Cross-domain collaborative network model construction method and system

Abstract

The invention discloses a cross-domain collaborative network model construction method and system, and relates to the technical field of data processing, the cross-domain collaborative network model construction system comprises a plurality of sub-network model generation modules and a splicing module, through the plurality of sub-network model generation modules, a network model of a plurality of domains is generated; splicing the networks of the plurality of domains through a splicing module to generate a cross-domain collaborative network model; the sub-network model generation module comprises a data acquisition module, a clustering module and a construction module; according to the cross-domain collaborative network model construction method and system, by using the technical scheme provided by the invention, the network model with high association degree by taking the command and control task as the center can be quickly generated, so that powerful support can be provided for subsequent quick command and control decision, and by adopting the scheme provided by the invention, the user experience is improved. And the advantages of high information transmission rate, high spectrum efficiency, high energy utilization rate and high confidentiality can be obtained.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for constructing a cross-domain collaborative network model. Background Art

[0002] Multi-domain command and control refers to a type of organizational leadership activity in which military commanders and their command organs exercise power and guidance over the operations and other military operations of their subordinate units and cross-domain units when completing multi-domain combat missions. In multi-domain command and control, cross-domain integration is the key. Multi-domain operations mainly achieve capability integration through five stages: preparation time, planning and execution time, duration, reset time and cycle. The boundaries of the multi-domain command and control system are fuzzy, uncertain, and game-adversarial. The number of component systems that constitute the system is large, and the interconnections and interactions are complex. Under limited resources, there are situations of mutual cooperation and competition. When there are problems in the cooperation between component systems, the interaction between component systems may change under the pressure of competition, causing changes in the system structure, and then causing changes in the system behavior. On the other hand, the future environment is uncertain. When the environment changes, it is inevitable that the system must adapt to the new changes, which requires the system to make corresponding structural and behavioral changes.

[0003] For example, the Chinese invention patent application with the publication number CN117149946A discloses a method and system for intelligent recommendation of highly dynamic situational spatiotemporal knowledge, including: first, constructing a knowledge graph using the static situational knowledge of the target, then using a relational graph convolutional neural network to embed the knowledge graph, and at the same time calculating the position code according to the spatiotemporal position of the target, splicing the embedding vector of the target entity with the position code, and finally splicing the retrieval target splicing vector and other target splicing vectors again, respectively, and inputting the multi-layer perceptron to obtain the recommendation result. This invention patent application introduces the spatiotemporal position information of the target into the situational knowledge graph, and adds the influence of the target spatiotemporal position relationship on the recommendation result when recommending situational knowledge, so that the recommendation result can adapt to the high dynamic changes of the target environment, and further improve the comprehensiveness and accuracy of situational knowledge recommendation. However, the current target environment is multi-domain coexisting, and the entity nodes are huge. The speed of generating the knowledge graph by the above method does not take into account the correlation between entities. Therefore, a method and system for constructing a cross-domain collaborative network model is proposed. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method and system for constructing a cross-domain collaborative network model, which solves the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for constructing a cross-domain collaborative network model, comprising a plurality of sub-network model generation modules and a splicing module, generating network models of multiple domains through the plurality of sub-network model generation modules; and splicing the networks of the multiple domains through the splicing module to generate a cross-domain collaborative network model; The sub-network model generation module includes a data acquisition module, a clustering module, and a construction module; The implementation process of each sub-network model generation module includes the following steps: S1: Receive the feature data of the decision node entity, monitoring node entity, and HL selected node entity of the domain in real time through the data acquisition module, and send the acquired entity set and all entity-related data in the entity set to the clustering module; S2: Cluster the entities in the entity set through the clustering module for the target command task, and select the central entity of the cluster center , and calculate the degree of association between it and other entities in the entity set, and generate The associated entity sequence; S3: Determined by building blocks In the associated entity sequence, the attribute relationship and weight of the edge connected to each entity are generated, and the entities are connected and entities associated with it; S4: Based on the target’s command mission, Connected EntitiesSelect an entity , and calculate the degree of association between it and the entities that have not been connected in the entity set, and generate the entity according to the degree of association The associated entity sequence; S5: Determined by building blocks The attribute relationship and weight of the edge connecting each entity in the associated entity sequence, and the connection entity and entities associated with it; S6: Repeat S4 and S5 until all entities in all entity sets are connected to form a cross-domain collaborative network model.

