Dynamic network system maintenance system and method under confrontation condition

Abstract

The invention discloses a dynamic network system maintenance system and method under confrontation conditions. The system comprises a strike prediction module, a decision updating module and a pre-deployment decision module, the strike prediction module is used for predicting an important node sequence that the opponent will strike the network system maintenance party at the next moment; the decision updating module is used for updating an opponent strike strategy and feeding back a result to the strike prediction module; and the pre-deployment decision module is used for performing deployment decision based on the prediction result of the network system maintainer. According to topological structure change conditions of a system at the (i-1) th moment and the ith moment, a decision-making architecture for predicting nodes attacked by the SoS and making judgment in advance is adopted, an importance sequence of communication nodes participating in SoS composition at the next moment is output according to an attack strategy, and the communication nodes in the sequence become attack targets at the next moment in a large probability mode, so that the aim of solving the problem is fulfilled. Based on this sequence, the SoS may make a pre-deployment decision such that communication nodes in the sequence maneuver in advance, or deploy redundant silent communication nodes in the vicinity in advance.

Description

Technical Field

[0001] The present invention belongs to the technical field of dynamic network system maintenance, and in particular relates to a dynamic network system maintenance system and method under confrontation conditions. Background Art

[0002] In the actual scenario where the network system maintainer is confronting the adversary, it is necessary to effectively unite various system units, such as command and control units, communication units, and attack units. These units are usually combined to perform tasks. However, the positions of various units in real applications are often randomly distributed. In order to effectively unite various system units, communication units will be used to provide communication services to system units within their coverage, connecting all units in the entire SoS (system of systems) into a mesh topology. In order to ensure that the communication between nodes in the SoS is always effective and uninterrupted, there are always redundant silent communication nodes in the SoS. These nodes are often deployed near communication nodes that are vulnerable to attacks, so as to ensure that these communication nodes can quickly take over and participate in the composition of the SoS after they are attacked by the opponent and exit the SoS.

[0003] Assuming that in practice, the opponent's attack strategy at a certain moment can be observed, it is very important to predict the sequence of important communication nodes in the SoS that may be attacked before the attack is implemented. The existing algorithms for calculating the importance of nodes in the mesh topology are difficult to apply to the adversarial SoS that needs to change rapidly and be connected at all times.

[0004] In current research, the main reason for the lack of flexibility and timeliness in attack prediction is that only the static topological structure of the network is considered. The commonly used methods for calculating the importance of nodes in the network are:

[0005] The shortcomings of the node neighbor-based sorting method and the path-based sorting method are as follows:

[0006] The network coverage method mainly considers how to quickly cover the network, but cannot predict the opponent's attack on the network system of the network system maintainer. It does not consider how to maintain the network system of the network system maintainer.

[0007] The CADer method mainly considers how to achieve network coverage in a dynamic network, but does not consider the node prediction of the opponent's attack on the network system maintenance party. It is impossible to maintain the system under dynamic confrontation conditions:

[0008] Although this method takes into account the coverage of dynamic networks and can perform system maintenance under non-dynamic confrontation conditions, it does not take into account the damage to the network of the network system maintainer caused by the opponent's attack, and therefore cannot perform system maintenance under dynamic confrontation conditions.

[0009] The HCNT method constructs a heterogeneous network to model the real network and identify important nodes, but this method does not consider the coverage problem and therefore cannot support system integrity maintenance.

[0010] This method can predict static attacks on nodes by identifying important nodes in a static network. However, it cannot predict dynamic attacks because it does not consider the damage to the network system maintained by the opponent's attacks and the dynamic changes in the network topology of the network system maintainer. This method does not consider the connectivity between nodes and network coverage issues, so it cannot perform system maintenance. Summary of the invention

[0011] In order to overcome the deficiencies of the above-mentioned prior art, the purpose of the present invention is to provide a dynamic network system maintenance system and method under confrontation conditions. The system and method design a strike prediction decision architecture based on a dynamic network and changes in strike strategies in a confrontation environment. The decision architecture can predict the nodes of the SoS that may be hit according to the changes in the topological structure of the system at the i-1th moment and the i-th moment, and make judgments in advance. The architecture will output the importance sequence of all communication nodes participating in the SoS composition in the next moment according to the opponent's strike strategy. The communication nodes in this sequence will most likely become the strike targets at the next moment. Based on this sequence, the SoS can make a pre-deployment decision, so that the communication nodes in the sequence can make maneuvers in advance, or deploy redundant silent communication nodes around them.

