44 results about "Vertex number" patented technology

Provided are a distributed parallel computing based graph computing method, a distributed parallel computing system and a computer readable medium. A graph computing method comprises: obtaining data of a graph to be computed, the graph comprising a plurality of vertices and edges; storing the graph according to a compressed sparse row format, and storing the graph in a sparse matrix form, wherein all adjacent edges having a same vertex are continuously stored, and an offset array is used to support an index function thereof; and for subgraphs in which the degrees of both source vertices and target vertices are greater than a predetermined threshold, in forward graphs, performing segmentation according to target vertices, i.e., the column of a matrix, and in reverse graphs, performing segmentation according to source vertices, i.e., the row of the matrix, and further dividing the subgraph into a plurality of segmented subgraph (SSG) components, so that a target vertex in each SSG is limited within a preset range. According to the SSG method of the embodiment for a high-degree vertex, the conventional problem that the space-time locality of accessing data of an adjacent vertex is poor due to a large vertex range of a single large graph is avoided or relieved.

The invention discloses a polygonal and quadrilateral mesh automatic subdivision method based on deep reinforcement learning, relates to the field of computational geometry and mesh generation, and constructs an automatic subdivision framework fusing geometric priori knowledge and a data driving strategy. According to the method, nine basic topology filling templates covering triangles to hexagons are predefined as discrete action spaces, and continuous geometric segmentation is converted into a sequence decision problem; learning a mapping relation between the polygon state characteristics and the optimal template selection strategy by using a deep Q network, and balancing exploration and utilization through a dynamic epsilon-greedy mechanism; and in combination with a vertex number priority scheduling strategy and a multi-dimensional quality award function, guiding an intelligent agent to adaptively generate a high-quality quadrilateral grid. According to the method, full-automatic and high-robustness subdivision of the complex polygon area is achieved, the grid orthogonality and the length-width ratio quality are remarkably improved, manual intervention is avoided, and the method is suitable for engineering scenes such as finite element analysis.

The invention discloses a quantum-classical collaborative wireless link scheduling optimization method based on a divide-and-conquer QUBO model and inverse non-adiabatic driving, and belongs to the technical field of quantum computing and wireless communication. According to the method, link interference in wireless communication is modeled into a conflict graph, namely, vertexes represent links, interference represents edges, utility functions on the links represent weights of the vertexes (numbers on the left of the vertexes are numbers of the vertexes, and numbers on the right of the vertexes are weights of the vertexes), and optimal wireless link scheduling is a weighted maximum independent set (WMIS) on the conflict graph. According to the method, a divide-and-conquer QUBO model is provided, the whole space is divided into Dicke state subspaces with different Hamming weights, XY-Mixer is adopted as a hybrid operator, and evolution of QAOA in the subspaces is limited. On the basis, a DQAOA algorithm is provided; a greedy algorithm and an expanded Lov-alpha-sz positive semidefinite plan are called to obtain a lower bound and an upper bound of a WMIS solution respectively; an anti-non-adiabatic driving item is introduced to modify a QAOA evolution formula, an XY-Mixer angatz is matched to enable the algorithm to be subjected to parallel evolution in a subspace, convergence is accelerated, and an optimal wireless link scheduling scheme is obtained.

The invention belongs to the technical field of channel allocation, and particularly relates to a full-duplex D2D (Device-to-Device) communication channel allocation algorithm based on novel graph coloring, which comprises the following steps of: constructing a bidirectional interference graph containing all potential interference links; expressing a channel allocation problem as an interference minimization graph coloring problem; distributing N independent communication channels for all cellular user equipment; a greedy graph coloring algorithm is adopted, and single colors are distributed to the D2D vertexes which are not colored one by one; if the number of D2D vertexes in a certain color set is increased or decreased, the interference level is evaluated again; and traversing each color, allocating a single color to the remaining D2D vertexes, terminating the algorithm after all D2D vertexes are traversed, and generating a conflict-free channel allocation scheme. According to the invention, the channel allocation can be optimized, and the interference in the system can be effectively managed and reduced, so that the access rate of the user and the overall spectrum efficiency are improved.

