30 results about "Shortest path search" patented technology

The invention discloses a translation memory bank retrieval and matching method based on semantic similarity, which relates to the technical field of computer-aided translation, and comprises the following steps: generating an internal edge and a cross-graph edge, calculating a one-time jump success probability, calculating edge and cross-graph edge cost based on cleanliness, correcting by using the one-time jump success probability to obtain an edge weight; generating a weighted graph, constructing an initial priority queue, searching by using a shortest path to obtain a distance matrix, calculating density based on a fuzzy similarity matrix, screening the density to obtain a genealogy center, calculating proximity, and generating candidate translations in combination with the distance matrix; static and dynamic fragment sets are constructed through a dependency syntax and BERT-NER, semantic matching precision and noise robustness are improved, a weighted graph is constructed in combination with cleanliness and transfer factors, shortest path matching is carried out, and the retrieval recall rate and quality of technical translation are improved.

The invention discloses an automatic control method, system and equipment based on industrial big data analysis, and relates to the technical field of industrial big data analysis. The automatic control method based on industrial big data analysis comprises the steps of receiving order data in real time, and generating first data corresponding to a target product based on the order data; executing shortest path search in a preset process database based on the first data; collecting multi-dimensional historical data, and inputting the multi-dimensional historical data into the trained time sequence prediction model to obtain a load prediction value of each station in a future time period; performing rolling scheduling on the process chain by adopting a reinforcement learning algorithm to generate a work order sequence with a timestamp; and issuing the work order sequence with the timestamp to an edge controller, and carrying out adaptive adjustment on field equipment. According to the technical scheme, automatic control and self-adaptive scheduling based on industrial big data can be realized in a multi-variety and small-batch mixed flow production scene, and the production flexibility and the resource utilization rate are improved.

The invention discloses a shortest path searching method and a social network relation analysis method, and belongs to the technical field of network information, the social network relation analysis method comprises the steps that a knowledge graph of a social network is constructed, vertexes in the knowledge graph comprise users or entities, and edges in the knowledge graph represent the interaction relation between the vertexes; calculating a shortest path from the source point to the target point based on a shortest path searching method; wherein the source point and the target point belong to a vertex set of the knowledge graph; and obtaining a relation analysis result of the source point and the target point according to the interaction relation on the shortest path from the source point to the target point. According to the method, large-scale social network graph data with a complex structure can be efficiently and stably processed.

The application discloses a kind of grain map neural network's weld ultrasonic beam path correction method, it is related to the field of welding structure nondestructive testing, by obtaining the EBSD data of weld area, extract grain orientation, grain boundary type and spatial distribution characteristics, construct grain map structure, and calculate initial propagation time weight based on anisotropic sound velocity model;Using graph neural network to learn end to end for grain node and grain boundary edge features, output corrected edge propagation time weight;The time of flight from element to imaging pixel is obtained by shortest path search, and the time of flight predicted by neural network surrogate model is fused and corrected;Finally, the corrected time of flight is used to perform full focus imaging, and the weld defect image after optimizing the beam path is obtained.The method combines material microstructure characteristics and machine learning model, solves the problem that the beam path description is not fine enough in existing weld ultrasonic testing, significantly improves the accuracy and reliability of defect detection.

The invention discloses an intelligent process generation method and device based on knowledge graph embedding similarity matching, and belongs to the technical field of intelligent manufacturing process planning, and the method comprises the steps: constructing an assembly process knowledge graph; an assembly process sub-graph is extracted with the process as the center and serialized; the method comprises the following steps of: obtaining node-level embedding through a twin Transform encoder with shared parameters; obtaining graph-level embedding through batch embedding splicing and a pooling layer; performing bilinear interaction on an assembly subject, a relationship and an object by using an assembly element interaction neural tensor network, and extracting deep semantic features; outputting a similarity score of the two sub-graphs; retrieving the most similar process node based on the score; and executing shortest path search on the directed acyclic graph to complement missing nodes, and generating a target assembly process path. According to the method, rapid matching and automatic generation of different and semantically similar processes are realized, and the process reuse rate and the generation efficiency in a multi-variety, small-batch and rapid iterative manufacturing environment are remarkably improved.