[0012] In order to achieve the above object, the technical solution adopted by the present invention is:

[0013] A dynamic network system maintenance system under confrontation conditions, including a strike prediction module, a decision update module and a pre-deployment decision module;

[0014] The attack prediction module is used to predict the important node sequence that the opponent will attack in the next moment.

[0015] The decision updating module is used to update the opponent's attack strategy and feed the result back to the attack prediction module;

[0016] The pre-deployment decision module is used to make deployment decisions based on the prediction results of the network system maintainer.

[0017] The attack prediction module predicts the attack nodes based on the changes in the topological structure of the network system maintenance party and the opponent's attack strategy; the decision update module observes whether the opponent's attack strategy changes. If it changes, the opponent's updated attack strategy is fed back to the attack prediction module for re-prediction. If it does not change, the pre-deployment decision module completes the decision, so that the network system maintenance party can make maneuvers in advance, or deploy redundant silent communication nodes next to the important nodes that will be attacked.

[0018] A method for using a dynamic network system maintenance system under confrontation conditions, comprising the following steps;

[0019] Step (1): construct an attack prediction module, establish an optimization model based on the opponent's current attack strategy and the node attributes of the network system maintenance party, take the attributes of each node in the network system maintenance party SoS as input, and use the A-CADer algorithm to solve the optimization model, and output the node sequence with the highest probability of being attacked by the opponent based on the SoS connected at the current moment;

[0020] Step (2): The decision update module observes whether the opponent's attack strategy has changed. If so, the updated attack strategy of the opponent is fed back to the attack prediction module for re-prediction. If not, the pre-deployment decision module completes the decision.

[0021] Step (3): The pre-deployment decision module makes advance maneuvers based on the prediction results of the network system maintenance party, or deploys redundant silent communication nodes next to the important nodes that will be attacked.

[0022] The attack prediction module completes the opponent's attack node sequence P{P1,P2,…,P n}, and then the decision update module updates the opponent's current attack strategy, and finally the pre-deployment decision module makes the final decision for the network system maintainer.

[0023] The step (1) is specifically:

[0024] In order to predict the attack target, it is necessary to assume that the opponent's attack strategy is to prioritize the attack on the communication nodes with high importance in the initial state. The importance of each communication node in SoS is set to k, and the scoring system indicators that constitute k are: the degree of the communication node d, the range of services that the communication node can provide r, the maximum number of system nodes that the communication node can provide n, and the attack frequency of the node f, that is,

[0025] k=a1d+a2r+a3n+a4f,

[0026] Where d is the degree of the communication node, r is the range of services that the communication node can provide, n is the maximum number of system nodes that the communication node can serve, f is the frequency of the communication node being attacked; a1~a4 are weighted average coefficients;

[0027] The specific steps are as follows:

[0028] Step 1: Set the initial attack strategy as follows: prioritize attacking communication nodes with high importance;

[0029] Step 2: Define the importance of communication nodes in SoS as

[0030] k=a1d+a2r+a3n+a4f,

[0031] Where d is the degree of the communication node, r is the range of services that the communication node can provide, n is the maximum number of system nodes that the communication node can provide services, and f is the frequency at which the node is attacked;

[0032] Step 3: Observe the topological structure changes of SoS between time i-1 and time i, and update the node attributes in SoS;

[0033] Step 4: Establish a strike prediction optimization model based on the attributes of communication nodes and system nodes within the SoS;

[0034] Step 5: Use the A-CADer algorithm to solve the optimization model and finally output the node sequence with the highest probability of being attacked by the opponent based on the currently connected SoS.

[0035] The step 3 determines the characteristic information contained in the node attributes based on the node state at the current moment: the degree d of the communication node, the range r of the service that the communication node can provide, the maximum number of system nodes n that the communication node can serve, the frequency f of the node being attacked, and the location information of each node.