The invention provides a layout splicing method. The layout splicing method comprises the following steps: acquiring a list of vertexes and a list of edges of a plurality of slices of a layout; according to the list of the vertexes of the multiple slices, a list of the vertexes of the layout is obtained, and the list of the vertexes of the layout comprises the vertexes of the multiple slices; according to a vertex of a first graph of the first slice, determining a tangency point of the first slice, the first graph being a graph with at least one edge overlapping with at least one boundary of the first slice in the first slice, and the tangency point being a vertex located on the boundary of the first slice in the first graph; according to the tangent points of the multiple slices, a cross-slice edge of the layout is determined, and the cross-slice edge comprises a first tangent point and a second tangent point which are located on a first boundary and a second boundary which are overlapped of the adjacent slices respectively; and determining a list of edges of the layout according to the list of the cross-slice edges and the plurality of slice edges. According to the method, the number of vertexes and the number of edges involved in a scanning line algorithm in the layout splicing process can be reduced, so that the layout splicing efficiency is improved.

The invention discloses a BFS-like algorithm-oriented graph processing system caching method, a controller and a system, and belongs to the technical field of graph processing system performance optimizing.The method comprises the steps that based on the relation between cache page access popularity and the number of unaccessed vertexes in a BFS-like algorithm, more accurate popularity level-to-level management is carried out on cache pages; when the cache is fully loaded, pages with low access potential are effectively eliminated, so that the cache hit rate is increased; fragmented effective edge data are extracted from a cache page and are independently stored in a cache of fine-grained management; the scheduling sequence of the vertexes in the task queue is adjusted by obtaining the proportion of the edge lists of the vertexes in the BFS-like algorithm task queue residing in the cache and combining the attributes such as the out-degree of the vertexes, and the data residing in the cache is utilized to the maximum extent. According to the method, the cache hit rate of the graph processing system when the graph processing system processes the BFS-like algorithm task can be effectively improved, so that the access performance of the graph processing system is improved.

The present invention relates to the field of graphics rendering technology, and in particular to a vertex shading method based on a vertex buffer. It includes the following steps: S1. After the vertex data preparation stage, the hardware module obtains the vertex data that needs to be shaded and the vertex index corresponding to the vertex data; S2. The vertex index that needs to be shaded is sent to the vertex shading module and the subsequent processing module; S3. The subsequent processing module queries the vertex buffer area according to the vertex index. If the vertex in the required state is queried, the subsequent processing module directly obtains the vertex and ends the process; S4. The vertex shading module queries the vertices in the vertex buffer area according to the received vertex index and renders the unshaded vertices; S5. The subsequent processing module queries the vertices in the vertex buffer area according to the vertex index and obtains the vertices after rendering to perform the subsequent processing process. Adding a storage area for vertex buffering solves the problem of repeated rendering of vertices and poor performance.

The application discloses an individualized partition method and device based on brain surface geometric features, a medium and a product, and relates to the technical field of neuroimaging. The method comprises the following steps: dividing each vertex into sulcus seed points, gyrus seed points and remaining vertices based on the shape index of each vertex, thereby obtaining a plurality of sulcus seed initial regions and gyrus seed initial regions; merging each connected sulcus seed initial region and each gyrus seed initial region, thereby obtaining a plurality of merged regions; determining each initial region that is not connected as an independent region; determining all vertices in the merged region and the independent region and all remaining vertices that have a number less than a preset number as transition vertices; determining the merged region and the independent region that have a number greater than and / or equal to the preset number as initial inclusion regions; iteratively including each transition vertex, thereby obtaining a personalized brain partition map, and achieving the partition of the brain. The application improves the accuracy of brain region partitioning.

The invention relates to a method and device for eliminating graphic quantization sawteeth in GDI and a medium, and the method comprises the steps: obtaining vertex data of a to-be-processed graphic, and caching the vertex data after affine transformation; constructing an integer straight line according to the slope and intercept of a straight line formed by two adjacent vertexes, and selecting the first vertex and the last vertex to calculate an initial vertex coordinate correction vector; traversing each vertex, circularly correcting the adjacent next vertex coordinate according to a vertex coordinate correction vector calculated according to the previous vertex coordinate, rounding off the corrected vertex coordinate to obtain an integer vertex coordinate, and updating the vertex coordinate correction vector for correcting the next vertex; and if the traversal is completed, writing the integer vertex coordinate value back to the graphic data, and cleaning the cache. Compared with the prior art, the method has the advantages of automatically eliminating the sawtooth deformation of the graph, improving the drawing efficiency and the like.