[0006] Optionally, the HL selects a node It includes a contribution calculation module and a request module. The contribution calculation module includes a data selection module and a graph attention mechanism model layer. The data selection module selects a target node and Neighbors HL selects the node, HL selects the node and The neighbor HL selected nodes form the HL selected node set ; Select the node of HL and The feature vectors of the nodes selected by the neighbor HL to complete the goal are input into each graph attention mechanism model of the graph attention mechanism model layer; the graph attention mechanism model layer includes N graph attention mechanism models, The graph attention mechanism model outputs the nodes selected by HL respectively. Contribution ; The request module selects nodes according to the neighbor HL selected nodes The contribution of each node is used to request the neighbors who have completed the goal.

[0007] Optionally, based on the above content, we can obtain: (1) In formula (1), Select the node set for HL The The feature vector of the node selected by HL to complete the goal Select the node for HL The contribution of the feature vector hn to the goal; is the activation function; Parameters from the input layer to the hidden layer of the graph attention mechanism model; represents the parameter matrix; Indicates that and Put it together.

[0008] Optionally, the decision node, the monitoring node and the HL selected node all include a wireless transmitter and a wireless receiver, and the wireless transmitters all include a first measurement unit, a transmission controller and a transmission unit; The first measuring unit may measure, according to each wireless resource, an electromagnetic wave environment between a wireless transmitter of a first node and a wireless receiver of a second node that communicate with each other; The transmission controller determines the carrier frequency combination and parameters used in symbol transmission based on the measurement result of the first measurement unit; The sending unit sends symbols of the bit group using a carrier frequency combination; The wireless receivers all include a second measuring unit, a receiving controller, a receiving unit and a decoding unit; The second measuring unit measures the electromagnetic wave environment for each wireless resource; The receiving controller estimates the parameter based on the measurement result of the second measurement unit, The receiving unit receives the symbol of the bit group sent by the wireless transmitter using the measurement result of the second measuring unit; The decoding unit decodes the sign of the received bit group based on the parameter.

[0009] Optionally, the transmitting unit includes an input source module, a bit grouping module, a time slot division module, a mapper and an adder, wherein the input source module is used to receive a bit information sequence to be modulated after source coding; the bit grouping module converts the bit information sequence into Bits are grouped The time slot division module is used to divide the symbol time length into Divide into time slots; the mapper maps each bit group of the bit information sequence into a symbol according to the following steps; the adder adds the bit group The mapping of the time slots is added to generate the symbol of the bit group.

[0010] Optionally, the mapper maps each bit group of the bit information sequence into a symbol according to the following method: Select from the radio resources measured by the first measurement unit mutually orthogonal frequency combinations , and the bit group is mapped in the first time slot according to the following formula: (2) Select the frequency cycle step size as ; According to the following formula, find the bit group at the back In the time slot Mapping of time slots: (3) According to formula (4), the bit group The mapping of time slots is added to generate the symbol of the bit group: . (4) Optionally, the wireless receiver includes a serial-to-parallel conversion module, a multi-carrier incoherent receiving module, a multi-time slot merging module, a calculation module and a decoding module; The serial-to-parallel conversion module divides the received data into multiple symbol data segments according to the symbol length, and each symbol data segment is further divided into multiple symbol data segments according to the time slot length. timeslot data segment; The multi-carrier non-coherent receiving module The time slot data segment is based on The carrier frequency is non-coherently received to form The output value of each channel is then Output value and other Time slot corresponding to The output values ​​of each channel are shifted and restored respectively; The multi-time slot merging module combines the first time slot corresponding to Output value and other The time slot is superimposed on the J-channel output value after shift recovery and merged into the J-channel output. , and provide it to the calculation module; The calculation module will The output of the circuit is divided into two parts: and , the transmitted symbol has carrier frequency to carry information, then the received bit group The corresponding log-likelihood ratio is: (5) is the zero-order modified Bessel function of the first kind; and is the estimated parameter; According to formula (6), the bit group The confidence probability of bits being "0": (6) The decoding module, according to the confidence probability of all bits Decode to get bit information.

[0011] A cross-domain collaborative network model construction system includes a storage medium and several processors, wherein the storage medium stores computer program codes, and the processors call the computer program codes in the storage medium to implement the above method.

