48 results about "Algorithmic efficiency" patented technology

The invention discloses a multi-target road transportation path planning method and system based on multiple parameters A * and NSGA-II, and belongs to the technical field of road transportation path optimization. The method comprises the following steps: constructing a large-scale road network directed graph structure; establishing a satellite orbit model and a risk distribution model, calculating a dynamic risk value of a roadside, and forming dynamic road network data; a three-stage initialization strategy is adopted to generate an initial population, transportation time and reconnaissance risks are taken as double targets, evolutionary optimization is carried out through an NSGA-II algorithm, and a Pareto optimal path set is output. The system comprises a corresponding road network module, a risk module and a planning module. According to the method, the technical problems of multi-target conflict, low initial population quality and satellite risk avoidance of a traditional method are solved, and the timeliness, the concealment and the algorithm efficiency of path planning are remarkably improved.

The invention discloses an unmanned aerial vehicle cluster task allocation method based on a large language model optimization genetic algorithm, and belongs to the field of computers. The method comprises the following steps: setting a specific chromosome coding mode; generating a multi-constraint initial population; calculating fitness to quantify the advantages and disadvantages of individual genes of the population; when the optimal individual meets the requirement or the maximum iteration round is reached, ending; retaining the optimal individual as a filial generation; generating a batch of new filial generation individuals by the large language model, and fusing the new filial generation individuals with the current filial generation population; calling an optimized large language model to analyze individual chromosome semantics, and outputting an evolutionary potential score; obtaining an individual comprehensive selection probability by integrating the fitness and the evolution potential score, and executing a selection operation; and selecting individuals based on the individual comprehensive selection probability to carry out crossover and mutation operation to generate offspring. The large language model is embedded into the core link of the genetic algorithm, and the algorithm efficiency is improved by improving the population diversity of the genetic algorithm in the unmanned aerial vehicle cluster task allocation scene.

The invention discloses a heavy-load freight direct-current electric locomotive on-board platform and a train operation curve calculation method. Step-by-step iterative optimization is carried out through a dynamic programming algorithm. Presetting operation steps and step lengths, and defining a selectable action set in each step length; and with the current state of the train as a starting point, evaluating each action in the action set in each iteration step, and calculating the next train state which may be transferred to. And evaluating all possible train state changes by adopting a target cost function, sorting according to the cost values from small to large, and preferentially selecting the action with the minimum cost value for subsequent calculation. And if all the action cost values of the current step exceed the limit, backtracking to the previous iteration step for reselection. And after iteration to the maximum step number, storing all train state changes, and finally obtaining an optimal train operation curve under the constraint of the target function. The method optimizes a control strategy through a dynamic programming algorithm, improves prediction precision, optimizes action selection, accelerates convergence, reduces redundancy, and improves operation stability, safety and algorithm efficiency.

The invention discloses a distributed heterogeneous flexible flow shop batch processing scheduling method and system, relates to the technical field of distributed production scheduling in the manufacturing industry, and aims to solve the problems that an existing scheduling method is not comprehensive in constraint consideration, poor in energy consumption optimization and low in algorithm efficiency. According to the method, a mixed integer linear programming model containing multiple constraints such as release time and sequence-related preparation time is constructed, a learning-assisted dual-objective co-evolution framework is established, and the maximum completion time and the total energy consumption are synchronously optimized by combining mixed initialization, global-local search collaboration, decision reinforcement learning operator selection and a collaborative energy-saving strategy. The release time, the sequence-related preparation time, the inter-stage transportation time and the batch processing scheduling are simultaneously considered in the distributed heterogeneous flexible flow shop scheduling for the first time, the established mixed integer linear programming model better fits the actual production scene, and the method fits the actual production scene, is good in energy consumption optimization effect and can be adapted to the non-ferrous metal metallurgy aluminum production process.

The invention belongs to the technical field of power grid dispatching, and discloses an energy system low-carbon dispatching method and device based on deep reinforcement learning, and the method comprises the steps: changing the input structure of a decision network through the extraction of Riemannian flow pattern features and the analysis of group distribution; the input of a global decision network in the prior art is changed from environment state information to global state information and streaming feature group distribution and prediction group distribution at the next moment. The input of a single local decision network in the prior art is changed from local state information to flow pattern feature group distribution, predicted group distribution at the next moment and a corresponding unit state; according to the method, the input dimension of the decision network and the complexity of decision search are remarkably reduced, so that the algorithm efficiency is effectively improved, and rapid scheduling of the energy system is realized.