The application discloses a graph data partitioning method based on a cache mechanism, comprising the following steps: obtaining graph data, obtaining the total number of graph vertices, the neighbor list of each vertex, the initial vertex partition set and the initial partition number according to the graph data; calculating the upper limit of the load; calculating the correlation score of the vertex in each vertex set which does not reach the upper limit of the load according to the vertex and the neighbor list of the vertex; judging whether the vertex meets the cache condition, if yes, partitioning according to the correlation score; if not, the vertex is partitioned into the cache module, when the cache module reaches the maximum capacity limit, the cache module returns the vertex and the neighbor list which are most suitable for the current partitioning, and the process returns; judging whether all vertices have been partitioned, if yes, the final graph data partition set is obtained; if not, the process returns. The application improves the partitioning precision of the Fennel algorithm under the condition that the size of the graph data is too large.

The invention discloses a dynamic modeling method and system for a three-dimensional model, and the method comprises the steps: recognizing an interactive operation, and obtaining coordinate point data in a three-dimensional scene in real time according to the interactive operation, so as to complete vertex assignment and form a vertex list; calculating a geometric center point of the vertex list based on the vertex list; connecting the vertexes in the vertex list with the geometric center point in a clockwise sequence to generate a planar triangular surface index array, and forming an initial plane; copying a vertex list of the initial plane and offsetting a target numerical value along a specified direction to generate a target plane; and connecting the vertex of the initial plane and the vertex of the target plane in sequence to generate a side triangular surface so as to complete modeling of the three-dimensional model. According to the method, the user interaction track is directly converted into the vertex data and the geometric calculation logic, so that the user can independently complete the whole-process modeling from plane to three-dimensional without relying on a modeler, the Unity scene development efficiency is remarkably improved, and a low-cost and high-flexibility solution is provided for interactive three-dimensional application.

The invention discloses a quantum bit mapping method and device, equipment and a storage medium, and belongs to the technical field of quantum. The method comprises the steps that a coupling graph corresponding to quantum equipment is acquired, each vertex in the coupling graph represents one quantum bit in the quantum equipment, and an edge between the two vertexes represents that a double-quantum-bit gate is allowed to directly act on the quantum bits corresponding to the two vertexes; at least two connected sub-graphs are generated according to the coupled graph, the number of vertexes of each connected sub-graph is larger than or equal to 2 and smaller than or equal to the routing number of the coupled graph, and the sum of the number of the vertexes of the at least two connected sub-graphs is equal to a constant multiple of the number of the vertexes of the coupled graph; and based on the at least two connected subgraphs, performing quantum bit mapping on the logic circuit to obtain a hardware compilable circuit equivalent to the logic circuit. According to the method, the problem that the depth is too large due to quantum bit permutation under the limitation of a coupling graph is solved, and the technical effect of effectively reducing the time complexity of quantum bit mapping is achieved.

Provided are a distributed parallel computing based graph computing method, a distributed parallel computing system and a computer readable medium. A graph computing method comprises: obtaining data of a graph to be computed, the graph comprising a plurality of vertices and edges; storing the graph according to a compressed sparse row format, and storing the graph in a sparse matrix form, wherein all adjacent edges having a same vertex are continuously stored, and an offset array is used to support an index function thereof; and for subgraphs in which the degrees of both source vertices and target vertices are greater than a predetermined threshold, in forward graphs, performing segmentation according to target vertices, i.e., the column of a matrix, and in reverse graphs, performing segmentation according to source vertices, i.e., the row of the matrix, and further dividing the subgraph into a plurality of segmented subgraph (SSG) components, so that a target vertex in each SSG is limited within a preset range. According to the SSG method of the embodiment for a high-degree vertex, the conventional problem that the space-time locality of accessing data of an adjacent vertex is poor due to a large vertex range of a single large graph is avoided or relieved.

The invention discloses a semi-supervised segmentation method based on graph prototype learning, and the method comprises the steps: collecting a data set, carrying out the preprocessing of the data set, carrying out the mapping of a pixel space to a graph semantic space for the bottom features of an encoder, carrying out the multi-stage graph convolution compression of the number of vertexes, and carrying out the alignment to a labeled image category prototype; sending an unlabeled image into the segmentation network to obtain a category prototype of a pseudo label, and aligning the centroid of the most similar subclass after the category prototype of the unlabeled image is subjected to self-clustering with the category prototype of the labeled image; fusing the structure information of the image with the label and the style information of the image without the label to obtain a composite image; screening prediction results of a non-label image and a plurality of weak disturbance images through a voting strategy, determining reliable and fuzzy pseudo labels, and optimizing a segmentation model and category prototype sampling; and updating network parameters through training to obtain a final segmentation network. According to the method, accurate segmentation can be realized in scenes of scarcity of high-quality labeled data, complex structure and shape of a target and the like.

