22 results about "Vector optimization" patented technology

The application relates to a merging vector optimization method in a mobile cell-free scenario, wherein the method comprises the following steps: generating and dividing a data sample set based on a pre-established full connection CF mMIMO network architecture; performing graph construction and initial graph embedding on random channel state information in a training set to obtain an initial graph embedding vector; performing forward propagation calculation on the initial graph embedding vector in combination with a pre-established graph neural network to obtain a current round merging vector; determining an initial graph neural network model when a training bit error rate meets a first bit error rate requirement; testing a test set to obtain a test bit error rate; determining a final graph neural network model when the test bit error rate meets a second bit error rate requirement; and deploying a final full connection CF mMIMO network architecture and performing network scheduling. Thus, the problems of high calculation complexity, low real-time performance and low reliability in the related art are solved, the calculation complexity is reduced, and the robustness, scalability and system fairness are improved.

This invention discloses a GNSS interference source localization method based on ADS-B data. The method involves acquiring ADS-B data, saving data points affected by GNSS interference, and, if no NIC is available, constructing ADS-B data with a NIC. The message data is then sorted by ICAO and timestamp. A diagonal matrix W is calculated based on the ICAO, timestamp, NIC, a table showing the correspondence between NIC and received interference power, and the hyperparameter value τ. A residual vector R is calculated based on the message data and the optimization vector, and the optimized vector is then output. This invention analyzes large-scale aviation ADS-B data, sorts and classifies the original message data to determine whether flights are subject to GNSS interference, and abstracts the interference source localization problem into a least-squares problem. An optimization function is constructed, and an iterative optimization algorithm is used to calculate the possible GNSS interference source locations and transmission power, thus more efficiently locating GNSS interference sources.

The invention discloses a representation vector optimization method, system and device for sequence data prediction and a medium, and particularly relates to the technical field of representation vector optimization. According to the method, a hidden variable is introduced into each variable, meanwhile, hidden state coding is carried out on original observation variables full of uncertainty, probability distribution obeyed by the hidden variables is estimated by using a neural network, and parameters of the distribution serve as more robust representation vectors of the variables; the method comprises the following steps: adaptively screening out an invariant feature code which is most stable in association with a predicted target from a variable sequence by using a gating mechanism so as to reduce the uncertainty of the invariant feature code; meanwhile, through an asynchronous loop optimization strategy, information gains provided by the node variables and the information variables in the information spreading process are continuously obtained, and representation information of the node variables and the information variables under information coupling is enhanced.

The present application relates to the technical field of forklift motor, and particularly relates to a forklift motor energy-saving operation method and system based on reluctance torque vector optimization, which collects motor operation parameters through a sensor, fuses multi-source dynamic characteristics by using a parallel time sequence feature extraction network, extracts working condition characteristics in combination with bidirectional domain adaptive transfer learning, determines the number of key factors based on a dynamic working condition correlation matrix, generates a motor control characteristic vector through convex optimization and probability density estimation, generates a control scene in combination with a reluctance torque model and a phase current coupling relationship, calculates an energy efficiency index based on torque response and energy consumption distribution matching degree, and solves optimal control parameters by using a multi-objective optimization algorithm. The parameters are input into a grouping collaborative control unit, control groups are divided according to phase space distribution, optimal reluctance torque vectors are calculated by using an energy consumption dynamic prediction model through an intra-group competition and inter-group cooperation mechanism, and high-efficiency energy-saving operation under low-speed large-torque and frequent start-stop working conditions is realized.

