WO2015027412A1 - 一种复杂图的处理方法和设备 - Google Patents

一种复杂图的处理方法和设备 Download PDF

Info

Publication number
WO2015027412A1
WO2015027412A1 PCT/CN2013/082468 CN2013082468W WO2015027412A1 WO 2015027412 A1 WO2015027412 A1 WO 2015027412A1 CN 2013082468 W CN2013082468 W CN 2013082468W WO 2015027412 A1 WO2015027412 A1 WO 2015027412A1
Authority
WO
WIPO (PCT)
Prior art keywords
graph
vertex
complex
vertices
new
Prior art date
Application number
PCT/CN2013/082468
Other languages
English (en)
French (fr)
Inventor
杨思晓
颜友亮
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to PCT/CN2013/082468 priority Critical patent/WO2015027412A1/zh
Priority to CN201380001321.7A priority patent/CN104737151A/zh
Publication of WO2015027412A1 publication Critical patent/WO2015027412A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/002D [Two Dimensional] image generation
    • G06T11/20Drawing from basic elements, e.g. lines or circles
    • G06T11/206Drawing of charts or graphs

Definitions

  • the present invention relates to the field of computer technologies, and in particular, to a method and a device for processing a complex graph.
  • the method for reducing the size of a complex graph is as shown in FIG. 1 , and the complex graph is roughened by merging adjacent points in a greatly matching region of the complex graph in the direction of the arrow, thereby reducing the scale of the complex graph and repeating
  • the rough processing process reduces the scale of the complex graph to the preset target, thereby obtaining a simple graph of the complex graph; then, by performing graph segmentation calculation on the simple graph, and expanding the segmentation result of the simple graph according to the inverse process of the rough processing process And the correction, mapping the segmentation result of the simple graph onto the original complex graph, and finally obtaining the segmentation result of the original complex graph.
  • finding a great match brings a large computational overhead, and is only applicable to the problem of graph segmentation.
  • Embodiments of the present invention provide a processing method and device for a complex graph, which implements a reduction in overhead in a computing process and applies multiple graph algorithms.
  • the embodiment of the present invention adopts the following technical solutions:
  • the embodiment of the present invention provides a processing method for a complex graph, including: deleting a first vertex in the complex graph, and the foregoing a first new map is obtained by the edge of the apex adjacent to the first vertex, wherein the first vertex is a vertex of degree 1, and the first new graph includes the complex graph a first vertex and a vertex and an edge other than an edge of the first vertex adjacent to the first vertex; Adding a weight of the first vertex in the complex graph to a weight of a neighboring point of the first vertex to obtain a weight of a neighboring point of the first vertex in the first new graph; Determining that the number of vertices of the first new map does
  • the method further includes: using the first new map as the first complex map; a first new vertex in the first complex graph, and an edge of the first vertex in the first complex graph connected to the first vertex in the first complex graph, to obtain a second new graph, where The second new map includes vertices in the first complex map except for the first vertex and the adjacent point of the first vertex and the edge connected to the first vertex in the first complex graph.
  • the second new map is equivalent to the complex graph Simple picture.
  • the method further includes: performing a graph operation on the second vertex in the simple graph to obtain an operation of the second vertex Resulting in the complex graph, performing a breadth-first search on each vertex of the complex graph, and obtaining a second vertex corresponding to each vertex of the complex graph, wherein the second vertex corresponding to the vertex is in the a second vertex closest to the vertex in the complex graph; a result of performing the graph operation on each vertex of the complex graph, respectively corresponding to an operation result of the second vertex corresponding to the vertex, to obtain the complex graph The result of the operation of each vertex.
  • the method further includes: performing a graph operation on each vertex of the simple graph to obtain an operation result of each vertex of the simple graph;
  • each vertex of the simple graph is traversed by a breadth-first search, and the vertex in the deleted complex graph corresponding to each vertex of the simple graph is obtained;
  • the operation results of the respective vertices of the simple graph are assigned to the vertices in the deleted complex map corresponding to the respective vertices of the simple graph.
  • the complex map comprises a scale-free complex graph.
  • the processing device of the complex image provided by the embodiment of the present invention includes: a deleting unit, configured to delete a first vertex in the complex graph, and an adjacent point of the first vertex and the first a first new map is obtained, wherein the first vertex is a vertex of degree 1, and the first new graph includes the first vertex and the first vertex in the complex graph. a vertex and an edge other than the edge to which the first vertex is connected;
  • a weight obtaining unit configured to add a weight of the first vertex in the complex graph to a weight of a neighboring point of the first vertex, and obtain an adjacency point to the first vertex at the first a weight in the new map
  • a determining unit configured to determine the first new map as a simple map equivalent to the complex map if it is determined that the number of vertices of the first new map does not exceed a preset threshold.
  • the determining unit is further configured to: if the determining unit determines that the number of vertices of the new map exceeds a preset threshold, using the first new map as First complex map;
  • the deleting unit is further configured to: delete a first vertex in the first complex graph, and connect a neighboring point of the first vertex in the first complex graph to a first vertex in the first complex graph;
  • a second new map is obtained, wherein the second new map includes a neighboring point of the first complex graph except the first vertex and the first vertex and a first one of the first complex graph a vertex and an edge other than the edge to which the vertex is connected;
  • the weight obtaining unit is further configured to: use the weight of the first vertex in the first complex graph and the first vertex in the first complex graph The weights of the neighboring points are added to obtain the weights of the neighboring points of the first vertex in the first complex graph in the second new graph;
  • the determining unit is further configured to: if
  • the method further includes: a graph operation unit, configured to perform a graph operation on the second vertex in the simple graph to obtain the second a result of the operation of the vertex; a search unit, configured to perform a breadth-first search on each vertex of the complex graph in the complex graph, to obtain a second vertex corresponding to each vertex of the complex graph, where the vertex Corresponding second vertex is a second vertex closest to the vertex in the complex graph; a result obtaining unit, configured to perform a result of the graph operation on each vertex of the complex graph, respectively corresponding to the vertex Corresponding to the operation result of the second vertex, the operation result of each vertex of the complex graph is obtained.
  • a graph operation unit configured to perform a graph operation on the second vertex in the simple graph to obtain the second a result of the operation of the vertex
  • a search unit configured to perform a breadth-first search on each vertex of the complex graph in the complex graph, to obtain a second vertex corresponding to each vertex of
  • the graph operation unit is further configured to perform a graph operation on each vertex of the simple graph to obtain the simple graph.
  • the operation result of each vertex; the device further includes an acquiring unit, configured to acquire a deleted complex graph corresponding to each vertex of the simple graph, where the deleted complex graph corresponding to the vertex of the simple graph is passed in the complex In the figure, the vertices other than the vertices in the simple graph are deleted;
  • the searching unit is further configured to: delete a complex corresponding to each vertex of the simple graph
  • the vertices of the simple graph are traversed by the breadth-first search, respectively, to obtain the vertices in the deleted complex graph corresponding to the vertices of the simple graph;
  • the device further includes an allocating unit,
  • the operation results of the respective vertices of the graph are assigned to the vertices in the deleted complex map corresponding to the respective vertices of the simple graph.
  • the complex graph includes a scale-free complex graph.
  • a processing device for a complex graph includes: a processor, configured to delete a first vertex in the complex graph, and a neighboring point of the first vertex and the first vertex a first new map is obtained, wherein the first vertex is a vertex of degree 1, and the first new graph includes adjacency of the complex graph except the first vertex and the first vertex a vertex and an edge other than an edge connected to the first vertex; and adding a weight of the first vertex in the complex graph to a weight of an adjacent point of the first vertex a weight of a neighboring point of the first vertex in the first new map;
  • the processor is further configured to: if the determining unit determines that the number of vertices of the new map exceeds a preset threshold, using the first new map as a first complex map; and deleting a first vertex in the first complex graph, and an edge of the first vertex in the first complex graph connected to a first vertex in the first complex graph, Obtaining a second new map, wherein the second new map includes an adjacent point in the first complex graph except the first vertex and the first vertex is connected to a first vertex in the first complex graph Adding vertices and edges outside the edges; and adding the weights of the first vertices in the first complex map to the weights of the contiguous points of the first vertices in the first complex map The first in the first complex map a weight of
  • the processor is further configured to: perform a graph operation on the second vertex in the simple graph to obtain the second vertex Operation result
  • the processor is further configured to perform a graph operation on each vertex of the simple graph to obtain the simple graph. The result of the operation of each vertex;
  • each vertex of the simple graph is traversed by a breadth-first search, respectively, to obtain a vertex in the deleted complex graph corresponding to each vertex of the simple graph;
  • the operation results of the respective vertices of the simple graph are assigned to the vertices in the deleted complex map corresponding to the respective vertices of the simple graph.
  • the complex graph includes a scale-free complex graph.
  • the embodiment of the invention provides a processing method and device for a complex graph. According to the scale-freeness of the complex graph, the vertices with a degree of deletion of 1 reduce the scale of the complex graph, and obtain and complex
  • the simple diagram of the equivalent diagram does not need to find the maximum matching of the complex graph, which reduces the overhead of the calculation process.
  • the process of reducing the scale of the complex graph is independent of the process of the graph operation, and can be based on the operation result and simple graph of the simple graph.
  • the equivalent relationship of complex graphs results in the operation of complex graphs, making it suitable for a variety of graph algorithms.
  • FIG. 2(1) is a first aspect of the present invention provided by the embodiment of the present invention
  • FIG. 2(2) is a first new diagram of the embodiment of the present invention
  • FIG. 3 is a flowchart of a method for processing a complex graph according to an embodiment of the present invention
  • FIG. 4 is a flowchart of a graph for obtaining a complex graph according to an operation result of a simple graph according to an embodiment of the present invention
  • Figure 5 (1) is a result of the operation of a simple diagram of the embodiment of the present invention
  • Figure 5 (2) is a result of the operation of a complex diagram according to an embodiment of the present invention
  • FIG. 7 (1) is a complex complex diagram corresponding to the vertices of a simple graph according to an embodiment of the present invention
  • FIG. 7 (2) is a deletion complex diagram corresponding to a vertex of another simple graph according to an embodiment of the present invention.
  • FIG. 7 is a flowchart of another complicated graph according to an embodiment of the present invention
  • FIG. 8 is a flowchart of another complex graph processing method according to an embodiment of the present invention
  • FIG. 9 is a schematic diagram of an apparatus for processing a complex diagram according to an embodiment of the present invention
  • FIG. 10 is a schematic diagram of a device of a processing device of another complicated figure according to an embodiment of the present invention.
  • FIG. 11 is a hardware structural diagram of a processing device of a complex graph according to an embodiment of the present invention.
  • the technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. example. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.
  • a basic property of a complex network graph is scale-free, which means that the degree of vertices in the network graph satisfies the power law distribution, while the network graph that satisfies the law distribution has the following properties: 1.
  • FIG. 2 (1) there is a scale-free complex graph in which an open circle represents a vertex; a solid line in black represents an edge; and a number in a vertex represents identification information of a vertex.
  • This embodiment is preferably identified by an identity number (ID, referred to as ID), and the number above the vertex indicates the weight of the corresponding vertex.
  • ID identity number
  • FIG. 2 (1) is set as an undirected graph. And the weight of each vertex is set to 1.
  • a method for processing a complex graph may include:
  • S301 deleting a first vertex in the complex graph, and an edge of the first vertex adjacent to the first vertex, to obtain a first new graph, where the first vertex is a vertex with a degree of 1, and the first new graph includes a complex
  • the complex graph includes a scale-free complex graph.
  • the complex graph and the first new graph belong to the corresponding mapping relationship, and the two are equivalent.
  • the first new map is taken as a simple graph equivalent to the complex graph.