The present invention provides a group discovery method based on searching conditional dense subgraphs in a large symbolic graph, and relates to the field of information processing technology. The method includes: generating an initial symbolic graph based on social data to be searched; sorting the vertices in the initial symbolic graph to obtain the sorting results, and sequentially generating the total branches corresponding to each vertex according to the sorting results; branching the total branch at least once to obtain a target sub-branch, and the target sub-branch meets the preset conditions. At each branch, the pivot point is determined based on the number of second symbolic neighbors of the vertex in the union of the candidate set and the required set of the parent branch, and the sub-branch is generated based on the pivot point; based on the target sub-branch with the largest number of vertices, the target group in the social data to be searched is determined. The present invention can find groups with close relationships and consistent relationship types from a large amount of social data, thereby improving the accuracy of content recommendation based on the group.

The present disclosure relates to a method, apparatus, device and medium for identifying characteristic regions of a vehicle model file. The method comprises: obtaining geometric information of each triangular facet in the vehicle model file; numbering the vertices of each triangular facet based on vertex coordinates to obtain three vertex numbers corresponding to each triangular facet, with the same vertex coordinates corresponding to the same vertex number; determining the shared edges between each triangular facet and adjacent triangular facets and the first characteristic edge of the vehicle model file based on the three vertex numbers corresponding to each triangular facet; identifying the shared edges that do not meet preset requirements as second characteristic edges; determining the characteristic region to which each triangular facet belongs based on the first characteristic edge and the second characteristic edge, with the characteristic region being a closed figure formed by connecting multiple first characteristic edges and / or multiple second characteristic edges end to end. When identifying characteristic regions, the present disclosure does not need to calculate the curvature of the surface at the vertex, which greatly saves computational effort and improves computational efficiency.

The invention relates to route planning and optimization. In some embodiments, the invention includes obtaining vertex data describing target attributes, including at least one target location, and process dynamics data describing system constraints of a process: defining vertex functionals, describing attribute requirements, and edge functionals, which use the system constraints to determine cost requirements, for cost objectives; and using the data, the functionals, and the cost objectives to construct a T-graph that includes a label space representing the target attributes. Next, the T-graph is solved to construct a label space trajectory by determining waypoints such that the process (1) travels through all target locations and (2) satisfies the attribute requirements, dynamically calculating travel cost by applying the cost requirements to each intervening edge, and minimizing the travel cost according to the cost objectives. The optimal space trajectory is executed to navigate the process through the target locations.

This invention discloses a maximal biclique enumeration method based on hybrid storage. It obtains a bipartite graph G(U,V,E), where U and V are two disjoint vertex sets and E is an edge set. A threshold t is set to determine the storage method of the subgraph during runtime. If the number of vertices in U is greater than t, G0 is stored using an adjacency list; otherwise, G0 is stored using a bitmap. The BicliqueFind(L,R,C,G) function is recursively called with U, V, and G0 as initial parameters. The count results of all maximal bicliques are output to the user. This invention employs a hybrid storage method, retaining the low storage overhead of adjacency list storage while combining the computational efficiency of bitmap storage. For tasks in biclique enumeration that require calculations only on small graphs, this invention significantly improves the execution speed of the biclique enumeration task by dynamically constructing sub-bitmaps.

The application discloses a kind of polygon quadrilateral grid automatic subdivision methods based on deep reinforcement learning, it is related to the field of computational geometry and grid generation, the application constructs a kind of automatic subdivision framework of fusion geometry priori knowledge and data-driven strategy.The method predefines 9 kinds of basic topological filling templates covering triangle to hexagon as discrete action space, converts continuous geometry segmentation into sequence decision problem;The mapping relationship between polygon state characteristics and optimal template selection strategy is learned using deep Q network, and exploration and utilization are balanced through dynamic epsilon-greedy mechanism;Combining vertex number priority scheduling strategy and multidimensional quality reward function, guide agent to adaptively generate high-quality quadrilateral grid.The application realizes the full-automatic, high robustness subdivision of complex polygon region, significantly improves the grid orthogonality and aspect ratio quality, avoids manual intervention, and is suitable for finite element analysis and other engineering scenarios.

