CN117061365A

CN117061365A - Node selection method, device, equipment and readable storage medium

Info

Publication number: CN117061365A
Application number: CN202311308673.9A
Authority: CN
Inventors: 李仁刚; 闫瑞栋; 郭振华; 刘璐; 金良; 徐聪
Original assignee: Inspur Electronic Information Industry Co Ltd
Current assignee: Inspur Electronic Information Industry Co Ltd
Priority date: 2023-10-11
Filing date: 2023-10-11
Publication date: 2023-11-14
Anticipated expiration: 2043-10-11
Also published as: CN117061365B

Abstract

The invention discloses a node selection method, a node selection device, node selection equipment and a readable storage medium in the technical field of computers. According to the method, the node is evaluated by taking the degree of the node and the structural bridge in the target network as structural factors, the node is evaluated by taking the propagation range of the node under the constraint of the preset propagation distance as unstructured factors, and the propagation force of the node is comprehensively evaluated by combining the structural factors and the unstructured factors, so that the node evaluation accuracy can be improved, and reliable data support is provided for the selection of the target node; after the target network is divided to obtain a plurality of subnets, if any one of the subnets comprises a target node, a management node is selected in the current subnet according to the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet and the real-time resource quantity, and the management node can be selected in different subnets by combining the propagation characteristics of the multi-source information in the same network.

Description

Node selection method, device, equipment and readable storage medium

Technical Field

The present invention relates to the field of computer technologies, and in particular, to a node selection method, device, apparatus, and readable storage medium.

Background

Currently, when nodes such as switching equipment and management equipment are addressed in a network, it is required to evaluate whether the location of the network node is suitable for deploying the switching equipment and the management equipment. The current node selection mainly adopts a centralized strategy, such as: node selection is performed throughout the network using a greedy algorithm or heuristic algorithm. Among other things, the high computational complexity of the greedy algorithm makes it difficult to scale to large-scale networks. In order to achieve the scalability of the algorithm, a series of heuristic algorithms such as Degree Discount (Degre discover), shortest Path (short Path), and betweenness centrality (BetwennessCentral) are sequentially proposed. However, heuristic algorithms lack a good theoretical guarantee, and the performance of the algorithms cannot be guaranteed.

Therefore, how to quickly and accurately select the locations of nodes such as a switching device and a management device is a problem that needs to be solved by those skilled in the art.

Disclosure of Invention

In view of the above, an object of the present invention is to provide a node selection method, apparatus, device and readable storage medium, so as to quickly and accurately select the positions of nodes such as a switching device and a management device. The specific scheme is as follows:

in a first aspect, the present invention provides a node selection method, including:

Determining the degree of each node in a target network, the number of structural bridges and the propagation range under the constraint of a preset propagation distance;

evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range, and selecting a target node in the target network according to the propagation force;

dividing the target network into a plurality of subnets, if any of the subnets comprises the target node, calculating the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet, and obtaining the propagation probability of each node in the current subnet; and determining the real-time resource quantity of each node in the current subnet, and selecting a management node in the current subnet according to the propagation probability and the real-time resource quantity of each node in the current subnet.

Optionally, determining the degree of each node in the target network includes:

constructing an N multiplied by N adjacency matrix corresponding to the target network; n is the total number of nodes in the target network, and the j-th column element of the i-th row in the adjacency matrix represents: node i points to the number of outputs of node j;

and determining the total number of the outgoing degrees and the total number of the incoming degrees of each node in the target network according to the adjacency matrix, and determining the sum of the total number of the outgoing degrees and the total number of the incoming degrees of the same node as the degree of the corresponding node.

Optionally, determining the number of structural bridges of each node in the target network includes:

aiming at each node in the target network, neighbor nodes taking the current node as propagation bridges form a neighbor group, and the number of the neighbor groups is taken as the structural bridge number of the current node; or for each node in the target network, taking the number of times of the current node serving as a propagation bridge of the neighbor node as the structural bridge number of the current node.

Optionally, determining the propagation range under the preset propagation distance constraint includes:

and calculating the number of nodes to which information is transmitted by each node in the target network under the constraint of a preset propagation distance as the propagation range of the corresponding node.

Optionally, the estimating the propagation force of each node according to the degree, the number of structural bridges and the propagation range includes:

performing standardized operation on the degree, the number of structural bridges and the propagation range;

and evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range after the standardized operation.

Optionally, the estimating the propagation force of each node according to the degree, the number of structural bridges and the propagation range after the standardized operation includes:

according to the first formula Calculating to obtain the propagation force of each node, wherein the first formula is as follows: e (E) _i =α×D _i +β×B _i +S _i ；E _i Representing the propagation force of node i in the target network, D _i Representing the degree after the standardized operation of node i, B _i Representing the number of structural bridges after standardized operation of node i, S _i And (3) representing the propagation range of the node i after the standardized operation, wherein alpha and beta are preset weight coefficients.

Optionally, the selecting a target node in the target network according to the propagation force includes:

arranging all nodes in the target network from large to small according to the propagation force, and taking the first M nodes as the target nodes; or arranging all nodes in the target network from small to large according to the propagation force, and taking the last M nodes as the target nodes; m is a preset value.

Optionally, the dividing the target network into a plurality of subnets includes:

constructing a eigenvalue calculation formula by using the adjacent matrix and the N multiplied by N identity matrix, and solving the eigenvalue calculation formula to obtain N eigenvalues;

Constructing N multiplied by N eigenvalue matrixes by taking the N eigenvalues as diagonal elements;

the degree of each node in the target network is used as a diagonal element, and an NxN degree matrix is constructed;

adding diagonal line elements of the eigenvalue matrix and the degree matrix in alignment to obtain a division index value of each node in the target network;

dividing each node in the target network into a plurality of subnets according to the division index value.

Optionally, the dividing each node in the target network into a plurality of subnets according to the division index value includes:

calculating the similarity between any two nodes in the target network according to the corresponding division index values;

and if the similarity is smaller than a preset similarity threshold, dividing the two nodes into the same subnet.

Optionally, the calculating the propagation probability of each node in the current subnet to propagate the multi-source information to other nodes in the current subnet includes:

calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet according to a second formula;

wherein the second formula is: ；/>Representing the propagation probability of the node u in the current subnet to propagate multi-source information to the node v in the current subnet, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, and +.>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet according to a third formula;

wherein the third formula is:；/>representing the propagation probability of the node u in the current subnet to propagate multi-source information to the node v in the current subnet, D (t) representing the delay function, Q representing the number of multi-source information, t representing the propagation of the information to the node uTime required by node v, < >>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

Optionally, the calculation formula of D (t) is: d (t) =t ² +t or D (t) =e ^t 。

Optionally, the selecting a management node in the current subnet according to the propagation probability and the real-time resource amount of each node in the current subnet includes:

And if the propagation probability of any node is greater than a preset probability threshold and the real-time resource quantity of the node is greater than a preset resource threshold, taking the node as the management node.