This invention provides an iterative measurement device and method for the surface shape of a cylindrical axially stressed component, belonging to the field of component measurement technology. It solves the problems of low measurement accuracy, single measurement method, low algorithm efficiency, and insufficient precision of existing measurement devices. This measurement device can detect the perpendicularity of the cylindrical axially stressed component, ensuring its vertical placement. It can stably adjust the height of the probe detection mechanism and the laser measurement mechanism, maintaining stable positions without wobbling, thus satisfying the requirement of measuring the outer contour of the cylindrical axially stressed component's surface twice. By controlling the main unit in conjunction with the probe detection mechanism and the laser measurement mechanism, it achieves rapid and accurate measurement of the cylindrical axially stressed component's appearance shape. The measurement method of this invention does not cause errors in the measured specimen's appearance due to the irregular shape of the track itself. Utilizing an iterative algorithm, it can achieve precise measurement of the appearance of the cylindrical axially stressed component and the track.

The application discloses a kind of based on the energy-saving method of urban rail comprehensive system optimization of multi-vehicle cooperation, comprising the following steps: according to basic input data, establish train-line-net-graph comprehensive system coupling model;Establish the time-space-speed three-dimensional state network of train operation;According to train motion characteristics, determine the train control operating mode transfer rule between adjacent time stages in time-space-speed three-dimensional state network;The train state under each time stage of time-space-speed three-dimensional state network is reduced dimension, and effective state space domain is obtained;According to multiple body dynamic programming algorithm, train-line-net-graph comprehensive system coupling model, train control operating mode transfer rule and effective state space domain, output optimal collaborative multi-train three-dimensional trajectory.The application can significantly reduce the iteration number of tidal flow calculation, greatly improve the algorithm efficiency, can simultaneously optimize the running process of all trains in multi-train system, realize system state optimization.

The invention provides a scheduling method for dual-objective optimization of a multi-agent flow shop under time-of-use electricity price, and relates to the technical field of intelligent manufacturing and production scheduling. According to the method, unified modeling is carried out on the total power cost TEC and the total customer completion time TCTA, and a dual-target mixed integer programming model is constructed; a double-node decision tree structure is constructed, and synchronous processing of task sorting and processing time interval distribution is achieved; analyzing the structural characteristics of the multi-agent flow shop under the constraint of time-of-use electricity price, designing a series of inequality constraints and multiple pruning rules, and compressing a search space on the premise of ensuring the integrity of a feasible solution; the method comprises the following steps of: performing dual-objective cooperative solution, embedding a branch and bound algorithm into an improved epsilon-constraint framework to perform iterative solution, accurately obtaining a Pareto frontier, and providing multiple optimal scheduling schemes balanced between cost and service quality for decision makers. According to the method, dual decisions of task sorting and cycle allocation can be realized, and the problems of decision model splitting, low solving algorithm efficiency and the like are solved.

The application discloses a kind of incremental unsupervised domain adaptive image recognition methods based on gradual calibration, comprising: obtaining unmarked target data set and labeled source data set;Using source domain pre-training model to predict target data set to generate target class, and generate pseudo-label for the sample in target data set, store target class and pseudo-label in memory library;Target data set is divided into sample with high confidence and sample with low confidence, and target model is trained in conjunction with source data set;And a variety of constraint losses are applied in the training process to carry out target class level calibration and target level calibration simultaneously;After each iteration training, update the data in memory library until the loss function converges;Using the trained target model to process the image to be identified, the final recognition result is obtained.The method reasonably utilizes the knowledge from the source field, balances and mitigates the relationship between negative transfer and catastrophic forgetting, and improves the algorithm efficiency and accuracy.

