93 results about "Optimal weight" patented technology

The invention belongs to the field of ocean internal temperature detection, and particularly relates to a method for optimizing a multi-source constraint weight in ocean thermohaline profile variational inversion, which comprises the following steps of: acquiring a historical actually measured thermohaline profile and matched satellite remote sensing sea surface data, constructing a training set and counting a climate state background field; constructing a weighted variational cost function containing multiple constraint terms, and introducing an adjustable weight parameter for a key constraint term; based on the training set, constructing an external optimization function taking inversion error minimization as a target; weight parameters are iteratively learned through a double-layer optimization framework, the weight is updated by adopting methods such as Bayesian optimization in outer-layer optimization, and the optimal profile is solved under the given weight in inner-layer optimization; and finally, applying the optimal weight obtained by learning to variation inversion of new observation data to realize high-precision reconstruction of the thermohaline profile. According to the method, objective and adaptive optimization of the weight is realized, the inversion precision and robustness are remarkably improved, and the method has good physical interpretability and system universality.

The invention relates to the technical field of microbial fermentation, in particular to a biological fermentation monitoring method and system based on multi-source data fusion. The method comprises the following steps: acquiring multi-source monitoring data in a fermentation tank and constructing a multi-dimensional state vector; on the basis of data of the multi-dimensional state vector in a preset time window, constructing a metabolic entropy of the current fermentation process; a simulated annealing algorithm is adopted to optimize control parameters, the temperature of the simulated annealing algorithm is positively correlated with the metabolic entropy, and minimization of a second energy function is taken as a target; a catastrophe reignition mechanism is set in the optimization process, fermentation abnormity is diagnosed based on triggering of the catastrophe reignition mechanism, and abnormal state characteristics are matched with a preset abnormity database through iterative search of an optimal weight distribution scheme so as to determine the abnormity type with the highest matching degree. By introducing the metabolic entropy and the catastrophe reignition mechanism, the sensitivity, the accuracy and the robustness of monitoring the biological fermentation process are remarkably improved.

The invention discloses a power distribution network load access point dynamic weight preferential determination method considering new energy access, and belongs to the technical field of electric power. The method solves the problems that an existing load access point determination method depends on expert experience or static empowerment and is difficult to adapt to multi-source time-varying data of the power distribution network under new energy access, weight distribution lacks dynamic response, and consequently the deviation between a decision and actual operation is large. According to the technical scheme, the method comprises the steps of collecting multi-source data and preprocessing; constructing an evaluation index system, and performing dimension alignment based on the standardized data to form a feature tensor; generating an initial dynamic weight of a load access point by using a physical information space-time diagram network; optimizing the initial dynamic weight by adopting a constrained multi-objective Bayesian optimization method to obtain an optimal dynamic weight; and calculating a comprehensive score of each load access point based on the optimal dynamic weight and determining an optimal access point. According to the invention, the safety and operation efficiency of the power distribution network under new energy access are improved.

The invention discloses a bearing fault diagnosis method and system based on a simulated physical neural network. The method comprises the following steps: S1, preprocessing a collected vibration signal to obtain a to-be-diagnosed signal; s2, executing a hybrid optimization algorithm to obtain an optimal transmission band parameter and an optimal weight parameter; s3, configuring a feature extraction module according to the optimal transmission band parameter, and writing the optimal weight parameter into a classification module; s4, inputting the to-be-diagnosed signal into a feature extraction module to obtain a simulation feature vector; s5, inputting the simulation feature vector into a classification module, and outputting a classification voltage signal; and S6, determining a fault type, and outputting a result. According to the method, a hybrid optimization algorithm is adopted, the ratio of the inter-class dispersion degree to the intra-class polymerization degree of the feature vectors is calculated, optimal adaptation of feature extraction and classification tasks is achieved, and the diagnosis precision and generalization ability are improved; by constructing a full analog domain signal processing architecture, analog-to-digital conversion and digital calculation are not needed, and microwatt-level power consumption and microsecond-level real-time response are realized.