Optionally, the determining the real-time resource amount of each node in the current subnet includes:

calculating the real-time resource quantity of each node in the current subnet according to a fourth formula; the fourth formula is: t (T) _u =a×X _u +b×Y _u +c×Z _u ，T _u Representing the real-time resource quantity of node u in the current subnet, X _u Representing the real-time computing power of a node u in the current subnet, Y _u Representing the real-time memory quantity of node u in the current subnet, Z _u And the real-time bandwidth of the node u in the current subnet is represented, and a, b and c are preset parameters.

Optionally, the method further comprises:

and if the target node is not included in any subnet, selecting a management node in the subnet.

Optionally, the method further comprises:

deploying data exchange equipment at the position of the management node; the data exchange device includes: switches and/or routers.

Optionally, the method further comprises:

deploying a dispatching center at the position of the management node;

and collecting node information of each node in the subnet where the management node is located by using the dispatching center.

Optionally, the collecting, by the scheduling center, node information of each node in the subnet where the management node is located includes:

The dispatching center is utilized to send information collection instructions to all nodes in the subnet where the management node is located at regular time, and node information returned by all nodes in the subnet where the management node is located according to the information collection instructions is received; or the dispatching center is utilized to receive node information sent by each node in the subnet where the management node is located at fixed time.

Optionally, the target network is: a distributed computing network, a wireless sensor network, or a data storage network.

In a second aspect, the present invention provides a node selection apparatus comprising:

the determining module is used for determining the degree of each node in the target network, the number of structural bridges and the propagation range under the constraint of a preset propagation distance;

the first selection module is used for evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range and selecting a target node in the target network according to the propagation force;

the second selection module is used for dividing the target network into a plurality of subnets, if any subnet comprises the target node, calculating the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet, and obtaining the propagation probability of each node in the current subnet; and determining the real-time resource quantity of each node in the current subnet, and selecting a management node in the current subnet according to the propagation probability and the real-time resource quantity of each node in the current subnet.

Optionally, the determining module is specifically configured to:

Optionally, the first selection module includes:

the standardized operation unit is used for carrying out standardized operation on the degree, the number of structural bridges and the propagation range;

And the evaluation unit is used for evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range after the standardized operation.

Optionally, the evaluation unit is specifically configured to:

the propagation force of each node is calculated according to a first formula, wherein the first formula is as follows: e (E) _i =α×D _i +β×B _i +S _i ；E _i Representing the propagation force of node i in the target network, D _i Representing the degree after the standardized operation of node i, B _i Representing the number of structural bridges after standardized operation of node i, S _i And (3) representing the propagation range of the node i after the standardized operation, wherein alpha and beta are preset weight coefficients.

Optionally, the first selection module includes:

a target node selection unit, configured to arrange all nodes in the target network from large to small according to the propagation force, and use the first M nodes as the target nodes; or arranging all nodes in the target network from small to large according to the propagation force, and taking the last M nodes as the target nodes; m is a preset value.

Optionally, the second selection module includes:

a first construction unit, configured to construct an n×n adjacency matrix corresponding to the target network; n is the total number of nodes in the target network, and the j-th column element of the i-th row in the adjacency matrix represents: node i points to the number of outputs of node j;

The first calculation unit is used for constructing a characteristic value calculation formula by utilizing the adjacent matrix and the N multiplied by N identity matrix, and solving the characteristic value calculation formula to obtain N characteristic values;

the second construction unit is used for constructing N characteristic value matrixes by taking the N characteristic values as diagonal elements;

a third construction unit, configured to make the degree of each node in the target network be a diagonal element, and construct an n×n degree matrix;

the second calculation unit is used for adding diagonal line elements of the eigenvalue matrix and the degree matrix in a counterpoint mode to obtain a division index value of each node in the target network;

and the dividing unit is used for dividing each node in the target network into a plurality of subnets according to the division index value.

Optionally, the dividing unit is specifically configured to:

Optionally, the second selection module includes:

the probability calculation unit is used for calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet according to a second formula;

Wherein the second formula is:；/>representing the propagation probability of a node u in the current subnet propagating multi-source information to a node v in the current subnetQ represents the number of multi-source information, t represents the time required for node u to propagate information to node v, +.>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

Optionally, the probability calculation unit is specifically configured to:

wherein the third formula is:；/>representing the propagation probability of the node u in the current sub-network to propagate multi-source information to the node v in the current sub-network, D (t) representing the delay function, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, < >>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

Optionally, the second selection module includes:

and the management node selection unit is used for taking any node as the management node if the propagation probability of the node is larger than a preset probability threshold value and the real-time resource quantity of the node is larger than a preset resource threshold value.

Optionally, the second selection module includes: the resource amount calculating unit is used for calculating the real-time resource amount of each node in the current subnet according to a fourth formula; the fourth formula is: t (T) _u =a×X _u +b×Y _u +c×Z _u ，T _u Representing the real-time resource quantity of node u in the current subnet, X _u Representing current subnet middleknotReal-time calculation of the point u, Y _u Representing the real-time memory quantity of node u in the current subnet, Z _u And the real-time bandwidth of the node u in the current subnet is represented, and a, b and c are preset parameters.

Optionally, the method further comprises:

the deployment module is used for deploying the data exchange equipment at the position where the management node is located; the data exchange device includes: switches and/or routers.

Optionally, the deployment module is further configured to:

deploying a dispatching center at the position of the management node;

Optionally, the deployment module is further configured to:

In a third aspect, the present invention provides an electronic device, comprising:

a memory for storing a computer program;

a processor for executing the computer program to implement the node selection method disclosed previously.

In a fourth aspect, the present invention provides a readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the node selection method disclosed previously.

As can be seen from the above scheme, the present invention provides a node selection method, which includes: determining the degree of each node in a target network, the number of structural bridges and the propagation range under the constraint of a preset propagation distance; evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range, and selecting a target node in the target network according to the propagation force; dividing the target network into a plurality of subnets, if any of the subnets comprises the target node, calculating the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet, and obtaining the propagation probability of each node in the current subnet; and determining the real-time resource quantity of each node in the current subnet, and selecting a management node in the current subnet according to the propagation probability and the real-time resource quantity of each node in the current subnet.

The beneficial effects of the invention are as follows: the node is evaluated by taking the degree of the node in the target network and the structural bridge as structural factors, and the node is evaluated by taking the propagation range of the node under the constraint of the preset propagation distance as unstructured factors, and the propagation force of the node is comprehensively evaluated by combining the structural factors and the unstructured factors, so that the node evaluation accuracy can be improved, and reliable data support is provided for the selection of the target node; after the target network is divided to obtain a plurality of subnets, if any one of the subnets comprises a target node, a management node is selected in the current subnet according to the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet and the real-time resource quantity, the management node can be selected in different subnets by combining the propagation characteristics of the multi-source information in the same network, the defects of insufficient single information propagation model and centralized management node selection are overcome, and the positions of nodes such as switching equipment, management equipment and the like can be selected rapidly and accurately.