A method and device for interpreting a graph neural network based on FPGA acceleration propose to use FPGA hardware to accelerate interpretation process of the graph neural network oriented to node classification in parallel, and improve node traversal and shortest path search of BFS, thereby optimizing requirements of algorithm calculation and storage, and accelerating generation of interpretation results. During calculating HN values, the present disclosure optimizes multiplication operation using the matrix characteristics, transforms the dense matrix multiplication into sparse-dense matrix multiplication, and optimizes the resource occupation using multi-PE parallel processing, greatly improving performance of graph neural network interpretation acceleration. Moreover, an overall architecture based on FIFO storage calculation task distribution is designed to reduce calculation difference between nodes, solving the difficulty of high time complexity of the graph neural network interpretation method based on node classification in actual data applications and improving the time efficiency of interpretation.

The application discloses a land surveying method and device based on multi-source data fusion, and relates to the technical field of surveying, which comprises the following steps: collecting three-dimensional geodetic coordinate data and three-dimensional space point cloud data, obtaining starting three-dimensional coordinate and ending three-dimensional coordinate of surveying, constructing a triangulated irregular network (TIN) terrain model and a shortest path search model on the ground, searching for a starting node and a target node in a weighted undirected graph, calculating the shortest path on the ground and the ground distance, rendering the TIN terrain model to draw the shortest path on the ground, and setting through points and forbidden areas to update the shortest path on the ground and the ground distance. The application breaks through the dependence of traditional land surveying on the line-of-sight condition by fusing multi-source data and constructing continuous digital terrain, intelligently calculates and visualizes the shortest path on the ground, avoids the error accumulation of manual segmented operation, provides interactive path replanning capability, and significantly improves the accuracy, efficiency and scientific nature of planning and decision-making of ground surveying under complex terrain.

The application discloses a shortest path divide-and-conquer search method based on a condensation level, improves the operation efficiency of an algorithm by introducing network layering, limiting a search level, divide-and-conquer search and the like, wherein the network layering is adopted to divide network nodes according to a condensation level clustering method, a plurality of level subgraphs are obtained, and the complexity of the search is reduced; the search range is limited within a certain level by limiting the search level, and invalid calculation in the search process is avoided; meanwhile, in order to further improve the search efficiency, the divide-and-conquer search is adopted to divide the whole search process into a plurality of subtasks and perform simultaneously, and the whole search method can accurately and efficiently perform the shortest path search in an edge computing network.

This invention discloses a collaborative operation optimization method for an integrated bus and metro network, comprising the following steps: S1, collecting and standardizing topology data, train timetable data, and passenger travel demand data; S2, constructing a multi-layer coupled spatiotemporal network; S3, performing K-shortest path search to generate a set of passenger reachable paths; S4, verifying the reachability of the passenger path selection set under train operation rules; S5, constructing a scheduling scheme based on the allocation probabilities output by the Logit model; S6, inputting the time-series operation data into an improved Mamba network to constrain and correct the number of train departure pairs and the first station departure time, forming an evaluation index for the optimization objective function. This invention achieves collaborative optimization of train operation sequence and passenger travel demand in an integrated bus and metro network, and is applicable to intelligent scheduling and dynamic passenger flow control scenarios in large-scale urban rail transit systems.

The invention provides a multi-robot path searching method for an optical communication scene, and belongs to the field of robot path planning, and the method comprises the steps: obtaining an optical communication scene road network containing optical communication sites and non-optical communication sites, and obtaining a simplified road network and road network nodes through the fusion of the non-optical communication sites; on the basis of neglecting conflicts among robot paths, forming an initial path scheme through a shortest path search algorithm, judging whether space-time conflicts exist or not, generating corresponding space-time constraints according to a space-time conflict judgment result, and recalculating to form a new path scheme through the shortest path search algorithm with the space-time constraints; and repeatedly executing the shortest path search algorithm with the space-time constraint until an optimal path scheme is obtained. The problems that in the prior art, due to the fact that continuous real-time wireless communication is relied on, effective conflict detection and avoidance cannot be carried out in an optical communication discontinuous communication scene, the robot collision risk is high, and the planning efficiency is low are solved.

