Cascade failure analysis method and device for unmanned platform cross-domain collaborative network

Abstract

The invention relates to a cascade failure analysis method and device for an unmanned platform cross-domain collaborative network. The method comprises the following steps: constructing an unmanned platform cross-domain collaborative network; the unmanned platform cross-domain collaborative network comprises a guarantee domain and at least two operation domains, the number of maintenance chains and the number of material supply chains of guaranteed nodes in the operation domains are obtained, and if the number of the maintenance chains or the number of the material supply chains is zero, the current guaranteed node is in a guarantee chain missing failure mode; the in-degree number and the out-degree number of other nodes except the target node in the job domain and the guarantee domain are obtained, and if the in-degree number or the out-degree number is zero, the current node is in a task failure mode; and obtaining damaged nodes in a failure mode in the unmanned platform cross-domain collaborative network, and carrying out load redistribution on the unmanned platform cross-domain collaborative network after the failure nodes are removed. The method can be used for analyzing the cascade failure of the cross-domain collaborative system and improving the performance of the cross-domain collaborative system.

Description

Technical Field

[0001] The present application relates to the field of cross-domain collaborative technology, and in particular to a cascading failure analysis method and device for a cross-domain collaborative network of an unmanned platform. Background Art

[0002] With the development of unmanned platform technology in recent years, the concept of cross-domain collaborative operation has been proposed. In the air-ground unmanned platform collaborative system, the advantages of air unmanned platforms are fast movement speed, wide field of view, and little influence by terrain, but their load capacity is limited, the types of mission loads, performance and functions are relatively simple, and they are easily affected by weather. The advantages of ground unmanned platforms are large load capacity, strong endurance and concealment and penetration capabilities, but their movement speed is slow, the field of view is small, and communication transmission is easily affected by terrain. Both have certain advantages and limitations, and the application of air-ground collaboration is particularly important. Equipment systems in a single physical domain, such as drones and unmanned vehicles, work together to complete tasks, which is called cross-domain collaboration. Considering the close dynamics and high uncertainty of modern operating environments, failure modeling and analysis are crucial to the reliable and safe operation of the cross-domain collaborative system of unmanned platforms, which ultimately affects the success of the mission. Complex network methods play an important role in the modeling of the cross-domain collaborative system of unmanned platforms, and can closely characterize the cross-domain collaborative system architecture. Based on this, further modeling of the cascading failure process of the cross-domain collaborative system can help analyze and evaluate the resilience and anti-destruction capabilities of the system in complex environments.

[0003] The cross-domain collaborative system includes interconnected equipment and command communication paths. Failures triggered by a single event in one domain, such as communication failure or equipment failure, can propagate across domains through interconnection, leading to cascading failures. Failure modeling and analysis are crucial to the reliable and safe operation of the cross-domain collaborative system, ultimately affecting the success of the mission. The research on cascading failure models mainly focuses on two aspects, the impact of information interference strategies and load redistribution strategies, which are mutually conditional. The identification of key equipment and relationship links plays a great role in reflecting the dynamics and high uncertainty in modern operating environments to further analyze cascading failures. Different information interference strategies determine the failure of corresponding equipment. The existing operating system mostly stays at this stage for failure modeling and analysis to analyze characteristics such as anti-destruction, and rarely considers cascading failures. Network cascading failure research mostly focuses on homogeneous or single-domain systems, and cross-domain collaboration is still under exploration. Therefore, it is meaningful to propose an applicable load redistribution strategy for cascading failure analysis of cross-domain collaborative systems. Summary of the invention

[0004] Based on this, it is necessary to provide a cascading failure analysis method and device for cross-domain collaborative networks of unmanned platforms to address the above technical problems.

[0005] A cascading failure analysis method for an unmanned platform cross-domain collaborative network, the method comprising:

[0006] Constructing a cross-domain collaborative network for unmanned platforms; the cross-domain collaborative network for unmanned platforms includes a support domain and at least two operation domains, the operation domains are connected by inter-layer collaborative edges, and the operation domains and support domains are connected by inter-layer support edges; the support domain provides logistical support for multiple operation domains;

[0007] Obtain the number of maintenance chains and material supply chains of the supported node in the operation domain. If the number of maintenance chains or material supply chains is zero, the current supported node is in the support chain missing failure mode.

[0008] Get the number of in-degrees and out-degrees of other nodes in the job domain and guarantee domain except the target node. If the number of in-degrees or out-degrees is zero, the current node is in task failure mode.

[0009] Obtain failed nodes that are damaged by interference in the unmanned platform cross-domain collaborative network that takes into account the failure mode of logistics support, and redistribute the load of the unmanned platform cross-domain collaborative network after removing the failed nodes to analyze cascading failures; the failure modes that take into account logistics support include support chain missing failure mode and mission failure mode.