Correspondingly, the node selection device, the node selection equipment and the readable storage medium have the technical effects.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a node selection method of the present disclosure;

FIG. 2 is a schematic diagram of a network structure according to the present disclosure;

FIG. 3 is a schematic diagram of a solution framework of the present disclosure;

FIG. 4 is a schematic diagram of a node selection according to the present disclosure;

FIG. 5 is a schematic diagram of a node selection device according to the present disclosure;

FIG. 6 is a schematic diagram of an electronic device according to the present disclosure;

FIG. 7 is a diagram illustrating a server configuration according to the present invention;

fig. 8 is a diagram of a terminal structure according to the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Currently, node selection mainly adopts a centralized strategy, such as: node selection is performed throughout the network using a greedy algorithm or heuristic algorithm. Among other things, the high computational complexity of the greedy algorithm makes it difficult to scale to large-scale networks. In order to achieve the scalability of the algorithm, a series of heuristic algorithms such as Degree Discount (Degre discover), shortest Path (short Path), and betweenness centrality (BetwennessCentral) are sequentially proposed. However, heuristic algorithms lack a good theoretical guarantee, and the performance of the algorithms cannot be guaranteed. Therefore, the invention provides a node selection scheme which can rapidly and accurately select the positions of nodes such as switching equipment, management equipment and the like.

Referring to fig. 1, the embodiment of the invention discloses a node selection method, which comprises the following steps:

s101, determining the degree of each node in the target network, the number of structural bridges and the propagation range under the constraint of a preset propagation distance.

It should be noted that the target network may be a distributed computing network, a wireless sensor network, a data storage network, or the like. The degree of any node in the target network includes: the number of directed edges that the node points to other nodes (i.e., the sum of the number of outgoing degrees that the node points to other nodes), the number of directed edges that the other node points to the node (i.e., the sum of the number of incoming degrees that the other node points to the node). Referring to fig. 2, the other nodes pointed to by the node 1 are only the node 3, and then the sum of the output numbers pointed to by the node 1 to the other nodes is 1; the other nodes pointing to node 1 are only node 2, and then the sum of the ingress numbers of the other nodes pointing to the node is 1, whereby the number of degrees of node 1 is 1+1=2. The arrowed connection lines between the different nodes in fig. 2 represent directed edges.

Referring to fig. 2, node 1 serves as a bridge connecting node 2 and node 3, and after node 1 is removed, node 2 cannot point to node 3, at which point node 1 is referred to as a structural bridge from node 2 to node 3. Similarly, node 3 acts as a bridge connecting node 1 and node 2, and after node 3 is removed, node 1 cannot point to node 2, so node 3 is also a structural bridge. Furthermore, node 2 acts as a bridge connecting node 3 and node 1, and after node 2 is removed, node 3 cannot point to node 1, so node 2 is also a structural bridge. Accordingly, in one embodiment, determining the number of structural bridges of each node in the target network includes: aiming at each node in the target network, the neighbor nodes taking the current node as a propagation bridge form a neighbor group, the number of the neighbor groups is taken as the structural bridge number of the current node, and the neighbor groups are as follows: node 1 is used as a node 2 and a node 3 of a propagation bridge; or for each node in the target network, the number of times the current node serves as a propagation bridge of the neighbor node is taken as the structural bridge number of the current node.

In one embodiment, determining the degree of each node in the target network includes: constructing an N multiplied by N adjacency matrix corresponding to the target network; n is the total number of nodes in the target network, and the ith row and jth column elements in the adjacency matrix represent: node i points to the number of outputs of node j; and determining the total number of the outgoing degrees and the total number of the incoming degrees of each node in the target network according to the adjacency matrix, and determining the sum of the total number of the outgoing degrees and the total number of the incoming degrees of the same node as the degree of the corresponding node. The network shown in fig. 2 corresponds to an 11×11 adjacency matrix, the elements of which can be referred to in table 1. The first row and first behavior node identifications in table 1, element 1 in the second row represents: the directed edges of the node 1 pointing to the node 3 are 1; each element 0 in the second row represents: node 1 does not point to nodes 1, 2, 4, 5, 6, 7, 8, 9, 10 and 11, i.e., the directed edges between node 1 and these nodes are 0. The other elements are analogized in turn. As can be seen from table 1, the number of outgoing and incoming nodes in the target network can be directly seen by the adjacency matrix shown in table 1, and thus the number of degrees of each node in the target network can be determined based on the adjacency matrix.

Table 1: adjacency matrix

Referring to fig. 2, if the preset propagation distance constraint value is 2, from the node 1, under the constraint of the propagation distance constraint 2, the node 1 may transmit information to the node 3, the node 2 and the node 9, and then the number of nodes to which the node 1 transmits information under the constraint of the propagation distance constraint 2 is 3, so that 3 is the propagation range of the node 1. Accordingly, in one embodiment, determining a propagation range under a preset propagation distance constraint includes: and calculating the number of nodes to which information is transmitted by each node in the target network under the constraint of a preset propagation distance as the propagation range of the corresponding node.

S102, evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range, and selecting a target node in a target network according to the propagation force.

S103, dividing the target network into a plurality of subnets, if any of the subnets comprises a target node, calculating the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet, and obtaining the propagation probability of each node in the current subnet; and determining the real-time resource quantity of each node in the current subnet, and selecting a management node in the current subnet according to the propagation probability and the real-time resource quantity of each node in the current subnet.

In one embodiment, estimating propagation forces for each node based on degrees, number of structural bridges, and propagation ranges comprises: carrying out standardized operation on the degree, the number of structural bridges and the propagation range; and evaluating the propagation force of each node according to the degree, the number of structural bridges and the propagation range after the standardized operation. Normalization operations such as: normalization operation, etc. to make the number of degrees, the number of structural bridges and the propagation range take values under the same value standard. Wherein, evaluate the propagation force of each node according to the number of degrees, the number of structural bridges and the propagation range after standardized operation, include: the propagation force of each node is calculated according to a first formula, wherein the first formula is as follows: e (E) _i =α×D _i +β×B _i +S _i ；E _i Representing the propagation force of node i in the target network, D _i Representing the degree after the standardized operation of node i, B _i Representing the number of structural bridges after standardized operation of node i, S _i And (3) representing the propagation range of the node i after the standardized operation, wherein alpha and beta are preset weight coefficients.

In this embodiment, selecting a target node in a target network according to propagation forces includes: arranging all nodes in a target network from large to small according to propagation force, and taking the first M nodes as target nodes; or arranging all nodes in the target network from small to large according to the propagation force, and taking the last M nodes as target nodes; m is a preset value.

