Distributed wireless network key node identification method

Abstract

The invention discloses a distributed wireless network key node identification method, belongs to the technical field of wireless networks, and relates to a distributed wireless network key node identification method. The invention provides a distributed wireless network key node identification method so as to improve the accuracy of wireless network key node identification. The method comprises the following steps: step 1, determining the number of sub-networks in a region and sub-network composition nodes; 2, classifying network topology structures; and step 3, identifying key core nodes of different topological structures.

Description

Technical Field

[0001] The invention belongs to the technical field of wireless network and relates to a distributed method for identifying key nodes in a wireless network. Background Art

[0002] In recent years, information technology has developed rapidly, with the increasing prevalence of intelligent sensing, the Internet of Things, and robotics, driving a dramatic increase in wireless services. Wireless networks, which offer significant advantages over traditional wired networks, have become ubiquitous in everyday life, with a growing number of connected devices. Network attack and defense, as a new type of electronic countermeasures technology, enables timely, accurate, and effective identification of critical network nodes. This not only allows for effective protection measures to prevent widespread network failures, but also provides insights into network operations and facilitates the assessment and prediction of network structure evolution. Different nodes play distinct roles in wireless networks, and an attack on a specific node could potentially disrupt the entire network. Identifying critical wireless network nodes facilitates monitoring, jamming, and attacking the network, effectively seizing the initiative in electronic countermeasures.

[0003] Current research on identifying key network nodes primarily focuses on complex networks such as social networks, infectious disease networks, protein networks, transportation networks, power grids, and graph neural networks. These networks are typically converted into graphs for processing, employing methods such as image analysis, centrality metrics, and network topology analysis. These methods have achieved promising results in identifying key network nodes. Unlike these networks, wireless networks have specific topologies. Key nodes may be crucial for the operation of the entire network, and the loss of a single node can paralyze the entire network. Network nodes may play different roles in different topologies, and single key node identification methods have limitations.

[0004] Existing key node identification methods directly identify key nodes without considering the possibility of multiple sub-networks in the area, and without considering the impact of network topology type on key node identification; therefore, the recognition rate of key nodes in wireless networks needs to be further improved. Summary of the Invention

[0005] The purpose of the present invention is to overcome the deficiencies of the above-mentioned prior art and to propose a distributed wireless network key node identification method to improve the accuracy of wireless network key node identification.

[0006] To achieve the above object, the present invention adopts the following technical solution, which includes the following steps:

[0007] Step 1: Identify the number of sub-networks in the region and the nodes that make up the sub-networks;

[0008] Step 2: Classify the network topology;

[0009] Step 3: Identify key core nodes of different topological structures.

[0010] As a preferred embodiment, step 1 of the present invention specifically includes the following steps:

[0011] Number the nodes in the region [v1,v2,...,v n ], where v n Represents node n, and the network node adjacency matrix in the region is obtained;

[0012]

[0013] calculate:

[0014]

[0015] Take the index of the value not equal to 0 in R and get the index matrix INDX,

[0016]

[0017] The sub-network is obtained as:

[0018]

[0019] As another preferred embodiment, step 2 of the present invention specifically includes the following steps:

[0020] Calculate the degree of each subnetwork node:

[0021]

[0022] Get the degree matrix of the network:

[0023] D=[D(1),D(2),...,D(n)]

[0024] calculate:

[0025] D h =D-1

[0026] Using star topology D h Only one element is greater than 0; each element in the ring topology D is equal to 2; each element in the mesh topology is greater than or equal to 2, and not all elements are equal to 2; the tree topology satisfies The hybrid topology is a combination of the above topologies; it realizes the classification of network topology.

[0027] In addition, step 3 of the present invention includes identifying key nodes of star topology and tree topology and identifying key nodes of mesh topology and hybrid topology;

[0028] a. Identification of key nodes in star and tree topologies

[0029] O n×n is an n×n zero matrix, where n is the total number of network nodes and the network adjacency matrix is ​​adj.

