Dynamic disintegration method for unmanned aerial vehicle cluster information interaction network

Abstract

The invention relates to a dynamic disintegration method for an unmanned aerial vehicle cluster information interaction network. The method comprises the following steps: inputting network information of the unmanned aerial vehicle cluster information interaction network to be disintegrated; constructing a collapsing model used for collapsing the unmanned aerial vehicle cluster information interaction network; key nodes in the network information are acquired in a preset duration by adopting a dynamic network disintegration algorithm so as to construct an initial disintegration removal set, and the key nodes in the initial disintegration removal set are screened based on the disintegration model so as to obtain a target disintegration removal set; and removing nodes in the unmanned aerial vehicle cluster information interaction network based on the target disintegration removal set to complete disintegration of the unmanned aerial vehicle cluster information interaction network. The scheme has sufficient applicability for the dynamic transformation characteristic of the unmanned aerial vehicle cluster information interaction network, so that the scheme has an excellent disintegration effect on the dynamic network.

Description

Technical Field

[0001] The present invention relates to the field of network disassembly technology, and in particular to a dynamic disassembly method for a drone cluster information interaction network. Background Art

[0002] The technical basis of the drone swarm model is derived from group behavior in nature, such as swarms of bees and flocks of birds. Through simple individual behaviors and logical rules, complex group behavior patterns are formed, and swarm intelligence emerges. Cluster drones have strong adaptability and flexibility, and can cope with complex and changing environments. The cluster dynamic communication network, as the core of the drone cluster, realizes information exchange and collaborative work between drones through wireless communication technology. It has the following characteristics: Node dynamics: Drone nodes may join or leave the network at any time. Edge dynamics: Communication links change over time and may be disconnected due to interference or obstacles. Self-organization: The network can automatically adjust its structure according to mission requirements. Adaptability: The network can adapt to environmental changes and maintain the stability and reliability of communication.

[0003] Network collapse refers to the phenomenon in which a network loses functionality due to node or edge failures caused by external shocks or internal failures. Network collapse typically occurs through the removal of nodes or edges, leading to the failure of local network structure and functionality, ultimately causing the collapse of the entire network. Collapse models are important tools for studying network collapse. Common collapse models include topology-based models and dynamics-based models. Topology-based models primarily predict network collapse points by analyzing node degree distribution and connectivity; dynamics-based models consider the interactions and dynamic evolution of nodes. Traditional network collapse methods primarily target static networks, decomposing the network structure to reduce overall performance. Common methods include those based on centrality metrics such as node degree, betweenness, and neighborhood degree. These methods have shown good results in static networks, but struggle to cope with the complexity of dynamic networks. In recent years, with the in-depth study of dynamic networks, scholars have begun to focus on the collapse problem of dynamic networks. However, due to the highly dynamic, adaptive, and self-organizing nature of dynamic communication networks, network reconstruction during collapse can restore network functionality, reducing the effectiveness of the collapse.

[0004] It can be seen that how to improve the effect of disrupting dynamic networks has become an urgent problem to be solved. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method for dynamically dismantling an information interaction network of a drone cluster.

[0006] To achieve the above-mentioned object, the present invention provides a method for dynamically dismantling a drone cluster information interaction network, comprising the following steps: S1. Input the network information of the drone cluster information interaction network to be collapsed; wherein, the network information is the node information of the drone cluster information interaction network with dynamic timing; S2. Constructing a disintegration model for disintegrating the drone swarm information interaction network; wherein the disintegration model is one of a disintegration model based on historical information, a disintegration model based on predictive information, and a disintegration model based on relative resilience; S3. Using a dynamic network disintegration algorithm to obtain key nodes from the network information for a predetermined period of time to construct an initial disintegration removal set, and screening the key nodes in the initial disintegration removal set based on the disintegration model to obtain a target disintegration removal set; wherein the dynamic network disintegration algorithm is a dynamic network disintegration algorithm based on closely connected groups; S4. Based on the target collapse removal set, remove the nodes in the drone cluster information interaction network to complete the collapse of the drone cluster information interaction network.

[0007] According to one aspect of the present invention, in step S2, in the step of constructing a collapse model for collapsing the drone cluster information interaction network, if the collapse model is a collapse model based on historical information, it is expressed as:

[0008] in, Represents historical information The disintegration of moments removes the set; represents a dynamic network disintegration algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Indicates the step size of historical information, which is used to reflect the span of the referenced historical information and is expressed as , subscript Indicates the span of historical information; Indicates the reference probability of historical information, which is used to reflect the reference degree of historical information. Represents a collection-to-collection mapping.

[0009] According to one aspect of the present invention, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain the target collapse removal set, if the collapse model is a collapse model based on historical information, the target collapse removal set is expressed as:

[0010] in, Based on historical information The disintegration of moments removes the set; Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

[0011] According to one aspect of the present invention, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain the target collapse removal set, if the collapse model is a collapse model based on historical information, the collapse removal set is constructed based on the following steps, which include: Set the step size of historical information used for reference and historical information span ; Construct the nodes in the UAV cluster information interaction network The reference probability at each moment is based on historical information and is expressed as:

[0012] in, Representation node In the span of historical information The reference probability under Representation node In the span of historical information The set of removed states under , and it is expressed as:

[0013]

[0014] in, Indicated by Time to Time Node Each removal status, Representation node exist Always remove the status flag, and ; Get the key nodes in the initial collapse removal set by filtering The judgment condition of whether the moment is the solution in the target collapse removal set is expressed as:

[0015] in, express Time node In the span of historical information The judgment result under Representation node exist The target of the moment is to disintegrate and remove the collection. Representation node Not present The target of the moment collapses and is removed from the set; Indicates the judgment threshold; Based on the judgment condition, all nodes in the initial collapse removal set are judged respectively to obtain the The goal of the moment is to disintegrate the removal set.

[0016] According to one aspect of the present invention, in step S2, in the step of constructing a collapse model for collapsing the drone cluster information interaction network, if the collapse model is a collapse model based on prediction information, it is expressed as:

[0017] in, Indicates that based on the prediction information The disintegration of moments removes the set; represents a collapse algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Represents the step size of the prediction information, which is used to reflect the span of the reference prediction information and is expressed as , subscript Indicates the size of the prediction information span; Indicates the reference probability of the forecast information, which is used to reflect the reference degree of the forecast information; Represents a collection-to-collection mapping.