[0010] A cascading failure analysis device for an unmanned platform cross-domain collaborative network, the device comprising:

[0011] A network construction module, used to construct a cross-domain collaborative network for unmanned platforms; the cross-domain collaborative network for unmanned platforms includes a support domain and at least two operation domains, the operation domains are connected by inter-layer collaborative edges, and the operation domains are connected to the support domains by inter-layer support edges; the support domain provides logistical support for multiple operation domains;

[0012] The failure analysis module is used to obtain the number of maintenance chains and material supply chains of the supported node in the operation domain. If the number of maintenance chains or material supply chains is zero, the current supported node is in a support chain missing failure mode.

[0013] The failure analysis module is also used to obtain the number of in-degrees and out-degrees of other nodes in the operation domain and the guarantee domain except the target node. If the number of in-degrees or out-degrees is zero, the current node is in task failure mode;

[0014] The load redistribution module is used to obtain failed nodes in the unmanned platform cross-domain collaborative network that are damaged by interference, and to redistribute the load of the unmanned platform cross-domain collaborative network after removing the failed nodes to analyze cascading failures; the failure modes that consider logistics support include support chain missing failure modes and mission failure modes.

[0015] The above-mentioned cascading failure analysis method and device for the cross-domain collaborative network of unmanned platforms can be closer to the actual situation and more widely used by constructing a cross-domain collaborative network of unmanned platforms that considers cross-domain collaboration and logistical support at the same time. The number of maintenance chains and the number of material supply chains of the supported nodes in the operation domain are obtained. If the number of maintenance chains or the number of material supply chains is zero, the current supported node is in the guarantee chain missing failure mode. The number of in-degrees and out-degrees of other nodes in the operation domain and the guarantee domain except the target node are obtained. If the number of in-degrees or the number of out-degrees is zero, the current node is in the task failure mode; the damaged nodes in the failure mode in the cross-domain collaborative network of unmanned platforms are obtained, and the load of the cross-domain collaborative network of unmanned platforms after removing the failed nodes is redistributed. The embodiments of the present invention can be used to analyze the cascading failures of the cross-domain collaborative system and improve the performance of the cross-domain collaborative system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flowchart of a cascading failure analysis method for a cross-domain collaborative network of unmanned platforms in one embodiment;

[0017] Figure 2 A schematic diagram of a job chain of job domain 1 in an embodiment;

[0018] Figure 3 A schematic diagram of a guarantee chain of a guarantee domain in an embodiment;

[0019] Figure 4 A schematic diagram of cross-domain collaboration in an embodiment;

[0020] Figure 5 This is a schematic diagram of a cross-domain collaborative operation chain of operation domain 1 in an embodiment;

[0021] Figure 6 A schematic diagram of the structure of a cross-domain collaborative system network in one embodiment;

[0022] Figure 7 A schematic diagram of a situation where a missing guarantee chain fails in one embodiment;

[0023] Figure 8 A schematic diagram of task failure in an embodiment;

[0024] Fig. 9 A schematic diagram of a flow chart of load redistribution in one embodiment;

[0025] Fig.10 It is a schematic diagram of the change of the number of kill chains and load entropy with the proportion of failed units in a random interference mode in one embodiment;

[0026] Fig.11 It is a structural block diagram of a cascading failure analysis device for an unmanned platform cross-domain collaborative network in one embodiment. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] In one embodiment, Figure 1 As shown, a cascading failure analysis method for an unmanned platform cross-domain collaborative network is provided, comprising the following steps:

[0029] Step 102: Build a cross-domain collaborative network for unmanned platforms.

[0030] The cross-domain collaborative network of unmanned platforms includes a support domain and at least two operation domains. The operation domains are connected through inter-layer collaborative edges, and the operation domains and support domains are connected through inter-layer support edges. The support domain provides logistical support for multiple operation domains.

[0031] In the implementation of the present invention, multiple operation space domains are simplified to consider the relationship between two layers of networks to study the relationship between cross-domain space domains. Therefore, a multi-layer network including two layers of operation domains is given, the operation units are regarded as nodes, and the collaborative relationship between equipment is regarded as the edge between nodes. Logistics support is also included as a layer of network and will be further explained.

[0032] The operation domains in the multi-layer network are set as operation domain 1 and operation domain 2. The operation units in the operation layer represent the weapons and equipment belonging to the corresponding space domain. According to the theory related to the OODA combat loop, they can be divided into reconnaissance units, decision-making units, and strike units, and the operation interference target is defined as the target unit. Operation domain 1 is represented as According to the different types of tasks completed by the work units in the network, the work units can be divided into different sets, each set is responsible for different tasks, and the work domain 1 also becomes a heterogeneous network. They are reconnaissance unit I, decision-making unit I, strike unit I and target unit I. Total number of operation units Job domain 2 is represented as Operation unit type They are reconnaissance unit II, decision-making unit II, strike unit II, target unit II, and the total number of operation units

[0033] If a unit in each layer does not have the ability to complete a task, it needs to make a request to other units that have the ability. In this process, the units share capabilities and results. This kind of cooperation belongs to result sharing. If a unit lacks information to complete a task, it makes a sharing request to the unit that has the information. This kind of sharing is called information sharing. It is precisely because in the established network, each unit has only the ability to complete its corresponding task, so several types of units complete the task together. This kind of cooperation is called task sharing, thus forming different types of task chains.