  • the preset threshold is a maximum range used to describe the scale of the simple graph, which may be the maximum number of vertices that the simple graph can contain; exemplarily, if the number of vertices of the first new graph exceeds a preset threshold,
  • the size of the first new map can be further reduced as follows: the first new map is taken as the first complex graph; the first vertex in the first complex graph is deleted, and the neighboring point of the first vertex in the first complex graph is deleted Obtaining a second new map with the edge connected to the first vertex in the first complex graph, where the second new graph includes the neighboring points of the first complex graph except the first vertex and the first vertex and the first complex graph Vertices and edges outside the edges to which the first vertex is connected; as described earlier, the second new map is equivalent to the complex graph.
  • the second new map is taken as a simple map equivalent to the complex graph
  • the second new map is taken as the new first complex map, and the processing of the first complex graph is repeated until the number of vertices of the obtained second new graph is obtained. Not exceeding the preset threshold, it is understandable that the simple graph obtained from this is still equivalent to the complex graph.
  • the graph operation can be performed on the simple graph, and then the equivalent relationship between the simple graph and the complex graph can be used.
  • the result of the operation of the single image is the result of the operation of the complex graph.
  • S602 Obtain a deleted complex graph corresponding to each vertex of the simple graph, where the deleted complex graph corresponding to the vertex of the simple graph is obtained by deleting other vertices other than the vertex in the simple graph in the complex graph;
  • each vertex of the simple graph is traversed by the breadth-first search, and the vertex in the deleted complex graph corresponding to each vertex of the simple graph is obtained;
  • S604 Assign the operation result of each vertex of the simple graph to the vertex in the deleted complex graph corresponding to each vertex of the simple graph.
  • the embodiment of the invention provides a processing method for a complex graph. According to the scale-free property of the complex graph, the vertices with a degree of deletion 1 are used to reduce the scale of the complex graph, and a simple graph equivalent to the complex graph is obtained, and no complicated graph is needed.
  • the maximal matching reduces the overhead of the calculation process; and the process of reducing the complexity of the graph is independent of the process of the graph operation, and can obtain the operation of the complex graph according to the operation result of the simple graph and the equivalent relationship between the simple graph and the complex graph. As a result, it is applicable to a variety of graph algorithms.
  • a method for processing a complex graph provided by an embodiment of the present invention is further described in detail.
  • the complex graph includes a scale-free complex graph.
  • the vertex ID is in the set ⁇ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ⁇
  • the vertices are all first vertices, and the contiguous points of the first vertices connected to the first vertices may be represented by an ID tuple by the first vertex ID and the adjacent point ID of the first vertex, for example: (first vertex ID, second Vertex ID), so the set of deleted edges is ⁇ ( 1, 11 ), ( 2, 12), ( 3, 13 ), ( 4, 14), ( 5, 15 ), ( 6, 16), ( 7 , 17), ( 8, 18), ( 9, 19), ( 10, 20) ⁇ ;
  • the first new map obtained is shown in Figure 2 ( 2). Further, it can be understood that since the first vertex is only connected to its neighboring point, the complex graph and the first new graph belong to the corresponding mapping relationship, and the two are equivalent.
  • the first new map is taken as a simple graph equivalent to the complex graph.
  • a preset threshold is the maximum range used to describe the scale of the simple graph, which may be simple
  • the maximum number of vertices that can be included in the graph, specifically, the threshold value preset in this embodiment is 5. If the number of vertices in FIG. 2 (2) does not exceed 5, then FIG. 2 (2) is taken as FIG. 2 ( 1) Equivalent simple graph; exemplarily, in this embodiment, the number of vertices of FIG.
  • the first new map is taken as the first complex map;
  • Figure 2 (2) as the first complex map; deleting the first vertex in the first complex graph, and the first vertex in the first complex graph a second new map is obtained by connecting the edge of the adjacent point to the first vertex in the first complex image, wherein the second new image includes the adjacent point of the first complex image and the first vertex and the first complex Vertices and edges outside the edges to which the first vertex is connected;
  • deleting the first vertex in FIG. 2 ( 2 ) and the adjacent point of the first vertex in FIG. 2 ( 2 ) are connected to the first vertex in FIG. 2 ( 2 ), as shown in FIG. 2 ( 2 )
  • the set of the first vertex ID of Figure 2 ( 2 ) is ⁇ 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20 ⁇ , and the set of deleted edges is ⁇ ( 11, 21 ), ( 12, 21 ), ( 13, 21 ),
  • the second new map obtained is shown in Fig. 2 (3); from the foregoing, in the same way, the second new map is equivalent to the first new map, and then the complex map can be known. It is also equivalent. Adding the weight of the first vertex in the first complex graph to the weight of the neighboring point of the first vertex in the first complex graph to obtain the neighboring point of the first vertex in the first complex graph in the second new graph Specifically, in Figure 2 (2), adding the weight of the first vertex to the weight of its corresponding neighboring point, the weight of the neighboring point in the second new graph can be obtained, as shown in Fig. 2 (3) as shown. If it is determined that the number of vertices of the second new map does not exceed a preset threshold, the second new map is taken as a simple map equivalent to the complex map;
  • the second new map is used as a new first complex map, and the processing of the first complex graph is repeated until the second new map is obtained.
  • the number of vertices does not exceed a preset threshold; in this embodiment, the number of vertices of FIG. 2 ( 3 ) is 2, which is less than the preset threshold 5, and FIG. 2 ( 3 ) can be regarded as FIG. 2 ( 1 ) Simple diagram of effectiveness.
  • a graph operation can be performed on the simple graph. It can be understood that the process of obtaining a simple graph equivalent to the complex graph and the subsequent graph operation on the simple graph are understood.
  • the set map operation is a graph division operation
  • the result of graph division for FIG. 2 (3) is as shown in FIG. 5 (1)
  • the broken line is a division line of the graph division
  • the result of the calculation of Fig. 5 (1) is that the vertex 21 is the first segment and the vertex 22 is the second segment, which is represented as 21 ⁇ 1, 22 ⁇ 2,
  • S402 performing a breadth-first search on each vertex of the complex graph in the complex graph, and obtaining a second vertex corresponding to each vertex of the complex graph, where the second vertex corresponding to the vertex is the second closest to the vertex in the complex graph Vertex;
  • the corresponding second vertex obtained by performing a breadth-first search for each vertex in FIG. 2(1) is:
  • the vertex ID set is ⁇ 1, 2, 3, 4, 5
  • the second vertex corresponding to the vertex of 11, 12, 13, 14, 15 ⁇ is the vertex 21, and the vertex ID set is the vertex corresponding to ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇
  • the second vertex is the vertices 22.
  • the segmentation result of the vertex ID set is ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ is the same as the vertex 21
  • the segmentation can be expressed as ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ 1
  • the set of vertex IDs is ⁇ 6, 7, 8, 9, 10, 16, 17, 18,
  • the segmentation result of the vertices of 19, 20 ⁇ is the same as the vertices 22 in the second segment, which can be expressed as: ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ 2, therefore, right
  • the result of the graph segmentation in Fig. 2 (1) is shown in Fig. 5 (2).
  • the second method can be as shown in Figure 6:
  • step S601 Performing a graph operation on each vertex of the simple graph to obtain an operation result of each vertex of the simple graph; Specifically, as step S401, details are not described herein again.
  • S602 Obtain a deleted complex graph corresponding to each vertex of the simple graph, where the deleted complex graph corresponding to the vertex of the simple graph is obtained by deleting other vertices other than the vertex in the simple graph in the complex graph;
  • Fig. 2 (3) the deletion complex map corresponding to the vertex 21 and the vertex 22 is as shown in Fig. 7 (1) and Fig. 7 (2), respectively.
  • each vertex of the simple graph is traversed by the breadth-first search, and the vertex in the deleted complex graph corresponding to each vertex of the simple graph is obtained;
  • the vertex 21 and the vertex 22 are traversed by the breadth-first search, and the vertex set of the vertex 21 in FIG. 7(1) can be obtained, and the ID set is ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ ; Similarly, the set of vertices of vertices 22 in Figure 7 (2), whose ID set is ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ .
  • S604 assigning the operation result of each vertex of the simple graph to the vertex in the deleted complex graph corresponding to each vertex of the simple graph; specifically, the operation result is that the vertex 21 is the first segment, and the vertex 22 is the second segment, which is represented as 21 ⁇ 1, 22 ⁇ 2.
  • the result of the vertex of the ID set being ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ is that the vertex of the ID set is the same as the vertex 21, and can be expressed as ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ 1; and the set of IDs is the vertices of ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ The result is that the vertex of the ID set is the same as the vertex 22, and can be expressed as ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ 2;
  • the embodiment of the invention provides a processing method for a complex graph.
  • the vertices with a degree of deletion 1 are used to reduce the scale of the complex graph, and a simple graph equivalent to the complex graph is obtained, and no complicated graph is needed.
  • the maximal matching reduces the overhead of the calculation process; and the process of reducing the complexity of the graph is independent of the process of the graph operation, and can obtain the complex graph according to the operation result of the simple graph and the equivalent relationship between the simple graph and the complex graph.
  • the network diagram can be represented by a matrix or by multiple lists.
  • steps S301-S303 in the embodiment shown in FIG. 3 can be implemented according to list 1, list 2, and list 3, as shown in FIG.
  • S801 traversing list 1, obtaining an item of degree 1 and saving it in the list la, and deleting the item with a degree of 1 in the list 1 to obtain the list 1A;
  • the list la can be obtained as: [(1, 1), (2, 1), (3, 1), (4, 1), (5, 1), (6 , 1), (7, 1), (8, 1), (9, 1) , (10, 1)];
  • Listing 1A is [(11, 2), (12, 2), (13, 2), (14, 2), (15,
  • an isolated point of degree 0 can be saved in a separately opened list and deleted in list 1 and list 2, which can save the time cost of subsequent traversal.
  • the list 3A obtained after deleting the vertices in list 1A in Listing 3 is: [(11, 21), (12, 21), (13, 21), (14, 21), (15, 21 ), (16, 22), (17, 22), (18, 22), (19, 22), (20, 22), (21, 22)], then, List 1A and List 3A can get the first New picture 2 (2).
  • Fig. 2 (2) can be represented by list 1 A, list 2A, and list 3A.
  • S804 determining whether the number of vertices of the first new map does not exceed a preset threshold: if not, the first new map is regarded as a simple map equivalent to the complex map; if exceeded, the first new map is taken as the first In the complex graph, S801-804 is repeatedly executed until the number of vertices of the new graph obtained does not exceed the preset threshold; exemplarily, the preset threshold is 5, and the number of vertices in FIG. 2 (2) exceeds 5, The S 801 -S 804 is executed repeatedly until the figure 2 ( 3 ) is obtained, and the number of vertices does not exceed the preset threshold, so FIG. 2 ( 3 ) is an equivalent simple diagram of FIG. 2 ( 1 ).
  • the graph operation can be performed on the simple graph. Further, the operation result of the complex graph can be obtained according to the operation result of the simple graph, and the complex graph is obtained according to the operation result of the simple graph.
  • the result of the operation is as follows
  • steps S401 - S403 or steps S 601 - S604 those skilled in the art can implement according to the above-mentioned steps by using a list for representing a simple graph, which will not be described herein; preferably, for saving a list of isolated points It can be understood that the operation result of the isolated point needs to be determined according to the algorithm itself of the graph operation. In this embodiment, if there is an isolated point, the calculation result of the graph segmentation of the complex graph can ignore the influence of the isolated point.
  • the embodiment of the invention provides a processing method for a complex graph.
  • the vertices with a degree of deletion 1 are used to reduce the scale of the complex graph, and a simple graph equivalent to the complex graph is obtained, and no complicated graph is needed.
  • the maximal matching reduces the overhead of the calculation process; and the process of reducing the complexity of the graph is independent of the process of the graph operation, and can obtain the operation of the complex graph according to the operation result of the simple graph and the equivalent relationship between the simple graph and the complex graph. As a result, it is applicable to a variety of graph algorithms.
  • a processing device 90 for a complex graph may include: a deleting unit 901, configured to delete a first vertex in a complex graph, and a neighboring point of the first vertex is connected to the first vertex The first new image is obtained, wherein the first vertex is a vertex with a degree of 1, and the first new graph comprises a boundary of the complex graph except the first vertex and the adjacent point of the first vertex connected to the first vertex a vertices and edges; a weight obtaining unit 902, configured to add the weights of the first vertices in the complex graph to the weights of the contiguous points of the first vertices, and obtain the weights of the contiguous points of the first vertex in the first new graph Value
  • the determining unit 903 is configured to: when determining that the number of vertices of the first new map does not exceed a preset threshold, use the first new map as a simple graph equivalent to the complex graph.