[0018] According to one aspect of the present invention, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain the target collapse removal set, the collapse model is a collapse model based on prediction information, and the target collapse removal set is expressed as:

[0019] in, Based on predictive information The disintegration of moments removes the set; Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

[0020] According to one aspect of the present invention, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on prediction information, then the target collapse removal set is constructed based on the following steps, which include: Set the step size of the prediction information used for reference and predicted information span ; Construct the nodes in the UAV cluster information interaction network The reference probability of the moment is based on the predicted information and is expressed as:

[0021] in, Representation node In predicting information span The reference probability under Representation node In predicting information span The set of removed states under , and it is expressed as:

[0022]

[0023] in, Indicated by Time to Time Node Each removal status, Representation node exist Remove the status flag, and ; Get the key nodes in the initial collapse removal set by filtering The judgment condition of whether the target solution in the set is removed at the moment is expressed as:

[0024] in, express Time node In predicting information span The judgment result under Representation node exist The target of the moment is to disintegrate and remove the collection. Representation node Not present The target of the moment collapses and is removed from the set; Indicates the judgment threshold; Based on the judgment condition, all nodes in the initial collapse removal set are judged respectively to obtain the prediction information based on the prediction information. The goal of the moment is to disintegrate the removal set.

[0025] According to one aspect of the present invention, in step S2, in the step of constructing a collapse model for collapsing the drone cluster information interaction network, if the collapse model is a collapse model based on a relative resilience collapse model, it is expressed as:

[0026] in, Relative toughness based on The disintegration of moments removes the set; represents a collapse algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Indicates the step size of the information, which is used to reflect the span of the reference information. Indicates the size of the information span; Indicates the information span Sequential removal of the drone cluster information interaction network Removed node set The average relative toughness after Represents a collection-to-collection mapping.

[0027] According to one aspect of the present invention, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain the target collapse removal set, if the collapse model is a collapse model based on a relative toughness collapse model, the collapse removal set is expressed as:

[0028]

[0029] in, Relative toughness based on The disintegration of moments removes the set; Indicates the removal of a node set. Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

[0030] According to one aspect of the present invention, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on relative toughness, the target collapse removal set is constructed based on the following steps, which include: Set the step size of the information used for reference and information span ; The key nodes in the initial collapse removal set are screened by using a collapse model based on historical information or a collapse model based on predicted information. The first target collapse removal set at time t; wherein the first target collapse removal set is expressed as:

[0031] in, Indicates that the first target collapses and removes the set, Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes; Based on the first goal, the information span is constructed by collapsing and removing the set About Moment Removed node set ,and ; Build the removed node set in the UAV cluster information interaction network The relative toughness model after removal is expressed as:

[0032]

[0033]

[0034] in, Represents the set of removed nodes in the drone cluster information interaction network exist The relative toughness after the moment is removed, Indicates the nodes in the UAV cluster information interaction network The communication performance after the moment is removed, Represents the UAV cluster information interaction network at the reference time The communication performance size, It represents the communication performance of the UAV cluster information interaction network at the initial moment, represents the scaling factor, and , represents the rate parameter, Indicates the communication performance after the node is removed in the UAV cluster information interaction network. The derivative of time, Indicates relative time; Calculate the information span Remove each node set sequentially The average relative toughness after , where the average relative toughness is expressed as:

[0035] in, Indicates the information span Sequential removal of the drone cluster information interaction network Removed node set The average relative toughness after Represents the removal of each node set in the UAV cluster information interaction network Relative toughness after A condition for selecting and judging the first target collapse removal set is constructed, wherein the condition is constructed based on the lowest average resilience of the drone cluster information interaction network, and is used to select the set that minimizes the average relative resilience of the drone cluster from the first target collapse removal set to obtain the final target collapse removal set, and is expressed as:

[0036]

[0037] in, Indicates the removal of a node set After that, the average relative resilience of the network, Indicates the removal of a node set The lowest average relative resilience of the network afterwards; The set is removed for the final target collapse based on relative toughness.

[0038] According to a solution of the present invention, this solution is fully applicable to the dynamic transformation characteristics of the drone cluster information interaction network, so that this solution has an excellent disruption effect on the dynamic network. Among them, this solution can fully consider the dynamic changes of node positions under historical trends and / or future trends by introducing historical information and / or predictive information to construct a disruption model, which is more conducive to reducing the resilience of the drone cluster information interaction network itself and achieving efficient and lasting communication function destruction, and fully guarantees the disruption effect of the present invention.

[0039] According to a solution of the present invention, the historical information-based disruption model and the predictive information-based disruption model proposed in this solution combine historical information and predictive information, and achieve redundant attack reduction and forward-looking strikes through key node mining and future topology change prediction, thereby improving disruption efficiency and accuracy, and achieving forward-looking disruption based on historical information-based key node mining and trend prediction, with efficient and lasting communication function destruction capabilities.

[0040] According to a solution of the present invention, this solution introduces resilience theory to address the problem of collapse of dynamic drone cluster information interaction networks, and establishes a collapse model based on relative resilience, which can more effectively solve the problem of collapse of dynamic networks.

[0041] According to a solution of the present invention, the solution may further introduce resilience theory on the basis of introducing historical information and / or predictive information, thereby making the solution more applicable to the collapse of dynamic networks.

[0042] According to one solution of the present invention, this approach, while constructing a relatively resilient collapse model, offers enhanced dynamic responsiveness to dynamic networks. It also considers metrics such as performance differential, time efficiency, and rate of change, making its application in dynamic networks more reliable. This approach offers superior applicability to the complexity and reconfigurability of dynamic network collapse, effectively addressing the difficulties traditional approaches face in addressing the dynamic changes in nodes and topology in dynamic networks, as well as the exponentially increasing difficulty of collapse.

[0043] According to a solution of the present invention, this solution introduces relative resilience, which can be used to characterize the relative immediate resilience level of the network during the collapse process. It has the immediate capability at a certain moment, which not only reflects the relative proportion of the performance difference, but also further superimposes the comprehensive influence of time efficiency and change rate, essentially achieving an "instantaneous snapshot" of the dynamic response.

[0044] According to one solution of the present invention, the proposed disruption algorithm significantly reduces computational complexity, meeting the real-time requirements of large-scale dynamic networks. Furthermore, the proposed disruption model, based on relative resilience, does not require complete time series data, making it more suitable for real-time monitoring and rapid adjustment of dynamic systems. This solution not only achieves continuous and thorough disruption of dynamic networks, but also balances cost and effectiveness during the disruption process, avoiding excessive resource consumption.