[0036] The specific steps are as follows:

[0037] Step 3.1: Update the index d based on the degree of the nodes in the current network maintained by the network system;

[0038] Step 3.2: Update the indicator r based on the scope of services that can be provided by the communication nodes in the current network of the network system maintainer;

[0039] Step 3.3: Update the index n based on the maximum number of system nodes that the communication node in the current network of the network system maintainer can serve;

[0040] Step 3.4: Update the index f based on the attack frequency of nodes in the current network of the network system maintainer.

[0041] The step 4 is specifically as follows:

[0042] Step 4.1: Set up communication node C i The coverage range r, only within this range, the system node S i Access to communications services;

[0043] Step 4.2: Set up communication node C i For system node S i Various communication transmission indicators;

[0044] Step 4.3: Set up communication node C i The maximum number of services and the principle of nearest service, that is, a certain system node S i Exceeding a certain communication node C i If the maximum number of services is exceeded or the distance is further from the node, the communication node C cannot be accepted. i services;

[0045] Step 4.4: Set node x i C is a redundant communication node i The communication nodes that participate in forming the SoS;

[0046] Step 4.5: Set each communication node C i Service cost v i ;

[0047] Step 4.6: Using the above steps 4.1 to 4.5 and step 3 as constraints, establish the goal: Optimization model.

[0048] The step 5 is specifically as follows:

[0049] Step 5.1: Generate an embedding vector for each node in the encoding phase. The vector contains both static structure (node ​​attributes of the network system maintenance party) and dynamic processing feature information. The encoder consists of two codes: static and dynamic.

[0050] Step 5.2: Use the multi-head attention mechanism to extract the feature information of the network maintenance node and generate a static embedding, where the feature information of the communication node includes the location, coverage radius r, and cost information v i , and the importance of the communication node k, while the system node S i Contains location information;

[0051] Step 5.3: Set the vector For communication node C i The number of system nodes within the coverage area that are not covered by the selected communication node in step t;

[0052] Step 5.4: In the uncovered system node set, there are key system nodes of the entire solution. The dynamic encoding information is projected to the embedding vector by linear projection. superior;

[0053] Step 5.5: Combine the static coding information with the dynamic coding information to jointly construct the final embedding vector of the communication node;

[0054] Step 5.6: The decoder takes the embedding vector outputted in step 5.5 as input. The decoding phase consists of k steps. In each step, the agent selects an unselected communication node and adds the selected communication node to the output node sequence. Then, the dynamic embedding is updated according to the environment state.

[0055] Step 5.7: Repeat step 5.6 until the constructed output node sequence P{P1,P2,…,P n} or the maximum number of iterations is reached.

[0056] The step (2) is specifically:

[0057] As the attack strategy is updated, SoS can determine the weighted coefficients of each indicator in the importance k of the communication node by the distribution of the opponent's actual attack nodes. Before the pre-deployment decision module makes the final decision, the attack prediction module is executed again to obtain a new sequence P based on the updated opponent's attack strategy. ′ {P1,P2,…,P n}, as follows:

[0058] Step 1: If the attack strategy is updated to prioritize the communication node with the highest number of service system nodes in the SoS at a certain moment, the importance is adjusted by increasing the coefficient a1.

[0059] Step 2: If the attack strategy is updated to prioritize attacking the communication node with the largest service range at a certain moment in the network system maintenance party SoS, the importance is adjusted by increasing the coefficient a2.

[0060] Step 3: If the attack strategy is updated to prioritize the communication node with the strongest service capability at a certain moment in the SoS, the importance is adjusted by increasing the coefficient a3.

[0061] Step 4: If the attack strategy is updated to prioritize the communication nodes that have been attacked the most times in the SoS, the importance is adjusted by increasing the coefficient a4.

[0062] Through the above steps, the decision updating module can automatically adjust the model established by the attack prediction module according to the changes in the attack strategy and update the prediction sequence.