The application discloses a linear trajectory planning method for a space manipulator load operation task and relates to the technical field of space manipulator trajectory planning. A multi-objective optimization model is designed, and optimization targets include straightness vector optimization, obstacle avoidance optimization and jerk optimization; a trajectory planner is designed, an individual searched is a middle node configuration of a quintic polynomial trajectory, a quantum state coding mode is adopted to code the configuration of the manipulator, a QBSO algorithm is provided, a complete trajectory of the manipulator is sequentially connected by a plurality of middle node configurations, the next node configuration is searched according to a current configuration position and a target trajectory each time, and finally, a joint trajectory connected according to a quintic polynomial equation by n middle node configurations is formed. By designing a mathematical coding mode of the quintic polynomial trajectory, designing a linear trajectory planner framework based on a sampling algorithm and designing a straightness vector constraint model, an optimal trajectory meeting constraint requirements can be generated.

The application discloses a kind of flat bottom end face milling cutter tool shaft vector optimization method, tool shaft vector optimization is considered from processing interference, cutting force, rotation axis angular displacement and angular velocity etc. Multiple factors comprehensively. First, the improved space partition tool shaft vector global interference detection method based on surface principle is proposed, whether the interference detection algorithm is used to judge tool to occur processing interference and carry out interference-free tool shaft vector space calculation;Second, the relationship model of end milling cutter tool shaft vector and cutting force change is established, and the tool shaft vector optimization is carried out with the tool cutting force optimization as the target;Third, using Dijkstra algorithm, an optimal path algorithm-based machine tool rotation axis angular displacement and angular velocity optimization model is proposed;Using the optimization algorithm for complex surface machining, the machining quality of the surface workpiece after tool shaft vector optimization is obviously improved compared with that before tool shaft vector optimization.

The application discloses a kind of legal dossier intelligent retrieval method and system based on multimodal heterogeneous big model, comprising: obtaining original legal dossier data containing text, image and table, carrying out multimodal analysis and structured preprocessing;Through multimodal heterogeneous big model, each modal depth feature is extracted and is fused using multimodal cross attention mechanism;Case element sparse graph is constructed and is optimized by optimization algorithm;Robust learning real value multi-index model algorithm is applied to weight adjustment and vector optimization;Through random recovery sampling technology, candidate case subset is screened and multidimensional matching calculation is carried out;Similar historical case list is visualized and interactive analysis and processing are carried out.The application improves the accuracy and efficiency of legal dossier intelligent retrieval by multimodal fusion understanding, sparse graph optimization and robust learning technology, provides accurate case reference and decision support for legal workers.

The invention discloses a virtual resource dynamic scheduling method and system in a hybrid cloud environment, relates to the technical field of resource scheduling, and translates an abstract service demand protocol into a scheduling strategy object which can be understood and activated by a machine by capturing and analyzing the abstract service demand protocol. Meanwhile, multi-dimensional fusion is carried out on the resource state, the multi-cloud price model and the real-time service KPI flow in the underlying mixed cloud environment to form a globally unified system state view. On this basis, an adaptive mechanism based on a control loop is introduced, and continuous weight vector optimization and generation are carried out by taking an activation strategy analyzed by an intention as a guide and taking a multivariate optimization index as a decision factor capable of being dynamically deployed. And finally, the scheduler performs intelligent weighted scoring on the new task request by using the group of dynamic weights, and makes an optimal placement decision. Therefore, an intelligent decision closed loop from business intention definition to resource scheduling execution can be realized, so that the resource allocation efficiency and business agility of the hybrid cloud environment are improved.

The present application relates to a kind of demand side response potential evaluation and benefit optimization system based on load feature clustering, belong to electric power system and its automation technical field.Therein, the method includes: obtaining industrial park heterogeneous load data and preprocessing, extract load time dimension feature by wavelet transform and fuse static attribute, generate standardization load representation vector after feature screening;Based on vector optimization clustering core point, build three-dimensional clustering evaluation system, construct user portrait in combination with production constraint;Quantify production rigidity and flexible constraint, construct response willingness degree model and dynamically correct;Build multi-time scale optimization framework, scale to interface grid data, output optimal response strategy.Load feature accurate representation and response potential scientific evaluation are realized, the pertinence and feasibility of demand side response strategy are improved, grid dispatching adaptability and user benefit maximization are considered, and technical support is provided for industrial park demand side response large-scale promotion.