[0045] According to one solution of this invention, theoretically, this solution provides a new approach and theoretical framework for the study of dynamic network disruption. Practically, its research findings can be applied to enhance the security of drone swarm communication networks, providing technical support for drone swarm technology, and possessing significant theoretical and practical significance. Compared to traditional disruption methods, this solution not only focuses on key nodes in a dynamic network, but also on the overall resilience of the dynamic network. This approach can effectively achieve a thorough and sustained reduction in network performance, thereby fully disrupting the network. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flowchart of the steps of the dynamic disintegration method for the UAV cluster information interaction network of the present invention; Figure 2 Schematic diagram of the principle of relative toughness of the present invention, wherein: Figure 2 (a) indicates the case where the relative toughness is positive, Figure 2 (b) represents the case where the relative toughness is negative; Figure 3 is a schematic diagram of the toughening process of the present invention; Figure 4 Schematic diagram of the relative toughness of the present invention; Figure 5 This is a performance curve diagram of the initially formed drone cluster information interaction network in an embodiment of the present invention; Figure 6 is a relative resilience graph of the initially formed drone cluster information interaction network in an embodiment of the present invention, wherein: Figure 6 (a) shows the basic relative resilience graph of the initially formed UAV cluster information interaction network. Figure 6 (b) Relative resilience graph of the initially formed UAV cluster information interaction network based on relative time; Figure 7 This is a performance curve diagram of the drone cluster information interaction network in an embodiment of the present invention after adopting the traditional information-free collapse method; Figure 8 : This is a relative resilience graph of the drone cluster information interaction network in an embodiment of the present invention after adopting the traditional information-free collapse method, where: Figure 8 (a) shows the basic relative resilience graph of the UAV cluster information interaction network after the traditional information-free collapse method is used. Figure 8 (b) Relative resilience graph based on relative time after the traditional information-free collapse method is used for the UAV cluster information interaction network; Figure 9 This is a network performance curve diagram of the UAV cluster information interaction network after the collapse of the historical information-based collapse model in an embodiment of the present invention; Figure 10This is a network performance curve diagram of the UAV cluster information interaction network after the collapse of the prediction information-based collapse model in an embodiment of the present invention; Figure 11 This is a comparison chart of performance curves of the UAV cluster information interaction network using different dynamic collapse methods in an embodiment of the present invention; Figure 12 This is a network performance curve diagram of the UAV cluster information interaction network after the collapse of the collapse model based on relative resilience in an embodiment of the present invention; Figure 13 : is a relative resilience diagram of the UAV cluster information interaction network after the collapse of the collapse model based on relative resilience in an embodiment of the present invention, wherein: Figure 13 (a) shows the basic relative resilience diagram of the UAV cluster information interaction network after the collapse of the collapse model based on relative resilience. Figure 13 (b) Relative resilience diagram based on relative time after the collapse of the UAV cluster information interaction network based on relative resilience. DETAILED DESCRIPTION

[0047] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] like Figure 1 As shown, according to one embodiment of the present invention, a dynamic collapse method for a drone cluster information interaction network of the present invention includes the following steps: S1. Input the network information of the drone cluster information interaction network to be collapsed; wherein the network information is the node information of the drone cluster information interaction network with dynamic time sequence; S2. Construct a disruption model for disrupting the drone swarm information interaction network; wherein the disruption model is one of a historical information-based disruption model, a predictive information-based disruption model, and a relative resilience-based disruption model; S3. A dynamic network collapse algorithm is used to obtain key nodes in network information for a preset time to construct an initial collapse removal set, and the key nodes in the initial collapse removal set are screened based on the collapse model to obtain a target collapse removal set; wherein the dynamic network collapse algorithm is a dynamic network collapse algorithm based on a closely connected group; S4. Based on the target collapse removal set, remove the nodes in the drone cluster information interaction network to complete the collapse of the drone cluster information interaction network.

[0049] According to one embodiment of the present invention, in step S1, in the step of inputting the network information of the drone cluster information interaction network to be dismantled, if the drone cluster information interaction network is a dynamic network, then the network information is the node information of the drone cluster information interaction network with dynamic time series; furthermore, it can be used to describe a network whose structure changes over time, and the change of the topological structure is manifested as the increase or decrease of nodes or edges, thereby realizing the description of the node information of the dynamic time series. For a dynamic network, since the nodes in the network are moving, their positions change over time, and thus it can be obtained Position at the moment for:

[0050]

[0051] in, 、 Indicates that the node is Coordinate value of the moment; 、 Indicates the initial coordinates of the node; Indicates the initial velocity of the node; represents the acceleration of the node; express Velocity component function of direction; express Directional velocity component function; Furthermore, since the positions of nodes are changing dynamically, the distances between nodes are also changing dynamically, which leads to the dynamic change of the topological structure of the UAV cluster information interaction network. Therefore, the differential equation of the network state can be obtained as follows:

[0052] in, Indicates that the node is The position at the moment, and , express dimensional vector space; Indicates The characteristic parameters of the moment, and , express dimensional vector space; Indicates the connection rules of the network; represents the instantaneous change function of the control network dynamics; According to the connection rules Depend on The dynamic system composed of nodes in time Network status at the time; express dimensional vector space, Indicates the characteristic dimension of the node.

[0053] Furthermore, the UAV cluster information interaction network is dynamically changing over time because the position of the nodes changes over time. For the network nodes at a certain moment, the connection rules are A dynamic generation algorithm for the topology structure of the UAV swarm information interaction network is adopted that takes geographical distance into consideration.

[0054] Furthermore, the initial time is set to Initial nodes, all initial nodes are fully connected, each time a new node is added, the new node have The edge is connected to the existing node, and the newly added node is considered when the distance is affected. Each edge of The connected modified probability is expressed as:

[0055] in, Represents a new node With existing nodes The connection probability of represents the distance impact factor; Indicates the degrees of freedom of the nodes; represents a partial positive factor; Indicates the number of nodes in the UAV cluster information interaction network at the current time (excluding new nodes); Represents a new node With existing nodes Connection probability function; Represents an existing node With existing nodes Connection probability function; Indicates the degrees of freedom of the nodes.