Further, dividing the target network into a plurality of subnets includes: constructing an N multiplied by N adjacency matrix corresponding to the target network; n is the total number of nodes in the target network, and the ith row and jth column elements in the adjacency matrix represent: node i points to the number of outputs of node j; constructing a eigenvalue calculation formula by using the adjacent matrix and the N multiplied by N identity matrix, and solving the eigenvalue calculation formula to obtain N eigenvalues; using N eigenvalues as diagonal elements to construct an NxN eigenvalue matrix; the degree of each node in the target network is used as a diagonal element, and an NxN degree matrix is constructed; adding diagonal line elements of the eigenvalue matrix and the degree matrix to obtain a division index value of each node in the target network; dividing each node in the target network into a plurality of subnets according to the division index value. The characteristic value matrix and the degree matrix only have values on diagonal lines, the values of other position elements are 0, each diagonal line element of the characteristic value matrix represents the characteristics of the access degree of each node in the target network, and each diagonal line element of the degree matrix represents the access degree of each node in the target network, so that the connection trend among the nodes in the target network can be drawn by combining the characteristic value matrix and the degree matrix, and the characteristics of the network structure can be represented through the matrix.

In one embodiment, dividing each node in the target network into a plurality of subnets according to the division index value includes: calculating the similarity between any two nodes in the target network according to the corresponding division index values; if the similarity is smaller than the preset similarity threshold, dividing the two nodes into the same subnet. That is: if the similarity degree of the division index values of any two nodes is smaller than a preset similarity degree threshold value, dividing the two nodes into the same subnet.

In one embodiment, calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet includes: calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet according to a second formula; wherein, the second formula is:；/>representing the propagation probability of the node u in the current subnet to propagate multi-source information to the node v in the current subnet, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, and +.>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet. Wherein (1) >The method is used for representing the propagation attenuation introduced by the propagation path, so that the propagation probability of the network node u for propagating the multi-source information to the node v in the current subnet is considered by considering the propagation attenuation, and the calculation accuracy of the propagation probability can be improved.

In one embodiment, calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet includes: calculating the propagation probability of each node in the current subnet to propagate multi-source information to other nodes in the current subnet according to a third formula; wherein, the third formula is:；/>representing the propagation probability of the node u in the current sub-network to propagate multi-source information to the node v in the current sub-network, D (t) representing the delay function, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, < >>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet. D (t) is a function of t, and its calculation formula is: d (t) =t ² +t or D (t) =e ^t Or otherwise. Wherein (1)>The method is used for representing the propagation delay introduced by the propagation path, so that the propagation probability of the network node u for propagating the multi-source information to the node v in the current subnet by considering the propagation delay can be improved in calculation accuracy of the propagation probability.

In one embodiment, the management node is selected in the current subnet based on the propagation probability and the real-time resource amount of each node in the current subnetA dot, comprising: and if the propagation probability of any node is greater than a preset probability threshold and the real-time resource quantity of the node is greater than a preset resource threshold, taking the node as a management node. Wherein determining the real-time resource amount of each node in the current subnet comprises: calculating the real-time resource quantity of each node in the current subnet according to a fourth formula; the fourth formula is: t (T) _u =a×X _u +b×Y _u +c×Z _u ，T _u Representing the real-time resource quantity of node u in the current subnet, X _u Representing the real-time computing power of a node u in the current subnet, Y _u Representing the real-time memory quantity of node u in the current subnet, Z _u And the real-time bandwidth of the node u in the current subnet is represented, and a, b and c are preset parameters.

In this embodiment, if the target node is not included in any of the subnets, the management node is not selected in that subnet. If any of the subnets comprises a target node, a management node is selected in the current subnet according to the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet and the real-time resource quantity.

In one embodiment, after a management node is selected from a subnet with a target node, data exchange equipment is deployed at a position where the management node is located; the data exchange device includes: switches and/or routers. Or a dispatching center is deployed at the position where the management node is located; and collecting node information of each node in the subnet where the management node is located by using the dispatching center. The method for collecting node information of each node in the subnet where the management node is located by using the dispatching center comprises the following steps: the scheduling center is utilized to send information collection instructions to all nodes in the subnet where the management node is located at regular time, and node information returned by all nodes in the subnet where the management node is located according to the information collection instructions is received; or receiving node information sent by each node in the subnet where the management node is located at fixed time by using the dispatching center.

Therefore, in the embodiment, the node is evaluated by taking the degree of the node in the target network and the structural bridge as the structuring factor, the node is evaluated by taking the propagation range of the node under the constraint of the preset propagation distance as the unstructured factor, and the propagation force of the node is comprehensively evaluated by combining the structuring factor and the unstructured factor, so that the node evaluation accuracy can be improved, and reliable data support is provided for the selection of the target node; after the target network is divided to obtain a plurality of subnets, if any one of the subnets comprises a target node, a management node is selected in the current subnet according to the propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet and the real-time resource quantity, the management node can be selected in different subnets by combining the propagation characteristics of the multi-source information in the same network, the defects of insufficient single information propagation model and centralized management node selection are overcome, and the positions of nodes such as switching equipment, management equipment and the like can be selected rapidly and accurately.

Referring to fig. 3, the modules shown in fig. 3 may be implemented according to the present invention, including: the system comprises a target node evaluation index modeling module, a network dividing module, an information propagation mode modeling module and a seed node (i.e. management node) selection module.

The overall operating principles and flow are presented below, and the inputs and outputs of the modules and the inherent logical associations between the modules are described. First, an original network is given, comprising individual network nodes and their associated edges, as shown in particular in fig. 2.

The target node evaluation index modeling module is capable of performing importance evaluation on a given network node. The module considers two factors of the node: structured and unstructured factors. Structural factors mainly include node degrees, node structural bridges and other dimensions, while non-structural factors mainly refer to node propagation capabilities. The module can comprehensively evaluate each node in the network by integrating the structural factors and the unstructured factors. The input of the target node evaluation index modeling module is an original network, the output of the module is an evaluation index combining the structural factors and the non-structural factors according to the invention, and the output of the module is a target node which has higher potential value in the original network, for example: m target nodes are selected.

The traditional network node importance evaluation method mainly utilizes the position and structure information of the network node, the node evaluation method is simple, and unstructured factors of the node, such as the information propagation breadth of the node, are ignored. Therefore, in view of the shortcomings of the traditional node evaluation method, the invention provides a novel evaluation method combining structural factors with unstructured factors. The structuring factor, unstructured factor, and target node evaluation index definition are set forth below, respectively.

The structuring factors of the invention mainly consider the degree of the network node and two dimensions of the structural bridge. The degree of a network node refers to the total number of edges connected by the node, including the number of edges out degre (v) issued by node v and the number of edges pointed to by other nodes to the nodeThe number of edges indecree (v). Thus, network node->Degree (v) =outdegree (v) +indegree (v).

For a given networkWherein->Representation->The number of network nodes in the network,representing q edges. From the information of G, an adjacency matrix a of the network G can be defined. The adjacency matrix a of the network G shown in fig. 2 is shown in table 1. The network shown in fig. 2 includes 11 nodes and 18 edges. The first row and column of table 1 are node identifications, and the remaining positions 0, 1 are elements in the adjacency matrix a.