  • the complex graph includes a scale-free complex graph, and this embodiment passes through FIG. 2 (1).
  • the scheme is illustrated with a scale-free complex diagram. The description of Figure 2 (1) is as before, and will not be repeated here. Exemplarily, as shown in FIG.
  • the first vertex deleted by the deleting unit 901 is a vertex whose vertex ID is in the set ⁇ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ⁇
  • the edge of the first vertex adjacent to the first vertex may be represented by an ID tuple by the first vertex ID and the adjacent point ID of the first vertex, for example: (first vertex ID, second vertex ID),
  • the set of deleted edges is ⁇ ( 1, 11 ), ( 2, 12 ), ( 3, 13 ), ( 4, 14), ( 5, 15), ( 6, 16), ( 7, 17), ( 8, 18), ( 9, 19), ( 10, 20) ⁇ ;
  • the first new map obtained is shown in Figure 2 ( 2 ). Exemplarily, in FIG.
  • the weight obtaining unit 902 adds the weight of the first vertex to the weight of the corresponding neighboring point, and obtains the weight of the neighboring point in the first new map.
  • the determining unit 903 determines whether the number of vertices in the figure exceeds a preset threshold. Specifically, the threshold value preset in this embodiment is 5, if the determining unit 903 determines that the number of vertices in FIG. 2 ( 2 ) does not exceed 5, then the determining unit 903 takes FIG. 2 ( 2 ) as a simple diagram equivalent to FIG. 2 ( 1 ); in this embodiment, FIG.
  • the determining unit 903 uses the first new map as the first complex map. Specifically, the determining unit 903 uses FIG. 2 ( 2 ) as the first complex graph; the deleting unit 901 is further configured to: Deleting a first vertex in the first complex graph, and an edge of the first vertex in the first complex graph connected to the first vertex in the first complex graph, to obtain a second new graph, wherein the second new graph And including a vertex and an edge of the first complex image except the first vertex and the adjacent point of the first vertex and the edge connected to the first vertex in the first complex graph; specifically, the deleting unit 901 may delete the graph 2 ( 2 ) The first vertex in the middle and the adjacent point of the first vertex in Figure 2 (2) and Figure 2 (2) The edge connected to the first vertex in Fig. 2 (2) shows that the set of the first vertex ID of Fig. 2 (2) is ⁇ 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 ⁇
  • the weight obtaining unit 902 is further configured to: add the weight of the first vertex in the first complex graph to the weight of the neighboring point of the first vertex in the first complex graph to obtain the first vertex in the first complex graph The weight of the neighboring point in the second new map;
  • the weight of the first vertex is added to the weight of the corresponding neighboring point, and the weight of the neighboring point in the second new graph can be obtained, as shown in FIG. 2 ( 3 ). Shown.
  • the determining unit 903 is further configured to: if it is determined that the number of vertices of the second new map does not exceed a preset threshold, use the second new map as a simple graph equivalent to the complex graph;
  • the determining unit 903 determines that the number of vertices of the second new map exceeds a preset threshold, the second new map is used as a new first complex map, and the processing process of the first complex map by the three functional units is repeated. Until the number of vertices of the obtained second new image does not exceed the preset threshold; in this embodiment, the number of vertices of FIG. 2 ( 3 ) is 2, which is less than the preset threshold 5, and the determining unit 903 can display FIG. 2 ( 3) As a simple diagram equivalent to Figure 2 (1). For example, after obtaining a simple graph equivalent to a complex graph, the graph operation can be performed on the simple graph. Further, as shown in FIG. 10, the device 90 may further include: a graph operation unit 904 for the simple graph The second vertex in the graph performs a graph operation to obtain an operation result of the second vertex;
  • the set map operation is a graph division operation
  • the graph operation unit 904 performs graph segmentation on FIG. 2 ( 3 ) as shown in FIG. 5 ( 1 ), and the broken line is graph split.
  • the operation result of Fig. 5 (1) can be obtained as the first segment of the vertex 21, and the second segmentation of the vertex 22, which is represented as 21 ⁇ 1, 22 ⁇ 2, and the search unit 905 is used for complex in the complex graph.
  • Each of the vertices of the graph performs a breadth-first search to obtain a second vertex corresponding to each vertex of the complex graph, wherein the second vertex corresponding to the vertex is the second vertex closest to the vertex in the complex graph;
  • the search unit 905 performs the breadth-first search for each vertex in FIG. 2(1), and the corresponding second vertex result is: the vertex ID set is ⁇ 1, 2, 3, 4, 5, 11, 12, The second vertex corresponding to the vertex of 13, 14, 15 ⁇ is the vertex 21, and the vertex ID set is ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ The second vertex corresponding to the vertex of the vertex is the vertex 22.
  • the result obtaining unit 906 is configured to perform a graph operation on each vertex of the complex graph, and respectively corresponding to the operation result of the second vertex corresponding to the vertex, and obtain an operation result of each vertex of the complex graph.
  • the result obtaining unit 906 can obtain: the segmentation result of the vertex with the vertex ID set being ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ is the vertex 21 Same as the first segmentation, which can be expressed as ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ 1;
  • the set of vertex IDs is ⁇ 6, 7, 8, 9, 10, 16,
  • the segmentation result of the vertices of 17, 18, 19, 20 ⁇ is the same as the vertices 22 in the second segment, which can be expressed as: ⁇ 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ 2 Therefore, the result of the segmentation of the graph of Fig. 2 (1) is shown in Fig. 5 (2).
  • the graph operation unit 904 is further configured to perform a graph operation on each vertex of the simple graph to
  • the device further includes an obtaining unit 907, configured to acquire a deleted complex graph corresponding to each vertex of the simple graph, where the deleted complex graph corresponding to the vertex of the simple graph is obtained by deleting other vertices other than the vertex in the simple graph in the complex graph.
  • the deletion complex map corresponding to the vertex 21 and the vertex 22 acquired by the obtaining unit 907 in FIG. 2 ( 3 ) is also used for the search unit 905 as shown in FIG. 7 ( 1 ) and FIG. 7 ( 2 ), respectively.
  • each vertex of the simple graph is traversed by the breadth-first search, and the vertex in the complex graph corresponding to each vertex of the simple graph is obtained; specifically, the search unit 905 is respectively in FIG. 7 (1) and in Fig. 7(2), for the vertex 21 and the vertex 22 to traverse by the breadth-first search, the vertex set of the vertex 21 in Fig. 7 (1) can be obtained, and the ID set is ⁇ 1, 2, 3, 4 , 5, 11, 12, 13, 14, 15 ⁇ ; Similarly, the set of vertices of vertices 22 in Figure 7 ( 2 ), whose ID set is ⁇ 6, 7, 8, 9, 10, 16, 17, 18 , 19, 20 ⁇ .
  • the device 90 further includes an allocating unit 908, configured to allocate the operation result of each vertex of the simple graph to the vertex in the deleted complex graph corresponding to each vertex of the simple graph; specifically, the operation result is that the vertex 21 is the first segment, and the vertex 22 For the second segmentation, it is represented as 21 ⁇ 1, 22 ⁇ 2.
  • the allocating unit 908 assigns the segmentation result of the vertex 21 to the vertex of the ID set of the vertex to ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ , and the vertex of the ID set is the same as the vertex 21 In the first segmentation, it can be expressed as ⁇ 1, 2, 3, 4, 5, 11, 12, 13, 14, 15 ⁇ 1; the allocating unit 908 assigns the segmentation result of the vertex 22 to the ID set at the vertex to be ⁇
  • the vertex of 6, 7, 8, 9, 10, 16, 17, 18, 19, 20 ⁇ , the vertex of the ID set is the same as the vertex 22, and can be expressed as ⁇ 6, 7, 8, 9, 10 , 16, 17, 18, 19, 20 ⁇ 2; Therefore, the segmentation result of Fig. 2 (1) is shown in Fig. 7 (3).
  • the embodiment of the present invention provides a processing device 90 of a complex graph.
  • the vertices with a degree of deletion 1 are used to reduce the scale of the complex graph, and a simple graph equivalent to the complex graph is obtained, and no complicated search is needed.
  • the maximum matching of the graph reduces the overhead of the calculation process; and the process of reducing the scale of the complex graph is independent of the process of the graph operation, and can obtain the complex graph according to the operation result of the simple graph and the equivalent relationship between the simple graph and the complex graph.
  • the result of the operation makes it suitable for a variety of graph algorithms.
  • a processing device 90 of a complex graph is provided according to an embodiment of the present invention, and a complex graph may include a complex graph without scale.
  • This embodiment has a scale-free manner as shown in FIG. 2 (1).
  • the complex diagram is used to describe the scheme. The description of FIG. 2 (1) is as described above, and is not described here.
  • the method may include: at least one processor 1101; at least one input unit 1102, configured to input the complex graph to the device 90.
  • the specific form may be a scanner, which is not limited by the embodiment of the present invention; the memory 1103 and the communication bus 1104 are used to implement connection communication between these devices.
  • the communication bus 1104 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus, or an extended industry standard architecture ( Extended). Industry Standard Architecture, referred to as EISA) bus.
  • the bus 1104 can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 11, but it does not mean that there is only one bus or one type of bus.
  • the memory 1103 is for storing executable program code, the program code including computer operating instructions.
  • the memory 1103 may include a high speed RAM memory, and may also include a non-volatile memory such as at least one disk memory.
  • the processor 1101 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more configured to implement the embodiments of the present invention. integrated circuit.
  • the processor 1101 is configured to execute executable program code stored in the memory 1103, such as a computer program, to execute a program corresponding to the executable code.
  • the processor 1101 is configured to: the processor 1101, configured to delete the first vertex in the complex graph, and the edge of the first vertex adjacent to the first vertex to obtain the first new graph
  • the first vertex is a vertex of degree 1
  • the first new graph includes vertices and edges of the complex graph except the first vertex and the edge of the first vertex adjacent to the first vertex; and the complex graph
  • the weight of the first vertex is added to the weight of the adjacent point of the first vertex to obtain the weight of the adjacent point of the first vertex in the first new graph; and if the number of vertex of the first new graph is determined not to exceed
  • the threshold is set, and the first new map is taken as a simple graph equivalent to the complex graph.
  • the processor 1101 is further configured to: if the determining unit determines that the number of vertices of the new graph exceeds a preset threshold, using the first new map as the first complex map; and deleting the first vertex in the first complex graph, and a second new map is obtained by the adjacent point of the first vertex in the first complex graph and the edge connected to the first vertex in the first complex graph, wherein the second new graph includes the first vertex and the first vertex in the first complex graph a vertex and an edge of a vertex adjacent to the edge connected to the first vertex in the first complex graph; and adjacency of the first vertex in the first complex graph with the first vertex in the first complex graph The weights of the points are added to obtain the adjacent points of the first vertex in the first complex graph.
  • the weight in the second new map and if it is determined that the number of vertices of the second new map does not exceed a preset threshold, the second new map is taken as a simple map equivalent to the complex map.
  • the graph operation can be performed on the simple graph.
  • the processor 1 101 is further configured to: perform a graph operation on the second vertex in the simple graph to obtain a second graph The result of the operation of the vertex; and the breadth-first search for each vertex of the complex graph in the complex graph, and the second vertex corresponding to each vertex of the complex graph, wherein the second vertex corresponding to the vertex is the distance vertex in the complex graph The nearest second vertex; and the result of performing the graph operation on each vertex of the complex graph, respectively corresponding to the operation result of the second vertex corresponding to the vertex, and obtaining the operation result of each vertex of the complex graph.
  • the processor 1101 is further configured to perform a graph operation on each vertex of the simple graph to obtain an operation result of each vertex of the simple graph; and obtain a delete complex graph corresponding to each vertex of the simple graph, where the vertex of the simple graph Corresponding deletion complex graphs are obtained by deleting other vertices other than vertices in the simple graph in the complex graph; and in the delete complex graph corresponding to each vertex of the simple graph, traversing each vertex of the simple graph by breadth-first search Obtaining the vertices in the complex graph corresponding to the respective vertices of the simple graph; and assigning the operation results of the vertices of the simple graph to the vertices in the deleted complex graph corresponding to the respective vertices of the simple graph.
  • the embodiment of the present invention provides a processing device 90 for a complex graph.
  • the vertices with a degree of deletion of 1 reduce the scale of the complex graph, and obtain a simple graph equivalent to the complex graph, without further searching for complexity.
  • the maximum matching of the graph reduces the overhead of the calculation process; and the process of reducing the scale of the complex graph is independent of the process of the graph operation, and can obtain the complex graph according to the operation result of the simple graph and the equivalent relationship between the simple graph and the complex graph.
  • the result of the operation makes it suitable for a variety of graph algorithms.
  • the program when executed, performs the steps including the foregoing method embodiments; and the foregoing storage medium includes: a medium that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk. .