[0056] In this embodiment, when When , the probability of all existing nodes being connected is 0, then the new node Unable to connect to existing nodes Establish a connection, and then the new node It will remain active and will be added according to the number of edges after adding new nodes. Edges; the formula for adding edges is expressed as:

[0057] in, represents the node connection probability; is the node communication range radius; is the distance between nodes; is the distance impact factor. Repeat the above process until the number of nodes reaches the scale set by the network .

[0058] According to an embodiment of the present invention, for a dynamically changing network, since its structure is constantly changing, After an attack at a certain moment, due to its inherent motion, the vulnerability created by the attack may be patched at the next moment due to structural changes, thus weakening the attack's effectiveness. Meanwhile, although the network is constantly in motion, its structure at each moment is fixed. The removal solution at that moment can be calculated using a dynamic network disintegration algorithm based on closely connected groups (see: A Simple Algorithm for Disintegrating Information Exchange Network of UAV Swarm, 2023 9th International Symposium on System Security, Safety, and Reliability (ISSSR), Louzhaohan Wang, Yun Huang, Haifeng Dai, Guanghan Bai, and Junyong Tao). Since the network is dynamic, the network structure can be related to previous or future moments. Therefore, the network structure at previous or future moments and the targeted disintegration removal solution can serve as a reference for the current moment. Therefore, by calculating the removal solution set at the previous moment or the future moment, as the historical information or predicted information of the collapse of the drone cluster information interaction network, it can be used as a reference and supplement for the attack moment, and the proposed key node identification and collapse algorithm can be optimized. Then, the final removal solution under dynamic conditions can be obtained by combining the collapse information of historical information or predicted information.

[0059] Based on this, in step S2, in the step of constructing a collapse model for collapsing the UAV cluster information interaction network, if the collapse model is a collapse model based on historical information, then The collapse model equation at the moment is expressed as:

[0060] in, Represents historical information The disintegration of moments removes the set; represents a dynamic network disintegration algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Indicates the step size of historical information, which is used to reflect the span of the referenced historical information and is expressed as , subscript Indicates the span of historical information; Indicates the reference probability of historical information, which is used to reflect the reference degree of historical information. Represents a collection-to-collection mapping.

[0061] In this embodiment, the historical information is the step size and dynamic frequency changes The size and relative size of jointly determine the time accuracy of the reference historical information. When the network changes rapidly, a shorter step size should be selected. , because the information of distant time has less reference value to the current moment; reference probability The size setting determines the accuracy of the reference degree. When the setting is large, it means that it will only be used as a reference when the overlap degree is high, that is, the node has a greater impact on the collapse and the impact lasts longer.

[0062] Therefore, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain the target collapse removal set, if the collapse model is a collapse model based on historical information, the target collapse removal set is constructed based on the following steps, which include: Set the step size of historical information used for reference and historical information span ; Constructing the nodes in the UAV cluster information interaction network The reference probability at each moment is based on historical information and is expressed as:

[0063] in, Representation node In the span of historical information The reference probability under Representation node In the span of historical information The set of removed states under , and it is expressed as:

[0064]

[0065] in, Indicated by Time to Time Node Each removal status, Representation node A flag for each removal status, and ; Get the key nodes in the filter initial collapse removal set The target of whether the moment is the solution of the set of collapse removal is determined by the condition, and is expressed as:

[0066] in, express Time node In the span of historical information The judgment result under Representation node exist The target of the moment is to disintegrate and remove the set, Representation node Not present The target of the moment collapses and is removed from the set; Indicates the judgment threshold; Based on the judgment conditions, all nodes in the initial collapse removal set are judged separately to obtain the historical information-based In this embodiment, all nodes in the initial collapse removal set are judged based on the judgment conditions obtained above, and the nodes Reference probability The expression for other nodes in the time period The possibility of removing the solution is also universal, and the same calculation method can be used for different nodes to obtain the corresponding reference probability. If the reference probability is higher, the impact of removing the node on the network will be greater.

[0067] Furthermore, based on the reference probability Each node is judged separately, and the reference probability obtained is used as the removal probability to sort in ascending order. The nodes with larger probabilities are the alternative solution sets for collapse removal based on historical information. Further, by comparing with the judgment threshold Compare, if the node removal probability (ie reference probability) is greater than , then the probability of the node being removed from the historical information is high, and it can be used as a supplementary solution to the historical information solution to formulate a collapse strategy. The collapse removal set at the moment, and the collapse removal set is expressed as:

[0068] in, Based on historical information The disintegration of moments removes the set; Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

[0069] According to another embodiment of the present invention, the collapse model based on historical information considers the impact of historical information of network structure on the current moment, uses it as a reference for current analysis, and takes into account the existing properties of the network. Compared with the collapse method that only analyzes the network structure at the current moment, the model based on historical information is more suitable for dynamic networks. At the same time, dynamic networks not only have existing properties, but also have a kind of inertia. For dynamic networks with certain movement rules, their network changes have certain trends, that is, the possible network structures at future moments. Therefore, for the possible structures of future changes, in t After the attack is launched at a certain moment, the changes in the network structure caused by the movement of the network itself may lead to an increase or decrease in the subsequent collapse effect. Since the network is dynamically changing, the network structure at this moment will have an impact on the subsequent network structure. Therefore, based on the collapse removal solution and collapse effect at future moments (i.e., predicted information), the optimal collapse removal solution at the current moment that will lead to an increase in the collapse effect at future moments is found.

[0070] Due to the dynamic nature of networks, historical information and predicted information have different impacts on the current situation. Current changes will affect predicted information, so when performing disruption based on predicted information, it is necessary to consider the changing trend of the disruption effect and use this information to conduct disruption analysis at the current moment.

[0071] Furthermore, by calculating the removal solution set and removal disruption effect at subsequent moments as the prediction information of the UAV cluster's disruption, the disruption strategy obtained at the attack moment can be used as a reference and supplement, the proposed key node identification and disruption algorithm is optimized, and the final removal solution under dynamic conditions is obtained by combining the disruption information of the future state.

[0072] Therefore, in step S2, in the step of constructing a disintegration model for disintegrating the UAV cluster information interaction network, if the disintegration model is a disintegration model based on prediction information, it can be expressed as:

[0073] in, Indicates that based on the prediction information The disintegration of moments removes the set; represents a collapse algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Represents the step size of the prediction information, which is used to reflect the span of the reference prediction information and is expressed as , subscript Indicates the size of the prediction information span; Indicates the reference probability of the forecast information, which is used to reflect the reference degree of the forecast information; Represents a collection-to-collection mapping.