[0034] Therefore, according to the result sharing and information sharing between the work units, the sharing set between the work units in work domain 1 can be expressed as Its quantity is expressed as The shared set between the job units of job domain 2 is expressed as Its quantity is expressed as

[0035] For task sharing, the task chain formed by the operation domain is similar to the function of the kill chain in the actual unmanned platform operation. The operation chain is defined as a series of actions taken by non-friendly targets to interfere with the operation unit through a series of activities such as "observation-identification-tracking-operation-evaluation". There will be multiple operation chains with different functions according to the difference in the meaning of unit sharing. Taking operation domain 1 as an example, Figure 2 shown.

[0036] From the perspective of the physical domain, most logistical support must be completed on the ground. At the same time, the division of the domain also includes capability dimensions such as situational awareness and integrated support. Therefore, it is more appropriate to consider logistical support as a capability or business and build it into a layer of network. The tasks of the support domain are assumed to be equipment maintenance and material supply. At the same time, equipment and materials need to be transported to the operating unit through the physical equipment and material transportation. This type of equipment is represented by equipment transportation and material transportation units. In combination with the command and control mechanism in cross-domain operations, a command and control center that can jointly command these corresponding units needs to be added, represented as a joint command unit. Therefore, the support domain can be represented as Supply unit type They are joint command unit, equipment maintenance unit, material storage unit, equipment transportation unit and material transportation unit, and the total number of support units

[0037] According to the result sharing and information sharing between assurance units, the shared set between assurance units in the assurance domain is expressed as Its quantity is expressed as Job Domain and security domain The dependency sharing between Its quantity is expressed as The protected units i=1,2,…,N F ,j=1,2,…,N C ,N W oeLh V . Job Domain and security domain The inter-dependency sharing is Its quantity is expressed as in i=1,2,…,N F ,j=1,2,…,N C ,N W or N V .

[0038] After the operation, the target and the own units will be assessed for damage. In addition to the damage and failure, the reconnaissance unit, decision-making unit and strike unit in the operation domain will become the supported unit due to the need for maintenance and material supply. Therefore, the supported unit will apply for maintenance and material support from the support domain. The completion of this task requires that various types of units in the support domain be shared in a certain order, thus forming a corresponding support chain. Depending on the difference in the meaning of sharing of each unit, a support chain with different functions will be formed, such as Figure 3 shown.

[0039] The equipment in Operation Domain 1 and Operation Domain 2 can switch between supporting and being supported roles in different situations. Through communication and coordination, the operation units of the two layers of the network can work together by sharing. In addition, the difference in the settings of the inter-layer relationship between Operation Domain 1 and Operation Domain 2 is that when an operation unit in one layer fails, the priority of the reconnaissance unit in the other layer receiving the information of its target unit for coordinated interference is higher than the priority of the same type of node in the same layer of the failed unit to replace the failed unit to undertake its network task for interference after task redistribution, thereby achieving the purpose of reorganizing the operation chain. The cross-domain collaboration diagram is shown in the figure below. Figure 4 shown.

[0040] When the operating units of operating domain 1 cooperate with the operating units of operating domain 2 to interfere with their target units, after receiving the information of the target units of domain 2, the units of domain 1 use the operating chain connection method of domain 1, and finally the attack units of domain 1 interfere with these target units; similarly, when the operating units of operating domain 2 cooperate with operating domain 1 to interfere with their target units, after receiving the information of the target units of domain 1, the units of domain 2 use the operating chain connection method of domain 2, and finally the attack units of domain 2 interfere with these target units. Through cross-domain collaborative tasks, a new operating chain will be formed between operating domain 1 and operating domain 2. Taking operating domain 1 as an example, the cross-domain collaborative operating chain is as follows: Figure 5 shown.

[0041] Based on the above analysis, the structural diagram of the established cross-domain collaborative system network is as follows: Figure 6 shown.

[0042] Step 104, obtaining the number of maintenance chains and the number of material supply chains of the supported node in the operation domain. If the number of maintenance chains or the number of material supply chains is zero, the current supported node is in a support chain missing failure mode.

[0043] Step 106, obtaining the in-degree number and out-degree number of other nodes in the operation domain and the guarantee domain except the target node. If the in-degree number or the out-degree number is zero, the current node is in task failure mode.

[0044] Step 108, obtaining failed nodes that are damaged by interference in the unmanned platform cross-domain collaborative network that takes into account the failure mode of logistics support, and performing load redistribution on the unmanned platform cross-domain collaborative network after removing the failed nodes to analyze cascading failures.

[0045] The failure modes of logistics support considered include support chain loss failure mode and mission failure mode.