The invention discloses a concealed wireless communication beam forming optimization method under limited character input, and belongs to the technical field of wireless communication. According to the method, a transmitting power and beam forming vector optimization strategy for jointly optimizing a transmitting end under different modulation modes is constructed, and a positive semidefinite relaxation algorithm is designed on the basis of utilizing lower bound approximation of an objective function: an initial non-convex objective function and a hidden constraint are equivalently converted into a convex form which is easier to process; and solving by using a CVX toolkit in MATLAB to obtain a design which enables the problem to be optimal in beam forming vector. Furthermore, in order to reduce the calculation complexity, the invention provides a low-complexity optimization strategy, and a simulation result shows that the method can remarkably reduce the calculation complexity while meeting the covert communication performance.

The invention discloses a multi-modal abstract generation method based on dynamic retrieval and self-adaptive query, and relates to a multi-modal data processing technology, and the method comprises the following specific steps: preprocessing a text sequence and a video sequence, extracting features, and projecting the features to the same dimension; constructing a knowledge base, and retrieving and enhancing to obtain enhanced features; fusing the text and video features into an initial mixed query vector and optimizing the initial mixed query vector; then constructing a joint feature sequence, inputting the joint feature sequence into a multi-layer multi-mode Transform, and outputting fusion features; and finally, calculating a correlation score, and generating a multi-modal abstract according to a Top-K strategy. Through multi-modal feature unification and knowledge base enhancement, a modal barrier is broken, and feature pertinence is improved; in combination with query vector optimization and multi-layer cross-modal fusion, interaction between modals is enhanced, high correlation of abstract elements is ensured, the problems of information alignment deviation and content one-sided are effectively solved through the design, the abstract quality and practicability are improved, and the method has wide popularization value.

The invention discloses a workpiece three-dimensional model construction and processing method based on binocular vision detection data, and the method comprises the steps: firstly carrying out the multi-angle synchronous scanning of a workpiece through employing a combination of a binocular camera and a line laser, and obtaining the multi-view three-dimensional point cloud data of the surface of the workpiece; and constructing a density sensing type dynamic registration framework, and performing distance weighted fusion and outlier elimination on the registered multi-view point clouds in an overlapping region to generate a complete workpiece point cloud. A feature retention type point cloud direct modeling strategy is adopted, normal vector optimization is performed on a down-sampling point cloud, key feature marks are manufactured, a lightweight three-dimensional model is generated through color coding rendering, and a large amount of calculation of traditional surface reconstruction is avoided. And finally, based on model coordinate system calibration and point cloud fitting optimization, extracting functional characteristic parameters of edge straightness, datum plane flatness and hole system position degree, and completing workpiece quality evaluation. According to the method, the point cloud processing speed and the key feature analysis precision are improved, and the real-time requirement of industrial online detection is met.

The application belongs to the field of tool axis vector trajectory planning, and particularly discloses a tool axis vector optimization domain planning method and system considering workpiece machining flexibility, which comprises the following steps: establishing a workpiece finite element proxy model, loading cutting force at each tool point to obtain equivalent flexibility of each tool point, and determining a maximum flexibility point; obtaining a maximum cutting force according to a maximum deformation allowed by the workpiece and the equivalent flexibility of the maximum flexibility point, and combining a pre-established cutting force prediction model to obtain a constraint condition; based on the constraint condition and a preset range of machining parameters, performing multi-objective optimization of the maximum flexibility point in coupling with machining parameters to obtain a critical tool axis vector of the maximum flexibility point under critical machining parameters; determining critical tool axis vectors of other tool points under the critical machining parameters to obtain a tool axis vector optimization domain. The application can calculate the tool axis vector optimization domain division within the machining parameter limited range under the influence of force-induced errors and machining efficiency, and provide guidance for tool axis vector trajectory planning.