[0074] In this embodiment, the step size of the prediction information and dynamic frequency changes The size and relative size of the reference jointly determine the time accuracy of the reference. When the network changes rapidly, a shorter step size should be selected. , because the information of distant time has less reference value to the current moment; reference probability The size setting determines the accuracy of the reference degree. When the setting is large, it means that it will only be used as a reference when the overlap degree is high, that is, the node has a greater impact on the collapse and the impact lasts longer.

[0075] Furthermore, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on prediction information, the collapse removal set is constructed based on the following steps, which include: Set the step size of the prediction information used for reference and predicted information span ; Constructing the nodes in the UAV cluster information interaction network The reference probability of the moment is based on the predicted information and is expressed as:

[0076] in, Representation node In predicting information span The reference probability under Representation node In predicting information span The set of removed states under , and it is expressed as:

[0077]

[0078] in, Indicated by Time to Time Node Each removal status, Representation node exist Remove the status flag, and ; Get the key nodes in the initial collapse removal set of the filter The judgment condition of whether the moment is the solution of the collapse removal set is expressed as:

[0079] in, express Time node In predicting information span The judgment result under Representation node exist The target of the moment is to disintegrate and remove the collection. Representation node Not present The target of the moment collapses and is removed from the set; Indicates the judgment threshold; Based on the judgment conditions, all nodes in the initial collapse removal set are judged separately to obtain the prediction information based on the prediction information. In this embodiment, all nodes in the initial collapse removal set are judged based on the judgment conditions obtained above, and the nodes Reference probability The expression for other nodes in the time period The possibility of removing the solution is also universal, and the same calculation method can be used for different nodes to obtain the corresponding reference probability. Furthermore, if the reference probability is higher, the impact of removing the node on the network will be greater.

[0080] Furthermore, based on the reference probability Each node is judged separately, and the reference probability obtained is used as the removal probability to sort in descending order. The nodes with larger probabilities are the alternative solution sets for collapse removal based on the prediction information. Further, by comparing with the judgment threshold Compare, if the node removal probability (ie reference probability) is greater than , then the probability of the node being removed from the prediction information is high, and it can be used as a supplementary solution to the prediction information to formulate a collapse strategy to obtain the final solution based on the prediction trend information. The target collapse removal set at time t is expressed as:

[0081] in, Based on predictive information The disintegration of moments removes the set; Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

[0082] According to another embodiment of the present invention, for dynamically changing networks, reconnection and reconstruction may occur during movement and collapse, leading to recovery of network performance. Therefore, relative resilience can be introduced based on historical and predicted information to construct a collapse model based on relative resilience.

[0083] Based on this, in step S2, in the step of constructing a collapse model for collapsing the UAV cluster information interaction network, if the collapse model is a collapse model based on relative resilience, it can be expressed as:

[0084] in, Relative toughness based on The disintegration of moments removes the set; represents a collapse algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Indicates the step size of the information, which is used to reflect the span of the reference information (for example, if the information is historical information, it is represented as , if the information is predictive information, it is expressed as ), subscript Indicates the size of the prediction information span; Indicates the information span Sequential removal of the drone cluster information interaction network Removed node set The average relative toughness after Represents a collection-to-collection mapping.

[0085] In this embodiment, the step size of the information and dynamic frequency changes The size and relative size of the reference jointly determine the time accuracy of the reference. When the network changes rapidly, a shorter step size should be selected. , because the reference value of information from a distant time to the current moment is reduced; in addition, for the resilience of the drone cluster information interaction network, it is based on constructing its average resilience index to achieve the lowest average resilience while disintegrating and removing nodes to control the disintegration effect of the disintegration model based on relative resilience.

[0086] Furthermore, in step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain the target collapse removal set, if the collapse model is a collapse model based on relative toughness, the target collapse removal set is constructed based on the following steps, which include: Set the step size of the information used for reference and information span ; Use the collapse model based on historical information or the collapse model based on predicted information to screen the key nodes in the initial collapse removal set. The first target collapse removal set at time t is represented as:

[0087] in, Indicates that the first target collapses and removes the set, Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes; In this embodiment, the collapse model based on historical information or the collapse model based on predicted information has been described in the previous steps and will not be repeated here. The target collapse removal set obtained by the collapse model based on historical information is If the collapse model based on prediction information is used, the first goal is to collapse and remove the set The target collapse removal set can be obtained by the aforementioned collapse model based on prediction information Build; Based on the first goal, the information span is constructed by collapsing and removing the set About Moment Removed node set ,and ; Among them, remove the node set It can be expressed as:

[0088] A relative resilience model of nodes in the UAV cluster information interaction network after removal is constructed and expressed as:

[0089]

[0090]

[0091] in, Represents the set of removed nodes in the drone cluster information interaction network exist The relative toughness after the moment is removed, Indicates the nodes in the UAV cluster information interaction network The communication performance after the moment is removed, Represents the UAV cluster information interaction network at the reference time The communication performance size, It represents the communication performance of the UAV cluster information interaction network at the initial moment, represents the scaling factor, and , represents the rate parameter, Indicates the communication performance after the node is removed in the UAV cluster information interaction network. The derivative of time, Indicates relative time; In this implementation, the relative resilience model is used to reflect the instantaneous capability of the UAV cluster information interaction network at a certain moment, which is reflected as the relative ratio of the performance difference (current moment to reference moment). Its essence is an instantaneous snapshot of the system's dynamic response, rather than a time-accumulated effect. Figure 2 As shown, the basic relative toughness model without considering other parameters is expressed as:

[0092] in, Represents the set of removed nodes in the drone cluster information interaction network exist The communication performance after time removal is comparable to that in the UAV cluster information interaction network at the reference time. The difference in communication performance is It represents the communication performance of the UAV cluster information interaction network at the initial moment and the communication performance of the UAV cluster information interaction network at the reference moment. The difference in communication performance.