The invention relates to a laser scanning-linear growth depth synergistic pre-welding plate splicing seam visual center positioning line acquisition method, which comprises the following steps of: aiming at a pre-welding plate, acquiring laser scanning data of global scanning of a laser scanning module in real time, synchronously acquiring a visual detection image of a visual detection module, and acquiring a visual center positioning line of the pre-welding plate; obtaining global seed points by taking the scanning result as a dominant and a visual detection image as an auxiliary; determining an initial position of the seed point in the visual detection image based on the global seed point, carrying out dense growth by taking the initial position as a starting point, and carrying out cyclic tracking and repairing on a generated breakpoint to obtain a splicing seam selection point; and on the basis of the abutted seam selection points, screening processing is performed firstly, and then composite iteration trimming fitting processing is performed to obtain an abutted seam center positioning line. According to the method, the technical problems of insufficient ultra-narrow slit positioning precision, weak complex interference resistance, difficulty in breaking through meter-scale long-range one-time detection, poor algorithm efficiency and generalization adaptation and the like can be solved.

The invention relates to a block cipher algorithm architecture, an algorithm calling method and device and electronic equipment. The algorithm architecture comprises a quantum security layer which is used for taking a quantum true random number generator as a key entropy source and generating an unpredictable initialization vector and a dynamic key seed based on a quantum optical effect; the hybrid encryption layer is used for encrypting a session key by adopting a hybrid encryption mechanism of post quantum cryptography and SM4 and using a public key algorithm based on lattice cryptography, generating a master key through an anti-quantum algorithm, and generating a sub-key sequence in combination with an SM4 round key expansion algorithm; the core algorithm layer is used for optimizing a round function, dynamically generating an S-box replacement table based on quantum random numbers through an implicit coding technology of dynamic S-box generation and white-box protection, carrying out implicit coding on XOR and shift operations in the round function and injecting redundant noise data; and the protocol adaptation layer is used for integrating a dynamic defense protocol. The encryption security can be improved, and the algorithm efficiency can be remarkably improved.

This application relates to the field of performance testing technology, providing an adaptive sampling method and apparatus for mixed responses based on uncertainty reduction. By constructing a Gaussian process model of multiple types of mixed responses, it models and quantifies the uncertainty of these responses and their potential correlations, thereby solving the problem of difficulty in constructing surrogate models for mixed responses. By generating a reduced candidate sample set based on ROI, it screens experimental samples, improving algorithm efficiency. Adaptive sampling based on the uncertainty reduction of the forward error of the mixed response is used to comprehensively address the uncertainty of the predicted variance and the uncertainty of the predicted mean. The sampling process follows a dynamically adjusted boundary-variance-distance criterion, ensuring the auxiliary role of the boundary-variance-distance criterion in the initial sampling stage. As the number of samples increases, the model stability improves, and the selection of samples gradually becomes dominated by the uncertainty reduction of the forward error of the mixed response with forward capability.

The invention discloses a motion compensation imaging method based on Radon transformation and an SWO algorithm, and belongs to the technical field of laser radar signal processing. The method comprises the following steps: firstly, carrying out dechirp receiving and range pulse compression on an ISAL echo signal; detecting an energy peak value by using Radon transformation, and roughly estimating a target radial speed; constructing a local constraint search space based on the rough estimation speed; in the constraint space, the SWO algorithm is utilized, the image contrast is used as a fitness function, and speed and acceleration parameters are finely searched; and finally, constructing a full-aperture phase compensation function by using the optimal parameters for imaging. The problems that a traditional cross-correlation method is low in precision, and a direct optimization method is slow in iteration and prone to falling into local extremum are solved, and the imaging quality and the algorithm efficiency are remarkably improved.

The invention discloses a method and system for predicting the residual life of a fuse and a storage medium, and relates to a method for evaluating the life of the fuse, and the system comprises a data collection module, an SVM model building module, a model algorithm efficiency analysis module and a decision tree algorithm optimization module. The data is mapped to a space with a higher dimension, and an optimal hyperplane is found to divide different categories, so that prediction is carried out; meanwhile, the SVM has good generalization ability and robustness, a good prediction effect can still be obtained in the face of a small number of training samples and noise, and beneficial effects are obtained from historical data of the fuse; when the speed of the SVM algorithm is slow or the prediction precision is low, the decision tree algorithm is used to optimize the SVM algorithm, and the speed and the prediction precision of the model are improved.