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Generation (AREA)

Abstract

本发明实施例提供了一种复杂图的处理方法和设备,实现减少计算过程中的开销并适用多种图算法。该方法包括:删除复杂图中的第一顶点,以及第一顶点的邻接点与第一顶点相连的边,得到第一新图;将复杂图中第一顶点的权值与第一顶点的邻接点的权值相加得到第一顶点的邻接点在第一新图中的权值;若确定第一新图的顶点数目不超过预设的阈值,则将第一新图作为与复杂图等效的简单图。

Description

一种复杂图的处理方法和设备
技术领域
本发明涉及计算机技术领域, 尤其涉及一种复杂图的处理方法 和设备。
背景技术
目前, 降低复杂图规模的方法如图 1 所示, 按照箭头的方向通 过在复杂图的极大匹配区域内对相邻点进行合并的方式对复杂图进 行粗糙处理, 降低复杂图规模, 并且重复粗糙处理过程使得复杂图 的规模降低至预设的目标, 以此得到复杂图的简单图; 然后通过对 简单图进行图分割计算, 并对简单图的分割结果按照粗糙处理过程 的逆过程进行扩大和修正, 将简单图的分割结果映射到原复杂图上, 最终得到对原复杂图的图分割结果。 在实现上述现有技术的过程中, 发明人发现现有技术中至少存 在如下问题: 寻找极大匹配带来很大的计算开销, 而且仅适用于图 分割的问题。
发明内容 本发明实施例提供一种复杂图的处理方法和设备, 实现减少计 算过程中的开销并适用多种图算法。 为达到上述目 的, 本发明的实施例采用如下技术方案: 第一方面、 本发明实施例提供了一种复杂图的处理方法, 包括: 删除所述复杂图中的第一顶点, 以及所述第一顶点的邻接点与 所述第一顶点相连的边, 得到第一新图, 其中, 所述第一顶点为度 为 1 的顶点, 所述第一新图包含所述复杂图中除所述第一顶点以及 所述第一顶点的邻接点与所述第一顶点相连的边之外的顶点和边; 将所述复杂图中所述第一顶点的权值与所述第一顶点的邻接点 的权值相加得到所述第一顶点的邻接点在所述第一新图中的权值; 若确定所述第一新图的顶点数目不超过预设的阈值, 则将所述 第一新图作为与所述复杂图等效的简单图。 在第一种可能的实现方式中, 结合第一方面, 若确定所述新图 的顶点数目超过预设的阈值, 所述方法还包括: 将所述第一新图作为第一复杂图; 删除所述第一复杂图中的第一顶点, 以及所述第一复杂图中的 第一顶点的邻接点与所述第一复杂图中的第一顶点相连的边, 得到 第二新图, 其中, 所述第二新图包含所述第一复杂图中除所述第一 顶点以及所述第一顶点的邻接点与所述第一复杂图中的第一顶点相 连的边之外的顶点和边; 将所述第一复杂图中所述第一顶点的权值与所述第一复杂图中 的第一顶点的邻接点的权值相加得到所述第一复杂图中的第一顶点 的邻接点在所述第二新图中的权值; 若确定所述第二新图的顶点数目不超过预设的阈值, 则将所述 第二新图作为与所述复杂图等效的简单图。
在第二种可能的实现方式中, 结合第一方面或第一种可能的实 现方式, 所述方法还包括: 对所述简单图中的第二顶点进行图运算得到所述第二顶点的运 算结果; 在所述复杂图中对所述复杂图的各个顶点分别进行广度优先搜 索, 得到所述复杂图的各个顶点分别对应的第二顶点, 其中, 所述 顶点对应的第二顶点为在所述复杂图中距离所述顶点最近的第二顶 点; 将所述复杂图的各个顶点进行所述图运算的结果, 分别与所述 顶点对应的第二顶点的运算结果对应, 得到所述复杂图的各个顶点 的运算结果。
在第三种可能的实现方式中, 结合第一方面或者第一种可能的 实现方式, 所述方法还包括: 对所述简单图的各个顶点分别进行图运算, 得到所述简单图的 各个顶点的运算结果;
获取所述简单图的各个顶点对应的删除复杂图, 其中, 所述简 单图的顶点对应的删除复杂图是通过在所述复杂图中删除所述简单 图中除顶点以外的其他顶点得到的; 在所述简单图的各个顶点对应的删除复杂图中, 对所述简单图 的各个顶点分别通过广度优先搜索遍历, 得到所述简单图的各个顶 点对应的删除复杂图中的顶点; 将所述简单图的各个顶点的运算结果分配至所述简单图的各个 顶点对应的所述删除复杂图中的顶点。
在第四种可能的实现方式中, 结合第一方面、 第一种至第三种 可能的实现方式中的任一项, 所述复杂图包括无标度的复杂图。
第二方面, 本发明实施例提供的一种复杂图的处理设备, 包括: 删除单元, 用于删除所述复杂图中的第一顶点, 以及所述第一 顶点的邻接点与所述第一顶点相连的边, 得到第一新图, 其中, 所 述第一顶点为度为 1 的顶点, 所述第一新图包含所述复杂图中除所 述第一顶点以及所述第一顶点的邻接点与所述第一顶点相连的边之 外的顶点和边;
权值获取单元, 用于将所述复杂图中所述第一顶点的权值与所 述第一顶点的邻接点的权值相加获取到所述第一顶点的邻接点在所 述第一新图中的权值; 确定单元, 用于若确定所述第一新图的顶点数目不超过预设的 阈值, 则将所述第一新图作为与所述复杂图等效的简单图。 在第一种可能的实现方式中, 结合第二方面, 所述确定单元还 用于, 若所述确定单元确定所述新图的顶点数目超过预设的阈值, 将所述第一新图作为第一复杂图; 所述删除单元还用于, 删除所述第一复杂图中的第一顶点, 以 及所述第一复杂图中的第一顶点的邻接点与所述第一复杂图中的第 一顶点相连的边, 得到第二新图, 其中, 所述第二新图包含所述第 一复杂图中除所述第一顶点以及所述第一顶点的邻接点与所述第一 复杂图中的第一顶点相连的边之外的顶点和边; 所述权值获取单元还用于, 将所述第一复杂图中所述第一顶点 的权值与所述第一复杂图中的第一顶点的邻接点的权值相加获取到 所述第一复杂图中的第一顶点的邻接点在所述第二新图中的权值; 所述确定单元还用于, 若确定所述第二新图的顶点数目不超过 预设的阈值, 则将所述第二新图作为与所述复杂图等效的简单图。
在第二种可能的实现方式中, 结合第二方面或第一种可能的实 现方式, 还包括: 图运算单元, 用于对所述简单图中的第二顶点进行图运算得到 所述第二顶点的运算结果; 搜索单元, 用于在所述复杂图中对所述复杂图的各个顶点分别 进行广度优先搜索, 得到所述复杂图的各个顶点分别对应的第二顶 点, 其中, 所述顶点对应的第二顶点为在所述复杂图中距离所述顶 点最近的第二顶点; 结果获取单元, 用于将所述复杂图的各个顶点进行所述图运算 的结果, 分别与所述顶点对应的第二顶点的运算结果对应, 获取到 所述复杂图的各个顶点的运算结果。
在第三种可能的实现方式中, 结合第二方面或第一种可能的实 现方式, 所述图运算单元还用于, 对所述简单图的各个顶点分别进 行图运算, 得到所述简单图的各个顶点的运算结果; 所述设备还包括获取单元, 用于获取所述简单图的各个顶点对 应的删除复杂图, 其中, 所述简单图的顶点对应的删除复杂图是通 过在所述复杂图中删除所述简单图中除顶点以外的其他顶点得到 的;
所述搜索单元还用于, 在所述简单图的各个顶点对应的删除复 杂图中, 对所述简单图的各个顶点分别通过广度优先搜索遍历, 得 到所述简单图的各个顶点对应的删除复杂图中的顶点; 所述设备还包括分配单元, 用于将所述简单图的各个顶点的运 算结果分配至所述简单图的各个顶点对应的所述删除复杂图中的顶 点。 在第四种可能的实现方式中, 结合第二方面、 第一种至第三种 可能的实现方式中的任一项, 所述复杂图包括无标度的复杂图。
第三方面, 本发明实施提供的一种复杂图的处理设备, 包括: 处理器, 用于删除所述复杂图中的第一顶点, 以及所述第一顶 点的邻接点与所述第一顶点相连的边, 得到第一新图, 其中, 所述 第一顶点为度为 1 的顶点, 所述第一新图包含所述复杂图中除所述 第一顶点以及所述第一顶点的邻接点与所述第一顶点相连的边之外 的顶点和边; 以及将所述复杂图中所述第一顶点的权值与所述第一顶点的邻 接点的权值相加获取到所述第一顶点的邻接点在所述第一新图中的 权值;
以及若确定所述第一新图的顶点数目不超过预设的阈值, 则将 所述第一新图作为与所述复杂图等效的简单图。 在第一种可能的实现方式中, 结合第三方面, 所述处理器还用 于, 若所述确定单元确定所述新图的顶点数目超过预设的阈值, 将 所述第一新图作为第一复杂图; 以及删除所述第一复杂图中的第一顶点, 以及所述第一复杂图 中的第一顶点的邻接点与所述第一复杂图中的第一顶点相连的边, 得到第二新图, 其中, 所述第二新图包含所述第一复杂图中除所述 第一顶点以及所述第一顶点的邻接点与所述第一复杂图中的第一顶 点相连的边之外的顶点和边; 以及将所述第一复杂图中所述第一顶点的权值与所述第一复杂 图中的第一顶点的邻接点的权值相加获取到所述第一复杂图中的第 一顶点的邻接点在所述第二新图中的权值; 以及若确定所述第二新图的顶点数目不超过预设的阈值, 则将 所述第二新图作为与所述复杂图等效的简单图。
在第二种可能的实现方式中, 结合第三方面或第一种可能的实 现方式, 所述处理器还用于: 对所述简单图中的第二顶点进行图运算得到所述第二顶点的运 算结果;
以及在所述复杂图中对所述复杂图的各个顶点分别进行广度优 先搜索, 得到所述复杂图的各个顶点分别对应的第二顶点, 其中, 所述顶点对应的第二顶点为在所述复杂图中距离所述顶点最近的第 二顶点; 以及将所述复杂图的各个顶点进行所述图运算的结果, 分别与 所述顶点对应的第二顶点的运算结果对应, 获取到所述复杂图的各 个顶点的运算结果。
在第三种可能的实现方式中, 结合第三方面或第一种可能的实 现方式, 所述处理器还用于, 对所述简单图的各个顶点分别进行图 运算, 得到所述简单图的各个顶点的运算结果;
以及获取所述简单图的各个顶点对应的删除复杂图, 其中, 所 述简单图的顶点对应的删除复杂图是通过在所述复杂图中删除所述 简单图中除顶点以外的其他顶点得到的; 以及在所述简单图的各个顶点对应的删除复杂图中, 对所述简 单图的各个顶点分别通过广度优先搜索遍历, 得到所述简单图的各 个顶点对应的删除复杂图中的顶点; 以及将所述简单图的各个顶点的运算结果分配至所述简单图的 各个顶点对应的所述删除复杂图中的顶点。 在第四种可能的实现方式中, 结合第三方面、 第一种至第三种 可能的实现方式中的任一项, 所述复杂图包括无标度的复杂图。 本发明实施例提供了一种复杂图的处理方法和设备, 根据复杂 图的无标度性, 删除度为 1 的顶点来降低复杂图规模, 得到与复杂 图等效的简单图, 无需另外寻找复杂图的极大匹配, 减少计算过程 的开销; 而且降低复杂图规模的过程与图运算的过程相互独立, 并 且可以根据简单图的运算结果和简单图与复杂图的等效关系, 得到 复杂图的运算结果, 使得适用于多种图算法。
附图说明 为了更清楚地说明本发明实施例或现有技术中的技术方案, 下 面将对实施例或现有技术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面描述中的附图仅仅是本发明的一些实施例, 对于 本领域普通技术人员来讲, 在不付出创造性劳动的前提下, 还可以 根据这些附图获得其他的附图。 图 1 为现有技术的方法流程示意图;
图 2 ( 1 ) 为本发明实施例提供的一种无边度复杂图; 图 2 ( 2 ) 为本发明实施例提供的一种第一新图; 图 2 ( 3 ) 为本发明实施例提供的一种简单图; 图 3为本发明实施例提供的一种复杂图的处理方法流程图; 图 4 为本发明实施例提供的一种根据简单图的运算结果得到复 杂图的运算结果的过程图; 图 5 ( 1 ) 为本发明实施例的一种简单图的运算结果图; 图 5 ( 2 ) 为本发明实施例的一种复杂图的运算结果图; 图 6 为本发明实施例提供的另一种根据简单图的运算结果得到 复杂图的运算结果的过程图; 图 7 ( 1 ) 为本发明实施例的一种简单图的顶点对应的删除复杂 图;
图 7 ( 2 ) 为本发明实施例的另一种简单图的顶点对应的删除复 杂图;