The application provides an MCR-WPT system plane coil structure parameter optimization method based on an improved four-vector optimization algorithm; an S-S topology compensation structure network is analyzed, and the energy transmission efficiency of the MCR-WPT system is derived; the influence of the coil physical structure on the coil characteristic parameters is analyzed, and the self-inductance value of the coil is calculated; the mutual inductance coefficient value between the coupled coils is calculated; the alternating current resistance is calculated; the coupling coefficient between the coupled coils is calculated, the frequency splitting is analyzed to establish a constraint condition; the optimization variables are determined, and the energy transmission efficiency is taken as the objective function value and the optimization target; the constraint condition is set according to the critical coupling coefficient, and the parameter range is adjusted according to the constraint condition; and the improved four-vector optimization algorithm is used to optimize the energy transmission efficiency, that is, the objective function, according to the coil setting parameters.

The invention provides an InSAR deformation anomaly extraction method and system based on multi-modal fusion and vector optimization, and relates to the technical field of artificial intelligence and geological disasters, and the method comprises the steps: obtaining regional multi-source remote sensing data, calculating earth surface deformation phase data, topographic relief, gradient, vegetation index, water body index and soil index, and carrying out the calculation of the earth surface deformation phase data, the topographic relief, the gradient, the vegetation index, the water body index and the soil index; re-sampling to obtain multi-source remote sensing full-waveband data; performing manual interpretation on the InSAR surface deformation abnormal range of the region based on expert knowledge to obtain vector data and binary mask data of the InSAR surface deformation abnormal range; constructing a double-branch multi-modal feature fusion network, and extracting the surface deformation anomaly information of the InSAR based on the binary mask data and the multi-source remote sensing full-waveband data; and constructing a dynamic vector optimization network, performing boundary optimization on the InSAR earth surface deformation abnormal information based on the multi-source remote sensing full-waveband data, and dynamically converting the information into a vector result. According to the invention, seamless connection from the remote sensing image to the GIS vector is realized.

The invention provides a deep learning aided rotation vector optimization method and system for a high dynamic environment, the method is suitable for attitude calculation in the high dynamic environment, the method mainly comprises two parts of construction of a rotation vector prediction model based on Transform and data fusion, according to the strong correlation and the nonlinear coupling relationship between the angular velocity and the rotation vector, the rotation vector prediction model based on the Transform is constructed, and the rotation vector prediction model based on the Transform is obtained. A three-axis angular velocity output by a gyroscope is used as a Transform input, and a rotation vector is used as an output to deeply excavate a nonlinear coupling relationship between the angular velocity and the rotation vector; according to the method, a state equation is constructed according to a rotation vector differential equation, an obtained Transform prediction result is used as an observed quantity design measurement equation, deep fusion of deep learning and a traditional rotation vector is realized by adopting UKF, the influence of non-exchangeable errors can be effectively reduced, and the attitude calculation precision is improved. The method is suitable for attitude calculation in a high-dynamic environment.

The present invention relates to the field of three-level inverters, and more specifically discloses a model-free, high-robustness predictive control method and system for a three-level SNPC inverter, which includes: performing real-time sampling and coordinate transformation on the inverter to obtain current and voltage components in a two-phase stationary coordinate system; establishing a hyperlocal model, unifying and equating system parameter deviations and external disturbances into concentrated disturbances, designing a full-order sliding mode observer and performing real-time estimation; calculating a reference voltage vector based on disturbance estimates and a reference current obtained by extrapolation; selecting the optimal dual vector combination by spatial vector optimization, determining the action time, selecting redundant small vectors according to the midpoint voltage deviation, and generating a PWM drive signal.

The application discloses a covert wireless communication beamforming optimization method under limited character input and belongs to the technical field of wireless communication.The application constructs a joint optimization strategy of the transmission power and the beamforming vector optimization strategy of the transmission end under different modulation modes, designs a semi-positive definite relaxation algorithm on the basis of using the lower bound approximation of the objective function, equivalently converts the initial non-convex objective function and the covert constraint into a more easily handled convex form, solves by using the CVX tool package in MATLAB, and obtains the design of the optimal beamforming vector of the problem.Further, in order to reduce the calculation complexity, the application proposes a low-complexity optimization strategy, and simulation results show that the method can significantly reduce the calculation complexity while meeting the performance of covert communication.