The invention belongs to the technical field of virtual power plant operation evaluation, and particularly relates to a virtual power plant operation comprehensive evaluation method and system.The method comprises the steps that a three-level multi-dimensional operation risk evaluation index system is constructed, and three types of weight vectors are obtained through an analytic hierarchy process, an entropy weight method and an inter-index correlation and conflict objective weighting method; carrying out collaborative correction through a non-cooperative game model to obtain an optimal comprehensive weight vector; a standard cloud model of each risk level is constructed based on an extension cloud model, an actual evaluation cloud model is constructed in combination with the score of a professional scoring system and an optimal weight, the score weight of the professional scoring system is determined through a grey correlation degree, an evaluation value is corrected, the actual cloud model is updated, and finally the digital characteristics of the actual and standard cloud models are compared. And quantitatively judging the operation risk level of the virtual power plant. According to the method, the comprehensiveness, reliability and scientificity of evaluation are improved, and support is provided for risk prevention and control of the virtual power plant.

The application provides a broadband array electronic detection anti-interference method and device and a storage medium, and relates to the technical field of array signal processing.The method comprises the following steps: defining a signal receiving carrier core parameter and constructing an interference signal model, arranging the interference signal in a space-time two-dimensional manner to form a data matrix, and decomposing to obtain an interference characteristic vector set; constructing a target signal model and fusing the interference model to form a mixed signal matrix, and estimating a covariance matrix thereof; constructing a target direction vector based on a search direction, establishing an optimization constraint in combination with the interference characteristic and the covariance matrix, and solving to obtain an optimal weight vector; and performing weighted processing on the mixed signal through the optimal weight vector, and outputting a target signal after interference suppression.The application constructs a data adaptive anti-interference logic, the core of which is to separate the interference and the target signal characteristics, and the fixed filtering parameter is not needed, so that the complex electromagnetic interference can be adaptively adapted, the interference can be suppressed, and the target signal can be completely reserved.

The embodiment of the invention provides a power prediction method and device, computer equipment and a medium, and relates to the technical field of power prediction. The method comprises the following steps: acquiring target multi-source data; respectively inputting the target multi-source data into the plurality of power prediction models to obtain a plurality of initial powers in one-to-one correspondence with the plurality of power prediction models; performing weighted average processing on the plurality of initial powers based on an optimal weight combination in the plurality of weight combinations to obtain a target power; wherein each weight combination comprises weight coefficients of a plurality of power prediction models. Compared with other weight combinations except the optimal weight combination in the plurality of weight combinations, the optimal weight combination enables the error between the target power and the real power of the target multi-source data to be smaller than the preset error, that is, the obtained target power is close to the real power of the target multi-source data, so that the real power of the target multi-source data is obtained. And the accuracy of the obtained power is improved.

The invention discloses a seismic source positioning method considering the influence of an abnormal value and a dynamic wave velocity, and the method comprises the steps: arranging sensors to detect sound waves emitted by a seismic source, and observing the time when the sound waves reach each sensor; subtracting the control equation established based on the reference sensor from the control equation established based on the measurement sensor, and carrying out linearization and matrix representation on a nonlinear equation line obtained by subtraction; the matrix is solved, an analytic solution enabling the sum of squares of the residual errors to be minimum is found, the obtained analytic solution is substituted into an overdetermined matrix equation to obtain a fitting model, and weight estimation is conducted according to the relation between the residual errors and weights; iteration is carried out between weight estimation and overdetermined matrix solving, after optimal weight estimation is obtained, overdetermined matrix solving is completed according to the optimal weight, and a final seismic source positioning result is obtained based on an analytical solution. According to the method, the technical problem that the influence of abnormal values and wave velocity measurement errors on the seismic source positioning precision in a complex engineering environment is large is solved, and the fault tolerance and the real-time performance of seismic source positioning are improved.

The invention relates to an ultra-short-term photovoltaic power prediction method and system based on physical constraint and multi-source data space-time fusion, and the method comprises the steps: carrying out the feature extraction through ViT, and obtaining a cloud picture feature; performing feature extraction by using TCN to obtain time sequence features; fusing the cloud picture features and the time sequence features based on a cross attention mechanism to obtain predicted photovoltaic power, and calculating basic prediction loss; extracting an instantaneous motion vector field of the cloud picture data to calculate the acceleration, divergence and optical flow edge intensity of the cloud picture, and matching a grade coefficient corresponding to each physical quantity; on the premise that only the loss corresponding to the single physical quantity is activated, training is carried out, and the optimal weight coefficient of each physical loss is obtained; taking the optimal weight coefficient of each physical loss as an initial point of a search space, and obtaining an optimal combined weight coefficient through Bayesian search; and physical constraint loss is calculated, and training of the ultra-short-term photovoltaic power prediction model is realized in combination with the basic prediction loss.