The invention discloses a three-dimensional model representation method and device and a three-dimensional model reconstruction method and device, and relates to the technical field of computers.The method comprises the steps that firstly, a vertex array is generated according to a topological structure of an original three-dimensional model, and then generating point representation, edge representation and surface representation of the original three-dimensional model according to the vertex id in the vertex array to obtain polygon data geometric representation of the original three-dimensional model consistent with topology. The point representation comprises each point represented by each vertex id, the edge representation comprises each edge represented by a plurality of vertexes id in the vertex array, and the vertex id in the vertex array is unique and corresponds to the determined vertex information, so that the unique and determined point or edge on the original three-dimensional model can be reconstructed according to one or more vertexes id; and a determined surface can be reconstructed under the limitation of the edge, so that a model consistent with an original three-dimensional model can be reconstructed, and the problem of inconsistency during model reconstruction is solved to a certain extent.

The feature preserving surface reconstruction method based on optimal transport involves three-dimensional geometric processing. 1) input the three-dimensional point cloud model, construct the convex hull of the point cloud and re-mesh as the initial mesh; 2) sampling on the initial mesh, calculating the optimal transport plan between the point cloud and the sampling points according to the transport cost function; 3) calculating the transformation matrix between the sampling points of the triangular face and the point cloud corresponding to the transport, calculating the new position of the triangular mesh vertex according to the transformation matrix, and moving each vertex to the new position to obtain the initial fitting mesh; 4) point relocation, edge flipping and edge collapse operations are performed on the initial fitting mesh to reduce the optimal transport cost of the point cloud and the mesh, until the number of reconstructed mesh vertices reaches the target vertex number. The triangular mesh with sharp edges and sharp point features is obtained. By using the optimal transport theory, the feature preserving result can be reconstructed with less number of faces, and it is robust in the case of noisy and missing point cloud data.

The invention discloses an individualized partitioning method and device based on brain surface geometric features, a medium and a product, and relates to the technical field of neuroimagines.The method comprises the steps that all vertexes are divided into brain sulcus seed points, brain gyrus seed points and remaining vertexes based on shape indexes of all the vertexes, a plurality of cerebral sulcus seed initial areas and cerebral gyrus seed initial areas are obtained; respectively merging the communicated brain sulcus seed initial regions and brain gyrus seed initial regions to obtain a plurality of merged regions, and determining the non-communicated initial regions as independent regions; and determining all the vertexes and all the remaining vertexes in the merged region and the independent region of which the number of the vertexes is smaller than a preset number as transition vertexes. Determining the merging areas and the independent areas with the vertex number greater than and / or equal to the preset number as initial inclusion areas; and iteratively incorporating the transition vertexes to obtain a brain personalized partition map so as to realize the partition of the brain. According to the invention, the accuracy of brain region partitioning is improved.

The invention discloses an efficient sub-graph matching query method of a cross-database system. The method comprises the following steps: reducing the influence of vertex number and edge number increase on sub-graph cardinal number estimation precision by utilizing pattern graph cardinal number constraint information; the accuracy of matching cardinal number estimation of the query sub-graph with attribute filtering is improved by utilizing a sparse sampling graph technology; calculating execution overhead on the basis of a query sub-graph cardinal number estimated value, and searching an optimal execution plan tree from bottom to top; on the basis of the optimal execution plan tree, optimized query statements capable of guiding execution plans are generated for different database systems, and efficient sub-graph matching query across the database systems is achieved. According to the method, the problems that cross-database unified management and optimization are difficult to realize and the advantages of each database system cannot be brought into full play in the existing method are solved, the query efficiency when the existing database system executes complex graph query is effectively improved, and the query time is shortened.

The invention relates to the field of graph calculation, in particular to a data use method and device in a graph calculation hybrid calculation mode. The method mainly comprises the following steps: partitioning vertexes according to weights of the vertexes and edges, and dividing the vertexes in each partition to corresponding computational nodes; each computing node stores the divided vertexes, a row and column sparse compression matrix is used for storing the adjacency relation of the vertexes and the edges, and the vertexes and the edges stored in each computing node are segmented and reconstructed according to the adjacency relation, so that the reconstructed vertexes and the reconstructed edges can adapt to all computing modes; and obtaining a current calculation mode according to the number of the currently active vertexes and the number of the currently active edges, and performing data processing in the current calculation mode by using the reconstructed vertexes and the reconstructed edges. According to the method, the storage representation mode of the graph data can be flexibly and efficiently converted in a mixed graph calculation mode system, and the calculation preposition requirements of different graph calculation modes are met.