图 7 ( 3 ) 为本发明实施例的另一种复杂图的运算结果图; 图 8为本发明实施例提供的另一种复杂图的处理方法流程图; 图 9 为本发明实施例提供的一种复杂图的处理设备的装置示意 图;
图 10为本发明实施例提供的另一种复杂图的处理设备的装置示 意图;
图 1 1为本发明实施例提供的一种复杂图的处理设备的硬件结构 图。 具体实施方式 下面将结合本发明实施例中的附图, 对本发明实施例中的技术 方案进行清楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明 一部分实施例, 而不是全部的实施例。 基于本发明中的实施例, 本 领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他 实施例, 都属于本发明保护的范围。 复杂网络图的一个基本性质就是无标度性, 也就是指网络图中 的顶点的度满足 律分布(power law distribution) , 而满足 律分布 的网络图具有如下性质: 1 .少数顶点的度很大,大部分顶点的度很小; 2.在对数尺度的坐标系下复杂图的度-顶点数的关系近似为一条直 线。 如图 2 ( 1 ) 所示的具有无标度性的复杂图, 其中, 空心圆圈表 示顶点; 黑色的实线表示边; 顶点内的数字表示顶点的标识信息。 本实施例优选用身份标识号 ( Identity , 简称 ID ) 标识, 顶点上方的 数字表示对应顶点的权重, 为了能够简要清楚的说明本发明的实施 例, 图 2 ( 1 ) 设定为无向图, 且每个顶点的权值均设置为 1。
参见图 3 , 为本发明实施例提供的一种复杂图的处理方法, 可以 包括:
S301 : 删除复杂图中的第一顶点, 以及第一顶点的邻接点与第 一顶点相连的边, 得到第一新图, 其中, 第一顶点为度为 1 的顶点, 第一新图包含复杂图中除第一顶点以及第一顶点的邻接点与第一顶 点相连的边之外的顶点和边; 示例性的, 复杂图包括无标度的复杂图。 示例性的, 由于第一顶点只与其邻接点相连, 因此复杂图和第 一新图属于对应的映射关系, 两者是等效的。
S302 : 将复杂图中第一顶点的权值与第一顶点的邻接点的权值 相加得到第一顶点的邻接点在第一新图中的权值; 示例性的, 第一新图降低了复杂图的规模, 且与复杂图等效。
S303 : 若确定第一新图的顶点数目不超过预设的阈值, 则将第 一新图作为与复杂图等效的简单图。 示例性的, 预设的阈值是用来描述简单图规模的最大范围, 可 以是简单图所能够包含的最大的顶点数目 ; 示例性的, 若第一新图的顶点数目超过预设的阈值, 则可以进 一步的降低第一新图的规模, 过程如下: 将第一新图作为第一复杂图; 删除第一复杂图中的第一顶点, 以及第一复杂图中的第一顶点 的邻接点与第一复杂图中的第一顶点相连的边, 得到第二新图, 其 中, 第二新图包含第一复杂图中除第一顶点以及第一顶点的邻接点 与第一复杂图中的第一顶点相连的边之外的顶点和边; 同前所述, 第二新图与复杂图也是等效的。 将第一复杂图中第一顶点的权值与第一复杂图中的第一顶点的 邻接点的权值相加得到第一复杂图中的第一顶点的邻接点在第二新 图中的权值; 若确定第二新图的顶点数目 不超过预设的阈值, 则将第二新图 作为与复杂图等效的简单图;
若确定第二新图的顶点数目超过预设的阈值, 则将第二新图作 为新的第一复杂图, 重复上述对第一复杂图的处理过程, 直至得到 的第二新图的顶点数目不超过预设的阈值, 可以理解的, 以此得到 的简单图, 仍然与复杂图等效。
示例性的, 在得到了与复杂图等效的简单图之后, 就可以对简 单图进行图运算, 然后可以根据简单图与复杂图的等效关系, 由简 单图的运算结果得到复杂图的运算结果, 具体的方法有两种, 而方 法一的过程如图 4所示, 包括:
S401 : 对简单图中的第二顶点进行图运算得到第二顶点的运算 结果;
S402 : 在复杂图中对复杂图的各个顶点分别进行广度优先搜索, 得到复杂图的各个顶点分别对应的第二顶点, 其中, 顶点对应的第 二顶点为在复杂图中距离顶点最近的第二顶点;
S403 : 将复杂图的各个顶点进行图运算的结果, 分别与顶点对 应的第二顶点的运算结果对应, 得到复杂图的各个顶点的运算结果。 方法二的过程如图 6所示, 包括:
S601 : 对简单图的各个顶点分别进行图运算, 得到简单图的各 个顶点的运算结果;
S602 : 获取简单图的各个顶点对应的删除复杂图, 其中, 简单 图的顶点对应的删除复杂图是通过在复杂图中删除简单图中除顶点 以外的其他顶点得到的;
S603 : 在简单图的各个顶点对应的删除复杂图中, 对简单图的 各个顶点分别通过广度优先搜索遍历, 得到简单图的各个顶点对应 的删除复杂图中的顶点;
S604 : 将简单图的各个顶点的运算结果分配至简单图的各个顶 点对应的删除复杂图中的顶点。 本发明实施例提供了一种复杂图的处理方法, 根据复杂图的无 标度性, 删除度为 1 的顶点来降低复杂图规模, 得到与复杂图等效 的简单图, 无需另外寻找复杂图的极大匹配, 减少计算过程的开销; 而且降低复杂图规模的过程与图运算的过程相互独立, 并且可以根 据简单图的运算结果和简单图与复杂图的等效关系, 得到复杂图的 运算结果, 使得适用于多种图算法。
具体的, 在图 3、 4、 6 的基础上, 以图 2 为例, 对本发明实施 例提供的一种复杂图的处理方法进行进一步的的详细说明, S301: 删除复杂图中的第一顶点, 以及第一顶点的邻接点与第 一顶点相连的边, 得到第一新图, 其中, 第一顶点为度为 1 的顶点, 第一新图包含复杂图中除第一顶点以及第一顶点的邻接点与第一顶 点相连的边之外的顶点和边; 示例性的, 复杂图包括无标度的复杂图。
示例性的, 如图 2 ( 1 ) 所示, 当经过步骤 S301 处理之后, 可 以知道, 顶点 ID 在集合 {1, 2, 3, 4, 5, 6, 7, 8, 9, 10}中的顶 点均为第一顶点, 第一顶点的邻接点与第一顶点相连的边可以通过 第一顶点 ID和第一顶点的邻接点 ID组成 ID元组表示, 例如: (第 一顶点 ID, 第二顶点 ID), 因此删除的边的集合为 { ( 1, 11 ), ( 2, 12), ( 3, 13 ), ( 4, 14), ( 5, 15 ), ( 6, 16), ( 7, 17), ( 8, 18), ( 9, 19), ( 10, 20) }; 得到的第一新图如图 2 ( 2) 所示。 进一步的, 可以理解的, 由于第一顶点只与其邻接点相连, 因 此复杂图和第一新图是属于对应的映射关系, 两者是等效的。
S302: 将复杂图中第一顶点的权值与第一顶点的邻接点的权值 相加得到第一顶点的邻接点在第一新图中的权值;
示例性的, 在图 2 ( 1 ) 中, 将第一顶点的权值加到其对应的邻 接点的权值上, 得到邻接点在第一新图中的权值, 如图 2 ( 2) 所示。
S303: 若确定第一新图的顶点数目不超过预设的阈值, 则将第 一新图作为与复杂图等效的简单图。 示例性的, 如图 2 ( 2) 所示的新图, 确定该图中的顶点数目是 否超过预设的阈值, 其中, 预设的阈值是用来描述简单图规模的最 大范围, 可以是简单图所能够包含的最大的顶点数目 , 具体的, 本 实施例中预设的阈值为 5, 若图 2 ( 2) 中的顶点数目不超过 5, 则将 图 2 ( 2) 作为与图 2 ( 1 ) 等效的简单图; 示例性的, 在本实施例中, 图 2 ( 2) 的顶点数目超过了预设的 阈值 5, 则将第一新图作为第一复杂图; 具体的, 将图 2 ( 2) 作为第一复杂图; 删除第一复杂图中的第一顶点, 以及第一复杂图中的第一顶点 的邻接点与第一复杂图中的第一顶点相连的边, 得到第二新图, 其 中, 第二新图包含第一复杂图中除第一顶点以及第一顶点的邻接点 与第一复杂图中的第一顶点相连的边之外的顶点和边;
具体的, 删除图 2 ( 2 ) 中的第一顶点以及图 2 ( 2 ) 中的第一顶 点的邻接点与图 2 ( 2 ) 中的第一顶点相连的边, 由图 2 ( 2) 可知, 图 2 ( 2 ) 的第一顶点 ID的集合为 {11 , 12, 13, 14, 15, 16, 17, 18, 19, 20}, 删除的边的集合为 { ( 11, 21 ), ( 12, 21 ), ( 13, 21 ),
( 14, 21 ), ( 15, 21 ), ( 16, 22 ), ( 17, 22 ), ( 18, 22 ), ( 19, 22 ),
( 20, 22 ) }, 得到的第二新图如图 2 ( 3 ) 所示; 由前所述, 同理, 第二新图与第一新图是等效的, 进而可以知 道与复杂图也是等效的。 将第一复杂图中第一顶点的权值与第一复杂图中的第一顶点的 邻接点的权值相加得到第一复杂图中的第一顶点的邻接点在第二新 图中的权值; 具体的, 在图 2 ( 2 ) 中, 将第一顶点的权值加到其对应的邻接 点的权值上, 可以得到邻接点在第二新图中的权值, 如图 2 ( 3 ) 所 示。 若确定第二新图的顶点数目 不超过预设的阈值, 则将第二新图 作为与复杂图等效的简单图;
具体的, 若确定第二新图的顶点数目超过预设的阈值, 则将第 二新图作为新的第一复杂图, 重复上述对第一复杂图的处理过程, 直至得到的第二新图的顶点数目不超过预设的阈值; 在本实施例中, 图 2 ( 3 ) 的顶点数目为 2, 小于预设的阈值 5, 则可以将图 2 ( 3 ) 作为与图 2 ( 1 ) 等效的简单图。 示例性的, 对得到于复杂图等效的简单图之后, 就可以对简单 图进行图运算, 可以理解的, 由复杂图得到与其等效的简单图的过 程和后续对简单图进行图运算的过程两者是相互独立的, 因此, 根 据两图之间的等效关系和过程的独立性, 可以对简单图实施多种类 型的图运算, 并且根据两图之间的等效性可以从简单图的运算结果 得到复杂图的运算结果。 具体可以通过两种方法实现, 其中, 方法一可以如图 4所示:
S401: 对简单图中的第二顶点进行图运算得到第二顶点的运算 结果;
具体在本实施例中, 为了清楚的说明, 设定图运算为图分割运 算, 对于图 2 ( 3 ) 进行图分割的结果如图 5 ( 1 ) 所示, 虚线为图分 割的分割线, 可以得到图 5 ( 1 ) 的运算结果为顶点 21 为第一分割, 顶点 22为第二分割, 表示为 21→ 1, 22→ 2,
S402: 在复杂图中对复杂图的各个顶点分别进行广度优先搜索, 得到复杂图的各个顶点分别对应的第二顶点, 其中, 顶点对应的第 二顶点为在复杂图中距离顶点最近的第二顶点; 具体的在本实施例中, 在图 2 ( 1 ) 中对每个顶点分别进行广度 优先搜索得到的对应的第二顶点结果是: 顶点 ID集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15 }的顶点对应的第二顶点为顶点 21 , 顶点 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}的顶点对应的第 二顶点为顶点 22。
S403: 将复杂图的各个顶点进行图运算的结果, 分别与顶点对 应的第二顶点的运算结果对应, 得到复杂图的各个顶点的运算结果。 具体的, 在本实施例中, 可以得到: 顶点 ID 集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}的顶点 的分割结果是与顶点 21 同在第一分割, 可表示为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}→ 1; 顶点 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}的顶点 的分割结果是与顶点 22 同在第二分割, 可表示为: {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}→ 2, 因此, 对图 2 ( 1 ) 的图分割结果如图 5 ( 2) 所示。 此外, 方法二可以如图 6所示:
S601: 对简单图的各个顶点分别进行图运算, 得到简单图的各 个顶点的运算结果; 具体的如步骤 S401, 在此不再赘述。
S602: 获取简单图的各个顶点对应的删除复杂图, 其中, 简单 图的顶点对应的删除复杂图是通过在复杂图中删除简单图中除顶点 以外的其他顶点得到的;
具体的, 在图 2 ( 3 ) 中, 顶点 21和顶点 22对应的删除复杂图 分别如图 7 ( 1 ) 和图 7 ( 2) 所示。
S603: 在简单图的各个顶点对应的删除复杂图中, 对简单图的 各个顶点分别通过广度优先搜索遍历, 得到简单图的各个顶点对应 的删除复杂图中的顶点;
具体的, 分别在图 7 ( 1 ) 和图 7 ( 2 ) 中, 对顶点 21和顶点 22 通过广度优先搜索遍历,可以得到顶点 21在图 7( 1 )中的顶点集合, 其 ID 集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}; 同理, 顶点 22在图 7 ( 2) 中的顶点集合, 其 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}。
S604: 将简单图的各个顶点的运算结果分配至简单图的各个顶 点对应的删除复杂图中的顶点; 具体的, 运算结果为顶点 21 为第一分割, 顶点 22为第二分割, 表示为 21→ 1, 22→2。 ID集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}的顶点的运算结果为该 ID集合的顶点与顶点 21 同在第一分割, 可表示为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}→ 1; 而 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}的顶点的 运算结果为该 ID集合的顶点与顶点 22同在第二分割, 可表示为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}→ 2;
因此, 对图 2 ( 1 ) 的图分割结果如图 7 ( 3 ) 所示。 本发明实施例提供了一种复杂图的处理方法, 根据复杂图的无 标度性, 删除度为 1 的顶点来降低复杂图规模, 得到与复杂图等效 的简单图, 无需另外寻找复杂图的极大匹配, 减少计算过程的开销; 而且降低复杂图规模的过程与图运算的过程相互独立, 并且可以根 据简单图的运算结果和简单图与复杂图的等效关系, 得到复杂图的 本领域技术人员可以理解的, 网络图可以通过矩阵的形式进行 表示, 也可以通过多个列表进行表示, 对于本发明的实施例, 优选 的可以通过多个列表文件来表示复杂图图 2 ( 1 ); 具体的可以包括: 第一列表, 通过 (顶点 ID, 度) 这种元组的列表形式表示顶点 以及顶点的度, 如列表 1:
[(1, 1), (2, 1), (3, 1), (4, 1), (5, 1), (6, 1), (7, 1), (8, 1), (9, 1), (10, 1), (11, 2), (12, 2), (13, 2), (14, 2), (15, 2), (16, 2), (17, 2), (18, 2), (19, 2), (20, 2), (21, 6), (22, 6)]; 第二列表, 通过 (顶点 ID, 权值) 这种元组的列表形式表示顶 点以及顶点的权值, 如列表 2:
[(1, 1), (2, 1), (3, 1), (4, 1), (5, 1), (6, 1), (7, 1), (8, 1), (9, 1), (10, 1), (11, 1), (12, 1), (13, 1), (14, 1), (15, 1), (16, 1), (17, 1), (18, 1), (19, 1), (20, 1), (21, 1), (22, 1)]; 第三列表, 通过 (顶点 ID, 顶点 ID) 这种元组的列表形式表示 顶点之间相连接的边, 如列表 3:
[(1, 11), (2, 12), (3, 13), (4, 14), (5, 15), (6, 16), (7, 17), (8, 18), (9, 19), (10, 20), (11, 21), (12, 21), (13, 21), (14,
21) , (15, 21), (16, 22), (17, 22), (18, 22), (19, 22), (20,
22) , (21, 22)]。 由此, 可以根据列表 1、 列表 2以及列表 3来实现图 3所示的实 施例中, 步骤 S301-S303的过程, 如图 8所示:
S801: 遍历列表 1, 得到度为 1 的项并保存在列表 la中, 并将 列表 1 中度为 1 的项进行删除, 得到列表 1A;
示例性的, 通过对列表 1进行遍历, 可以得到列表 la为: [(1, 1), (2, 1), (3, 1), (4, 1), (5, 1), (6, 1), (7, 1), (8, 1), (9, 1) , (10, 1)]; 列表 1A为 [(11, 2), (12, 2), (13, 2), (14, 2), (15,
2) , (16, 2), (17, 2), (18, 2), (19, 2), (20, 2), (21, 6), (22, 6)];
优选的, 在该步骤之前, 可以将度为 0 的孤立点保存在另外开 辟的列表中, 并在列表 1、 列表 2 中进行删除, 可以节省后续遍历所 花费的时间代价。
S802: 在列表 3 中删除列表 1A中的顶点与其邻接点相连的边, 得到列表 3A;
示例性的, 在列表 3 中删除列表 1A 中的顶点之后得到的列表 3A为: [(11, 21), (12, 21), (13, 21), (14, 21), (15, 21), (16, 22), (17, 22), (18, 22), (19, 22), (20, 22), (21 , 22)], 于是, 列表 1A和列表 3A可以得到第一新图图 2 ( 2)。
S803: 将列表 2 中与列表 la相同 ID 的顶点的权值加到列表 3 中的包括列表 la的顶点的邻接点的权值上,并将列表 2中与列表 la 相同 ID的顶点的项进行删除, 得到列表 2A; 示例性的, 在列表 2 中, 与列表 la相同 ID的顶点的项为 [(1 , 1), (2, 1), (3, 1), (4, 1), (5, 1), (6, 1), (7, 1), (8, 1), (9, 1), (10, 1)], 将这些项中的权值, 加入到列表 2 中的并且在列表 3 中的包括列表 la的顶点的邻接点的权值上, 并将列表 2中与与列表 la相同 ID 的顶点的项进行删除, 可以得到第一新图图 2 ( 2) 的顶 点的权值如列表 2A为:
[(11, 2), (12, 2), (13, 2), (14, 2), (15, 2), (16, 2), (17, 2), (18, 2), (19, 2), (20, 2), (21, 1), (22, 1)]; 因此, 图 2 ( 2 ) 可以由列表 1 A、 列表 2A和列表 3A表示。
S804: 判断第一新图的顶点数目是否不超过预设的阈值: 若不超过, 则将第一新图作为与复杂图等效的简单图; 若超过, 则将第一新图作为第一复杂图, 重复执行 S801-804, 直至最后得到的新图的顶点数目不超过预设的阈值; 示例性的, 预设的阈值为 5, 图 2 ( 2 ) 的顶点数目超过 5, 则重 复执行 S 801 -S 804 , 直至得到图 2 ( 3 ) 时, 顶点数目不超过预设的 阈值, 因此图 2 ( 3 ) 为图 2 ( 1 ) 等效的简单图。 对得到于复杂图等效的简单图之后, 就可以对简单图进行图运 算, 进一步的, 可以根据简单图的运算结果得到复杂图的运算结果, 具体的根据简单图 的运算结果得到复杂图 的运算结果如步骤
S401 -S403或者步骤 S 601 -S604所述,本领域技术人员可以根据上述 的步骤通过已知的用于表示简单图的列表进行实施, 在此不再赘述; 优选的, 对于保存孤立点的列表, 可以理解的, 需要根据图运 算的算法本身确定孤立点的运算结果, 在本实施例中, 如果存在孤 立点, 则对复杂图进行图分割的计算结果可以忽略孤立点的影响。 本发明实施例提供了一种复杂图的处理方法, 根据复杂图的无 标度性, 删除度为 1 的顶点来降低复杂图规模, 得到与复杂图等效 的简单图, 无需另外寻找复杂图的极大匹配, 减少计算过程的开销; 而且降低复杂图规模的过程与图运算的过程相互独立, 并且可以根 据简单图的运算结果和简单图与复杂图的等效关系, 得到复杂图的 运算结果, 使得适用于多种图算法。
参见图 9 , 为本发明实施例提供的一种复杂图的处理设备 90 , 可以包括: 删除单元 901 , 用于删除复杂图中的第一顶点, 以及第一顶点的 邻接点与第一顶点相连的边, 得到第一新图, 其中, 第一顶点为度 为 1 的顶点, 第一新图包含复杂图中除第一顶点以及第一顶点的邻 接点与第一顶点相连的边之外的顶点和边; 权值获取单元 902 ,用于将复杂图中第一顶点的权值与第一顶点 的邻接点的权值相加获取到第一顶点的邻接点在第一新图中的权 值;
确定单元 903 ,用于若确定第一新图的顶点数目不超过预设的阈 值, 则将第一新图作为与复杂图等效的简单图。 示例性的, 复杂图包括无标度的复杂图, 本实施例通过图 2 ( 1 ) 所示的具有无标度性的复杂图来对方案进行说明, 具体对图 2 ( 1 ) 的描述如前, 在此不再赘述。 示例性的, 如图 2 ( 1 ) 所示, 删除单元 901删除的第一顶点为 顶点 ID 在集合 {1, 2, 3, 4, 5, 6, 7, 8, 9, 10}中的顶点, 而第 一顶点的邻接点与第一顶点相连的边可以通过第一顶点 ID 和第一 顶点的邻接点 ID组成 ID元组表示, 例如: (第一顶点 ID , 第二顶 点 ID), 因此删除的边的集合为 { ( 1, 11 ), ( 2, 12 ), ( 3, 13 ), ( 4, 14), ( 5, 15), ( 6, 16), ( 7, 17), ( 8, 18), ( 9, 19), ( 10, 20) }; 得到的第一新图如图 2 ( 2 ) 所示。 示例性的, 在图 2 ( 1 ) 中, 权值获取单元 902将第一顶点的权 值加到其对应的邻接点的权值上, 获取到邻接点在第一新图中的权 值, 如图 2 ( 2 ) 所示。 示例性的, 如图 2 ( 2) 所示的新图, 确定单元 903确定该图中 的顶点数目是否超过预设的阈值, 具体的, 本实施例中预设的阈值 为 5, 若确定单元 903确定图 2 ( 2 ) 中的顶点数目不超过 5, 则确定 单元 903将图 2 ( 2) 作为与图 2 ( 1 ) 等效的简单图; 在本实施例中, 图 2 ( 2) 的顶点数目超过了预设的阈值 5, 则 确定单元 903 将第一新图作为第一复杂图, 具体的, 确定单元 903 将图 2 ( 2) 作为第一复杂图; 删除单元 901 还用于, 删除第一复杂图中的第一顶点, 以及第 一复杂图中的第一顶点的邻接点与第一复杂图中的第一顶点相连的 边, 得到第二新图, 其中, 第二新图包含第一复杂图中除第一顶点 以及第一顶点的邻接点与第一复杂图中的第一顶点相连的边之外的 顶点和边; 具体的, 删除单元 901可以删除图 2 ( 2 ) 中的第一顶点以及图 2 ( 2) 中的第一顶点的邻接点与图 2 ( 2) 中的第一顶点相连的边, 由图 2 ( 2) 可知, 图 2 ( 2) 的第一顶点 ID的集合为 {11, 12, 13, 14, 15, 16, 17, 18, 19, 20}, 删除的边的集合为 { ( 11 , 21 ), ( 12,