The invention discloses an auditable asset allocation system and method based on time-varying risk preference estimation. The system comprises a data acquisition and preprocessing module; the time-varying risk preference estimation module is used for carrying out online Bayesian estimation on the risk preference time-varying hidden state by adopting a sequential Monte Carlo method based on the state space model; the dynamic asset allocation optimization module is used for taking the estimated value as an endogenous variable to be embedded into an optimization model to solve the optimal weight; and the compliance and auditing module is used for recording the full-link log in a tamper-proof manner and generating an interpretable report. According to the method, closed-loop optimization from dynamic risk preference estimation to asset allocation decision is realized, and auditing performance of the whole process is ensured.

The application discloses a fish coated product whole life cycle quality tracing method based on Internet of Things, and belongs to the technical field of food production, and specifically comprises the following steps: obtaining individual morphological parameter set and actual weight gain parameter set of each fish product to establish an individual coating powder characteristic database; extracting morphological feature vectors and matching optimal weight gain labels to construct a feedforward prediction model; extracting actual coating powder control parameters and slurry rheological parameters to form a deviation compensation parameter set; reversely propagating to calculate closed-loop feedback correction parameters and splicing the deviation compensation parameter set into a joint parameter correction set; training to generate an adaptive closed-loop control strategy model; obtaining coating process optimization instructions based on the morphological feature vectors of the current product; and generating equipment warning instructions when the numerical symbols of the coating process optimization instructions are the same and the difference between adjacent numerical values presents a continuous one-way change. The application solves the problems of uneven coating and quality fluctuation caused by constant process by time series correlation storage of individual level data.

The invention discloses a gas reservoir three-dimensional geological modeling method and system and a medium. The method comprises the following steps: acquiring multi-source data of a target work area; in combination with vertical seismic section data and virtual well data, depth correction is performed on the horizon after time-depth conversion, and a three-dimensional tectonic framework model is established based on the corrected horizon and fault data; fusing and generating a sand body space development probability trend body based on the seismic attribute body, the seismic inversion body and the determined optimal weight coefficients of the seismic attribute body and the seismic inversion body; a three-dimensional tectonic framework model is used as a space framework, a sand body space development probability trend body is used as a transverse soft constraint, a plurality of isoprobable lithofacies model initial implementation is established, and a three-dimensional lithofacies model is obtained; a three-dimensional distribution model of reservoir physical property parameters is generated; and importing the three-dimensional lithofacies model and the three-dimensional distribution model into a numerical simulator, performing historical fitting by using gas reservoir dynamic production data, analyzing fitting differences, and performing iterative modeling to obtain a high-precision and low-uncertainty three-dimensional gas reservoir geologic model.

This invention relates to a hysteresis modeling and loop-forming structured control method for piezoelectric actuators, comprising: constructing a piezoelectric actuator model using a Hammerstein structure; describing the uncertainty of the piezoelectric actuator model based on the v-gap metric, and setting open-loop amplitude constraints and the structure and parameter vector search domain of the controller based on the uncertainty description; setting initial weight functions, accuracy coefficients, and an initial controller, and constructing inequality constraints for the loop-forming system under the open-loop amplitude constraints; iteratively solving a first optimization problem and a second optimization problem under the inequality constraints until a preset iteration condition is met, obtaining the optimal weight function and the desired controller; if no solution to the optimization problem is obtained until the preset iteration condition is met, the open-loop amplitude constraints and the parameter vector search domain of the controller are reset, and the next round of iterative solution is performed until a solution is obtained. This invention achieves significant improvements in dynamic response, control accuracy, and robustness.

The application provides an MPC-LSTM air quality prediction method combined with adaptive weights, comprising the following steps: constructing a data set of air quality samples and preprocessing the data set; constructing an MPC-LSTM air quality prediction model; training the MPC-LSTM air quality prediction model based on the preprocessed data set; and obtaining predicted air quality data based on the trained MPC-LSTM air quality prediction model. The application combines a multi-scale parallel convolution fusion network, a long short-term memory neural network and an attention mechanism, realizes further mining of potential feature relationships of data, selects optimal weights, constructs an optimal air quality prediction model, and thus improves air quality prediction accuracy.