[0093] like Figure 3 As shown in the figure, the entire resilience process of the UAV cluster information interaction network can be described as: impact - performance degradation - recovery and adaptation - stable operation. Furthermore, the resilience process can be divided into four stages: impact stage, performance degradation stage, performance recovery stage, and performance stabilization stage. Furthermore, the entire resilience process has the following changes: Impact stage - performance degradation stage: In this process, the system performance gradually declines due to the impact. Resilience, as a property of the system, demonstrates its ability to resist and absorb impacts, that is, to resist changes caused by the impact.

[0094] Performance degradation phase - performance recovery phase: After the performance drops to the lowest value during this process, the system begins to gradually recover due to its resilience. In this phase, the system demonstrates its recovery capability, that is, it eliminates the changes caused by the shock.

[0095] Performance recovery phase - performance stabilization phase: During this process, the system experiences performance degradation caused by the shock, and the loss and recovery are balanced. In the recovery phase, when the recovery is greater than the loss, the system state will remain relatively stable at a performance level. In this phase, the system demonstrates its adaptability, that is, its ability to adapt to changes caused by the shock.

[0096] To this end, the reference moments for judging the performance change trend in the resilience process are selected based on the performance degradation stage and performance recovery stage in the resilience process. See also Figure 3 As shown in the figure, the switching between the performance degradation stage and the performance recovery stage shows the turning point of the system resilience change trend. Therefore, the reference time Select the intersection of the performance degradation phase and the performance recovery phase. Figure 4 , which can further facilitate and accurately analyze the performance change trend during the toughness process of performance.

[0097] At the reference time Previously, there were two types of analysis results, namely: If the performance change trend is upward, then at the reference moment Previously in the performance recovery phase, reference time The performance is the best, among which B = The value is negative, indicating that the reference time The closer, the better the performance; If the performance change trend is downward, then at the reference moment Previously, it was in the performance degradation stage, reference time The performance is the worst, among which B = The value is positive, indicating that The further away, the better the performance; At the reference time Afterwards, there are two types of analysis results, which are: If the performance change trend is upward, then at the reference moment After that, it is in the performance recovery phase, with reference to the time The performance is the worst, among whichB = The value is positive, indicating that The further away, the better the performance; If the performance change trend is downward, then at the reference moment After that, it is in the stage of performance degradation, reference time The performance is the best, among which B = The value is negative, indicating that the reference time The closer, the better the performance.

[0098] On this basis, Time and reference time The positive and negative changes in the relative time between the intervals can be used to functionally describe the performance changes in each interval, thereby modifying the basic relative toughness model and obtaining its functional expression:

[0099] ; On this basis, the nodes in the UAV cluster information interaction network are obtained. The rate of change of communication performance after time is removed is its rate parameter and is expressed as: ; Then, according to the positive and negative changes of the rate parameters in each interval and the positive and negative changes of the values ​​of the basic relative toughness model, the rate parameters can be introduced for functional description, thereby correcting the initial relative toughness model and obtaining its functional expression:

[0100]

[0101] .

[0102] Calculate the information span Remove each node set sequentially The average relative resilience after the information span is Can produce A relative toughness, thus, through The corresponding average relative toughness can be calculated by summing and averaging the relative toughnesses, and can be expressed as:

[0103] in, Indicates the information span Sequential removal in the information interaction network of UAV swarm Removed node set The average relative toughness after Represents the removal of each node set in the UAV cluster information interaction network Relative toughness after A condition for selecting and judging the first target collapse removal set is constructed, wherein the condition is constructed based on the lowest average resilience of the drone cluster information interaction network, and is used to select the set that minimizes the average relative resilience of the drone cluster in the first target collapse removal set to obtain the final target collapse removal set, and is expressed as:

[0104]

[0105] in, Indicates the removal of a node set After that, the average relative resilience of the network, Indicates the removal of a node set The lowest average relative resilience of the network afterwards; The set is removed for the final target collapse based on relative toughness; Therefore, based on the relative toughness The target collapse removal set at time t is expressed as:

[0106]

[0107] in, Relative toughness based on The disintegration of moments removes the set; Indicates the removal of a node set. Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

[0108] In order to further illustrate the effect of this solution, further examples are given to illustrate it.

[0109] Example Based on the aforementioned collapse model, a simulation analysis is performed on the dynamic collapse method of the drone cluster information interaction network. The drone cluster information interaction network is a dynamic network, and the parameter settings for the drone cluster information interaction network are shown in Table 1 below.

[0110] Table 1

[0111] The movement change rules of nodes in the UAV cluster information interaction network are as follows: The node moves on a dynamically changing circular orbit, with its speed and direction changing at every time point. This type of motion is common in vibration systems and dynamic equilibrium. Its path is a circular trajectory with a radius that changes with time. The radius change is related to time and frequency and is controlled by sine and cosine functions. The dynamic change frequency is . Among them, the dynamically changing radius is:

[0112] in, It's time The trajectory radius, is the frequency of change, the radius changes periodically, and its period is Frames. The radius varies between 0 and 50 over time.

[0113] Therefore, the position of a node in the network is calculated as follows:

[0114]

[0115] in, are the initial coordinates, and exist and The offset in the direction changes over time. The network edges are connected according to the connection model. Next, the performance and relative resilience of the dynamic network itself and different collapse methods are analyzed.

[0116] Initially formed drone swarm information interaction network The formation process of the initial UAV swarm information exchange network can be referenced in: A Simple Algorithm for Disintegrating Information Exchange Network of UAV Swarm, 2023 9th International Symposium on System Security, Safety, and Reliability (ISSSR), Louzhaohan Wang, Yun Huang, Haifeng Dai, Guanghan Bai, and Junyong Tao.

[0117] The formed UAV cluster information interaction network has a certain degree of self-adaptation and self-recovery capability due to the dynamic movement of its own nodes, dynamic reconnection of edges, and dynamic changes in network structure. The network structure at each moment of network movement without collapse is obtained through simulation, and its communication performance is calculated to obtain its performance curve as shown below. Figure 5 .

[0118] Depend on Figure 5 It can be seen that the communication performance of the network changes all the time, and the time period 2 to 7. The communication performance is low at time 10 and 13 to The performance is better at 15 seconds, which may be due to the adaptive characteristics of the network itself. At the same time, it is found that the performance fluctuates greatly at some time points, which may be related to the dynamic nature of the network. According to the basic relative resilience model and the relative resilience model based on relative time correction, the selection 5 as the reference moment, and calculate its basic relative toughness and relative toughness based on relative time to get Figure 6 shown.