21 ), ( 13, 21 ), ( 14, 21 ), ( 15, 21 ), ( 16, 22 ), ( 17, 22 ), ( 18,
22 ), ( 19, 22 ), ( 20, 22 ) }, 得到的第二新图如图 2 ( 3 ) 所示。 权值获取单元 902 还用于, 将第一复杂图中第一顶点的权值与 第一复杂图中的第一顶点的邻接点的权值相加获取到第一复杂图中 的第一顶点的邻接点在第二新图中的权值;
具体的, 在图 2 ( 2 ) 中, 将第一顶点的权值加到其对应的邻接 点的权值上, 可以得到邻接点在第二新图中的权值, 如图 2 ( 3 ) 所 示。
确定单元 903 还用于, 若确定第二新图的顶点数目不超过预设 的阈值, 则将第二新图作为与复杂图等效的简单图;
具体的, 若确定单元 903 确定第二新图的顶点数目超过预设的 阈值, 则将第二新图作为新的第一复杂图, 重复上述三个功能单元 对第一复杂图的处理过程, 直至得到的第二新图的顶点数目不超过 预设的阈值; 在本实施例中, 图 2 ( 3 ) 的顶点数目为 2, 小于预设的阈值 5, 则确定单元 903可以将图 2 ( 3 ) 作为与图 2 ( 1 ) 等效的简单图。 示例性的, 对得到于复杂图等效的简单图之后, 就可以对简单 图进行图运算, 进一步的, 如图 10所示, 设备 90还可以包括: 图运算单元 904,用于对简单图中的第二顶点进行图运算得到第 二顶点的运算结果;
具体在本实施例中, 为了清楚的说明, 设定图运算为图分割运 算, 图运算单元 904对于图 2 ( 3 ) 进行图分割的结果如图 5 ( 1 ) 所 示, 虚线为图分割的分割线, 可以得到图 5 ( 1 ) 的运算结果为顶点 21 为第一分割, 顶点 22为第二分割, 表示为 21→ 1, 22→ 2, 搜索单元 905,用于在复杂图中对复杂图的各个顶点分别进行广 度优先搜索, 得到复杂图的各个顶点分别对应的第二顶点, 其中, 顶点对应的第二顶点为在复杂图中距离顶点最近的第二顶点; 具体的在本实施例中, 搜索单元 905在图 2 ( 1 ) 中对每个顶点 分别进行广度优先搜索得到的对应的第二顶点结果是: 顶点 ID集合 为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15 }的顶点对应的第二顶点为 顶点 21, 顶点 ID 集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20} 的顶点对应的第二顶点为顶点 22。 结果获取单元 906, 用于将复杂图的各个顶点进行图运算的结 果, 分别与顶点对应的第二顶点的运算结果对应, 获取到复杂图的 各个顶点的运算结果。 具体的, 在本实施例中, 结果获取单元 906可以得到: 顶点 ID 集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}的顶点 的分割结果是与顶点 21 同在第一分割, 可表示为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}→ 1; 顶点 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}的顶点 的分割结果是与顶点 22 同在第二分割, 可表示为: {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}→ 2, 因此, 对图 2 ( 1 ) 的图分割结果如图 5 ( 2 ) 所示。 此外, 图运算单元 904 还用于, 对简单图的各个顶点分别进行 图运算, 得到简单图的各个顶点的运算结果;
具体的如前, 在此不再赘述。 设备还包括获取单元 907,用于获取简单图的各个顶点对应的删 除复杂图, 其中, 简单图的顶点对应的删除复杂图是通过在复杂图 中删除简单图中除顶点以外的其他顶点得到的; 具体的, 获取单元 907在图 2 ( 3 ) 中所获取的顶点 21 和顶点 22对应的删除复杂图分别如图 7 ( 1 ) 和图 7 ( 2 ) 所示 搜索单元 905 还用于, 在简单图的各个顶点对应的删除复杂图 中, 对简单图的各个顶点分别通过广度优先搜索遍历, 得到简单图 的各个顶点对应的删除复杂图中的顶点; 具体的, 搜索单元 905分别在图 7 ( 1 ) 和图 7 ( 2 ) 中, 对顶点 21和顶点 22通过广度优先搜索遍历, 可以得到顶点 21在图 7 ( 1 ) 中的顶点集合, 其 ID集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}; 同理, 顶点 22在图 7 ( 2 ) 中的顶点集合, 其 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}。 设备 90还包括分配单元 908, 用于将简单图的各个顶点的运算 结果分配至简单图的各个顶点对应的删除复杂图中的顶点; 具体的, 运算结果为顶点 21 为第一分割, 顶点 22为第二分割, 表示为 21→ 1, 22→ 2。 分配单元 908将顶点 21 的分割结果分配给在 顶点的 ID 集合为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}的顶点, 该 ID集合的顶点与顶点 21 同在第一分割, 可表示为 {1, 2, 3, 4, 5, 11, 12, 13, 14, 15}→ 1; 分配单元 908将顶点 22的分割结果分配给在顶点的 ID集合为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}的顶点, 该 ID集合的顶点 与顶点 22同在第二分割, 可表示为 {6, 7, 8, 9, 10, 16, 17, 18, 19, 20}→ 2; 因此, 对图 2 ( 1 ) 的图分割结果如图 7 ( 3 ) 所示。 本发明实施例提供了一种复杂图的处理设备 90, 根据复杂图的 无标度性, 删除度为 1 的顶点来降低复杂图规模, 得到与复杂图等 效的简单图, 无需另外寻找复杂图的极大匹配, 减少计算过程的开 销; 而且降低复杂图规模的过程与图运算的过程相互独立, 并且可 以根据简单图的运算结果和简单图与复杂图的等效关系, 得到复杂 图的运算结果, 使得适用于多种图算法。
参见图 11, 为本发明实施例提供的一种复杂图的处理设备 90, 而复杂图可以包括无标度的复杂图, 本实施例通过图 2 ( 1 ) 所示的 具有无标度性的复杂图来对方案进行说明, 具体对图 2 ( 1 ) 的描述 如前, 在此不再赘述, 可以包括: 至少一个处理器 1101; 至少一个输入单元 1102, 用于将复杂图 输入至设备 90, 具体形式可以是扫描仪, 本发明实施例对此不作任 何限定; 存储器 1103和通信总线 1104, 用于实现这些装置之间的连 接通信。
其中 , 通信总线 1104 可以是工业标准体系结构 ( Industry Standard Architecture, 简称为 ISA )总线、 夕卜部设备互连( Peripheral Component, 简称为 PCI) 总线或扩展工业标准体系结构 ( Extended Industry Standard Architecture, 简称为 EISA) 总线等。 该总线 1104 可以分为地址总线、 数据总线、 控制总线等。 为便于表示, 图 11 中 仅用一条粗线表示, 但并不表示仅有一根总线或一种类型的总线。 存储器 1103用于存储可执行程序代码, 该程序代码包括计算机 操作指令。 存储器 1103可能包含高速 RAM存储器, 也可能还包括 非易失性存储器 ( non- volatile memory), 例如至少一个磁盘存储器。 处理器 1101可能是一个中央处理器 ( Central Processing Unit, 简称为 CPU), 或者是特定集成电路( Application Specific Integrated Circuit, 简称为 ASIC ), 或者是被配置成实施本发明实施例的一个 或多个集成电路。 处理器 1101 用于执行存储器 1103 中存储的可执行程序代码, 例如计算机程序来运行与可执行代码对应的程序。 当复杂图输入至 设备 90后, 处理器 1101用于: 处理器 1101, 用于删除复杂图中的第一顶点, 以及第一顶点的 邻接点与第一顶点相连的边, 得到第一新图, 其中, 第一顶点为度 为 1 的顶点, 第一新图包含复杂图中除第一顶点以及第一顶点的邻 接点与第一顶点相连的边之外的顶点和边; 以及将复杂图中第一顶点的权值与第一顶点的邻接点的权值相 加获取到第一顶点的邻接点在第一新图中的权值; 以及若确定第一新图的顶点数目不超过预设的阈值, 则将第一 新图作为与复杂图等效的简单图。 示例性的, 处理器 1101还用于, 若确定单元确定新图的顶点数 目超过预设的阈值, 将第一新图作为第一复杂图; 以及删除第一复杂图中的第一顶点, 以及第一复杂图中的第一 顶点的邻接点与第一复杂图中的第一顶点相连的边, 得到第二新图, 其中, 第二新图包含第一复杂图中除第一顶点以及第一顶点的邻接 点与第一复杂图中的第一顶点相连的边之外的顶点和边; 以及将第一复杂图中第一顶点的权值与第一复杂图中的第一顶 点的邻接点的权值相加获取到第一复杂图中的第一顶点的邻接点在 第二新图中的权值; 以及若确定第二新图的顶点数目不超过预设的阈值, 则将第二 新图作为与复杂图等效的简单图。 示例性的, 对得到于复杂图等效的简单图之后, 就可以对简单 图进行图运算, 进一步的, 处理器 1 101还用于: 对简单图中的第二顶点进行图运算得到第二顶点的运算结果; 以及在复杂图中对复杂图的各个顶点分别进行广度优先搜索, 得到复杂图的各个顶点分别对应的第二顶点, 其中, 顶点对应的第 二顶点为在复杂图中距离顶点最近的第二顶点; 以及将复杂图的各个顶点进行图运算的结果, 分别与顶点对应 的第二顶点的运算结果对应, 获取到复杂图的各个顶点的运算结果。 此外, 处理器 1 101还用于, 对简单图的各个顶点分别进行图运 算, 得到简单图的各个顶点的运算结果; 以及获取简单图的各个顶点对应的删除复杂图, 其中, 简单图 的顶点对应的删除复杂图是通过在复杂图中删除简单图中除顶点以 外的其他顶点得到的; 以及在简单图的各个顶点对应的删除复杂图中, 对简单图的各 个顶点分别通过广度优先搜索遍历, 得到简单图的各个顶点对应的 删除复杂图中的顶点; 以及将简单图的各个顶点的运算结果分配至简单图的各个顶点 对应的删除复杂图中的顶点。 本发明实施例提供了一种复杂图的处理设备 90 , 根据复杂图的 无标度性, 删除度为 1 的顶点来降低复杂图规模, 得到与复杂图等 效的简单图, 无需另外寻找复杂图的极大匹配, 减少计算过程的开 销; 而且降低复杂图规模的过程与图运算的过程相互独立, 并且可 以根据简单图的运算结果和简单图与复杂图的等效关系, 得到复杂 图的运算结果, 使得适用于多种图算法。 本领域普通技术人员可以理解: 实现上述方法实施例的全部或 部分步骤可以通过程序指令相关的硬件来完成, 前述的程序可以存 储于一计算机可读取存储介质中, 该程序在执行时, 执行包括上述 方法实施例的步骤; 而前述的存储介质包括: ROM、 RAM , 磁碟或 者光盘等各种可以存储程序代码的介质。 最后应说明的是: 以上实施例仅用以说明本发明的技术方案, 而非对其限制; 尽管参照前述实施例对本发明进行了详细的说明, 本领域的普通技术人员应当理解: 其依然可以对前述各实施例所记 载的技术方案进行修改, 或者对其中部分技术特征进行等同替换; 而这些修改或者替换, 并不使相应技术方案的本质脱离本发明各实 施例技术方案的精神和范围。