The invention belongs to the technical field of dual-channel cross-correlation noise cancellation, and particularly relates to a dual-channel-based cross-correlation noise cancellation method and application thereof. The first channel and the second channel are mutually independent; an output signal of the tested equipment is divided into two paths of identical signals through the power divider, one path enters the first channel, and the other path enters the second channel; and the first channel signal and the second channel signal enter a signal processing unit, and are subjected to convolution cross-correlation and dynamic average processing, so that completely uncorrelated noise items are mutually counteracted, and output containing a target signal is obtained. According to the two-channel cross-correlation noise cancellation method, two channels are arranged, and signal processing is carried out by adopting convolution cross-correlation and dynamic averaging, so that the algorithm efficiency is improved, the signal quality is improved, and the method is suitable for more complex noise environments.

The invention discloses a lightweight video super-resolution method for a resource-constrained scene, and belongs to the field of video super-resolution reconstruction. The method comprises the following steps: constructing a training sample set; constructing a lightweight network model fused with an improved attention mechanism; training the lightweight network model by using the training sample set; deploying the trained lightweight network model at a video receiving end; and the video receiving end receives the low-resolution video sequence and the down-sampling multiple sent by the video sending end, the down-sampling multiple is used as a reasoning parameter of the lightweight network model, the low-resolution video frame is used as network reasoning input, a network reasoning process is started, and a reconstructed high-resolution video frame sequence is output. According to the method, the complexity of a super-resolution reconstruction algorithm is reduced, the parameter quantity of the algorithm is reduced, the algorithm efficiency is improved, and the real-time communication quality can be guaranteed in a resource limited scene.

The application discloses a power line detection method and system based on a maximum tree and graph signal processing, acquires original image data; according to the numerical law of the RGB channel of the original image data, a color filter is constructed, the color filter is used for retaining gray pixels; the original image data is input into the color filter to obtain a gray image; a Maxtree model is constructed, node attribute signals of the Maxtree model are set, and power line image data is obtained by segmenting the power line and background noise according to the Maxtree model; and a straight line equation of the power line is obtained through graph signal processing. The method utilizes the unique structure of the Maxtree and a graph signal processing method to realize power line detection, and the color difference between the power line and background noise is filtered to remove noise, improve algorithm efficiency, reduce memory loss, basically solve the misjudgment condition, and reduce the missed detection of the power line.

The invention relates to a public-water combined transport inventory path optimization method based on adaptive large-scale neighborhood search. The method comprises the following steps: S1, constructing an inventory routing decision model; s2, constructing a transportation volume model; s3, constructing an initial solution; s4, an operator selection mechanism; s5, repairing a judgment mechanism; s6, designing an operator; s7, a solution accepting strategy; s8, an algorithm stopping condition; according to the public-water combined transportation inventory path optimization method, the public-water combined transportation scene of the commercial vehicle, the route scheme of comprehensive transportation and inventory management are considered for overall optimization, and the inventory path optimization problem is divided into a transportation route optimization module and a transportation volume optimization module; a transportation route optimization module with high solving difficulty of an accurate algorithm is solved by using a heuristic method, and a transportation volume module with relatively easy solving is solved by using the accurate algorithm, so that the efficiency and quality of the algorithm are ensured.

The application provides an optimization processing method of a PCB panel, comprising: obtaining a first pattern of a panel target, and rotating the first pattern based on a discrete rotation angle set to construct a single pattern rotation set; obtaining a candidate combined pattern based on pattern splicing of the single pattern rotation set, and screening based on a multi-index evaluation system to construct an optimized combined pattern set; merging the two sets to construct a to-be-arranged pattern set; constructing a layer set corresponding to a mother board based on the mother board type and the to-be-arranged pattern set; determining an extended to-be-arranged pattern set based on the to-be-arranged pattern set and the layer set; constructing a mixed integer linear programming model with the extended to-be-arranged pattern set as input and the maximum mother board material utilization rate as output under the condition of meeting the preset constraint condition, so as to optimize the layout of the PCB panel. The application can significantly reduce the search space, avoid inefficient single arrangement attempts, improve the algorithm efficiency, and explore a compact layout that cannot be realized by single piece layout, thereby improving the space utilization rate.