This invention discloses a method for obtaining K shortest paths in a city that considers real-time traffic conditions, comprising: 1. Constructing a city road network based on traffic information at the initial time; 2. Defining and initializing parameters; 3. Searching for the k-th shortest path from the starting intersection node to the ending intersection node at the current time; 4. Searching for the (k+1)-th shortest path from the starting intersection node to the ending intersection node at the current time; 5. Obtaining the K shortest paths from the starting intersection node to the ending intersection node at the next time after updating traffic conditions, based on real-time traffic conditions. This invention can obtain K optimal travel paths that better reflect actual traffic conditions while improving computational efficiency, thereby ensuring the stable and efficient operation of social traffic.

The application discloses a routing method, medium, equipment and product of a low-orbit satellite constellation, relates to the technical field of satellite routing, and comprises the following steps: according to a quality of service parameter, performing snapshot freezing on a current network topology to construct a static topology graph structure; constructing a basic quality of service parameter cost function for each link; performing an offline ant colony optimization algorithm to globally evaluate a satellite topology library constructed in advance, constructing a heuristic function according to a link basic quality of service parameter cost; initializing a pheromone value for each link, and updating the pheromone in an iteration process according to the heuristic function; adjusting the cost function based on the pheromone; taking the cost of the discounted pheromone as the final link weight of the satellite network, constructing a weighted graph structure, and adopting a shortest path search algorithm to calculate a routing path between a source node and a target node, so that an optimal routing path is obtained. The application solves the problems of high-speed time-varying satellite link state adaptability and poor quality of service.

The application discloses a kind of vehicle operating mode transfer planning methods for fuel consumption optimization, comprising: step 1: according to vehicle relevant test operating mode acquisition driving data;Step 2: each variable in the historical data obtained is filtered as different dimensions and is clustered into partitions;Step 3: according to the information extraction of each zone divided by cluster, the relationship and transfer cost of each zone are obtained from continuous operating mode data;Step 4: according to the relationship between each zone, the operating mode transfer directed graph is constructed, and all the minimum cost transfer paths between different initial and final regions are obtained;Step 5: according to the current operating mode and target operating mode information, the region selected by each operating mode in the path is selected as the same as the minimum cost path in step 4 as the optimal transfer path;The operating mode information in the historical data is divided into regions, then each region is regarded as a vertex in the directed graph, the transfer cost between regions is regarded as the distance between each vertex, and the vehicle operating mode transfer path planning is changed into the shortest path search of directed graph.

The invention relates to an edge device LLM deployment system and method based on memory filling and game optimization, and the system comprises a network state synchronization module, a distributed model deployment module and a model reasoning module. Through dynamic model deployment, cold-start-free model loading and self-adaptive adjustment are realized, and reasoning scheduling is optimized through a layer-device execution graph and shortest path search.

Disclosed are embodiments of shortest path search in an obstacle environment. Provided is a shortest path search method including: creating an epsilon graph; and determining k closest points from a query point on the basis of the graph, wherein the creating of the graph includes: creating a node group of nodes on the basis of a point group of arrival points; creating an edge group of edges on the basis of the created nodes; setting, on the basis of an obstacle group of at least one obstacle, whether each of the nodes is a valid node or a non-valid node; setting, on the basis of the obstacle group, whether each of the edges is a valid or non-valid edge; and including, for each of the arrival points, a particular arrival point on an anchor list of a valid node closest from the particular arrival point among the valid nodes.