Claims

权 利 要 求 书
1、 一种复杂图的处理方法, 其特征在于, 包括:
删除所述复杂图中的第一顶点,以及所述第一顶点的邻接点与所述第 一顶点相连的边, 得到第一新图, 其中, 所述第一顶点为度为 1的顶点, 所述第一新图包含所述复杂图中除所述第一顶点以及所述第一顶点的邻 接点与所述第一顶点相连的边之外的顶点和边;
将所述复杂图中所述第一顶点的权值与所述第一顶点的邻接点的权 值相加获取到所述第一顶点的邻接点在所述第一新图中的权值;
若确定所述第一新图的顶点数目不超过预设的阈值,则将所述第一新 图作为与所述复杂图等效的简单图。
2、 根据权利要求 1所述的方法, 其特征在于, 若确定所述新图的顶 点数目超过预设的阈值, 所述方法还包括:
将所述第一新图作为第一复杂图;
删除所述第一复杂图中的第一顶点,以及所述第一复杂图中的第一顶 点的邻接点与所述第一复杂图中的第一顶点相连的边, 得到第二新图, 其 中,所述第二新图包含所述第一复杂图中除所述第一顶点以及所述第一顶 点的邻接点与所述第一复杂图中的第一顶点相连的边之外的顶点和边; 将所述第一复杂图中所述第一顶点的权值与所述第一复杂图中的第 一顶点的邻接点的权值相加获取到所述第一复杂图中的第一顶点的邻接 点在所述第二新图中的权值;
若确定所述第二新图的顶点数目不超过预设的阈值,则将所述第二新 图作为与所述复杂图等效的简单图。
3、 根据权利要求 1或 2所述的方法, 其特征在于, 所述方法还包括: 对所述简单图中的第二顶点进行图运算得到所述第二顶点的运算结 果;
在所述复杂图中对所述复杂图的各个顶点分别进行广度优先搜索,得 到所述复杂图的各个顶点分别对应的第二顶点, 其中, 所述顶点对应的第 二顶点为在所述复杂图中距离所述顶点最近的第二顶点;
将所述复杂图的各个顶点进行所述图运算的结果,分别与所述顶点对 应的第二顶点的运算结果对应, 获取到所述复杂图的各个顶点的运算结 果。
4、 根据权利要求 1或 2所述的方法, 其特征在于, 所述方法还包括: 对所述简单图的各个顶点分别进行图运算,得到所述简单图的各个顶 点的运算结果;
获取所述简单图的各个顶点对应的删除复杂图, 其中, 所述简单图的 顶点对应的删除复杂图是通过在所述复杂图中删除所述简单图中除顶点 以外的其他顶点得到的;
在所述简单图的各个顶点对应的删除复杂图中,对所述简单图的各个 顶点分别通过广度优先搜索遍历,得到所述简单图的各个顶点对应的删除 复杂图中的顶点;
将所述简单图的各个顶点的运算结果分配至所述简单图的各个顶点 对应的所述删除复杂图中的顶点。
5、 根据权利要求 1 -4任一项所述的方法, 其特征在于, 所述复杂图 包括无标度的复杂图。
6、 一种复杂图的处理设备, 其特征在于, 包括:
删除单元, 用于删除所述复杂图中的第一顶点, 以及所述第一顶点的 邻接点与所述第一顶点相连的边, 得到第一新图, 其中, 所述第一顶点为 度为 1的顶点,所述第一新图包含所述复杂图中除所述第一顶点以及所述 第一顶点的邻接点与所述第一顶点相连的边之外的顶点和边;
权值获取单元,用于将所述复杂图中所述第一顶点的权值与所述第一 顶点的邻接点的权值相加获取到所述第一顶点的邻接点在所述第一新图 中的权值;
确定单元, 用于若确定所述第一新图的顶点数目不超过预设的阈值, 则将所述第一新图作为与所述复杂图等效的简单图。
7、 根据权利要求 6所述的设备, 其特征在于, 所述确定单元还用于, 若所述确定单元确定所述新图的顶点数目超过预设的阈值,将所述第一新 图作为第一复杂图;
所述删除单元还用于, 删除所述第一复杂图中的第一顶点, 以及所述 第一复杂图中的第一顶点的邻接点与所述第一复杂图中的第一顶点相连 的边, 得到第二新图, 其中, 所述第二新图包含所述第一复杂图中除所述 第一顶点以及所述第一顶点的邻接点与所述第一复杂图中的第一顶点相 连的边之外的顶点和边;
所述权值获取单元还用于,将所述第一复杂图中所述第一顶点的权值 与所述第一复杂图中的第一顶点的邻接点的权值相加获取到所述第一复 杂图中的第一顶点的邻接点在所述第二新图中的权值;
所述确定单元还用于,若确定所述第二新图的顶点数目不超过预设的 阈值, 则将所述第二新图作为与所述复杂图等效的简单图。
8、 根据权利要求 6或 7所述的设备, 其特征在于, 还包括: 图运算单元,用于对所述简单图中的第二顶点进行图运算得到所述第 二顶点的运算结果;
搜索单元,用于在所述复杂图中对所述复杂图的各个顶点分别进行广 度优先搜索, 得到所述复杂图的各个顶点分别对应的第二顶点, 其中, 所 述顶点对应的第二顶点为在所述复杂图中距离所述顶点最近的第二顶点; 结果获取单元, 用于将所述复杂图的各个顶点进行所述图运算的结 果, 分别与所述顶点对应的第二顶点的运算结果对应, 获取到所述复杂图 的各个顶点的运算结果。
9、 根据权利要求 6或 7所述的设备, 其特征在于, 所述图运算单元 还用于, 对所述简单图的各个顶点分别进行图运算, 得到所述简单图的各 个顶点的运算结果;
所述设备还包括获取单元,用于获取所述简单图的各个顶点对应的删 除复杂图, 其中, 所述简单图的顶点对应的删除复杂图是通过在所述复杂 图中删除所述简单图中除顶点以外的其他顶点得到的;
所述搜索单元还用于, 在所述简单图的各个顶点对应的删除复杂图 中, 对所述简单图的各个顶点分别通过广度优先搜索遍历, 得到所述简单 图的各个顶点对应的删除复杂图中的顶点;
分配单元,用于将所述简单图的各个顶点的运算结果分配至所述简单 图的各个顶点对应的所述删除复杂图中的顶点。
10、 根据权利要求 6-9任一项所述的设备, 其特征在于, 所述复杂图 包括无标度的复杂图。
PCT/CN2013/082468 2013-08-28 2013-08-28 一种复杂图的处理方法和设备 WO2015027412A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2013/082468 WO2015027412A1 (zh) 2013-08-28 2013-08-28 一种复杂图的处理方法和设备
CN201380001321.7A CN104737151A (zh) 2013-08-28 2013-08-28 一种复杂图的处理方法和设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2013/082468 WO2015027412A1 (zh) 2013-08-28 2013-08-28 一种复杂图的处理方法和设备

Publications (1)

Publication Number Publication Date
WO2015027412A1 true WO2015027412A1 (zh) 2015-03-05

Family

ID=52585391

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/082468 WO2015027412A1 (zh) 2013-08-28 2013-08-28 一种复杂图的处理方法和设备

Country Status (2)

Country Link
CN (1) CN104737151A (zh)
WO (1) WO2015027412A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101196933A (zh) * 2008-01-09 2008-06-11 王珊 利用连接表压缩数据图的方法和设备
US20110238607A1 (en) * 2006-11-02 2011-09-29 D-Wave Systems Inc. Graph embedding techniques
CN102664787A (zh) * 2012-04-01 2012-09-12 华为技术有限公司 决策树的生成方法和装置
CN103198470A (zh) * 2013-02-26 2013-07-10 清华大学 图像的分割方法及系统
CA2808217A1 (en) * 2012-02-24 2013-08-24 Callum David Mcdonald Method of graph processing

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110238607A1 (en) * 2006-11-02 2011-09-29 D-Wave Systems Inc. Graph embedding techniques
CN101196933A (zh) * 2008-01-09 2008-06-11 王珊 利用连接表压缩数据图的方法和设备
CA2808217A1 (en) * 2012-02-24 2013-08-24 Callum David Mcdonald Method of graph processing
CN102664787A (zh) * 2012-04-01 2012-09-12 华为技术有限公司 决策树的生成方法和装置
CN103198470A (zh) * 2013-02-26 2013-07-10 清华大学 图像的分割方法及系统

Also Published As

Publication number Publication date
CN104737151A (zh) 2015-06-24

Similar Documents

Publication Publication Date Title
US20210004625A1 (en) Object Detection Model Training Method, Apparatus, and Device
WO2018214505A1 (zh) 立体匹配方法与系统
US9053540B2 (en) Stereo matching by census transform and support weight cost aggregation
CN108804226B (zh) 一种用于分布式图计算的图分割划分方法
WO2018099151A1 (zh) 一种指令生成过程的复用方法及装置
CN110297680B (zh) 一种传输虚拟桌面图像的方法及装置
WO2017076296A1 (zh) 处理图数据的方法和装置
WO2021052169A1 (zh) 分布式数据的均衡处理方法、装置、计算终端和存储介质
WO2021120410A1 (zh) 基于霍夫变换的绝对相位噪声去除方法、装置和存储介质
US20170041620A1 (en) Image Processing Methods and Image Processing Apparatuses
WO2022027917A1 (zh) 图像处理方法、装置、系统、电子设备及可读存储介质
WO2020024584A1 (zh) 一种训练物体检测模型的方法、装置以及设备
JP2015076093A5 (zh)
WO2019128459A1 (zh) 图像阴影消除方法和装置
WO2015090044A1 (zh) 一种数据扫描的方法及装置
KR102239588B1 (ko) 이미지 처리 방법 및 장치
US10311557B2 (en) Automated tonal balancing
US9875528B2 (en) Multi-frame patch correspondence identification in video
US11194861B2 (en) Graph partitioning method and apparatus
US10659304B2 (en) Method of allocating processes on node devices, apparatus, and storage medium
WO2015027412A1 (zh) 一种复杂图的处理方法和设备
US11755549B2 (en) Method, electronic device, and computer program product for data processing
WO2017031961A1 (zh) 数据处理方法和装置
CN110347511B (zh) 含隐私约束条件的地理分布式进程映射方法、装置及终端
CN104166959B (zh) 一种图像降噪的加速方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13892555

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13892555

Country of ref document: EP

Kind code of ref document: A1