[0119] Depend on Figure 6 It can be seen that at the time point with low performance, the relative resilience value of the network is indeed small or even negative, indicating that the resilience of the network at this moment is low. 10 to The network relative resilience value is high during the 15-hour period, indicating that the network has good resilience during this period.

[0120] a. The drone swarm information interaction network is disintegrated using traditional information-free disintegration methods According to the dynamic network disintegration algorithm (see: A Simple Algorithm for Disintegrating Information Exchange Network of UAV Swarm, 2023 9th International Symposium on System Security, Safety, and Reliability (ISSSR), Louzhaohan Wang, Yun Huang, Haifeng Dai, Guanghan Bai, and Junyong Tao), in By disintegrating the network at all times, we can obtain The sequence of nodes removed from the dynamic communication network is [13 34 47], and the network performance change curve is as follows: Figure 7 ,in, Cp 0 is the network performance change curve when the UAV cluster information interaction network does not collapse, Cp 1 is the UAV cluster information interaction network based on the dynamic network collapse algorithm Network performance change curve when the moment collapses: When choosing After carrying out disintegration strikes at all times, = = 6, 7, 8, 9, the performance has declined, but the trend of subsequent moments still has a certain recovery ability. According to the basic relative toughness model and the relative toughness model based on relative time correction, the basic relative toughness and the relative toughness based on relative time are calculated to obtain Figure 8 shown.

[0121] Depend on Figure 8 It can be seen that relative resilience changes at all times because the network structure changes at all times due to the dynamic movement of network nodes and the connection rules of the network. Therefore, the resilience capability as a network attribute changes at all times and is closely related to the choice of reference time. Time periods 7, 8, 9, and 10 all have high resilience values, indicating that the network was in a relatively good state of recovery during this time period (compared to time period 5). This also demonstrates that the impact of the information-free collapse strategy on the network is limited, and that the network can recover through its own adaptive and self-recovery capabilities. Therefore, it is essential to analyze the network in conjunction with relative resilience. By reducing performance resilience, the network's recovery and adaptability are continuously reduced until it loses its ability to self-recover, leading to complete collapse.

[0122] b. The UAV swarm information interaction network is collapsed using a collapse model based on historical information and predicted information. According to the above model, the collapse removal solution sets considering historical information and considering the forecast trend are:

[0123]

[0124] Based on the obtained collapse removal solution set, the UAV cluster information interaction network is collapsed and the performance curve of the collapse network based on historical information is obtained (see Figure 9 Network performance change curve Cp 2) and collapse network performance curves based on prediction information (see Figure 10 Network performance change curve Cp 3); Further, the network performance curves based on historical information and the network performance curves based on predicted information are summarized to obtain a performance curve comparison chart (see Figure 11 ).

[0125] Combine Figure 9 、 Figure 10 and Figure 11 It can be seen that the collapse model based on historical information is = 6, 7, 8, and 10, the network performance decreased significantly, but the network showed a certain recovery trend in the subsequent time periods. The prediction information method showed a significant decline in the subsequent time periods and no recovery trend was shown, which shows the effectiveness of this method.

[0126] This shows that both disruption models achieve improved disruption effectiveness compared to information-free disruption. This demonstrates that considering historical information and predicted trends is essential for addressing dynamic network disruption. Compared to attacks based solely on the network structure at a given moment, disruption strategies based on historical information and predicted trends can achieve longer-term impacts and maintain a sustained disruption effect. Therefore, dynamic models offer greater impact, effectively reducing disruption costs and increasing disruption effectiveness. Furthermore, considering predicted trends has a greater impact on the network's subsequent impact than models based on historical information.

[0127] c. The UAV swarm information interaction network disintegrates using a disintegration model based on relative resilience Based on the aforementioned collapse model based on relative toughness, The network at time t is collapsed. The collapse removal solution set is expressed as:

[0128] Based on the obtained collapse removal solution set, the UAV cluster information interaction network is collapsed and the network collapse performance curve of the collapse model based on relative resilience is obtained (see Figure 12 Network performance change curve Cp 4) and, based on the basic relative resilience model and the relative resilience model modified based on relative time, the basic relative resilience and the relative resilience based on relative time are calculated to obtain the network collapse relative resilience diagram of the collapse model based on relative resilience (see Figure 13 ).

[0129] Combine Figure 12 and Figure 13 As shown, when selecting =5 After the disintegration attack based on instantaneous toughness, it can be clearly seen that compared with the case without attack, the relative toughness at each moment decreases significantly, especially at the moment The performance and relative resilience values ​​for time periods 7, 8, 9, and 10 show a significant decrease, indicating that the network's recovery during this time period was average (compared to time 5). Furthermore, performance declined significantly at and after the collapse time, suggesting that the information-based collapse strategy is more effective and sustainable from a performance perspective. Furthermore, compared to collapse models based on historical information and predicted trends, this approach focuses not only on key network nodes but also on the network's overall resilience.

[0130] The above contents are merely examples of specific solutions of the present invention. For devices and structures not described in detail, it should be understood that they can be implemented by adopting general devices and methods available in the art.

Claims

1. A method for dynamically dismantling a UAV cluster information interaction network, characterized in that: The following steps are involved: S1. Input the network information of the drone cluster information interaction network to be collapsed; wherein, the network information is the node information of the drone cluster information interaction network with dynamic timing; S2. Constructing a disintegration model for disintegrating the drone swarm information interaction network; wherein the disintegration model is one of a disintegration model based on historical information, a disintegration model based on predictive information, and a disintegration model based on relative resilience; S3. Using a dynamic network disintegration algorithm to obtain key nodes from the network information for a predetermined period of time to construct an initial disintegration removal set, and screening the key nodes in the initial disintegration removal set based on the disintegration model to obtain a target disintegration removal set; wherein the dynamic network disintegration algorithm is a dynamic network disintegration algorithm based on closely connected groups; S4. Based on the target collapse removal set, remove the nodes in the drone cluster information interaction network to complete the collapse of the drone cluster information interaction network.

2. The method for dynamically dismantling a UAV cluster information interaction network according to claim 1, characterized in that: In step S2, in the step of constructing a collapse model for collapsing the drone cluster information interaction network, if the collapse model is a collapse model based on historical information, it can be expressed as: in, Represents historical information based on The disintegration of moments removes the set; represents a dynamic network disintegration algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Indicates the step size of historical information, which is used to reflect the span of the referenced historical information and is expressed as , subscript Indicates the span of historical information; Indicates the reference probability of historical information, which is used to reflect the reference degree of historical information. Represents a collection-to-collection mapping.

