet loss in a network control system. Under a data driving framework, a batch iterative prediction model based on a lifting technology is established to perform batch prediction and compensation on lost data, the calculation real-time performance is high, deterioration of control performance of a network system due to data loss can be effectively avoided, and the method can be applied to a linear system and can also be applied to a nonlinear system.

The invention discloses a data-driven nonlinear system actuator fault factor identification technology, which belongs to the field of intelligent control. Under a data-driven framework, aiming at a fault failure factor identification problem of a nonlinear system with an actuator fault, the identification method comprises the steps of establishing the nonlinear system; dynamically linearizing thenonlinear system into an equivalent linear data model with an actuator fault failure factor; designing an updating algorithm to estimate a pseudo partial derivative in the linear data model; designinga state observer to estimate the system state; and designing an online estimation algorithm of the fault factor to realize estimation of the fault factor. According to the data-driven nonlinear system actuator fault factor identification technology disclosed by the invention, the actuator fault failure factor in the nonlinear system can be estimated on line; a data driving method is used; the method does not depend on specific model information; and the requirement for estimating the actuator fault is met on the premise that good adaptability is achieved.

This invention concerns an iterative procedure for conversion of structured software objects into a raw data stream and vice versa, providing for their direct transfer using simple communication resources such as those of an embedded computer station, and reset of said software objects or reutilisation of memory space allocated to them.This procedure can be used by an embedded platform (2) or a portable object including at least a processor capable of exchanging information with a terminal in the form of linear data sequences. The procedure includes a step for conversion of a data set, in one direction or the other, between a linear data sequence arrangement on the one hand, and a structured arrangement describing or representing an object-oriented software object on the other hand.

The invention discloses a multivariate data structure automatic conversion method and system for large-screen chart adaptation. The method comprises the following steps: configuring a visual large-screen chart, and configuring the attributes and categories of the chart through a visual large screen; configuring a data interface service, including configuring a data conversion type and a data conversion process; accessing the data row and column conversion operator and the data structure conversion operator into a data conversion process; starting a data interface service to obtain converted data; and testing the data interface service by using the converted data until the verification result is correct. The data row and column conversion operator is specifically used for converting one-dimensional / multi-dimensional array rows and columns into linear data through a visual arrangement method; and the data structure conversion operator is specifically used for converting a one-dimensional / multi-dimensional array structure into linear data through a visual arrangement method. The adaptation cost of the large-screen data chart configuration process can be reduced, and the adaptation efficiency of the large-screen data chart configuration process is improved.

The invention provides a data table blood relationship analysis system and method, which are applied to the technical field of data blood relationship analysis, the system comprises a front end and arear end, the front end comprises a user operation area and a result display area; wherein the user operation area is used for displaying a classification directory tree containing a plurality of datatables, receiving a selection instruction and sending an identifier corresponding to the selected data table to the rear end based on the selection instruction; the result display area is used for receiving the analysis result sent by the rear end and displaying the analysis result; the back end receives and takes the data table corresponding to the identifier as a termination node or a start node of the data flow; and querying a target data table having an association relationship with the data table in a database by using a recursive algorithm, and arranging the target data table into a linear data stream according to a logic relationship between the target data tables. By applying the embodiment of the invention, the mapping of the internal relationship of the data table is realized, the data generation sequence is also combined, and the successive relationship can also reflect the association relationship of the data.

The invention discloses an information analysis and prediction method and device, a computer system and a readable storage medium, and the method comprises the steps that an operation server receivesto-be-tested information transmitted by a client, and obtains a prediction model matched with the to-be-tested information from a preset model database as a target model; the preprocessing server obtains to-be-tested information and a target model from the operation server, extracts data matched with index data in the target model from the to-be-tested information, and sets the data as linear data, wherein the index data is data corresponding to independent variables of a target function in the target model, and the target function is a linear regression equation; and the operation server obtains the linear data from the preprocessing server, calculates the linear data through the target model to obtain prediction information, and sends the prediction information to the client. According to the method, the situation that the prediction model is directly entered without distinguishing the linear relation and the nonlinear relation is avoided, and the accuracy and stability of the prediction result of the prediction model are guaranteed.

The invention provides an intelligent recognition control method for linear data of an ultra-wide and ultra-large-span bridge. The method comprises the steps that state variable data and observation variable data of the bridge are detected in real time, wherein the state variable data comprises a temperature change value, a deflection change value, a bracket deformation value and a finished bridge camber; state variable data and observation variable data in unit time are obtained, and integrity and abnormal data screening is carried out to obtain modeling data; fusion modeling is conducted on the state variable data and the observation variable data; reliability analysis and calculation are carried out on the modeling in combination with bridge foundation engineering parameters, and different control strategies areformulated for control, wherein the control strategies comprise concrete pouring adjustment, perpendicularity construction design and external stress reinforcement design. Reliability analysis and calculation are conducted on modeling in combination with bridge foundation engineering parameters, different control strategies are formulated for control, and the assessment accuracy of the bridge structure reliability can be improved.

The invention discloses a highway OD prediction method based on deep set empirical mode decomposition. The method comprises the steps of 1, extracting a highway OD sequence; 2, carrying out time feature extraction based on ensemble empirical mode decomposition; and 3, carrying out OD sequence prediction based on the GRU network. According to the prediction method provided by the invention, the ensemble empirical mode decomposition model and the deep learning network are combined, on one hand, the model can convert non-stationary nonlinear data into stationary linear data, a great auxiliary effect is achieved on the prediction model to mine a sequential relationship, and the prediction precision of the model is improved; and on the other hand, the model complexity can be reduced, and the model calculation efficiency is improved.