The invention provides a navigational speed optimization method and system based on a ship energy consumption prediction model, and relates to the technical field of industrial intelligent optimization. The method comprises the following steps: constructing a ship energy consumption prediction model; based on the ship energy consumption prediction model, establishing a navigational speed optimization model; and an SLSQP algorithm is adopted to solve the navigational speed optimization model. According to the invention, dynamic coupling of environmental factors is realized; and an SLSQP algorithm is adopted to convert a nonlinear optimization problem into a quadratic sub-problem for iterative solution, so that the calculation efficiency is remarkably improved, the blank of an existing method in the aspects of environmental adaptability, algorithm efficiency and real-time performance is filled up, and a key tool is provided for energy conservation and emission reduction of the shipping industry.

This application relates to the field of autonomous driving technology, and particularly to a pose estimation method, device, vehicle, and storage medium. The method includes: acquiring laser point cloud data of the vehicle's surrounding environment; filtering ground point cloud data from the laser point cloud data, dividing the ground point cloud data into grids, downsampling the point cloud data in each grid, and calculating a score for the point cloud data in each grid; based on the scores of the point cloud data in each grid, filtering out grid point cloud data with scores greater than a preset value, and calculating the pose of the LiDAR based on the grid point cloud data. Therefore, by downsampling the grids, the algorithm efficiency is improved, and statistical analysis from multiple dimensions can effectively filter out noise points on the ground. Furthermore, by calculating the LiDAR pose using the filtered grid point cloud data, the problem of unstable ground pose estimation and large errors in different scenarios is solved, ensuring high accuracy in the fitted ground.

This invention discloses a passive localization path planning algorithm for UAV swarms based on analytical global CRLB, comprising the following steps: initializing the target, UAVs, a position information uncertainty model, and measurement parameters; calculating the initial position information of the target using the Chan-TDOA algorithm; handling the uncertainty of the target position information by calculating the integral average of CRLB in the position information uncertainty region (global CRLB); and finding the optimal flight path point of the UAV swarm at the next moment by minimizing the positioning CRLB value at each moment, thereby obtaining the optimal flight trajectory and positioning accuracy. This invention incorporates the uncertainty of the target position information, uses the global CRLB algorithm to determine the flight position of the UAVs at the next moment, and derives and designs an analytical calculation method for global CRLB under a certain uncertainty distribution, greatly improving the algorithm efficiency and adapting to real-time requirements. This algorithm can adapt to complex electromagnetic environments, effectively optimize the flight trajectory of UAV swarms, and significantly improve the accuracy of target localization.

This invention discloses a path planning method and system based on an improved teaching and learning algorithm, relating to the field of autonomous navigation technology for intelligent robots. The method includes map construction and path node constraints. The improved teaching and learning algorithm is initialized based on the map and path node constraints, using Dijkstra's algorithm to generate all initial paths. Teacher-stage path updates are performed based on all initial paths, yielding teacher-stage path update results. Student-stage path updates are then performed based on the teacher-stage path update results, yielding student-stage path update results. A termination condition is determined; if the termination condition is met, the path with the largest fitness function value among the student-stage path update results is output as the optimal path; otherwise, teacher and student-stage path updates continue until the termination condition is met. This invention offers advantages such as superior path quality, higher algorithm efficiency, stronger environmental adaptability, and wider application scenarios.

The invention relates to the technical field of microwave and millimeter wave imaging, in particular to a microwave and millimeter wave sparse array wavenumber domain algorithm combined compressed sensing imaging method and system. According to the technical scheme, the method comprises the following steps: obtaining a backscattering echo signal from a target, and carrying out digital down-conversion processing to obtain a digital IQ demodulation signal; according to the method, the wavenumber domain algorithm and the compressed sensing technology are combined in a cross-dimension mode, efficient focusing and data decoupling are achieved through the wavenumber domain algorithm in the motion scanning dimension, high-quality reconstruction is conducted through the compressed sensing algorithm in the array-distance dimension, and therefore image sidelobes are remarkably restrained, the imaging quality is improved, and meanwhile the image quality is improved. The dimension and calculation complexity of the sensing matrix are greatly reduced, the algorithm efficiency and practicability are improved, and the method can flexibly adapt to various sparse array configurations such as straight lines, broken lines and arc lines.