[0030] Find a row or column A in adj where only one node is 1, and record the node.

[0031] Assign A to O n×n .

[0032] Set the iteration condition to continue until there is no row or column in adj with only one node set to 1. The update equation is:

[0033]

[0034] num is the importance of the node.

[0035] The last iteration is for nodes with zero remaining rows or columns, num+1.

[0036] Sort the importance of nodes according to the size of num to obtain the final key node identification result.

[0037] b. Identification of key nodes in mesh and hybrid topologies

[0038] Determine whether there is a node with degree 1 in the adjacency matrix;

[0039] If it does not exist, skip a and do not remove the branch; if it exists, use the above method a to pre-process the adjacency matrix and remove the branch structure;

[0040] Get the branch node sorting matrix:

[0041] BN=[a1,a2,...,a m ]

[0042] Define the node importance matrix:

[0043]

[0044] in, d ij represents the shortest path from node i to node j.

[0045] The importance of the node is calculated according to the node importance matrix:

[0046]

[0047] Define the support of node i as:

[0048]

[0049] Where NN=[A1,A2,...,A g ], A i is a node set consisting of the neighborhood of node i and the neighborhood of its neighbors.

[0050] Define the key node identification scale function:

[0051] NI i =NIZ i ×e -ISN

[0052] Where, ISN=[B1,B2,...,B g ], the formula of B is:

[0053]

[0054] k is the number of overlaps.

[0055] The network node ranking MN is obtained based on the calculation result of the scaling function, and the final ranking of the importance of the network nodes is:

[0056]

[0057] in, After removing the nodes containing BN in MN, the set of BN and MN is merged.

[0058] The present invention has beneficial effects.

[0059] The present invention obtains a wireless network adjacency matrix and uses network adjacency matrix operations to clarify the number of subnetworks and the nodes that make up the subnetworks; combines wireless network topology characteristics to design a topology type discrimination method and complete network topology type classification; classifies network topology types according to different topology type characteristics and designs a method for identifying key nodes of different topology types.

[0060] The present invention adopts a scaling function to achieve network node importance ranking, avoids the limitation of a single key node in key node identification, and makes wireless network key node identification more accurate.

[0061] Based on the analysis of the distribution of key nodes in wireless networks, the present invention solves the problem that the identification of key nodes in the network does not take into account the impact of multiple subnetworks and different topology types in the perception area on the identification of key nodes. At the same time, the method of the present invention combines the traditional degree centrality indicator to judge the importance of nodes from both global and local aspects, thereby improving the recognition rate of key nodes in wireless networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention is further described below with reference to the accompanying drawings and specific embodiments. The scope of protection of the present invention is not limited to the following description.

[0063] Figure 1 It is a flow chart of the present invention.

[0064] Figure 2 A randomly generated network topology graph.

[0065] Figure 3 Graph showing the effect of sub-network identification.

[0066] Figure 4 Identify renderings for different network topology types.

[0067] Figure 5 These are different randomly generated network topologies, where (a) is a tree topology and (b) is a hybrid topology.

[0068] Figure 6 Comparison of the effects of different methods on key node identification, where (a), (c), and (e) are tree topology comparison diagrams, and (b), (d), and (f) are hybrid topology comparison diagrams.

[0069] For the sake of clarity, the drawings of the present invention retain color. DETAILED DESCRIPTION

[0070] Reference Figure 1 , the specific steps for implementing the present invention are further described in detail.

[0071] Step 1: Clarify the number of sub-networks and the nodes that make up the sub-networks.

[0072] The first step is to number the nodes in the region [v1,v2,...,v n ], where v n Represents node n, and the adjacency matrix of network nodes in the region is obtained.

[0073]

[0074] The second step is to calculate:

[0075]

[0076] The third step is to take the index of the value in R that is not equal to 0 and obtain the index matrix INDX.

[0077]

[0078] The fourth step is to divide the subnetwork into:

[0079]

[0080] Step 2: Classify network topology types.