As shown in Table 1, the adjacency matrix indicates the connection relationship between nodes, and if there is a connection edge between node v and node u, the corresponding position of the adjacency matrix is marked as 1, otherwise, the adjacency matrix is marked as 0. Specifically, the first row of the adjacency matrix represents the case where node 1 points to all nodes. For example, in fig. 2 there is only one piece from node 1 to node 3The side, therefore, is marked 1 at the 1 st row and 3 rd column positions of the adjacent matrix a, and 0 at the other positions of the first row. Similarly, the second row represents the link case where node 2 points to all nodes. By analogy, the number of 1 s in the v-th row in the adjacency matrix represents the output number of the node v, outdegreey (v), e.g., 1 in the first row is only 1, then =1。

Furthermore, a certain column of the adjacency matrix represents the case where all nodes point to that node. For example, the first column in the adjacency matrix has only 1. Thus, the ingress of node 1=1. In view of the above-mentioned, it is desirable,=1+1=2。

similarly, the degree of all nodes can be calculated from the adjacency matrixThe method comprises the following steps:. For the convenience of subsequent calculation, the degree of the node v is subjected to standardized operation, namely: />. Wherein (1)>Indicating the degree before the normalized operation of node v, +.>Representing the sum of degrees of all nodes of the network, +.>Representing the number of degrees after the normalized operation of node v.

Network nodeThe structural bridge of (2) is referred to as the node +.>As a bridge connecting two nodes x and y. For example, in FIG. 2, node 1 serves as a bridge connecting node 2 and node 3, and after node 1 is removed, node 2 cannot point to node 3, at which point node 1 is referred to as a link (node 2->Node 3). Similarly, node 3 acts as a bridge connecting node 1 and node 2, and after node 3 is removed, node 1 cannot point to node 2, so node 3 is also a structural bridge. Furthermore, node 2 acts as a bridge connecting node 3 and node 1, and after node 2 is removed, node 3 cannot point to node 1, so node 2 is also a structural bridge.

Based on the definition of the structural bridges described above, the number of structural bridges per node as a link can be calculated. Taking fig. 2 as an example, the number of structural bridges of all nodes in the network is calculated. Node 1The method comprises the steps of carrying out a first treatment on the surface of the Node 2 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 3The method comprises the steps of carrying out a first treatment on the surface of the Node 4 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 5 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 6 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 7The method comprises the steps of carrying out a first treatment on the surface of the Node 8 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 9The method comprises the steps of carrying out a first treatment on the surface of the Node 10 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 11。

According to the above-mentioned individual nodesCan calculate the total nodes of the network GThe method comprises the following steps: />。

For convenience of subsequent calculation, the number of structural bridges of the node vAnd (3) performing standardized operation, namely: />The method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Representing the number of structural bridges before the normalization operation of node v,/->Representing the sum of the number of structural bridges of all nodes of the network, < >>Representing the number of structural bridges after the standardized operation of node v.

With reference to the foregoing, the present invention proposes the following structural factor evaluation index. For any node v in the network G, its structural factor evaluation index:the method comprises the steps of carrying out a first treatment on the surface of the Wherein 0 is</>Two coefficients or weights between (0, 1) are represented to characterize the specific gravity of the factor of the degree and the factor of the structural bridge, and the default value is set to 0.5.

The non-structural factors of the network node provided by the invention mainly refer to the information propagation breadth of the node. In particular, the information propagation breadth of a node v should be measured by the number of nodes it can reach. The invention provides a propagation hop threshold value To control the number of nodes that the node can reach for the information. The larger the threshold value, the stronger the information propagation breadth of the node v, but the greater the computational complexity is in general. Taking the original network given in fig. 2 as an example and let +.>The information propagation breadth (i.e., propagation range under propagation distance constraint 2) of each node in the network G shown in fig. 2 is as follows: node 1 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 2The method comprises the steps of carrying out a first treatment on the surface of the Node 3 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 4 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 5The method comprises the steps of carrying out a first treatment on the surface of the Node 6 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 7 +.>The method comprises the steps of carrying out a first treatment on the surface of the NodePoint 8The method comprises the steps of carrying out a first treatment on the surface of the Node 9 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 10 +.>The method comprises the steps of carrying out a first treatment on the surface of the Node 11。

Thus, the information propagation breadth of the nodes of the network G as a wholeThe definition is as follows: />。

For convenience of subsequent calculation, information propagation breadth of node vAnd (3) performing standardized operation, namely: />. Wherein,representing the extent of information propagation prior to the normalization operation of node v,representing the breadth of information propagation for all nodes of the network,representing the extent of information propagation after the normalization operation of node v.

In summary, the invention provides a target node evaluation method combining a structural factor and a non-structural factor, which is specifically defined as follows:. Wherein S->Representing the structural factor evaluation result of the node v,unstructured factor evaluation result (i.e., information propagation breadth of node) representing node v, +. >Representing the propagation force of node v.

Referring to the foregoing, an algorithm for selecting M target nodes in the network G may be designed to: algorithm input: network g= (V, E), target node number M, thresholdParameter->The method comprises the steps of carrying out a first treatment on the surface of the Algorithm output: first M->The node with the higher value is taken as the target node set T. The algorithm process comprises the following steps: step 1, computing +.f. for all nodes v in network G>A value; step 2, all +.>The values are arranged in descending order; step 3, selecting the first M->The node with the higher score is taken as a target node set T.

The network dividing module is defined with a characteristic value matrix and a degree matrix of the network so as to represent the structural characteristics and distribution conditions of all nodes in the network. And combining the eigenvalue matrix and the degree matrix to divide the original network into a plurality of sub-networks. The network dividing module divides the original network into a plurality of subnets with smaller scale by utilizing the novel network dividing rule provided by the invention on the basis of the previous module. The output of the module is thus the respective subnetwork.

In consideration of factors such as the scale of network nodes, the distribution condition of geographic positions, the individual difference of the nodes, the calculation cost of algorithm execution and the like, the method effectively splits the original large-scale network and provides a sub-network division strategy. The advantages of the subnet partitioning are the following two aspects: firstly, the original large-scale network is split into a plurality of sub-networks, so that the calculation complexity of seed selection can be effectively reduced; secondly, the multiple sub-networks are mutually independent and do not interfere with each other, so that distributed parallel computing can be realized, the parallelism of a subsequent seed selection algorithm is improved, and the algorithm execution efficiency is accelerated.

The invention provides a partitioning rule based on a characteristic value Matrix Eigenvalue Matrix (EM) and a Degree Matrix Degre Matrix (DM) of an adjacent Matrix A. The definition of the eigenvalue matrix and the definition of the degree matrix are described below, respectively.

The eigenvalue matrix EM is defined as solving determinantAnd sequentially storing the n eigenvalues at diagonal positions of the EM matrix. Specifically, for an n×n adjacency matrix a, the following can be obtained by solvingA matrix of eigenvalues EM is determined. />Representing an N identity matrix, ">An N-dimensional vector is represented and each component represents a corresponding eigenvalue. Expansion->The following characteristic polynomial equation is obtained: />. The characteristic multipleThe equation can in turn be converted into the following form: />…/>=0. Thus (S)>Has n values, respectively +.>=/>、…。

Will first characteristic valuePlaced in the first row and first column of the eigenvalue matrix EM, the second eigenvalue +.>Placed in the second row and second column of the eigenvalue matrix EM, and so on, finally the nth eigenvalue +.>Placed in the nth row and column of the eigenvalue matrix EM. In summary, the eigenvalue matrix EM constructed from eigenvalues is: />. In EM->Characterizing node 1, +.>Characterizing node 2 and so on, < - >The characteristics of node n are described. In summary, the characteristic value matrix EM reflects the characteristics and attributes of the network structure to a certain extent, so EM is selected as an important reference index for subnet division.