The application discloses a virtual resource dynamic scheduling method and system in a hybrid cloud environment, and relates to the technical field of resource scheduling. The method captures and analyzes abstract business requirement specifications, translates them into scheduling strategy objects that can be understood and activated by machines. At the same time, the resource state in the underlying hybrid cloud environment, multi-cloud price models, and real-time business KPI flows are multi-dimensionally fused to form a globally unified system state view. Based on this, an adaptive mechanism based on a control loop is introduced, with the activated strategy derived from the intention analysis as the guide, and multi-element optimization indicators as the decision factors that can be dynamically allocated, for continuous weight vector optimization and generation. Finally, the scheduler uses the dynamic weights to intelligently score new task requests and make optimal placement decisions. In this way, an intelligent decision-making closed loop from business intention definition to resource scheduling execution can be achieved, thereby improving the resource allocation efficiency and business agility of the hybrid cloud environment.

The invention discloses a multi-quadratic constraint flexible attachment system optimal state estimation method, and belongs to the field of deep space exploration. The implementation method comprises the following steps: constructing a node kinetic equation, an observation equation and a state constraint equation of the flexible lander; the minimum mean square error is used as a performance index, an optimal filter is designed as a prediction-correction structure, and a minimum mean square error estimation problem under state constraint is converted into a filtering gain optimization problem. An original optimization problem is equivalently decomposed to obtain three sub-optimization problems. And for an equality constraint optimization problem, constructing a corresponding Lagrange function, and solving a KKT conditional solution of the Lagrange function. And determining an optimal solution by judging whether the corresponding Hessian matrix is positive definite or not. For an inequality constraint optimization problem, an unconstrained part and a constrained increment part in a filtering gain are separated, and an increment vector is introduced to convert the filtering gain optimization problem into an increment vector optimization problem. Solving an incremental vector optimization problem by adopting a sequence convex optimization method; and updating the optimal state estimation of the flexible attachment system.

The application discloses a kind of based on digital image's modeling optimization processing method and system, it is related to three-dimensional modeling technical field, including using lightweight Vision Transformer extraction global feature, using Harris corner point detection and identification feature point, through Laplace decomposition generates multi-scale feature map, combined with DeepLabv3+ Generation semantic segmentation chart, constructs hybrid Poisson equation, carries out solving optimization, through Marching-Cubes algorithm generates smooth three-dimensional surface model, adjusts initial contour, combined with dynamic change vector optimization, by completing, denoising, simplifying, quantizing and compressing processing, generates final optimization model.The application is enhanced feature extraction quality by standardization processing and Vision Transformer, improves initial model precision by multi-scale fusion and semantic segmentation, realizes efficient refined modeling by hybrid Poisson optimization and multi-scale adjustment, generates high-quality and low-complexity three-dimensional model.

The invention provides a pre-process-considered heavy industry product assembly scheduling optimization method, which comprises the following steps of: constructing a scheduling model in which a processing stage and an assembly stage are coordinated, part processing completion is taken as a pre-condition for starting a corresponding assembly body, the assembly operation is executed by a worker team and group, and the scheduling model comprises a neat set constraint and a team and group configuration constraint; the neat set constraint ensures that any assembly starts to be assembled only after all required parts are machined; setting a multi-objective optimization function including the minimum maximum completion time, the minimum machining energy consumption and the minimum chemical worker team process time balance rate; solving the scheduling model by adopting an improved non-dominated sorting genetic algorithm, wherein the improvement comprises an elite pool strategy, a virtual target vector optimization strategy and a differential evolution operator introduction strategy; and outputting an optimal scheduling scheme including a part processing sequence, machine distribution, an assembly body assembly sequence and team personnel configuration.