[0046] In the above-mentioned cascading failure analysis method for the cross-domain collaborative network of unmanned platforms, by constructing a cross-domain collaborative network of unmanned platforms that considers both cross-domain collaboration and logistical support, it can be closer to the actual situation and more widely used. The number of maintenance chains and the number of material supply chains of the supported nodes in the operation domain are obtained. If the number of maintenance chains or the number of material supply chains is zero, the current supported node is in the guarantee chain missing failure mode. The number of in-degrees and out-degrees of other nodes in the operation domain and the guarantee domain except the target node are obtained. If the number of in-degrees or the number of out-degrees is zero, the current node is in the task failure mode; the damaged nodes in the failure mode in the cross-domain collaborative network of unmanned platforms are obtained, and the load of the cross-domain collaborative network of unmanned platforms after removing the failed nodes is redistributed. The embodiment of the present invention can be used to analyze the cascading failure of the cross-domain collaborative system and improve the performance of the cross-domain collaborative system.

[0047] Referring to the definition of kill chains and the calculation method of their number, the number of operation chains in operation domain 1 and operation domain 2 is calculated to obtain the number of kill chains as a performance measurement indicator. and reconnaissance units For example, the adjacency matrix of two types of cells Represents each unit and If Can reach Then the element a in the adjacency matrix ij =1, otherwise a ij = 0. There is also an adjacency matrix between units of the same type. Taking the classic operation chain of operation domain 1 as an example, we can get the target unit Arrival matrix under the classic job chain type Calculation formula:

[0048]

[0049] According to the physical meaning of the arrival matrix, the number of job chains of this type in job domain 1 is:

[0050]

[0051] in is the number of target units in job domain 1.

[0052] There are 7 types of job chains for job domain 1 to its target unit, so the calculation formula for the number of all types of job chains contained in job domain 1 is:

[0053]

[0054] There are also 7 types of collaborative operation chains between operation domain 1 and domain 2 target units. Therefore, the calculation formula for the number of collaborative operation chains of all types contained in operation domain 1 is:

[0055]

[0056] The sum of the number of job chains and collaborative job chains in job domain 1 is taken as the number of kill chains in job domain 1:

[0057]

[0058] The same method can be used to calculate the intra-layer operation chain of operation domain 2 Collaborative operation chain And the number of kill chains

[0059] Therefore, the number of kill chains for the entire network is calculated as the sum of the number of kill chains in Operational Domain 1 and the number of kill chains in Operational Domain 2:

[0060]

[0061] In one embodiment, the method also includes: obtaining the impact efficiency when the operation node fails, and obtaining the intentional interference importance of each operation node in the operation domain according to the information interference importance, security interference importance and impact efficiency of each operation node in the operation domain; obtaining the impact efficiency when the security node fails, and obtaining the intentional interference importance of each security node in the security domain according to the information interference importance and impact efficiency of each security node in the security domain; obtaining the key nodes in the unmanned platform cross-domain collaborative network according to the intentional interference importance of each node in the operation domain and the security domain.

[0062] Specifically, for joint command units, equipment maintenance units, material storage units, equipment transportation units, and material transportation units, there is an information transmission relationship between these units, and each unit needs to receive and send information. The information to be transmitted between different support units is different, and the importance is also different. Different weights are given to the number of different types of sharing to define the importance of information interference of the unit. In the support domain, Indicates that the protected unit L is transferred to unit H i The number of shared links for which support is requested; Representation unit H i Issue supply command information sharing numbers to other units; Representation unit H i Provide equipment maintenance information sharing to other units; Representation unit H i Provide material supply information sharing number to other units; Representation unit H i The number of information sharing on transporting equipment to supported units; Representation unit H i The number of shared information on transporting materials to the supported units. 1 ,α 2 ,α 3 is the importance weight of different types of information flows, let α 1 >α 2 >α 3 , that is, the importance of each type of information flow is ranked as follows: application support information = joint command information > equipment maintenance information = material supply information > equipment / material transportation information. Since the value of the unit information interference importance is between 0 and 1, in order to normalize the information interference importance, the denominator introduces the unit H i Degree The importance of unit information interference in the security domain for:

[0063]

[0064] During the operation, the network unit H i After the interference fails, the number of guarantee chains will decrease. The ratio of the difference between the number of guarantee chains before the interference and the number of guarantee chains after the interference to the number of guarantee chains before the interference is used to represent the unit H. i Impact efficiency when failure occurs

[0065]

[0066] The importance of information interference of the unit is assigned an information interference importance weight θ, and the product of the sum of the importance after the weight coefficient is assigned and the change in the impact efficiency is used as the intentional interference importance index of the security unit.