[0063] The step (3) is specifically:

[0064] The prediction attack module outputs the node sequence P{P1,P2,…,P n}, make decisions and deployments based on the confrontation situation at that moment;

[0065] The details are as follows:

[0066] Step 1: Observe whether the opponent's attack strategy is updated;

[0067] Step 2: If the attack strategy has been updated, execute the decision update module and output a new sequence P′{P1,P2,…,P n}; otherwise, follow the existing sequence P{P1,P2,…,P n} Make decisions;

[0068] Step 3: Based on this sequence and the actual confrontation situation, the SoS can make pre-deployment decisions so that the communication nodes in the sequence can make maneuvers in advance or deploy redundant silent communication nodes around them.

[0069] After the above process, the decision-making framework can predict to a certain extent the node sequence that the opponent may attack at the next moment according to the opponent's attack strategy and the topological structure changes of the SoS network maintained by the network system, so as to make decisions in advance.

[0070] Beneficial effects of the present invention:

[0071] The present invention is based on a strike prediction decision framework based on dynamic networks and changes in opponent strike strategies. It can predict the important node sequence of the opponent's strike at the next moment while maintaining the overall network connectivity according to the changes in the topological structure of the network system maintainer's SoS and the adjustment of the opponent's strike strategy, so that the network system maintainer can make more favorable decisions. The method proposed by the present invention has the following advantages:

[0072] The present invention can complete the attack prediction and network coverage at the same time, so that the nodes in the network system remain connected.

[0073] This method can support static calculation of important node sequences based on network topology and complete static attack prediction.

[0074] This method can support dynamic attack prediction based on the dynamic changes of the network topology structure of the network system maintainer and the opponent's attack strategy.

[0075] In a non-confrontational environment, this method can perform network coverage and system maintenance in a dynamic environment.

[0076] This method takes into account the impact of the opponent's attack on the network system of the network system maintainer, can make attack predictions based on the adjustment of the opponent's attack strategy, and can perform system maintenance in a dynamic confrontation environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is the overall flow chart of the present invention.

[0078] Figure 2 It is a flowchart of the attack prediction module. DETAILED DESCRIPTION

[0079] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0080] like Figure 1 As shown, the present invention discloses a dynamic network system maintenance system under confrontation conditions, including a strike prediction module, a decision update module and a pre-deployment decision module;

[0081] The attack prediction module is used to predict the important node sequence that the opponent will attack in the next moment.

[0082] The decision updating module is used to update the opponent's attack strategy and feed the result back to the attack prediction module;

[0083] The pre-deployment decision module is used to make deployment decisions based on the prediction results of the network system maintainer.

[0084] The attack prediction module predicts the attack nodes based on the changes in the topological structure of the network system maintenance party and the opponent's attack strategy; the decision update module observes whether the opponent's attack strategy changes. If it changes, the opponent's updated attack strategy is fed back to the attack prediction module for re-prediction. If it does not change, the pre-deployment decision module completes the decision, so that the network system maintenance party can make maneuvers in advance, or deploy redundant silent communication nodes next to the important nodes that will be attacked.

[0085] The present invention discloses a dynamic network system maintenance method under confrontation conditions, and designs a decision-making framework that can predict the nodes that may be hit by the network system maintenance party SoS under confrontation conditions and make judgments in advance. Figure 1 Process, the specific implementation method of the present invention is as follows:

[0086] Step 1: Set the opponent's initial attack strategy to prioritize attacking communication nodes with high importance to the network system maintainer.

[0087] Step 2: Define the importance of the communication node in the network system maintenance party SoS as k=a1d+a2r+a3n+a4f, where d is the degree of the communication node, r is the range of services that the communication node can provide, n is the maximum number of system nodes that the communication node can serve, and f is the frequency of the node being attacked.

[0088] Step 3: Observe the topological structure changes of SoS at time i-1 and time i, and determine the node attributes in SoS based on the node status at the current time. The characteristic information included includes: the four indicators d, r, n, f that constitute the importance k of the communication node, and the location information of each node. (e.g. Figure 2(shown)

[0089] Step 4: Establish a strike prediction optimization model based on the attributes of communication nodes and system nodes within the SoS; (e.g. Figure 2 (shown)

[0090] Step 4.1: Set up communication node C i The coverage range r, only within this range, the system node S i Access to communications services;

[0091] Step 4.2: Set up communication node C i For system node S i Various communication transmission indicators; communication node C i Signal transmission power and communication quality between nodes;

[0092] Step 4.3: Set up communication node C i The maximum number of services and the principle of nearest service, that is, a certain system node S i Exceeding a certain communication node C i If the maximum number of services is exceeded or the distance is further from the node, the communication node C cannot be accepted. i services;

[0093] Step 4.4: Set node x i C is a redundant communication node i The communication nodes that participate in forming the SoS;

[0094] Step 4.5: Set each communication node C i Service cost v i ;

[0095] Step 4.6: Using the above steps 4.1 to 4.5 and step 3 as constraints, establish the goal: Optimization model.