The invention discloses a method for calculating the minimum thickness of a karst cave top plate of a karst pile foundation. The method comprises the following steps: step 1, drawing up basic parameters for calculating the minimum thickness of the karst cave top plate of the karst pile foundation; step 2, carrying out stress analysis on the karst pile foundation; 3, establishing a nonlinear mathematical programming model for solving the minimum thickness of the karst pile foundation karst cave roof; and 4, solving the nonlinear mathematical programming model of the minimum thickness of the karst pile foundation karst cave top plate to obtain the minimum thickness of the karst pile foundation top plate. According to the method, a nonlinear mathematical programming model for solving the minimum thickness of the karst pile foundation karst cave top plate is established, shear failure and tensile failure of a karst cave top plate failure surface are considered at the same time, and the minimum thickness of the karst pile foundation karst cave top plate is sought in combination with a nonlinear mathematical programming algorithm; compared with a traditional method, the method is rigorous in theory, the solved result is more accurate, and programming is simple.

The invention discloses a super computer operation failure active prediction method based on application similarity, and belongs to the field of super computers, and the method comprises the steps: S1, extracting feature data from an operation log, adding operation path data, preprocessing the feature data and the operation path data, and taking the preprocessed data as input features of a machine learning algorithm model; and S2, after the machine learning algorithm model processes the input feature data, the operation failure state is actively predicted. According to the method, the characteristics capable of accurately describing the job application attributes are mined, and a good prediction improvement effect is achieved; a machine learning algorithm is adopted to find an operation failure prediction method, the robustness of a prediction model is improved, and the method is especially suitable for nonlinear data; the clustering calculation overhead is remarkably reduced and the error is reduced by adopting a clustering method for the job application attributes; the prediction efficiency is high, and the method can be practically applied to large supercomputers.

The invention discloses a data-driven dynamic internal model control technology, and belongs to the field of intelligent control. Under a data-driven framework, aiming at the problem that the non-linear non-affine system needs to improve the processing capability for uncertainty and disturbance, the control scheme provided by the invention comprises the following steps of: converting the unknown non-linear non-affine system into an equivalent linear data model by using a dynamic linearization method, and regarding the linear data model as an internal model; designing a parameter updating law to estimate unknown pseudo partial derivative parameters in the linear data model; designing a nominal controller capable of enabling the system to output a tracking target trajectory; and designing an uncertainty compensation controller to compensate disturbance and mismatch between an actual system and the internal model. The data-driven dynamic internal model control technology disclosed by the invention is simple and convenient in implementation mode, small in operand, convenient for engineering implementation, and capable of effectively inhibiting disturbance and realizing a relatively good tracking effect under the condition that only input and output data can be obtained.

A symbol-determining device according to an embodiment includes: a transmission line shortening unit that multiplies each symbol value of a symbol array that is part of an input signal by a tap gain of a linear digital filter and outputs a symbol array representing a sum of values acquired through the multiplication; a transmission line estimating unit that estimates a transfer function of a transmission line using an adaptive nonlinear digital filter on the basis of a symbol array representing a state of the transmission line; an addition comparison processing unit that calculates a minimum value of a distance function in a Viterbi algorithm on the basis of a metric that is calculated on the basis of the output of the transmission line shortening unit and the transfer function; and a path tracing-back determination unit that performs symbol determination by tracing back a trellis path in the Viterbi algorithm on the basis of the minimum value of the distance function.

The invention relates to an upper-bound limit analysis method of ultimate bearing capacity of rock slope considering rock translation and rotation effects at the same time, and belongs to the technical field of rock slope bearing capacity analysis. According to the method, based on a plastic upper-bound limit analysis theory, by means of multiple freedom degree rigid block unit discrete jointed rock slope, the translational speed and rotational speed of the rigid block unit centroid are used as unknown quantities, the translation and rotation mechanical effects of rocks are considered at the same time, the overload coefficient of the slope is used as a target function, and plastic flowing conditions satisfying shear failure, tensile failure and rotational failure of a structural plane are constructed; then combined with a virtual work principle and speed boundary conditions, a linear mathematic programming model for working out the ultimate bearing capacity of the jointed rock slope is built, the linear mathematic programming model is solved by means of a simplex method, a strength reserve coefficient is solved by means of an iterative method at the same time, and finally the upper-bound solution and corresponding failure mechanism of the ultimate bearing capacity of the jointed rock slope are acquired. The method has the advantages that the concept is clear and the computational accuracy is high.

A device for processing a digital input value includes a first memory portion having stored thereon Q delimiter values where the Q delimiter values divide the range of the N-bit digital input value into Q+1 regions and at least a first region and a second region are of unequal sizes, a second memory portion has stored thereon a look-up table storing Q+1 sets of coefficients for performing numerical value conversion of the digital input value to a digital output value in a second, natural unit, and an arithmetic logic performing numerical value conversion using the N-bit digital input value and the selected coefficient pair. The N-bit digital input value is compared with the Q delimiter values to determine a respective one of the Q+1 regions in which the N-bit digital input value lies. The look-up table is indexed according to the respective region to provide the selected coefficient pair.

The invention discloses a landslide stability evaluation method based on an LSTM algorithm, and the method comprises the steps: obtaining a plurality of landslide basic parameters, and carrying out the normalization processing of all landslide basic parameters; dividing the normalized landslide basic parameters into a training set and a test set, and respectively training and verifying the constructed LSTM model; and inputting the re-acquired landslide basic parameters into the verified and qualified LSTM model to obtain a corresponding landslide stability state, The problems that in an existing landslide stability evaluation method, the number of selected evaluation factors is small, existing landslide related data are not considered too much, and linear mathematical analysis is difficult to carry out due to the complex geological environment are solved, and the reliability of landslide stability evaluation is improved.