The invention discloses a concealed pipe drainage design method and system for saline-alkali soil. The method comprises the following steps: firstly, acquiring a farmland or saline-alkali soil surface image, processing the image by adopting a machine vision technology, and automatically generating a mask pattern of a water accumulation area; then, according to the area of a mask connected domain, the number of drainage points is calculated in a self-adaptive mode, and drainage point coordinates are extracted through a clustering algorithm; a cost field is constructed in combination with a mask graph, an eight-neighborhood weighted grid graph is established, and a direction penalty function and an acute angle suppression rule are introduced in shortest path search, so that the generated drainage path is smoother, has no sawteeth and meets the construction requirements of concealed conduits. Automatic grouping is achieved through the shortest path distance between terminal points, and a multi-concealed-pipe layout scheme with the minimum total cost is obtained in combination with automatic selection of exit points and multi-threshold global search. According to the invention, full-process automation from image recognition to drainage path generation can be realized, and the intelligent level of saline-alkali soil treatment and farmland drainage design is improved.

The present application relates to the field of rail transit, and specifically provides a regional multi-mode rail transit network passenger flow distribution prediction method, medium and computer device, wherein the regional multi-mode rail transit network passenger flow distribution prediction method comprises the following steps: improving the K-shortest path search algorithm according to the hierarchical characteristics of the pre-constructed regional multi-mode rail transit network; constructing a set of feasible paths between any OD based on the pre-generated passenger flow OD matrix of the "region + urban area" and the improved K-shortest path search algorithm; and integrating the set of feasible paths between any OD into a travel path probability selection model to realize the passenger flow distribution of the regional multi-mode rail transit network. Through such a structure, more scientific integrated analysis of the passenger flow demand of the regional rail transit network can be provided, so as to meet the diversified travel demands of passengers in different regions, with different functions and different speeds.

The invention belongs to the technical field of emergency command and dispatch management evaluation, and particularly discloses and provides a multi-dimensional dynamic perception integrated cooperative command and dispatch system, which comprises a traffic weight updating module for updating the real-time traffic weight of each link of a whole road network according to the real-time vehicle traffic speed and the real-time ponding depth; the ponding trend prediction module determines a ponding prediction area based on the starting point and the ending point of the scheduling task, and predicts the ponding depth change of each link in the area in a future time period through hydrological deduction; the passing time window judgment module is used for judging a time window for keeping a passable state of each link in a future time period in combination with the accumulated water depth change; and the passing plan generation module is used for generating a scheduling path and a time plan by adopting a time-dependent shortest path search algorithm. According to the invention, the problems that the scheduling instruction is disjointed from the dynamic environment and the rescue deployment is in a passive response state are solved, and the accuracy and response efficiency of emergency command are improved.

This invention discloses an AI large-scale intelligent question cloud analysis method and system, comprising: a question preprocessing module for acquiring and preprocessing natural language question requests, and simultaneously establishing a data asset set in the cloud; a semantic retrieval module for recalling and re-ranking candidate semantic elements and constructing a training sample set; a model fine-tuning module for performing multiple rounds of fine-tuning on the LLaMA model, and performing semantic parsing and multiple rounds of planning to form a question intent field; a path selection module for performing bidirectional shortest path search on the data asset set in the cloud and selecting the optimal execution path as the question execution path; an operator compilation module for cost optimization compilation to generate a cloud-executable task instruction sequence; and a feedback optimization module for collecting feedback data and performing incremental fine-tuning updates. This invention enables the understanding of natural language question intent and improves the accuracy of question intent parsing.

The invention belongs to the related technical field of energy systems, and provides an energy router optimization method and system based on information physical fusion, an energy router system is constructed into a directed graph model, a node set represents various physical devices, and an edge set represents an energy flow path between nodes; carrying out shortest path search by adopting a path search algorithm, optimizing a minimum loss energy path for a source-load in real time, and generating a dynamic routing table; on the basis of economic flow and power flow, an energy storage carbon collaborative optimization target is introduced, a multi-target optimization model is constructed, and dynamic optimization of an energy transmission path, rapid stabilization of power fluctuation and unified optimization of economic and carbon emission targets are realized.