[0096] Step 5: Use the A-CADer algorithm to solve the model and finally output the node sequence P{P1,P2,…,P n};

[0097] Step 5.1: Generate an embedding vector for each node in the encoding phase. The vector contains both static structure (node ​​attributes of the network system maintenance party) and dynamic processing feature information. The encoder consists of two codes: static and dynamic.

[0098] Step 5.2: Use the multi-head attention mechanism to extract the feature information of the network maintenance node and generate a static embedding, where the feature information of the communication node includes the location, coverage radius r, and cost information v i , and the importance of the communication node k, while the system node S i Contains location information;

[0099] Step 5.3: Set the vector For communication node C i The number of system nodes within the coverage area that are not covered by the selected communication node in step t;

[0100] Step 5.4: In the uncovered system node set, there are key system nodes of the entire solution. The dynamic encoding information is projected to the embedding vector by linear projection. superior;

[0101] Step 5.5: Combine the static coding information with the dynamic coding information to jointly construct the final embedding vector of the communication node;

[0102] Step 5.6: The decoder takes the embedding vector outputted in step 5.5 as input. The decoding phase consists of k steps. In each step, the agent selects an unselected communication node and adds the selected communication node to the output node sequence. Then, the dynamic embedding is updated according to the environment state.

[0103] Step 5.7: Repeat step 5.6 until the constructed output node sequence P{P1,P2,…,P n} or the maximum number of iterations is reached.

[0104] Step 6: Execute the pre-deployment decision module to observe the opponent's attack target to determine whether it updates the attack strategy. If the opponent updates the attack strategy, proceed to step 7. Otherwise, use the existing sequence P{P1,P2,…,P n} Make deployment decisions.

[0105] Step 7: Execute the decision update module, adjust the importance k of the communication node, and output a new sequence P′{P1,P2,…,P n}.

[0106] Based on this sequence and the actual confrontation situation, the network system maintainer SoS can make pre-deployment decisions so that the communication nodes in the sequence can maneuver in advance, or deploy redundant silent communication nodes around them.

[0107] The multi-head attention mechanism described in the above steps is an extended form of the attention mechanism widely used in the Transformer model. It obtains the attention distribution of different subspaces of the input sequence by running multiple independent attention mechanisms in parallel, thereby more comprehensively capturing the potential multiple semantic associations in the sequence. The present invention uses this method to extract the feature information of the nodes of the network system maintenance party.

[0108] The main purpose of the static encoding described in the above steps is to extract the static structural information of the input nodes, which remains unchanged as the decoding step proceeds.

[0109] The dynamic coding described in the above steps is that the agent selects a communication node as output at each decoding step, and the coverage of the selected communication node changes dynamically as the decoding process changes. This dynamically changing state affects the selection of the next decoding step.

[0110] The embedding vector described in the above steps Defined as communication node C i The number of system nodes within the coverage area that are not covered by the selected communication node in step t.

[0111] The update of the opponent's attack prediction in the decision update module described in the above steps is mainly based on observing the changes in the frequency of the opponent's attacks on the nodes in the network system maintenance party. The opponent's attack frequency on a certain node in the system increases around a certain time, indicating that the opponent's attack strategy during this period tends to be nodes with this feature, and this feature corresponds to d, r, and n in the node importance scoring index.

[0112] Table 1 compares the differences between the method of the present invention and existing methods. Network coverage methods mainly consider how to quickly cover, and some methods such as CADer can be used for system maintenance under non-dynamic confrontation conditions. The node importance evaluation method does not consider the coverage problem, so it cannot support system integrity maintenance. The solution of the present invention considers the problem of how to quickly maintain the system under dynamic confrontation conditions, and can simultaneously support static attack prediction based on network topology and attack prediction based on the opponent's attack strategy. Therefore, the solution of the present invention is superior to existing solutions in solving the problem of maintaining and commanding network systems under dynamic system confrontation conditions.