[0012] The present invention provides a cross-domain collaborative network model construction method and system, which has the following beneficial effects: The cross-domain collaborative network model construction method and system, by using the technical solution provided by the present invention, can quickly generate a network model with a high correlation centered on completing the command task, thereby providing strong support for subsequent rapid command decisions. Through the solution provided by the present invention, the advantages of high information transmission rate, high spectrum efficiency, high energy utilization and high confidentiality can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 is a flow chart of the cross-domain collaborative network model construction method of the present invention; Figure 2 It is a block diagram of the composition of the wireless transmitter of each node of the present invention; Figure 3 It is a block diagram of the composition of the wireless receiver of each node of the present invention. DETAILED DESCRIPTION

[0014] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0015] Embodiment 1 See also Figures 1 to 3 , the present invention provides a technical solution: a method for constructing a cross-domain collaborative network model, comprising a plurality of sub-network model generation modules and a splicing module, generating network models of multiple domains through the plurality of sub-network model generation modules; and splicing the networks of the plurality of domains through the splicing module to generate a cross-domain collaborative network model; The sub-network model generation module includes a data acquisition module, a clustering module, and a construction module; The implementation process of each sub-network model generation module includes the following steps: S1: Receive the feature data of the decision node entity, monitoring node entity, and HL selected node entity of the domain in real time through the data acquisition module, and send the acquired entity set and all entity-related data in the entity set to the clustering module; S2: Cluster the entities in the entity set through the clustering module for the target command task, and select the central entity of the cluster center , and calculate the association between it and other entities in the entity set, and generate The associated entity sequence; S3: Determined by building blocks In the associated entity sequence, the attribute relationship and weight of the edge connected to each entity are generated, and the entities are connected and entities associated with it; S4: Based on the target’s command mission, Connected EntitiesSelect an entity , and calculate the degree of association between it and the entities that have not been connected in the entity set, and generate the entity according to the degree of association The associated entity sequence; S5: Determined by building blocks The attribute relationship and weight of the edge connected to each entity in the associated entity sequence, and the connection entity and entities associated with it; S6: Repeat S4 and S5 until all entities in all entity sets are connected to form a cross-domain collaborative network model.

[0016] Among them, the association degree associated with the computing entity includes: the aggregation ability evaluation value of the computing entity and other entities, the correlation between the entity and other entities based on topology and interaction behavior, and the contribution of other entities to the completion of the entity's tasks.

[0017] Among them, HL selects the node It includes a contribution calculation module and a request module. The contribution calculation module includes a data selection module and a graph attention mechanism model layer. The data selection module selects a target node and Neighbors HL selects the node, HL selects the node and The neighbor HL selected nodes form the HL selected node set ; Select the node of HL and The feature vectors of the nodes selected by the neighbor HL to complete the goal are input into each graph attention mechanism model of the graph attention mechanism model layer; the graph attention mechanism model layer includes N graph attention mechanism models, The graph attention mechanism model outputs the nodes selected by HL respectively. Contribution ; The request module selects nodes according to the neighbor HL selected nodes The contribution of each node is used to request the neighbors who have completed the goal.

[0018] Based on the above, we can get: (1) In the formula, Select the node set for HL The The feature vector of the node selected by HL to complete the goal Select the node for HL The contribution of the feature vector hn to the goal; is the activation function; Parameters from the input layer to the hidden layer of the graph attention mechanism model; represents the parameter matrix; Indicates that and Put it together.

[0019] The method and system provided by the present invention can quickly generate a highly correlated network model centered on completing the command task, providing strong support for subsequent rapid command decision-making.