The degree matrix DM of the network G is defined as:. DM is represented as an N x N matrix, the diagonal values represent the degrees of each node, and the other positions are all 0.

Based on the eigenvalue matrix and the degree matrix, the method can calculate:. Split (G) is a diagonal matrix, and its first row diagonal element represents the attribute +_of node 1>(i.e., the division index value of node 1), the second row diagonal element thereof represents the attribute of node 2, and so on.

The invention provides a subnet division algorithm, which is input by the following algorithm: network g= (V, E) and its dividing condition matrix Split (G), distance parameter d; algorithm output: a plurality of subnets obtained after dividing network G,…/>. The process comprises the following steps: step 1, determining the Split (v) value of each node v in the original network G, wherein the Split (v) value of each node v is each diagonal element of the matrix Split (G). Step 2, defining the distance between any two nodes u and v, namely:. Step 3, according to the distance formula, calculating the distance between each node v and other nodes u, and if the distance between the nodes v and the other nodes u is +. >Then node v and node u are partitioned into the same subnetwork. Step 4, repeating the above process until splitting the node in the original network G into +.>Sub-network,…/>。

The input of the subsequent seed node selection module is each subnet, not the original network, so as to parallelize different subnets.

The information propagation mode modeling module defines a propagation mode of the information. Specifically, the invention researches an information propagation model, analyzes and reveals the principle and mechanism of the information propagation model and provides preconditions for subsequent algorithm design. In order to solve the defects of the existing independent cascade IC (Independent Cascade) model and the linear threshold LT (Linear Threshold) model, the invention considers the multisource property, the delay property and the attenuation property of information propagation and provides a multisource information delay model and an attenuation model. The independent cascade IC model and the linear threshold LT model rely on different rules to effect the transfer of information, and once activated (information received) the node will remain active until the information propagation process is completed. The IC model considers that the node activation process is independent, and the activation of the linear threshold LT model node needs to meet the influence degree of the neighbor node exceeding a preset threshold. The models are difficult to reveal the information propagation modes and general rules of the real network influence, so that the invention analyzes and builds a novel network influence propagation model to reveal the dynamic evolution mechanism of influence propagation. The information propagation mode modeling module provides a multi-source information delay and attenuation model, and can calculate propagation probability in each subnet according to the propagation mode, delay and attenuation model of the multi-source information.

The conventional IC model is used to describe single source information, i.e. only the same information exists in one network. Obviously, this is not consistent with the true network propagation content, and therefore the present inventionLet us consider firstly the attenuation model of multiple information sources. It is assumed that Q information sources exist simultaneously in the network and that one node can only be activated by one type of information. The invention is defined inTime network node->Activating node->The probability of (i.e., the probability of node u passing information to node v)) is as follows: />，/>Representing node->Activating node->T represents the propagation time, ++>Representing the attenuation factor. The propagation probability between two nodes (e.g., node 6 and node 9 shown in fig. 2) connected to different subnets is calculated based on this. That is: the propagation probability between each node in the entire network can be calculated from this formula.

The information propagation process is developed according to the following discrete time slice rule: assume that n initial seed sets areAnd they are in an active state (the nodes obtain information) and the rest of the nodes are in an inactive state (the nodes do not obtain information). Let->Is indicated at->Time active node set, and- >Is>All have a single opportunity to go through the directed edge +.>With probability->At->Every inactive neighbor deactivating it at the moment +.>. The above process is repeated until no new node can be activated and the information dissemination process ends. Therefore, after any node obtains information at the time t, the process of the node transmitting information to the inactive neighbor node is as follows: the node is +.>With probability->This information is propagated to its inactive neighbor nodes. If a node is required to be able to transition from an inactive state to an active state without reverse transition, and if a node has been activated by some information, the rest of the information no longer plays a role on it, then when more information propagates to the same node, the node randomly selects one of the information and receives it.

Besides attenuation factors, the invention also considers that certain delay of information can appear in the propagation process by using a functionAnd (3) representing. Therefore, will->After substitution into formula (15), it is defined as +.>Time network node->Activating node->The probability of (2) is as follows:. The delay function D (t) may be generally defined as a polynomial function or an exponential function, e.g., D (t) = +.>+t，D(t)=/>Etc. Optionally at least one of the two above modes can be used for computing the network node +. >Activating node->Is a probability of (2). If two modes are used simultaneously for calculating propagation probability +.>The calculated average of the two propagation probabilities may be taken as the propagation probability for this node.

The existing information propagation model, such as an independent cascade IC model or a linear threshold LT model, does not consider the characteristics of information, and also does not consider factors such as attenuation, delay and the like in the information propagation process. The invention takes the IC and LT models as the basis, considers the multisource of information propagation (namely, multiple information propagates in the same network at the same time), delay (namely, delay or hysteresis of information propagation caused by the influence of factors such as network transmission or physical hardware) and attenuation (namely, attenuation effect of information propagation caused by factors such as network environment) on the information propagation model, reveals the inherent objective rule and mechanism of information propagation, and establishes a mathematical model for accurately reflecting the true influence propagation.

The seed node selection module is responsible for simultaneously carrying out seed node selection operation in each subnet. In order to realize efficient selection of the seed nodes, the invention can design an expansion algorithm under different constraint conditions (such as selecting K seed nodes) and finally output an optimal seed set. The input of the seed node selection module is each subnet with target nodes and an information propagation model in the subnet, and the module selects seed sets meeting certain constraint conditions (such as K seed nodes) by utilizing the expansion algorithm provided by the invention, so that the number of the seed nodes is maximized.

The seed node selection module is used for respectively selecting seed nodes for each subgraph so as to finish the selection of K seed nodes. The input of the module is each subnet, and the seed set S which can maximize the activation probability of the target node is searched by utilizing the information propagation model defined by the invention.

Input of a seed node selection method:subnet->,…/>Constraint condition K, information propagation model. And (3) outputting: a set of seed nodes S, the process comprising: step 0, initializing a seed node set S to be an empty set. Step 1, for each subnet +.>The following operations are performed in parallel. Step 2, if the sub-network does not contain the target node, the sub-network is p +.>No operation is performed. Step 3, if the subnet +>Contains the target node v, then the maximum +.>Node->As seed nodes;. Step 4, updating the seed set, namely +.>. And 5, comparing the size of the seed set S with the size of K. If |S|<K, the algorithm repeatedly performs the above-mentioned process, if +.>The algorithm stops.

If one seed node is selected in each sub-network with target nodes, the number of the seed nodes is still not K, the process is repeatedly executed, and then the seed nodes are selected in each sub-network with target nodes.