This application relates to the field of shortwave ad hoc network routing optimization technology, and in particular to a multipath routing method for shortwave ad hoc networks based on node disjointness and adaptive traffic splitting. The method includes: each node exchanging messages according to a preset period to obtain one-hop neighbor information and two-hop neighbor information; constructing a comprehensive link cost model based on the current node's one-hop symmetric neighbors; modeling the shortwave ad hoc network as an original topology graph; performing a shortest path search on the original topology graph based on the comprehensive link cost to obtain primary and backup paths based on node disjointness constraints; and allocating service traffic according to the comprehensive link cost of the primary and backup paths. The method designed in this application can improve the service continuity, throughput performance, latency performance, and network robustness of shortwave ad hoc networks under conditions of sudden link degradation and rapid topology changes, while ensuring controllable routing computation complexity.

The invention discloses a tunnel image crack semi-automatic labeling method and system based on shortest path search, and relates to the technical field of image labeling. The method comprises the following steps: acquiring a tunnel crack image sample, calculating an initial optimal path with an initial label in the tunnel crack image sample by using a minimum path search method, modeling the tunnel crack image sample into a graph structure, defining a corresponding cost function according to the graph structure, and calculating the initial optimal path with the initial label in the tunnel crack image sample. Searching a path with the minimum cost between an appointed starting point and an appointed ending point; introducing crack width information, and further optimizing the fine cracks in the initial optimal path to obtain a final optimal path; and performing interactive image annotation according to the final optimal path. According to the method, efficient and accurate marking of the fine crack is realized, so that the core problem that high-quality marking data is difficult to obtain is solved.

This invention belongs to the field of aircraft landing gear position detection and retraction control technology, and discloses a landing gear position status detection method, system, and electronic equipment based on a ring bus link. The method is executed by the current node in a redundant landing gear retraction control system: receiving neighboring node data in the current cycle, updating the communication weight matrix based on the actual reception, and performing a shortest path search based on the updated communication weight matrix; filtering non-neighboring node data from the neighboring node data according to the shortest path; performing a majority vote on data representing the same physical state among the current node data, neighboring node data, and the filtered non-neighboring node data; jointly determining the majority vote result with the landing gear physical mutual exclusion state, and outputting the landing gear position status. This invention can reduce duplicate data and state conflicts in scenarios with link anomalies and multi-path forwarding.

The application relates to the technical field of computer vision, and discloses a picture splicing method based on multiple cameras. The method comprises the following steps: acquiring multiple image frames collected synchronously; correcting and aligning the images; constructing an initial pixel-level difference cost map; combining a previous frame splicing seam to generate a space-time guide weight map; fusing to obtain a dynamic optimization cost map; performing shortest path search in a limited area to determine a current best splicing seam; and generating a seamless splicing image by weighted fusion according to the current best splicing seam. By introducing a space-time guide mechanism and an edge preservation factor, the calculation complexity is reduced, the splicing continuity and visual quality are guaranteed, and the method is suitable for high-dynamic scenes such as vehicle-mounted surround view and unmanned aerial vehicle aerial photography.

The application belongs to the technical field of building engineering quality detection and intelligent planning, and discloses an engineering structure entity quality detection point intelligent planning method, which comprises the following steps: analyzing component geometry and position parameters based on a BIM model, and generating a detection point candidate set according to detection specification clauses; removing redundant points through centroid calculation, normalization and symmetry analysis; constructing a walkable space constraint in combination with the shape size and safety gap of a detection robot, and removing unreachable points; establishing a space topology structure according to the connectivity relationship between rooms and doors, identifying key door nodes by using an intermediate index, and dividing a suite type and a public area by clustering; calculating the shortest travel distance between suite doors under the configuration space constraint, constructing a traveling salesman problem model to solve the public area sequence. The shortest path search is performed in the suite, and the door nodes and the public area sequence are connected, so that a global detection sequence and a task instruction sequence are generated to guide the detection robot to implement detection.