3. The method for dynamically dismantling a UAV cluster information interaction network according to claim 2, characterized in that: In step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on historical information, the target collapse removal set is expressed as: in, Based on historical information The disintegration of moments removes the set; Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

4. The method for dynamically dismantling a UAV cluster information interaction network according to claim 3, characterized in that: In step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on historical information, the collapse removal set is constructed based on the following steps, which include: Set the step size of historical information used for reference and historical information span ; Construct the nodes in the UAV cluster information interaction network The reference probability at each moment is based on historical information and is expressed as: in, Representation node In the span of historical information The reference probability under Representation node In the span of historical information The set of removed states under , and it is expressed as: in, Indicated by Time to Time Node Each removal status, Representation node exist Always remove the status flag, and ; Get the key nodes in the initial collapse removal set by filtering The judgment condition of whether the moment is the solution in the target collapse removal set is expressed as: in, express Time node In the span of historical information The judgment result under Representation node exist The target of the moment is to disintegrate and remove the collection. Representation node Not present The target of the moment collapses and is removed from the set; Indicates the judgment threshold; Based on the judgment condition, all nodes in the initial collapse removal set are judged respectively to obtain the The goal of the moment is to disintegrate the removal set.

5. The method for dynamically dismantling a UAV cluster information interaction network according to claim 1, characterized in that: In step S2, in the step of constructing a collapse model for collapsing the drone cluster information interaction network, if the collapse model is a collapse model based on prediction information, it is expressed as: in, Indicates that based on the prediction information The disintegration of moments removes the set; represents a collapse algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Represents the step size of the prediction information, which is used to reflect the span of the reference prediction information and is expressed as , subscript Indicates the size of the prediction information span; Indicates the reference probability of the forecast information, which is used to reflect the reference degree of the forecast information; Represents a collection-to-collection mapping.

6. The method for dynamically dismantling a UAV cluster information interaction network according to claim 5, characterized in that: In step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, the collapse model is a collapse model based on prediction information, and the target collapse removal set is expressed as: in, Based on predictive information The disintegration of moments removes the set; Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

7. The method for dynamically dismantling a UAV cluster information interaction network according to claim 6, characterized in that: In step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on prediction information, the target collapse removal set is constructed based on the following steps, which include: Set the step size of the prediction information used for reference and predicted information span ; Construct the nodes in the UAV cluster information interaction network The reference probability of the moment is based on the predicted information and is expressed as: in, Representation node In predicting information span The reference probability under Representation node In predicting information span The set of removed states under , and it is expressed as: in, Indicated by Time to Time Node Each removal status, Representation node exist Remove the status flag, and ; Get the key nodes in the initial collapse removal set by filtering The judgment condition of whether the target solution in the set is removed at the moment is expressed as: in, express Time node In predicting information span The judgment result under Representation node exist The target of the moment is to disintegrate and remove the collection. Representation node Not present The target of the moment collapses and is removed from the set; Indicates the judgment threshold; Based on the judgment condition, all nodes in the initial collapse removal set are judged respectively to obtain the prediction information based on the prediction information. The goal of the moment is to disintegrate the removal set.

8. The method for dynamically dismantling a UAV cluster information interaction network according to claim 1, characterized in that: In step S2, in the step of constructing a collapse model for collapsing the UAV cluster information interaction network, if the collapse model is a collapse model based on relative resilience, it is expressed as: in, Relative toughness based on The disintegration of moments removes the set; represents a collapse algorithm based on closely connected groups; Indicates the dynamic change frequency of the UAV cluster information interaction network, which is used to reflect the speed of the change of the UAV cluster information interaction network over time; Indicates the step size of the information, used to reflect the span of the reference information, subscript Indicates the size of the information span; Indicates the information span Sequential removal of the drone cluster information interaction network Removed node set The average relative toughness after Represents a collection-to-collection mapping.

9. The method for dynamically dismantling a UAV cluster information interaction network according to claim 8, characterized in that: In step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on a relative toughness collapse model, the collapse removal set is expressed as: in, Relative toughness based on The disintegration of moments removes the set; Indicates the removal of a node set. Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes.

10. The method for dynamically dismantling a UAV cluster information interaction network according to claim 9, characterized in that: In step S3, in the step of screening the key nodes in the initial collapse removal set based on the collapse model to obtain a target collapse removal set, if the collapse model is a collapse model based on relative toughness, the target collapse removal set is constructed based on the following steps, which include: Set the step size of the information used for reference and information span ; The key nodes in the initial collapse removal set are screened by using a collapse model based on historical information or a collapse model based on predicted information. The first target collapse removal set at time t; wherein the first target collapse removal set is expressed as: in, Indicates that the first target collapses and removes the set, Indicates the nodes removed from the UAV cluster information interaction network, with the superscript 、 、 Labels representing different nodes; Based on the first goal, the information span is constructed by collapsing and removing the set About Moment Removed node set ,and ; Build the removed node set in the UAV cluster information interaction network The relative toughness model after removal is expressed as: in, Represents the set of removed nodes in the drone cluster information interaction network exist The relative toughness after the moment is removed, Indicates the nodes in the UAV cluster information interaction network The communication performance after the moment is removed, Represents the UAV cluster information interaction network at the reference time The communication performance size, It represents the communication performance of the UAV cluster information interaction network at the initial moment, represents the scaling factor, and , represents the rate parameter, Indicates the communication performance after the node is removed in the UAV cluster information interaction network. The derivative of time, Indicates relative time; Calculate the information span Remove each node set sequentially The average relative toughness after , where the average relative toughness is expressed as: in, Indicates the information span Sequential removal of the drone cluster information interaction network Removed node set The average relative toughness after Represents the removal of each node set in the UAV cluster information interaction network Relative toughness after A condition for selecting and judging the first target collapse removal set is constructed, wherein the condition is constructed based on the lowest average resilience of the drone cluster information interaction network, and is used to select the set that minimizes the average relative resilience of the drone cluster from the first target collapse removal set to obtain the final target collapse removal set, and is expressed as: in, Indicates the removal of a node set After that, the average relative resilience of the network, Indicates the removal of a node set The lowest average relative resilience of the network afterwards; The set is removed for the final target collapse based on relative toughness.

Images