[0020] Among them, the policy node, monitoring node and HL selected node all include wireless transmitters and wireless receivers; like Figure 2 As shown, the wireless transmitter includes: a first receiving antenna, a first measuring unit, a transmitting controller, a transmitting unit and a first transmitting antenna; The first receiving antenna receives electromagnetic waves of a channel that can be used by the wireless transmitter of the first node and the wireless receiver of the second node in an electric wave environment between the wireless transmitter of the first node and the wireless receiver of the second node that communicate with each other; The first measuring unit measures the electromagnetic wave environment between the wireless transmitter of the first node and the wireless receiver of the second node communicating with each other according to the electromagnetic wave received by the first receiving antenna; A transmission controller, based on the measurement result of the radio wave environment, determines the radio resources and parameters used in the transmission of the bit information sequence; a transmitting unit, using the determined wireless resources to transmit information indicating the determined wireless resources and a bit information sequence; The first transmitting antenna is used to convert the symbols of the bit group into electromagnetic waves and send them into space; The sending unit includes an input source module, a bit grouping module, a time slot division module, a mapper and an adder; the input source module is used to receive the bit information sequence to be modulated after source coding; The bit grouping module converts the bit information sequence into Bits are grouped The time slot division module is used to divide the symbol duration into Divide into time slots; A mapper maps each bit group of the bit information sequence into a symbol according to the following steps; an adder is used to add the bit group into a symbol according to the following formula The mapping of time slots is added to generate the symbol of the bit group: Select from the radio resources measured by the first measurement unit mutually orthogonal frequency combinations , and the bit group is mapped in the first time slot according to the following formula: ; (2) Select the frequency cycle step size as ; According to the following formula, find the bit group at the back In the time slot Mapping of time slots: (3) The bit groups are grouped according to the following formula The mapping of time slots is added to generate the symbol of the bit group: (4) like Figure 3 As shown, the wireless receiver includes: a second receiving antenna, a second measuring unit, a receiving controller, a receiving unit and a third receiving antenna. The second receiving antenna receives electromagnetic waves of a channel usable by the wireless transmitter of the first node and the wireless receiver of the second node in an electric wave environment between the wireless transmitter of the first node and the wireless receiver of the second node communicating with each other; The second measuring unit measures the radio wave environment for each wireless resource; A receiving controller estimates parameters based on the measurement results of the radio wave environment; A third receiving antenna receives symbols of the bit group; The receiving unit receives information indicating the determined wireless resource and a symbol of the bit group using the determined wireless resource; The decoding unit decodes the sign of the received bit group based on the parameter; The decoding unit includes a serial-to-parallel conversion module, a multi-carrier incoherent receiving module, a multi-time slot merging module, a calculation module and a decoding module; The serial-to-parallel conversion module divides the received data into multiple symbol data segments according to the symbol length, and each symbol data segment is further divided into timeslot data segment; The multi-carrier non-coherent receiving module first The time slot data segment is based on The carrier frequency is non-coherently received to form The output value of each channel is then Output value and other Time slot corresponding to The output values ​​of each channel are shifted and restored respectively; The multi-slot merging module combines the first time slot corresponding to Output value and other Time slots vs. The output values ​​of the channels are shifted and restored to form Output , and provide it to the calculation module; The calculation module will The output of the circuit is divided into two parts: and , the transmitted symbol has carrier frequency to carry information, then the received bit group The corresponding log-likelihood ratio is: (5) is the zero-order modified Bessel function of the first kind; and is the estimated parameter; The bit group number is calculated according to the following formula The confidence probability of bits being "0": (6) The decoding module calculates the confidence probability of all bits. Decode to get bit information.

[0021] The solution proposed in the present invention can achieve high information transmission rate, high spectrum efficiency, high energy utilization rate and high confidentiality.

[0022] Embodiment 2 A cross-domain collaborative network model construction system includes a storage medium and several processors. The storage medium stores computer program codes, and the processors call the computer program codes in the storage medium to implement the above method.

[0023] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for constructing a cross-domain collaborative network model, characterized in that: It includes multiple sub-network model generation modules and splicing modules. Through the multiple sub-network model generation modules, network models of multiple domains are generated; through the splicing module, networks of multiple domains are spliced ​​to generate a cross-domain collaborative network model; The sub-network model generation module includes a data acquisition module, a clustering module, and a construction module; The implementation process of each sub-network model generation module includes the following steps: S1: Receive the feature data of the decision node entity, monitoring node entity, and HL selected node entity of the domain in real time through the data acquisition module, and send the acquired entity set and all entity-related data in the entity set to the clustering module; S2: Cluster the entities in the entity set through the clustering module for the target command task, and select the central entity v at the cluster center k , and calculate the association between it and other entities in the entity set, and generate the entity v according to the association k The associated entity sequence; S3: Determine v through building blocks k In the associated entity sequence, the attribute relationship and weight of the edge connected to each entity are generated, and the entity v is connected k and entities associated with it; S4: According to the target's command task, the central entity v k Connected Entities Select an entity v j , and calculate the degree of association between it and the entities that have not been connected in the entity set, and generate the entity v according to the degree of association j The associated entity sequence; S5: Determine v through building blocks j The attribute relationship and weight of the edge connected to each entity in the associated entity sequence, and the connection entity v j and entities associated with it; S6: Repeat S4 and S5 until all entities in all entity sets are connected to form a cross-domain collaborative network model.

2. A cross-domain collaborative network model construction method according to claim 1, characterized in that: The HL selects the node It includes a contribution calculation module and a request module. The contribution calculation module includes a data selection module and a graph attention mechanism model layer. The data selection module selects a target node and N-1 neighbor HL selected nodes from the network model of the domain. The HL selected node and N-1 neighbor HL selected nodes form the HL selected node set V N ; Select the node of HL The feature vectors of the target completed by the HL selected node and its N-1 neighbors are respectively input into each graph attention mechanism model of the graph attention mechanism model layer; the graph attention mechanism model layer includes N graph attention mechanism models, and the N graph attention mechanism models respectively output the target completed by the HL selected node Contribution α kj ; α request module selects nodes according to neighbor HL selected nodes The contribution of the decision node is used to request the neighbors who have completed the goal.