The invention discloses an infrastructure whole process data management method and system based on digital twin and block chains, and relates to the technical field of infrastructure project management, and the method comprises the steps: collecting original data of an infrastructure project, carrying out the preprocessing, obtaining normalized data, carrying out the encryption and hashing, and forming a binding data package for storage; extracting the stored binding data packet for decryption and verification to obtain normalized data, setting a mapping relation table to map the normalized data, constructing a digital twin model to simulate the mapped normalized data, and outputting simulation results for weighted summation to obtain a predicted risk score; based on the predicted risk score, constructing an objective function, using a hybrid optimization algorithm to iteratively update the objective function, generating an optimal weight set, and feeding back the optimal weight set to weighted summation to obtain a final risk score; the data management efficiency and the safety operation and maintenance level of the capital construction project are remarkably improved.

The invention provides a buried pipe network optical fiber sensing signal classification method and system by using an optimized and improved extreme learning machine. The method comprises the following steps: firstly, converting a vibration signal sample into a normalized one-dimensional matrix and generating a corresponding one-hot coding label; then network parameters are initialized, an optimization model with classification accuracy and recall ratio as objective functions is constructed, optimal parameters are searched through collaborative iteration of an exploration group, a mining group and an alternative group, the exploration group is responsible for expanding a search range around the whole space, the mining group is responsible for searching for a better solution between member ranges, and the alternative group is responsible for providing an optional solution; and finally, training a classifier by using the optimal parameter, and verifying the model performance on a test set. According to the method, the network weight is iteratively updated according to the rule, the purpose of quickly searching the optimal weight is achieved, and the recognition rate of the algorithm can be remarkably improved.

The application provides a space-time model direct positioning method and system based on medium and low orbit satellite fusion, comprising: position rough measurement; constructing a direct positioning scene and a satellite receiving data model; constructing a space-time model of satellite receiving data; deriving a maximum likelihood estimator target function based on a least square estimation criterion; introducing an optimal weight vector, solving the weight vector according to a minimum variance distortionless response criterion, and deriving a target function based on the minimum variance distortionless criterion; drawing a space spectrum diagram according to the target function value, performing spectrum peak search, and determining the position of a ground radiation source. The application reduces the search range and the calculation amount by using the rough estimation result of single satellite direction finding positioning to obtain the direct positioning area to be searched; the space-time direct positioning model based on the angle of arrival and the Doppler frequency shift is established by using Doppler information, high-precision positioning can be realized under low signal-to-noise ratio, and high resolution can still be realized when the ground radiation sources are very close.

The application provides a multimode blind equalization method based on a swallow line function and a variable fractional order gradient. The algorithm uses statistical characteristics of an equalizer output signal and a transmitting end signal to construct a multimode cost function based on a swallow line function, so that phase errors in a channel equalization process can be compensated without an additional carrier recovery loop. Subsequently, a variable fractional order gradient method is used to replace a traditional fixed order gradient updating mode to adaptively update equalizer weight vectors, so that optimal weight vectors are obtained. Compared with the prior art, the application can not only realize channel equalization and phase recovery, but also effectively reduce residual inter-symbol interference and bit error rate in a strong impulse noise environment, while the convergence speed is accelerated, and higher equalization efficiency and robustness are shown.

The application discloses a YOLOv5s light-weight sheep breed identification method and system based on knowledge distillation, and the identification method comprises the following steps: S1: collecting videos of multiple breeds of sheep inside a breeding farm; S2: extracting image frames in the videos as original images, and performing data preprocessing on the original images, and dividing the processed images into a training set, a verification set and a test set; S3: constructing a light-weight sheep breed identification neural network model based on knowledge distillation of YOLOv5s; S4: training the sheep breed identification neural network model by using the training set and the verification set, and obtaining optimal weights; and S5: inputting the test set into the trained sheep breed identification network model and evaluating the performance of the model. The method provided by the application utilizes knowledge distillation to transfer effective features learned by a teacher network with large parameters and high recognition accuracy to a student network, so that the recognition accuracy of the breed identification network is improved, and the purpose of small network model parameters and light weight is achieved.