[0067]

[0068] Following the calculation method of the importance of the above support unit, there is also information and material exchange between each unit. The importance of each type of information flow is ranked as follows: load allocation application information = operation decision information = operation interference information = collaborative operation interference information > equipment transportation information = material transportation information > operation reconnaissance information = target locking information = collaborative target locking information. The information interference importance of the operation unit in operation domain 1 is:

[0069]

[0070] The information interference importance of the operation unit in operation domain 2 is:

[0071]

[0072] The supported units of the two layers of the network need to receive the supported loads transported by the equipment or material transport units, and the equipment transport units share the transported equipment information with the supported units. Assign maintenance information a weight δ, and (1-δ) is the number of shared information on materials transported by the material transport unit to the supported unit. The sum of the two is normalized by the unit degree to express the importance of the support interference of the operation unit:

[0073]

[0074] In the process of calculating the importance of intentional interference of the units in these two layers of the network, the importance of information interference must be considered at the same time to ensure the importance of interference and the impact on efficiency. The importance index of intentional interference of the operating unit is:

[0075]

[0076]

[0077] In one embodiment, for a unit, the guarantee chain loss failure mode is:

[0078]

[0079]

[0080] Where N LCMWL and N LCCMWL is the number of maintenance chains; N LCUVL and N LCCUVL is the number of supply chains. In this embodiment, the guaranteed unit Number of repair task chains Or the number of material supply task chains If the maintenance chain or supply chain is 0, the corresponding operation unit fails. Figure 7 shown.

[0081] In one embodiment, for a unit, the task failure mode is:

[0082] k in (H j )*k out (H j )=0 (17)

[0083]

[0084]

[0085] Among them, H i is the i-th node, H j is the jth node, H m is the mth node, k in (H j ) represents the number of in-degrees of the node, k out (H j ) represents the number of out-degrees of the node, N H is the number of corresponding nodes. In this embodiment, except for the specific protected unit, which will fail due to loss of protection, all unit types except the target unit If the number of in-degrees or out-degrees of a unit is equal to 0, the corresponding unit fails, which is called task failure. Figure 8 Schematic diagram of task failure.

[0086] In one embodiment, the flow chart of load redistribution is as follows: Fig. 9As shown, the load redistribution of the unmanned platform cross-domain collaborative network after removing the failed nodes includes: obtaining the failed nodes after being interfered and damaged; if the current failed node is a protection node in the protection domain, the load of the current failed node is redistributed according to the preset capability load redistribution strategy; if the current failed node is an operation node in the operation domain, the load of the current failed node is redistributed according to the preset cross-domain collaborative priority redistribution strategy; and updating the load of the non-failed nodes.

[0087] The load capacity model of the cross-domain collaborative system network is constructed with reference to the load-based ML model. Taking into account the heterogeneity of network units and the different tasks and functions performed by the network, it is also necessary to use the proposed importance index to calculate the corresponding load and capacity of the unit. This model can also be used to analyze the failure process of the network.

[0088] The operation unit and the support unit have specific functions and carry certain tasks, such as confrontation tasks and support tasks. The initial load is defined and the capacity is calculated based on the initial load.

[0089] For the operation unit, the initial load is calculated by combining the unit degree, unit betweenness, and the amount of information and materials received. The sum of the internal and external unit degrees and unit betweenness is assigned to the network topology load weight ε, and the product of the sum of the unit information interference importance and the security interference importance and the unit degree is assigned to the network task load weight (1-ε). The sum of these two parts is used as the unit initial load.

[0090]

[0091]

[0092] Which sets Since the unit externality represents the relationship between the unit and other layers, and involves tasks such as information transmission, material transportation, and supply, the unit externality should be more important than the internality, and the load it carries is greater. The larger the unit betweenness, the heavier the task the unit bears, and the greater the load. The second half of the formula considers the unit function and task factors to measure the load, including information interference importance and security interference importance, and introduces the unit degree Reduce the importance of each unit to the above three quantities.

[0093] The initial load of the security unit is:

[0094]

[0095]

[0096] In one embodiment, load redistribution of the current failed node according to a preset capacity load redistribution strategy includes: calculating the initial load of each security node in the security domain according to the unit degree, unit betweenness and unit information interference importance of each security node in the security domain; calculating the unit capacity of each security node according to the initial load of each security node, and obtaining the real-time processing capability index of each security node according to the unit capacity and the initial load; and load distribution to neighboring nodes of the same type according to the ratio of the real-time processing capability index of neighboring nodes of the current failed node to the real-time processing capability index of all neighboring units of the same type.

[0097] According to the load capacity model, the specific unit capacity is set in direct proportion to the unit initial load. for:

[0098]

[0099] Among them, β is a tolerance parameter, which mainly represents the unit H i The greater the tolerance to load, the larger the unit capacity is and the stronger its ability to carry load is.