The disclosure provides a social network data analysis method and system based on limited k-means, and relates to the technical field of social network data processing. In the method, the initialization center selection stage mainly considers the connected constraint. After the first center is randomly selected, the remaining clustering centers are selected by cyclically calculating the weight probability affected by the connected constraint. Due to the influence of the connected constraint, the centroid of each connected set is used to represent the data points in the connected set to solve the limited k-means problem. Then, in the assignment step of the algorithm iteration stage, for the two types of data constraints, the strategy of preferentially processing the disjoint non-connected set and preferentially considering the intersection of the disjoint non-connected set and the connected set is adopted, and the constraint points are classified and processed, so that higher algorithm efficiency is achieved. The disclosure solves the problem of inaccurate data processing of the connected constraint and the disjoint non-connected constraint in the clustering process.

The invention belongs to the technical field of gene sequence comparison, and particularly relates to a haplotype sensing sequence-to-graph comparison method, system, device and medium, and the method directly embeds haplotype information into an edge correlation structure of a generic genome graph, so that the graph structure has the ability of expressing a haplotype path. And then, an improved haplotype perception partial sequence comparison algorithm is adopted to map the sequence to the graph, so that the graph can identify and utilize embedded haplotype information, and the global consistency of a comparison result is ensured. And finally, an optimal comparison path is obtained by adopting path multiplexing backtracking based on global cache, so that the algorithm efficiency is remarkably improved while the precision is ensured.

The application discloses a kind of gas film cooling measuring point layout optimization method and system based on expected information gain, it is related to the field of aero-engine hot test and testing technology.The method comprises the following steps: constructing the prior model of the temperature field of the component to be measured;Establish the information gain criterion of measuring point optimization;Based on the greedy algorithm, the marginal reduction of the posterior estimation uncertainty after the introduction of the candidate point is calculated, and the candidate point with the maximum marginal reduction is selected to join the measuring point set;Temperature field reconstruction is carried out based on the optimized measuring point layout.The application quantifies the information value of the measuring point, and realizes high-precision reconstruction of the temperature field under the constraint of limited measuring points, has the advantages of high reconstruction precision, strong robustness and high algorithm efficiency, and can significantly reduce the number of sensors required in aero-engine hot test.

The invention belongs to the technical field of communication measurement noise cancellation, and particularly relates to a noise cancellation method based on a WIFI7 standard protocol and application thereof. A real-time measurement average noise cancellation algorithm is adopted, and a noise coefficient is calculated by capturing internal noise when no signal is input and capturing average extracted total noise signals for multiple times, so that external noise is estimated, and single-time captured signals are corrected. According to the invention, a noise cancellation method based on measurement averaging is adopted, so that not only can inherent noise of the signal analyzer be eliminated, but also ircoherent noise in the signal can be eliminated; a convolution correlation algorithm is adopted for synchronization, and the method is suitable for more complex noise environments; and a real-time dynamic averaging method is adopted, so that the stability of the offset effect can be ensured, and the algorithm efficiency is improved. In addition, the method is low in implementation cost and complexity, can be implemented in an FPGA (Field Programmable Gate Array) or a DSP (Digital Signal Processor) in a software mode, has good reconstruction capability and is very suitable for various application environments.

This invention belongs to the field of gene sequence alignment technology, specifically relating to a haplotype-aware sequence-to-graph alignment method, system, device, and medium. This method directly embeds haplotype information into the edge association structure of a pan-genome graph, enabling the graph structure itself to express haplotype paths. Then, an improved haplotype-aware partial order alignment algorithm is used to map the sequence onto this graph, allowing it to identify and utilize the embedded haplotype information, ensuring global consistency of the alignment results. Finally, a path reuse backtracking based on global caching is used to obtain the optimal alignment path, significantly improving algorithm efficiency while maintaining accuracy.

This application relates to a method, device, vehicle, and storage medium for predicting road lane lines. The method includes: collecting the current vehicle's location information and lane line information; determining a target boundary point based on the location and lane line information, and predicting the trajectory of a first lane line within the target boundary point using a preset Kalman filter model, and predicting the trajectory of a second lane line outside the target boundary point using a preset data-driven neural network model; using the first lane line trajectory as a reference, identifying when the deviation between the curvature of the trajectory points of the target boundary point and the second lane line trajectory and the curvature of the trajectory points of the first lane line trajectory exceeds a preset threshold, correcting the trajectory points of the second lane line trajectory, and obtaining the final predicted lane line based on the corrected second lane line trajectory and the first lane line trajectory. Therefore, by using different processing methods for lane lines in different areas, computational resources are saved and algorithm efficiency is improved.