It should be noted that, when selecting a seed node in a subnet with a target node, the following procedure may be referred to: if the propagation probability of the node is greater than the preset probability threshold value and the real-time resource quantity of the node is also greater than the preset resource threshold value, selecting the node as a seed node, thereby comprehensively considering the capability of the node for propagating information and the inherent attribute T of the node _u =a×X _u +b×Y _u +c×Z _u Selecting seed nodes, T _u Representing the real-time resource quantity of node u in the current subnet, X _u Representing the real-time computing power (such as the main frequency size of a processor) of a node u in the current subnet, Y _u Representing the real-time memory quantity of node u in the current subnet, Z _u Representing the real-time bandwidth of node u in the current subnet, a, b and c beingAnd presetting parameters.

Under the traditional task application scene, the seed node is mainly selected by adopting a centralized strategy, namely, the seed node is selected from a given whole network. The disadvantage of the centralized approach is the following two aspects. Firstly, the calculation amount is large. Since the seed selection is directed to the whole network, when the number of network nodes is n, the complexity of sequentially traversing the whole network to select the seed set is at least O (n), so that the seed set cannot be effectively expanded. And secondly, the parallel calculation is difficult. The seed nodes are selected by iteration of the traditional centralized algorithm, and the algorithm is difficult to realize parallel calculation due to the sequence problem in the iteration process. Aiming at the defects of the traditional centralized scheme, the invention adopts a distributed strategy to improve the expandability and parallelism of the algorithm, divides the original large-scale network into a plurality of subnets with smaller scale through effectively dividing the original network, and respectively and simultaneously carries out seed screening operation aiming at each subnet, thereby reducing the calculation complexity of the algorithm and improving the parallelism of the algorithm so as to optimize the integral calculation time of the algorithm.

Therefore, the invention provides a distributed cross-domain node collaborative computing method, which solves the problem of maximizing the number of specific target nodes. Firstly, modeling potential values of network nodes, and proposing a modeling strategy of target nodes to determine a target node group. Then, the invention provides a novel information propagation model aiming at the problem of inaccuracy of the conventional information propagation model. Secondly, considering the deficiency of the existing centralized strategy and the distribution condition of the target node group in the network, the invention provides a distributed seed screening strategy to reduce the calculation complexity of the algorithm and improve the availability and expandability of the algorithm. The invention may refer to fig. 4 for the overall process of fig. 2.

Moreover, the invention can be widely applied to application scenes of intensive tasks and intensive computation, such as artificial intelligent model training networks, data centers, intelligent network systems, distributed computing systems, wireless sensor networks and the like.

A node selection device provided in the embodiments of the present invention is described below, and a node selection device described below may refer to other embodiments described herein.

Referring to fig. 5, an embodiment of the present invention discloses a node selection device, including:

The determining module 501 is configured to determine a degree of each node in the target network, a number of structural bridges, and a propagation range under a preset propagation distance constraint;

the first selection module 502 is configured to evaluate a propagation force of each node according to the degree, the number of structural bridges, and the propagation range, and select a target node in the target network according to the propagation force;

a second selecting module 503, configured to divide the target network into a plurality of subnets, and if any of the subnets includes a target node, calculate a propagation probability of each node in the current subnet for propagating multi-source information to other nodes in the current subnet, so as to obtain a propagation probability of each node in the current subnet; and determining the real-time resource quantity of each node in the current subnet, and selecting a management node in the current subnet according to the propagation probability and the real-time resource quantity of each node in the current subnet.

In one embodiment, the determining module is specifically configured to:

constructing an N multiplied by N adjacency matrix corresponding to the target network; n is the total number of nodes in the target network, and the ith row and jth column elements in the adjacency matrix represent: node i points to the number of outputs of node j;

In one embodiment, the determining module is specifically configured to:

aiming at each node in the target network, neighbor nodes taking the current node as propagation bridges form neighbor groups, and the number of the neighbor groups is taken as the structural bridge number of the current node; or for each node in the target network, the number of times the current node serves as a propagation bridge of the neighbor node is taken as the structural bridge number of the current node.

In one embodiment, the determining module is specifically configured to:

In one embodiment, the first selection module includes:

In one embodiment, the evaluation unit is specifically configured to:

In one embodiment, the first selection module includes:

the target node selection unit is used for arranging all nodes in a target network from large to small according to the propagation force, and taking the first M nodes as target nodes; or arranging all nodes in the target network from small to large according to the propagation force, and taking the last M nodes as target nodes; m is a preset value.

In one embodiment, the second selection module includes:

a first construction unit, configured to construct an n×n adjacency matrix corresponding to the target network; n is the total number of nodes in the target network, and the ith row and jth column elements in the adjacency matrix represent: node i points to the number of outputs of node j;

the second construction unit is used for constructing N characteristic values serving as diagonal elements to obtain an N multiplied by N characteristic value matrix;

the third construction unit is used for enabling the degrees of all nodes in the target network to serve as diagonal elements and constructing an NxN degree matrix;

The second calculation unit is used for adding diagonal line elements of the eigenvalue matrix and the degree matrix to each other to obtain a division index value of each node in the target network;

In one embodiment, the dividing unit is specifically configured to:

if the similarity is smaller than the preset similarity threshold, dividing the two nodes into the same subnet.

In one embodiment, the second selection module includes:

wherein, the second formula is:；/>representing the propagation probability of the node u in the current subnet to propagate multi-source information to the node v in the current subnet, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, and +.>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

In one embodiment, the probability calculation unit is specifically configured to:

wherein, the third formula is:；/>representing the propagation probability of the node u in the current sub-network to propagate multi-source information to the node v in the current sub-network, D (t) representing the delay function, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, < >>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

In one embodiment, the formula for D (t) is: d (t) =t ² +t or D (t) =e ^t 。

In one embodiment, the second selection module includes:

and the management node selection unit is used for taking any node as a management node if the propagation probability of the node is larger than a preset probability threshold value and the real-time resource quantity of the node is larger than a preset resource threshold value.

In one embodiment, the second selection module includes a resource amount calculation unit, configured to calculate a real-time resource amount of each node in the current subnet according to a fourth formula; the fourth formula is: t (T) _u =a×X _u +b×Y _u +c×Z _u ，T _u Representing the real-time resource quantity of node u in the current subnet, X _u Representing the real-time computing power of a node u in the current subnet, Y _u Representing the real-time memory quantity of node u in the current subnet, Z _u And the real-time bandwidth of the node u in the current subnet is represented, and a, b and c are preset parameters.

In one embodiment, the method further comprises:

if the target node is not included in any sub-network, the management node is not selected in the sub-network.

In one embodiment, the method further comprises:

In one embodiment, the deployment module is further to:

deploying a dispatching center at the position of the management node;

In one embodiment, the deployment module is further to:

the scheduling center is utilized to send information collection instructions to all nodes in the subnet where the management node is located at regular time, and node information returned by all nodes in the subnet where the management node is located according to the information collection instructions is received; or receiving node information sent by each node in the subnet where the management node is located at fixed time by using the dispatching center.