[0100] Introduction Representation unit H i The real-time processing capability indicators are:

[0101]

[0102] The load redistribution strategy is defined as distributing the load of the failed unit after being damaged by non-self-interference according to the ratio of the current real-time processing capability index value of the neighboring units of the same type to the current real-time processing capability index value of all neighboring units of the same type. This is called the capacity load redistribution strategy. According to this strategy, the failed unit H i Load Distribute to neighboring units of the same type in a certain proportion. The load distribution formula is:

[0103]

[0104] in, For unit H i After failure, the neighbor unit H j The allocated load; ρ represents the consistency coefficient of the unit task type. Since the cross-domain collaborative system network involves different unit types, only units of the same type can bear the load allocated by the failed unit. When ρ = 1, it means that the failed unit H i The unit H connected to it j are the same type of units, ρ=0 means they are not the same type of units; φ i Refers to the failure unit H iThe set of directly connected neighbor cells of all types.

[0105] In one embodiment, load redistribution of the current failed node according to a preset cross-domain collaborative priority redistribution strategy includes: calculating the initial load of each job node in the job domain according to the unit degree, unit betweenness, unit information interference importance and guarantee interference importance of each job node in the job domain; calculating the priority of the same type of neighboring nodes of the current failed node in each job domain according to a preset cross-domain collaborative priority judgment method; and distributing the initial load of the current failed node to the same type of neighboring nodes according to the priority of the same type of neighboring nodes.

[0106] According to the cross-domain collaborative operation mode, load distribution can also be performed between units of the same task type between operation domain 1 and operation domain 2. Compared with the load distribution between units on the same layer, load distribution needs to be performed according to a certain priority. Taking operation domain 2 as an example, for the failed unit The load transfer unit priority determination method is as follows:

[0107]

[0108]

[0109] Among them, y is the priority, χ i is the failed unit H i The set of all task type neighbor units directly connected in the two-layer operation domain, r (r∈[0,1]) is the priority coefficient, when r<0.5, the unit load requirement has a great impact on the system, and the load requirement is prioritized; on the contrary, the cross-domain coordination requirement is prioritized. The cross-domain coordination requirement part is represented by the ratio of the number of units in domain 1 to the sum of the number of remaining units in domain 1 and domain 2, and the numerator is given an appropriately large proportion x to reflect the coordination priority, and the load requirement part is represented by the ratio of the real-time load processing capability of the computing unit to the sum of the load processing capabilities of the two-layer units.

[0110] When y is the same, the cross-domain collaboration priority is high when r>0.5, and vice versa when the load is relatively large. The corresponding layer gives priority to the collaboration of the task, so r>0.5 is set. The load is distributed according to the ratio of the priority of the same type of units in the operation domain 1 and the operation domain 2 to all the same type of units in the two layers, which is called the cross-domain collaboration priority redistribution strategy. According to the above priority judgment method, the interference failure unit H in the operation domain 1 and the operation domain 2 i Load The load will be distributed to neighboring units of the same type in a certain proportion. The load distribution formula is:

[0111]

[0112] Unit Hj The load changes to:

[0113]

[0114] According to the above failed unit load redistribution strategy, the loads of all non-failed units in the network will be updated.

[0115] If the failed unit H i The same type of neighbor unit H j Share the additional load increase After that, meet:

[0116]

[0117] Then unit H j It also immediately became invalid.

[0118] In one embodiment, the method further includes: evaluating the performance of the unmanned platform cross-domain collaborative network according to the number of damaged nodes.

[0119] In one embodiment, the network structure entropy can describe the heterogeneity of the interactive operation system. When the cross-domain collaborative network is orderly, the degree distribution of the units tends to be the same and the structure is stable; when the cross-domain collaborative network is disordered, the degree distribution of the network units is inconsistent and the structure is unstable. The network structure entropy is a description of the order or disorder of the cross-domain collaborative network. When the network changes from an ordered state to a disordered state, the network structure entropy increases, and vice versa. The calculation formula for the cross-domain collaborative network structure entropy is:

[0120]

[0121] In the formula Network unit H i The structural importance of The calculation method is as follows:

[0122]

[0123] Referring to the calculation method of network structure entropy, load entropy is an indicator to measure the degree of load balance of units in the network. Its calculation formula is:

[0124]

[0125] in, For unit H i The ratio of the load at the moment to the total network load is:

[0126]

[0127] In order to study the changes of the two indicators of kill chain number and load entropy as the air-ground collaborative network is blocked and fails and causes cascade failure, and reflect its reflection effect on network performance, the unit interference attack method selects random attack, and obtains the changes of the two indicators of kill chain number and load entropy with the proportion of failed units under random attack mode. Fig.10 As shown. Fig.10 It can be seen that, in general, the number of kill chains decreases as the proportion of failed units gradually increases during the process of unit failure and cascading failure in the air-ground collaborative operation network, and becomes 0 at a certain failure ratio, while the load entropy continues to increase as the proportion of failed units increases. From a local observation, when the proportion of failed units begins to increase, the ratio of the number of kill chains to the initial value begins to decrease from 1, and becomes 0 when the proportion of failed nodes reaches 0.86, which means that the number of kill chains becomes 0 at this proportion of failed units, and the network has lost its corresponding mission capability and tends to collapse. The load entropy ratio increases with the increase of the proportion of failed units and can reach 13.57. Under the game between network unit load and capacity, the load is redistributed to make up for the load capacity of the failed unit, resulting in uneven load distribution between units and disorder. However, under the continuous attack, even if the network task performance collapses, the load distribution process between units is still carried out, and the load entropy continues to increase accordingly, which can reflect that the kill chain number indicator reflects the network task performance better than the load entropy. Compared with it, the load entropy can better reflect the game process between load and capacity in the cascading failure process, as well as the process of other units bearing their capacity load after the unit fails. The changes in these two can reflect the anti-destruction ability of the network to maintain a certain operating capacity without collapse after being attacked.