To effectively reduce the number of apexes of a mesh of an object.SOLUTION: An information processing system simplifies a plurality of apexes to each of which at least one kind of material is set, which are apexes constituting a mesh defined in a virtual space, and draws a mesh based on the apexes in which simplification is reflected, on the basis of drawing settings based on a material of each apex. The simplification is achieved when a plurality of apexes of an apex group are replaced by one apex to which a total number of kinds of materials are set when a simplification condition at least including a fact that a total number of kinds of materials of each apex included in the apex group is a first number or less is satisfied for each of a plurality of apex groups including a plurality of adjacent apexes respectively.SELECTED DRAWING: Figure 19

The invention provides a group search method and device based on maximum group search on an uncertain graph, and relates to the field of data processing, and the method comprises the steps: obtaining an initial graph corresponding to a to-be-processed social network; determining a reference value of each vertex in the initial image; deleting vertexes of which the corresponding reference values are not greater than the heuristic solution values in the initial image to obtain a vertex set corresponding to the heuristic solution values, and determining a middle sub-image based on the vertex set; all the vertexes in the middle sub-graph are traversed to execute multiple search operations, a target sub-graph with the maximum number of the vertexes is determined based on candidate sub-graphs obtained through the search operations, an edge exists between any two different vertexes in the target sub-graph, and the product of the existence probabilities of all the edges is larger than a preset probability value; and determining a target social user group in the to-be-processed social network based on the target sub-graph. According to the method, the social user group can be searched based on the uncertain associated data between the social users.

The present disclosure relates to an induction range calibration method, device, ceiling appliance, equipment and storage medium. The induction range calibration method determines the standing position of the target object as the vertex of the required calibration area, obtains the position of the vertex and determines the coordinates of the horizontal plane projection of the point. When the number of vertices obtained is equal to 2, the rectangle surrounded by the straight line where the horizontal coordinates of the two vertices are located and the straight line where the vertical coordinates are located is taken as the induction working range or the induction non-working range. When the number of vertices obtained is greater than or equal to 3, the polygon surrounded by connecting all the vertices in turn is taken as the induction working range or the induction non-working range. In this way, the range that needs to be calibrated can be determined, and the user can be prevented from accidentally triggering the induction device, thereby causing resource waste. At the same time, this method is simple and does not require manual measurement and calculation, thereby improving the user experience.

The invention provides an incremental calculation optimization method for dynamic graph point pair analysis. The method comprises the following steps: executing edge adding and edge deleting operations through a preset hardware accelerator; the hardware accelerator comprises a prefetching module, an identification module, a scheduling module and a propagation module, and adopts an assembly line technology to process multi-batch dynamic changes in parallel. During edge adding operation, whether a source point can reach an edge adding starting point or not is judged, only edge topology changes are recorded if the source point cannot reach the edge adding starting point, and state calculation is not conducted; if yes, recording the topology change and updating the reachable vertex state of the edge-adding end point topology; during edge deletion operation, judging whether the edge deletion end point state depends on the edge deletion starting point state, if not, only recording edge deletion, and not triggering state update; and if yes, recording the change of the edge topology, and updating the edge deletion end point topology to reach a vertex state. The invention further provides a corresponding hardware accelerator. Therefore, the number of activated (traversed) vertexes can be effectively reduced, so that the calculation overhead is reduced to a great extent; and the processing efficiency can be further improved based on the preset hardware accelerator.

The present invention discloses a cost-balanced gate-level circuit diagram segmentation method and a computer storage medium. The cost-balanced gate-level circuit diagram segmentation method includes obtaining the symmetric adjacency matrix of the undirected graph corresponding to the gate-level circuit, the cost matrix of the vertices, the number of sub-graph blocks to be segmented, and the allowed deviation ratio; calculating the average cost of the sub-graph and the deviation threshold range; randomly selecting a deviation threshold from the deviation threshold range, combining the average cost, and obtaining the current cost range of the sub-graph; using iterative operations to segment the undirected graph multiple times, selecting a round of segmentation results with the least number of isolated vertices, and recording the segmentation results; segmenting the isolated vertices into the sub-graphs adjacent to them according to preset rules, so that the number of sub-graph blocks after segmentation is consistent with the number of sub-graph blocks to be segmented. The present invention realizes the reasonable segmentation of the gate-level circuit diagram, there are no isolated vertices or vertex groups, and the sub-graph costs are balanced, which improves the simulation efficiency.