The application discloses a kind of multi-edge device task assignment methods based on dynamic resource perception, purpose is to solve the imbalance between inference delay and inference energy consumption problem.The multi-edge device task assignment system based on task splitting containing delay prediction model is constructed, the delay prediction model is trained, the inference time of different task fragments running on the resource-limited, highly heterogeneous edge device cluster is predicted, the resource condition of edge device is monitored in real time, and the memory capacity of edge device is introduced as constraint, the delay and energy consumption model of multi-device collaborative computing based on task splitting is constructed, to minimize the total weighted cost of task assignment system as optimization goal, construct collaborative inference joint optimization problem, convert the optimization problem into hierarchical directed graph shortest path search problem, determine optimal task allocation and routing, realize task splitting and multi-device dynamic efficient execution, so that inference delay and inference energy consumption are balanced, greatly shorten task inference time.

The embodiment of the invention provides an abnormal root cause determination method and device, and is applied to the technical field of micro-service systems. The method can quantify the time delay and influence rule of the abnormal information of the upstream node through the propagation kernel of the upstream node, and ensures that dictionary atoms are consistent with actual propagation characteristics. Through a full time sequence response mode that a dictionary atom covers an upstream node in an abnormal window, a full link process of abnormal excitation, time sequence propagation and effect display can be captured, and a globally optimal target excitation vector can be obtained through one-time solution in combination with optimization solution logic of sparse excitation inversion. Therefore, compared with graph search in which the time sequence is added, for example, abnormal peak values are sequenced firstly, and then random walk or shortest path search is performed in the calling graph, the probability of missing detection and false alarm can be reduced, and the accuracy of abnormal root cause positioning can be improved.

The invention relates to the technical field of power system dispatching operation control, and discloses a power transmission section identification method and system based on dynamic matching of fragile lines in a power transmission channel. Constructing an improved PageRank algorithm to calculate a key degree index of the node; calculating a transaction volume sequence between a generator and a load node by using the PTDF, screening out a key generator-load node pair, extracting a key line set in a corresponding source-load channel from the updated topological graph by means of network conversion and shortest path search, and giving a vulnerability index of the line in combination with the LODF and the real-time power flow of the line; and matching the highly fragile line identified on line with the minimum cut candidate section obtained by the offline search, and dynamically screening to obtain the optimal power transmission section in the current operation state. The method can adapt to the dynamic change of the operation state of the power grid, quantifies the key of the nodes in real time, has high accuracy and high calculation efficiency, and provides powerful support for safe and stable operation of the power grid.

The invention provides a condyle process CBCT medical image segmentation method based on deep learning, and relates to the technical field of medical image segmentation, and the method specifically comprises the steps: cutting a condyle process image from an original CBCT image of a to-be-diagnosed patient, collecting the original clinical data of the patient, converting the original clinical data into a clinical feature vector, and obtaining a condyle process image; combining the multi-scale features of the encoder and the multi-scale features of the decoder to obtain fusion features, obtaining a segmentation probability graph by using a multi-modal fusion model, extracting a segmentation boundary by applying an edge detection algorithm, identifying and connecting fracture end points, obtaining a continuously closed condyle segmentation mask, and multiplying the condyle segmentation mask and an original CBCT image point by point to obtain a condyle segmentation mask. And an optimized condylar process image is obtained. According to the method, by accurately marking the fracture end points, constructing the distance transformation diagram and adopting the shortest path search algorithm, the accuracy of the initial contour is reserved, the complete and closed requirements are met, and the boundary is prevented from being disjointed from the real physiological morphology after repairing.

The invention belongs to the technical field of constructional engineering quality detection and intelligent planning, and discloses an engineering structure entity quality detection point intelligent planning method, which comprises the following steps: analyzing component geometry and position parameters based on a BIM model, and generating a detection point candidate set according to detection specification terms; screening redundant points through centroid calculation, normalization and symmetry analysis; a walking space constraint is constructed by combining the boundary dimension of the detection robot and the safety gap, and unreachable point locations are removed; establishing a spatial topological structure according to a communication relationship between a room and a door, identifying key door nodes by using betweenness indexes, and dividing a cover type and a public area through clustering; and the shortest passing distance between the suit type doors is calculated under the constraint of the configuration space, and a traveling salesman problem model is constructed to solve the sequence of the public areas. And executing shortest path search in the condom and sequentially connecting the gate nodes with the public areas to generate a global detection sequence and a task instruction sequence for guiding the detection robot to implement detection.