This invention relates to the field of portrait generation technology, specifically to a method and system for adaptive portrait generation based on multi-dimensional fusion of hybrid factors. To address the problems of large fitting deviations and low scene adaptability in traditional portrait generation, this invention first decomposes the target user information into multi-dimensional α-sensory factors and... β The rational factor is used to extract overlapping feature points through dual computation and assign initial weights. Then, the initial weight factors are classified and core factors are selected. Through bidirectional convergence verification of dual computation fusion and natural index fitting, the optimal weight factors are obtained. Finally, the optimal weight factors are weighted with dimensional information to obtain multi-dimensional features. The core factors are matched to generate personality profiles. Then, the profiles are graded and output through multi-dimensional aggregation. The generated standardized graded profiles not only fit the real characteristics of the target users, but also accurately match the needs of different application scenarios, which greatly improves the practical application value and adaptability of the personality profile generation method.

This invention provides a convex optimization design method and system for eVTOL based on power kit selection, belonging to the field of vertical takeoff and landing aircraft design technology. The convex optimization design method includes: obtaining the parameters of the power kit and continuous flight dynamic parameters; constructing aerodynamic models, load characteristic models, battery discharge models, and efficiency models for the eVTOL cruise phase based on the parameters to determine constraints; constructing a convex optimization objective function that minimizes energy consumption per unit range based on the models; solving the convex optimization objective function using the interior-point method to obtain optimal weights and continuous parameters; and obtaining the optimal power kit and mass parameters of each system based on the constraints and optimal weights. This invention achieves simultaneous optimization of discrete power kit selection and continuous parameters under complex constraints, significantly improving optimization accuracy while meeting the strict safety and performance constraints of the entire eVTOL flight profile.

The invention discloses a method for predicting the optimal rotating speed of a motor by optimizing a neural network based on a whale optimization algorithm, and relates to the technical field of whale optimization algorithms. The method comprises the following steps: inputting data and an optimal rotating speed, and carrying out normalization processing on the data; designing a BP neural network topological structure, and initializing parameters of the BP neural network topological structure; initializing parameters of a whale optimization algorithm, improving the whale optimization algorithm by introducing a weight strategy and a nonlinear convergence factor, and optimizing the BP neural network by improving the whale optimization algorithm; outputting an optimal weight and a threshold value of the BP neural network, and carrying out a training test; and constructing a model for predicting the optimal rotating speed by optimizing the BP neural network. According to the method, the whale optimization algorithm is improved based on adaptive weight and nonlinear convergence, the optimization precision and the optimization speed are improved, errors are reduced, the BP neural network is optimized by using the improved whale optimization algorithm, and the accuracy of predicting the optimal rotating speed of the motor is improved.

The application provides a target selection and motion trend prediction method, device and terminal equipment, which comprises the following steps: acquiring a dangerous characteristic value in real time, and constructing a data set; adaptively learning a weight value corresponding to the dangerous characteristic value, acquiring an optimal weight value, and constructing a relationship model A; calculating an arithmetic mean of an optimal dangerous coefficient, comparing the size, and determining a target to be tracked; constructing a relationship model B between the dangerous characteristic value increment change of the target to be tracked and the PTZ increment change of a video monitoring device, and predicting the motion trend of the target to be tracked. Through multi-target adaptive screening of a radar-video linkage system, an adaptive learning model is constructed, which can be copied to a similar geographical environment scene, reduces the trial and error cost of subsequent construction, improves the user experience of the system, quickly and accurately locks the target to be tracked in multiple moving targets, and synchronously rotates the linkage video monitoring device to the predicted position, thereby improving the capture rate and accuracy of the video monitoring device on the target.

The present application belongs to the technical field of federated learning, and discloses a federated learning non-commutative conformal risk control method. The method mainly comprises two key modules: at the client level, a RACE kernel density estimation method based on local sensitive hashing is used to capture local data characteristics and generate a distribution sketch matrix; at the server level, the maximum entropy principle and the grid search method are used to solve the optimal weight of the sample and the minimum conservative value of the conformal risk control, so that the prediction set constructed by the client meets the given risk control requirements. In theory, it is proved that the risk of the method can be controlled within a given range, providing a solid guarantee for the effectiveness of the algorithm. In the experiment, compared with a variety of baseline methods on multiple real and synthetic data sets, the results show that Fed-Non-X-CRC performs outstandingly on different data sets.