In one embodiment, the target network is: a distributed computing network, a wireless sensor network, or a data storage network.

The more specific working process of each module and unit in this embodiment may refer to the corresponding content disclosed in the foregoing embodiment, and will not be described herein.

It can be seen that the present embodiment provides a node selection apparatus, which can combine propagation characteristics of multi-source information in the same network to perform management node selection in different subnets.

An electronic device provided in the embodiments of the present invention is described below, and an electronic device described below may refer to other embodiments described herein.

Referring to fig. 6, an embodiment of the present invention discloses an electronic device, including:

a memory 601 for storing a computer program;

a processor 602 for executing the computer program to implement the method disclosed in any of the embodiments above.

Further, the embodiment of the invention also provides electronic equipment. The electronic device may be a server as shown in fig. 7 or a terminal as shown in fig. 8. Fig. 7 and 8 are each a block diagram of an electronic device according to an exemplary embodiment, and the contents of the drawings should not be construed as any limitation on the scope of use of the present invention.

Fig. 7 is a schematic structural diagram of a server according to an embodiment of the present invention. The server specifically may include: at least one processor, at least one memory, a power supply, a communication interface, an input-output interface, and a communication bus. Wherein the memory is configured to store a computer program that is loaded and executed by the processor to implement the relevant steps in node selection disclosed in any of the foregoing embodiments.

In this embodiment, the power supply is configured to provide a working voltage for each hardware device on the server; the communication interface can create a data transmission channel between the server and external equipment, and the communication protocol to be followed by the communication interface is any communication protocol applicable to the technical scheme of the invention, and the communication protocol is not particularly limited; the input/output interface is used for acquiring external input data or outputting data to the external, and the specific interface type can be selected according to the specific application requirement, and is not limited in detail herein.

In addition, the memory may be a read-only memory, a random access memory, a magnetic disk, an optical disk, or the like as a carrier for storing resources, where the resources stored include an operating system, a computer program, data, and the like, and the storage mode may be transient storage or permanent storage.

The operating system is used for managing and controlling each hardware device and computer program on the Server to realize the operation and processing of the processor on the data in the memory, and the operation and processing can be Windows Server, netware, unix, linux and the like. The computer program may further comprise a computer program capable of being used to perform other specific tasks in addition to the computer program capable of being used to perform the node selection method disclosed in any of the previous embodiments. The data may include data such as information on a developer of the application program in addition to data such as update information of the application program.

Fig. 8 is a schematic structural diagram of a terminal according to an embodiment of the present invention, where the terminal may specifically include, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, or the like.

Generally, the terminal in this embodiment includes: a processor and a memory.

The processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented in at least one hardware form of DSP (Digital Signal Processing ), FPGA (Field-Programmable Gate Array, field programmable gate array), PLA (Programmable Logic Array ). The processor may also include a main processor, which is a processor for processing data in an awake state, also called a CPU (Central Processing Unit ), and a coprocessor; a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor may incorporate a GPU (Graphics Processing Unit, image processor) for rendering and rendering of content required to be displayed by the display screen. In some embodiments, the processor may also include an AI (Artificial Intelligence ) processor for processing computing operations related to machine learning.

The memory may include one or more computer-readable storage media, which may be non-transitory. The memory may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In this embodiment, the memory is at least configured to store a computer program, where the computer program, when loaded and executed by the processor, is capable of implementing relevant steps in the node selection method performed by the terminal side as disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory can also comprise an operating system, data and the like, and the storage mode can be short-term storage or permanent storage. The operating system may include Windows, unix, linux, among others. The data may include, but is not limited to, update information for the application.

In some embodiments, the terminal may further include a display screen, an input-output interface, a communication interface, a sensor, a power supply, and a communication bus.

Those skilled in the art will appreciate that the structure shown in fig. 8 is not limiting of the terminal and may include more or fewer components than shown.

A readable storage medium provided by embodiments of the present invention is described below, and the readable storage medium described below may be referred to with respect to other embodiments described herein.

A readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the node selection method disclosed in the foregoing embodiments. The readable storage medium is a computer readable storage medium, and can be used as a carrier for storing resources, such as read-only memory, random access memory, magnetic disk or optical disk, wherein the resources stored on the readable storage medium comprise an operating system, a computer program, data and the like, and the storage mode can be transient storage or permanent storage.

In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, so that the same or similar parts between the embodiments are referred to each other.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of readable storage medium known in the art.

The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.

Claims

1. A method of node selection, comprising:

2. The method of claim 1, wherein determining the degree of each node in the target network comprises:

3. The method of claim 1, wherein determining the number of structural bridges for each node in the target network comprises:

4. The method of claim 1, wherein determining the propagation range under the preset propagation distance constraint comprises:

5. The method of claim 1, wherein said evaluating the propagation forces of each node based on said degree, said number of structural bridges, and said propagation range comprises:

6. The method of claim 5, wherein the evaluating the propagation forces of each node based on the degree, the number of structural bridges, and the propagation range after the normalization operation comprises:

7. The method of claim 1, wherein said selecting a target node in said target network based on said propagation force comprises:

8. The method according to any one of claims 1 to 7, wherein the dividing the target network into a plurality of sub-networks comprises:

9. The method of claim 8, wherein dividing each node in the target network into a plurality of subnets according to the division index value comprises:

10. The method according to any one of claims 1 to 7, wherein calculating the propagation probability of each node in the current subnet to propagate the multi-source information to other nodes in the current subnet comprises:

wherein the second formula is:；/>representing the propagation probability of the node u in the current subnet to propagate multi-source information to the node v in the current subnet, Q representing the number of multi-source information, t representing the time required for the node u to propagate information to the node v, and +.>Representing an initial propagation probability of the node u in the current subnet propagating the multi-source information to the node v in the current subnet.

11. The method according to any one of claims 1 to 7, wherein calculating the propagation probability of each node in the current subnet to propagate the multi-source information to other nodes in the current subnet comprises:

12. The method of claim 11, wherein D (t) is calculated as: d (t) =t ² +t or D (t) =e ^t 。

13. The method according to any one of claims 1 to 7, wherein selecting the management node in the current subnet according to the propagation probability and the real-time resource amount of each node in the current subnet comprises:

14. The method according to any one of claims 1 to 7, wherein determining the real-time resource amount of each node in the current subnet comprises:

15. The method according to any one of claims 1 to 7, further comprising:

16. The method according to any one of claims 1 to 7, further comprising:

17. The method according to any one of claims 1 to 7, further comprising:

deploying a dispatching center at the position of the management node;

18. The method of claim 17, wherein the collecting, by the dispatch center, node information of each node in the subnet where the management node is located, includes:

19. The method according to any one of claims 1 to 7, wherein the target network is: a distributed computing network, a wireless sensor network, or a data storage network.

20. A node selection apparatus, comprising:

21. An electronic device, comprising:

a memory for storing a computer program;

a processor for executing the computer program to implement the method of any one of claims 1 to 19.

22. A readable storage medium for storing a computer program, wherein the computer program when executed by a processor implements the method of any one of claims 1 to 19.