[0128] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0129] In one embodiment, Fig.11 As shown, a cascading failure analysis device for an unmanned platform cross-domain collaborative network is provided, comprising:

[0130] The network construction module 1102 is used to construct a cross-domain collaborative network for unmanned platforms; the cross-domain collaborative network for unmanned platforms includes a support domain and at least two operation domains, the operation domains are connected through inter-layer collaborative edges, and the operation domains and the support domains are connected through inter-layer support edges; the support domain provides logistical support for multiple operation domains;

[0131] The failure analysis module 1104 is used to obtain the number of maintenance chains and the number of material supply chains of the supported node in the operation domain. If the number of maintenance chains or the number of material supply chains is zero, the current supported node is in a support chain missing failure mode;

[0132] The failure analysis module 1106 is further used to obtain the number of in-degrees and out-degrees of other nodes in the operation domain and the guarantee domain except the target node. If the number of in-degrees or the number of out-degrees is zero, the current node is in task failure mode;

[0133] The load redistribution module 1108 is used to obtain failed nodes that have been damaged by interference in the unmanned platform cross-domain collaborative network that takes into account the failure mode of logistics support, and to redistribute the load of the unmanned platform cross-domain collaborative network after removing the failed nodes to analyze cascading failures; the failure modes considered for logistics support include the failure mode of missing support chain and the failure mode of mission.

[0134] In one of the embodiments, the task failure mode is:

[0135] k in (H j )*k out (H j )=0

[0136]

[0137]

[0138] Among them, H i is the i-th node, H j is the jth node, H m is the mth node, k in (H j ) represents the number of in-degrees of the node, k out (H j ) represents the number of out-degrees of the node, N H is the number of corresponding nodes.

[0139] In one embodiment, the failure mode of the guarantee chain loss is:

[0140]

[0141]

[0142] Among them, N LCMWL and N LCCMWL is the number of maintenance chains; N LCUVL and N LCCUVL is the number of supply chains.

[0143] In one of the embodiments, it is also used to obtain a failed node after being damaged by interference; if the current failed node is a protection node in the protection domain, the load of the current failed node is redistributed according to a preset capacity load redistribution strategy; if the current failed node is an operating node in the operating domain, the load of the current failed node is redistributed according to a preset cross-domain collaborative priority redistribution strategy; and the load of the non-failed nodes is updated.

[0144] In one of the embodiments, it is also used to calculate the initial load of each security node in the security domain according to the unit degree, unit betweenness and unit information interference importance of each security node in the security domain; calculate the unit capacity of each security node according to the initial load of each security node, and obtain the real-time processing capability index of each security node according to the unit capacity and the initial load; and distribute the load to the neighboring nodes of the same type according to the ratio of the real-time processing capability index of the neighboring nodes of the current failed node to the real-time processing capability index of all neighboring units of the same type.

[0145] In one of the embodiments, it is also used to calculate the initial load of each work node in the work domain based on the unit degree, unit betweenness, unit information interference importance and guarantee interference importance of each work node in the work domain; calculate the priority of the same type of neighboring nodes of the current failed node in each work domain according to a pre-set cross-domain collaborative priority judgment method; and distribute the initial load of the current failed node to the same type of neighboring nodes according to the priority of the same type of neighboring nodes.

[0146] In one of the embodiments, it is also used to obtain the impact efficiency when the operation node fails, and obtain the intentional interference importance of each operation node in the operation domain according to the information interference importance, security interference importance and impact efficiency of each operation node in the operation domain; obtain the impact efficiency when the security node fails, and obtain the intentional interference importance of each security node in the security domain according to the information interference importance and impact efficiency of each security node in the security domain; obtain the key nodes in the unmanned platform cross-domain collaborative network according to the intentional interference importance of each node in the operation domain and the security domain.

[0147] In one of the embodiments, it is also used to evaluate the performance of the cross-domain collaborative network of the unmanned platform based on the number of damaged nodes.