3. The method for constructing a cross-domain collaborative network model according to claim 1, characterized in that: Based on the above, we can get: In formula (1), α kj Select node set V for HL N The characteristic vector h of the jth HL selected node to complete the goal j Select the node for HL The contribution of the feature vector hn to the target; LeakyReLU is the activation function; ρ is the parameter from the input layer to the hidden layer of the graph attention mechanism model; W represents the parameter matrix; || represents the conversion of Wh k and Wh j Put it together.

4. The method for constructing a cross-domain collaborative network model according to claim 1, characterized in that: The decision node, monitoring node and HL selected node all include a wireless transmitter and a wireless receiver, and the wireless transmitters all include a first measurement unit, a transmission controller and a transmission unit; The first measuring unit may measure, according to each wireless resource, an electromagnetic wave environment between a wireless transmitter of a first node and a wireless receiver of a second node that communicate with each other; The transmission controller determines the carrier frequency combination and parameters used in symbol transmission based on the measurement result of the first measurement unit; The sending unit sends symbols of the bit group using a carrier frequency combination; The wireless receivers all include a second measuring unit, a receiving controller, a receiving unit and a decoding unit; The second measuring unit measures the electromagnetic wave environment for each wireless resource; The receiving controller estimates the parameter based on the measurement result of the second measurement unit, The receiving unit receives the symbol of the bit group sent by the wireless transmitter using the measurement result of the second measuring unit; The decoding unit decodes the sign of the received bit group based on the parameter.

5. The method for constructing a cross-domain collaborative network model according to claim 1, characterized in that: The transmitting unit comprises an input source module, a bit grouping module, a time slot division module, a mapper and an adder. The input source module is used to receive a bit information sequence to be modulated after source coding; the bit grouping module groups the bit information sequence into groups B={b1,…,b j ,…b J } is divided and provided to the symbol mapper; the time slot division module is used to divide the symbol time length T symbol The time slot is divided into N time slots; the mapper maps each bit group of the bit information sequence into a symbol according to the following steps; the adder adds the mappings of the bit group in N time slots to generate the symbol of the bit group.

6. A cross-domain collaborative network model construction method according to claim 1, characterized in that: The mapper maps each bit group of the bit information sequence into a symbol according to the following method: Select J mutually orthogonal frequency combinations F = {f1, ..., f j ,…f J }, and the bit group is mapped in the first time slot according to the following formula: Select the frequency cycle step size as a; find the mapping of the bit group in the nth time slot in the next N-1 time slots according to the following formula: According to formula (4), the mapping of the bit group N time slots is added to generate the symbol of the bit group:

7. A cross-domain collaborative network model construction method according to claim 1, characterized in that: The wireless receiver includes a serial-to-parallel conversion module, a multi-carrier incoherent receiving module, a multi-time slot combining module, a calculation module and a decoding module; The serial-to-parallel conversion module divides the received data into multiple symbol data segments according to the symbol length, and each symbol data segment is further divided into N time slot data segments according to the time slot length; The multi-carrier non-coherent receiving module performs non-coherent reception on N time slot data segments based on J carrier frequencies to form NJ output values, and then performs shift recovery on the J output values ​​corresponding to the first time slot and the J output values ​​corresponding to other N-1 time slots respectively; The multi-time slot merging module superimposes the J-channel output value corresponding to the first time slot and the J-channel output value corresponding to other N-1 time slots after shift recovery to form a J-channel output r = {r1, ..., r j ,…,r J }, and provide it to the calculation module; The calculation module divides the J-path output into two parts, namely and The transmitted symbol has J carrier frequencies per time slot to carry information, so the received bit group b j The corresponding log-likelihood ratio is: I0 is the first kind of zero-order modified Bessel function; a and σ are estimated parameters; The confidence probability that the kth bit of the bit group is "0" is calculated according to formula (6): The decoding module, according to the confidence probability p of all bits i (0) Decode to obtain bit information.

8. A cross-domain collaborative network model construction system, characterized by: The method comprises a storage medium and several processors, wherein the storage medium stores computer program codes, and the processors call the computer program codes in the storage medium to implement any one of the methods described in claims 1-3.

Images