The invention relates to the field of power systems, and discloses a power system operation risk assessment model construction method, a power system operation risk assessment model, equipment and a medium, and the method comprises the steps: obtaining historical operation data of a power system; according to the historical operation data, determining an evaluation index by a plurality of evaluation dimensions, and determining an initial weight of the evaluation index by adopting an analytic hierarchy process for the evaluation index; and carrying out particle swarm optimization on the initial weight to obtain an optimal weight of the evaluation index so as to obtain a power system operation risk evaluation model. According to the method, the evaluation indexes can be determined according to a plurality of evaluation dimensions, and the weight of each evaluation index is dynamically optimized, so that a model capable of accurately evaluating the operation risk of the power system is constructed.

The application discloses a differential privacy model training method based on a buffer mechanism, a medium and a system, wherein the method comprises the following steps: sampling a training data set to obtain small-batch training data; calculating a gradient value and processing the gradient value to generate a private gradient, which is applied to a current model parameter to obtain a preselected weight; sampling a verification data set to obtain small-batch verification data; calculating a loss change value, and performing clipping and noise adding processing on the loss change value to generate a noisy loss change value; judging whether the noisy loss change value is less than a preset rejection threshold; if yes, adding the candidate weight to the buffer; when the number of candidate weights is equal to a preset number threshold, determining the optimal weight according to the relative noisy loss change value, and updating the current model parameter based on the optimal weight; the method can effectively protect privacy, improve the updating quality of the model in the training process, and improve the accuracy of the final prediction result.

This invention belongs to the technical field of hydrometeorological extreme value theory, and provides an unbiased estimation method for rainfall probability distribution based on least squares fitting. Existing parameter estimation methods are numerous, each with its own specific applicable scope and advantages and disadvantages. The choice of method significantly affects the accuracy of generalized extreme value distribution parameter estimation. This invention comprehensively considers the advantages of five parameter estimation methods: Bayesian, GMLE, L-moments, MLE, and MPS. Based on least squares fitting, it determines the optimal weight coefficients of a single method, establishing an unbiased and efficient ensemble parameter estimation framework. By integrating multiple parameter estimation methods, the bias and variance of individual methods can be reduced, improving the robustness and reliability of parameter estimation, and providing a reference for accurately determining the rainfall frequency curve and design rainfall.

The invention provides a wafer asymmetric deformation mark alignment correction method and device, equipment and a medium. The wafer asymmetric deformation mark alignment correction method comprises the steps of obtaining structure parameters of an asymmetric deformation mark on a target wafer; wherein the structure parameters at least comprise a marked geometrical shape, a material attribute and process historical information; inputting the structure parameters into a pre-established asymmetric mark simulation model, and simulating to generate alignment position deviation values under each wavelength; and according to a pre-trained multi-wavelength optimal weight coefficient, performing weighted fusion on the alignment position deviation value corresponding to each wavelength, and outputting a corrected final alignment position of the target wafer to compensate a measurement error caused by asymmetric deformation of the mark. No matter what complex process steps are experienced by the mark, the corresponding corrected final alignment position can be generated in real time only by inputting the actual structural parameters of the mark, and large-scale experimental calibration does not need to be carried out again, so that self-adaptive response to various process conditions is realized, and the production efficiency of wafers is greatly improved.

The invention relates to the technical field of space-ground integrated power communication networks, and discloses a service suitability determination method for a space-ground integrated power communication network, which comprises the following steps of: constructing an index system, and further decomposing the index system into communication sub-indexes which can be directly quantized; calculating subjective weights of the communication sub-indexes based on a fuzzy analytic hierarchy process, calculating objective weights of the communication sub-indexes based on an entropy weight method, and introducing a game theory model to obtain an optimal weight combination of the subjective weights and the objective weights; classifying the power business, constructing a demand scoring model, and quantifying the demand of the power business for the communication sub-indexes; taking the communication sub-index demand as an ideal solution, setting a negative ideal solution with the lowest tolerance limit, and calculating the MARCOS utility degree and grey correlation degree of each communication sub-index; an optimization model is constructed based on the minimum information entropy principle to determine the combined weight of the utility degree and the grey correlation degree, the comprehensive utility degree is calculated, suitability sorting is carried out according to the comprehensive utility degree, and the accuracy is improved.