[0148] For the specific limitations of the cascading failure analysis device for the cross-domain collaborative network of unmanned platforms, please refer to the limitations of the cascading failure analysis method for the cross-domain collaborative network of unmanned platforms mentioned above, which will not be repeated here. Each module in the above-mentioned cascading failure analysis device for the cross-domain collaborative network of unmanned platforms can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A cascading failure analysis method for unmanned platform cross-domain collaborative network, characterized in that: The method comprises: Constructing a cross-domain collaborative network for unmanned platforms; the cross-domain collaborative network for unmanned platforms includes a support domain and at least two operation domains, the operation domains are connected by inter-layer collaborative edges, and the operation domains and support domains are connected by inter-layer support edges; the support domain provides logistical support for multiple operation domains; Obtain the number of maintenance chains and material supply chains of the supported node in the operation domain. If the number of maintenance chains or material supply chains is zero, the current supported node is in the support chain missing failure mode. Get the number of in-degrees and out-degrees of other nodes in the job domain and guarantee domain except the target node. If the number of in-degrees or out-degrees is zero, the current node is in task failure mode. Obtain failed nodes that are damaged by interference in the unmanned platform cross-domain collaborative network that takes into account the failure mode of logistics support, and redistribute the load of the unmanned platform cross-domain collaborative network after removing the failed nodes to analyze cascading failures; the failure modes that take into account logistics support include support chain missing failure mode and mission failure mode.

2. The method according to claim 1, characterized in that The task failure modes are: k in (H j )*k out (H j )=0 Among them, H i is the i-th node, H j is the jth node, H m is the mth node, k in (H j ) represents the number of in-degrees of the node, k out (H j ) represents the number of out-degrees of the node, N H is the number of corresponding nodes.

3. The method according to claim 1, characterized in that The failure mode of the missing guarantee chain is: Among them, N LCMWL and N LCCMWL is the number of maintenance chains; N LCUVL and N LCCUVL is the number of supply chains.

4. The method according to claim 2, characterized in that: Load redistribution of the unmanned platform cross-domain collaborative network after removing the failed node includes: Obtain failed nodes after being damaged by interference; If the current failed node is a guarantee node in the guarantee domain, the load of the current failed node is redistributed according to the pre-set capacity load redistribution strategy; If the current failed node is an operating node in the operating domain, the load of the current failed node is redistributed according to the pre-set cross-domain collaborative priority redistribution strategy; Update the load of the nodes that have not failed.

5. The method according to claim 2, characterized in that: The performing load redistribution on the currently failed node according to the preset capacity load redistribution strategy includes: The initial load of each security node in the security domain is calculated according to the unit degree, unit betweenness and unit information interference importance of each security node in the security domain; The unit capacity of each security node is calculated according to the initial load of each security node, and the real-time processing capability index of each security node is obtained according to the unit capacity and the initial load; The load of neighboring nodes of the same type is distributed according to the ratio of the real-time processing capability index of the neighboring nodes of the current failed node to the real-time processing capability index of all neighboring units of the same type.

6. The method according to claim 2, characterized in that The load redistribution of the current failed node according to the preset cross-domain collaborative priority redistribution strategy includes: The initial load of each operation node in the operation domain is calculated according to the unit degree, unit betweenness, unit information interference importance and guarantee interference importance of each operation node in the operation domain; Calculate the priority of the same type of neighbor nodes of the current failed node in each operation domain according to the preset cross-domain collaboration priority judgment method; The initial load of the currently failed node is distributed to neighboring nodes of the same type according to the priorities of the neighboring nodes of the same type.

7. The method according to claim 1, characterized in that The method further comprises: Obtain the impact efficiency when the operation node fails, and obtain the intentional interference importance of each operation node in the operation domain according to the information interference importance, security interference importance and impact efficiency of each operation node in the operation domain; Obtain the impact efficiency when the security node fails, and obtain the intentional interference importance of each security node in the security domain according to the information interference importance and impact efficiency of each security node in the security domain; According to the importance of intentional interference of each node in the operation domain and the support domain, the key nodes in the cross-domain collaborative network of the unmanned platform are obtained.

8. The method according to claim 1, characterized in that The method further comprises: Evaluate the performance of the cross-domain collaborative network of unmanned platforms based on the number of damaged nodes.

9. A cascading failure analysis device for unmanned platform cross-domain collaborative network, characterized in that: The device comprises: A network construction module, used to construct a cross-domain collaborative network for unmanned platforms; the cross-domain collaborative network for unmanned platforms includes a support domain and at least two operation domains, the operation domains are connected by inter-layer collaborative edges, and the operation domains are connected to the support domains by inter-layer support edges; the support domain provides logistical support for multiple operation domains; The failure analysis module is used to obtain the number of maintenance chains and material supply chains of the supported node in the operation domain. If the number of maintenance chains or material supply chains is zero, the current supported node is in a support chain missing failure mode. The failure analysis module is also used to obtain the number of in-degrees and out-degrees of other nodes in the operation domain and the guarantee domain except the target node. If the number of in-degrees or out-degrees is zero, the current node is in task failure mode; The load redistribution module is used to obtain failed nodes in the unmanned platform cross-domain collaborative network that are damaged by interference, and to redistribute the load of the unmanned platform cross-domain collaborative network after removing the failed nodes to analyze cascading failures; the failure modes that consider logistics support include support chain missing failure modes and mission failure modes.

Images