[0113] Table 1 Comparison of functions of different methods

[0114]

[0115] This method can complete network coverage in a dynamic confrontation environment, so that each node in the network system maintenance SoS remains connected.

[0116] This method can predict the opponent's attack nodes in a dynamic confrontation environment based on the static attributes of the network system maintenance party's nodes and the opponent's dynamic attack strategy.

[0117] This method can quickly update the attack prediction model based on changes in the opponent's attack strategy to ensure the accuracy of the attack prediction sequence.

[0118] This method can predict attacks in a dynamic environment based on the topological changes of the network system maintained by the network system and the opponent's attack strategy.

[0119] This method takes into account the impact of the opponent's attack on the network system of the network system maintainer, can make attack predictions based on the adjustment of the opponent's attack strategy, and can perform system maintenance in a dynamic confrontation environment.

[0120] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. A dynamic network system maintenance system under confrontation conditions, characterized in that: It includes a strike prediction module, a decision update module, and a pre-deployment decision module; The attack prediction module is used to predict the important node sequence that the opponent will attack in the next moment. The decision updating module is used to update the opponent's attack strategy and feed the result back to the attack prediction module; The pre-deployment decision module is used to make deployment decisions based on the prediction results of the network system maintainer.

2. According to claim 1, a dynamic network system maintenance system under confrontation conditions is characterized in that: The attack prediction module predicts the attack nodes based on the changes in the topological structure of the network system maintenance party and the opponent's attack strategy; the decision update module observes whether the opponent's attack strategy changes. If it changes, the opponent's updated attack strategy is fed back to the attack prediction module for re-prediction; if it does not change, the pre-deployment decision module completes the decision, so that the network system maintenance party can make maneuvers in advance, or deploy redundant silent communication nodes next to the important nodes that will be attacked.

3. A method for using a dynamic network system maintenance system under confrontation conditions according to claim 1 or 2, characterized in that: The steps include: Step (1): Constructing the attack prediction module: An optimization model is established based on the opponent's current attack strategy and the node attributes of the network system maintainer, and the attributes of each node in the network system maintainer SoS are used as input. The optimization model is solved using the A-CADer algorithm to output a node sequence with the highest probability of being attacked by the opponent in the currently connected SoS; Step (2): Decision update module: Observe whether the opponent's attack strategy has changed. If so, the updated attack strategy of the opponent is fed back to the attack prediction module for re-prediction. If not, the pre-deployment decision module completes the decision; Step (3): Pre-deployment decision module: Based on the prediction results of the network system maintainer, maneuvers can be made in advance, or redundant silent communication nodes can be deployed next to important nodes that will be attacked.

4. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 3, characterized in that: In the step (1), the specific steps are as follows: Step 1: Set the initial attack strategy as follows: prioritize attacking communication nodes with high importance; Step 2: Define the importance of communication nodes in SoS as k=a1d+a2r+a3n+a4f, Where d is the degree of the communication node, r is the range of services that the communication node can provide, n is the maximum number of system nodes that the communication node can serve, f is the frequency of the communication node being attacked; a1~a4 are weighted average coefficients; Step 3: Observe the topological structure changes of SoS between time i-1 and time i, and update the node attributes in SoS; Step 4: Establish a strike prediction optimization model based on the attributes of communication nodes and system nodes within the SoS; Step 5: Use the A-CADer algorithm to solve the optimization model and finally output the node sequence with the highest probability of being attacked by the opponent based on the currently connected SoS.

5. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 4, characterized in that: In the step 3, the characteristic information contained in the node attributes is determined based on the node state at the current moment: the degree d of the communication node, the range r of the service that the communication node can provide, the maximum number of system nodes n that the communication node can serve, the frequency f of the node being attacked, and the location information of each node. The specific steps are as follows: Step 3.1: Update the index d based on the degree of the nodes in the current network maintained by the network system; Step 3.2: Update the indicator r based on the scope of services that can be provided by the communication nodes in the current network of the network system maintainer; Step 3.3: Update the index n based on the maximum number of system nodes that the communication node in the current network of the network system maintainer can serve; Step 3.4: Update the index f based on the attack frequency of nodes in the current network of the network system maintainer.

6. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 4, characterized in that: The step 4 is specifically as follows: Step 4.1: Set up communication node C i The coverage range r, only within this range, the system node S i Obtaining communications services; Step 4.2: Set up communication node C i For system node S i Various communication transmission indicators; Step 4.3: Set up communication node C i The maximum number of services and the principle of nearest service, that is, a certain system node S i Exceeding a certain communication node C i If the maximum number of services is exceeded or the distance is further from the node, the communication node C cannot be accepted. i services; Step 4.4: Set node x i C is a redundant communication node i The communication nodes that participate in forming the SoS; Step 4.5: Set each communication node C i Service cost v i ; Step 4.6: Using the above steps 4.1 to 4.5 and step 3 as constraints, establish the goal: Optimization model.

7. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 4, characterized in that: The step 5 is specifically as follows: Step 5.1: Generate an embedding vector for each node in the encoding stage. The vector contains the feature information of static structure and dynamic processing. The encoder consists of two codes: static and dynamic. Step 5.2: Use the multi-head attention mechanism to extract the feature information of the network system maintenance node and generate a static embedding, where the feature information of the communication node includes location, coverage radius r, cost information v i , and the importance of the communication node k, while the system node S i Contains location information; Step 5.3: Set the vector For communication node C i The number of system nodes within the coverage area that are not covered by the selected communication node in step t; Step 5.4: In the uncovered system node set, there are key system nodes of the entire solution. The dynamic encoding information is projected to the embedding vector by linear projection. superior; Step 5.5: Combine the static coding information with the dynamic coding information to jointly construct the final embedding vector of the communication node; Step 5.6: The decoder takes the embedding vector outputted in step 5.5 as input. The decoding phase consists of k steps. In each step, the agent selects an unselected communication node and adds the selected communication node to the output node sequence. Then, the dynamic embedding is updated according to the environment state. Step 5.7: Repeat step 5.6 until the constructed output node sequence P{P1, P2, ..., P n } or the maximum number of iterations is reached.

8. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 4, characterized in that: In step (2), as the attack strategy is updated, the SoS determines the value of k of the weighted coefficient of each indicator in the importance k of the communication node through the distribution of the opponent's actual attack nodes, and executes the attack prediction module again before the pre-deployment decision module makes a final decision, and obtains a new sequence P′{P1, P2, ..., P n }.

9. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 8, characterized in that: The step (2) is specifically as follows: Step 1: If the attack strategy is updated to prioritize the communication node with the highest number of service system nodes in the SoS at a certain moment, the importance is adjusted by increasing the coefficient a1; Step 2: If the attack strategy is updated to prioritize attacking the communication node with the largest service range at a certain moment in the network system maintenance party SoS, the importance is adjusted by increasing the coefficient a2; Step 3: If the attack strategy is updated to prioritize the communication node with the strongest service capability at a certain moment in the SoS, the importance is adjusted by increasing the coefficient a3; Step 4: If the attack strategy is updated to prioritize the communication nodes that have been attacked the most times in the SoS, the importance is adjusted by increasing the coefficient a4; Through the above steps, the decision updating module automatically adjusts the model established by the attack prediction module according to the changes in the attack strategy and updates the prediction sequence.

10. The method for using a dynamic network system maintenance system under confrontation conditions according to claim 8, characterized in that: The step (3) is specifically: The prediction attack module outputs the node sequence P{P1, P2, ..., P n }, make decisions and deployments based on the confrontation situation at that moment; The details are as follows: Step 1: Observe whether the opponent's attack strategy is updated; Step 2: If the attack strategy has been updated, execute the decision update module and output a new sequence P′{P1, P2, ..., P n }; otherwise, follow the existing sequence P{P1, P2, ..., P n } Make decisions; Step 3: Based on this sequence and the actual confrontation situation, the SoS makes a pre-deployment decision so that the communication nodes in the sequence can make maneuvers in advance or deploy redundant silent communication nodes around them.

Images