[0081] The first step is to calculate the degree of each sub-network node:

[0082]

[0083] The second step is to get the degree matrix of the network:

[0084] D=[D(1),D(2),...,D(n)]

[0085] The third step is to calculate:

[0086] D h =D-1

[0087] Step 4: Network topology classification:

[0088]

[0089] Step 3: Identify key core nodes of different topological structures.

[0090] a. Identification of key nodes in star and tree topologies

[0091] Step 1: O n×n is an n×n zero matrix, where n is the total number of network nodes and the network adjacency matrix is ​​adj.

[0092] Step 2: adj searches for a row or column A with only one node set to 1, and records the node.

[0093] Step 3: Assign A to O n×n .

[0094] Step 4: Set the iteration condition to continue until there is no row or column in adj with only one node set to 1. The update equation is:

[0095]

[0096] num is the importance of the node.

[0097] Step 5: Iterate for the last time over the nodes with zero rows or columns, num+1.

[0098] Step 6: Sort the importance of nodes according to the size of num to obtain the final key node identification result.

[0099] b. Identification of key nodes in mesh and hybrid topologies

[0100] The first step is to determine whether there are nodes with degree 1 in the adjacency matrix. If so, use the above method a to pre-process the adjacency matrix and remove the branch structure. The branch node sorting matrix is ​​obtained:

[0101] BN=[a1,a2,...,a m ]

[0102] The second step is to define the node importance matrix:

[0103]

[0104] in, d ij represents the shortest path from node i to node j.

[0105] The third step is to calculate the importance of the node according to the node importance matrix:

[0106]

[0107] The fourth step is to define the support of node i as:

[0108]

[0109] Where NN=[A1,A2,...,A g ], A i is a node set consisting of the neighborhood of node i and the neighborhood of its neighbors.

[0110] The fifth step is to define the key node identification scale function:

[0111] NI i =NIZ i ×e -ISN

[0112] Where, ISN=[B1,B2,...,B g ], the formula of B is:

[0113]

[0114] k is the number of overlaps.

[0115] Step 6: The network node ranking MN is obtained based on the calculation result of the scaling function, and the final ranking of the network node importance is:

[0116]

[0117] in, It is defined as the set of nodes that contain BN in MN and then merge BN with MN.

[0118] The effects of the present invention are further described below in conjunction with simulation experiments:

[0119] Simulation experiment conditions:

[0120] The hardware platform of the simulation experiment of the present invention is: the processor is Intel i9-12900H CPU, the main frequency is 2.5GHz, and the memory is 16GB.

[0121] The software platform for the simulation experiment of the present invention is: Windows 11 Home Chinese version, and Matlab 2020a is used for experimental simulation to obtain the adjacency matrix of different topological structures.

[0122] Simulation content and result analysis:

[0123] like Figure 2 As shown in Figure 2, the randomly generated network topology diagram is composed of 7 independent sub-network modules. Each sub-network uses a fully connected mesh topology to connect nodes. The entire network contains a total of 97 nodes.

[0124] like Figure 3 As shown in the figure, the sub-network identification effect diagram. Figure 2 The network shown is divided into 7 sub-networks, 6 of which have 13 nodes and the remaining sub-network contains 19 nodes. The total number of nodes is strictly maintained at 97, which is consistent with the node numbering and distribution of the original network topology.

[0125] like Figure 4 As shown in the figure, the recognition effect diagram of different network topology types is designed. An experiment with 1000 independent simulations was designed. Each simulation was performed through a topology generation method (the topology generation method can first set the main program, and then design corresponding sub-functions for each topology and call them in the main program. The generation method of each topology is to process the adjacency matrix, and can use existing related programs or existing artificial intelligence to write it). Five typical network topologies are randomly constructed - star topology, tree topology, ring topology, mesh topology and hybrid topology. Figure 4 It can be seen from the experimental results that the present invention can accurately identify the network topology with a low false recognition rate.

[0126] like Figure 5 As shown in Figure 2, different randomly generated network topologies are constructed. A simulation environment containing 100 nodes is constructed to generate two typical topologies: tree topology and hybrid topology network.

[0127] like Figure 6As shown in the figure, the effects of different methods on key node identification are compared. Harmonic, K-Shell, SIR, CC, EC and the method proposed in the present invention are used to identify key nodes respectively, and the importance of the nodes is ranked. Subsequently, the key nodes are deleted in descending order according to the importance of the ranked nodes, and the impact of different methods on network stability is analyzed. The three aspects of the percentage of the largest remaining connected subgraph in the network, network efficiency and the number of subgraphs are compared to verify the effectiveness of different methods in key node identification. The effectiveness of the present invention in key node identification has been fully verified, and it is superior to other methods in reducing network connectivity, accelerating network efficiency decay and increasing the number of subnetworks.

[0128] It can be understood that the above specific description of the present invention is only used to illustrate the present invention and is not limited to the technical solutions described in the embodiments of the present invention. Those skilled in the art should understand that the present invention can still be modified or replaced by equivalents to achieve the same technical effects; as long as the use requirements are met, they are within the scope of protection of the present invention.

Claims

1. A distributed wireless network key node identification method, characterized in that The following steps are involved: Step 1: Identify the number of sub-networks in the region and the nodes that make up the sub-networks; Step 2: Classify the network topology; Step 3: Identify key core nodes of different topological structures.

2. A distributed wireless network key node identification method according to claim 1, characterized in that The step 1 specifically includes the following steps: Number the nodes in the region [v1,v2,...,v n ], where v n Represents node n, and the network node adjacency matrix in the region is obtained; calculate: Take the index of the value not equal to 0 in R and get the index matrix INDX, The sub-network is obtained as:

3. A distributed wireless network key node identification method according to claim 1, characterized in that The step 2 specifically includes the following steps: Calculate the degree of each subnetwork node: Get the degree matrix of the network: D=[D(1),D(2),...,D(n)] calculate: D h =D-1 Using star topology D h Only one element is greater than 0; each element in the ring topology D is equal to 2; each element in the mesh topology is greater than or equal to 2, and not all elements are equal to 2; the tree topology satisfies The hybrid topology is a combination of the above topologies; it realizes the classification of network topology structures.

4. A distributed wireless network key node identification method according to claim 1, characterized in that Step 3 includes identifying key nodes of star topology and tree topology and identifying key nodes of mesh topology and hybrid topology; a. Identification of key nodes in star and tree topologies O n×n is an n×n zero matrix, where n is the total number of network nodes and the network adjacency matrix is ​​adj; Find a row or column A in adj where only one node is 1, and record the node; Assign A to O n×n ; Set the iteration condition to continue until there is no row or column in adj with only one node set to 1; the update equation is: num is the importance of the node; The last iteration has nodes with zero remaining rows or columns, num+1; Sort the importance of nodes according to the size of num to obtain the final key node identification result; b. Identification of key nodes in mesh and hybrid topologies Determine whether there is a node with degree 1 in the adjacency matrix; If it does not exist, skip a and do not remove the branch; if it exists, use the above method a to pre-process the adjacency matrix and remove the branch structure; Get the branch node sorting matrix: <h2 style=";text-align:left;direction:ltr">BN=[a1,a2,...,a<h2 style=";text-align:left;direction:ltr"> m <h2 style=";text-align:left;direction:ltr"> ] Define the node importance matrix: in, d ij represents the shortest path from node i to node j; The importance of the node is calculated according to the node importance matrix: Define the support of node i as: Where NN=[A1,A2,...,A g ], A i is the node set consisting of the neighborhood of node i and the neighborhood of the neighborhood; Define the key node identification scale function: NO i =SET i ×e -ISN Where, ISN=[B1,B2,...,B g ], the formula of B is: k is the number of overlaps; The network node ranking MN is obtained based on the calculation result of the scaling function, and the final ranking of the importance of the network nodes is: SN=MN⊕BN Among them, ⊕ is the set of BN and MN after removing the nodes containing BN in MN.

Images