System toughness dynamic evaluation method based on multi-agent collaborative reconstruction

Abstract

The invention discloses a system toughness dynamic evaluation method based on multi-agent collaborative reconstruction, and belongs to the technical field of system engineering and intelligent decision making. The problem that an existing toughness evaluation method and evaluation index are single, the dynamic evolution process of function recovery cannot be reflected, and consequently evaluation accuracy is low is solved. According to the method, system toughness evaluation indexes are established from four dimensions of absorption capability characteristics, recovery capability characteristics, structure reconstruction and function reconstruction, and all influence factors in a dynamic evolution process of realizing function recovery through multi-agent collaborative reconstruction after the system is disturbed are considered; the task performance maintenance and recovery capability quantitative evaluation of the complex system under the disturbance condition is realized, and the system toughness evaluation result is accurate. The method is mainly applied to dynamic evaluation of the toughness of the intelligent agent system.

Description

Technical Field

[0001] This invention belongs to the field of systems engineering and intelligent decision-making technology. Background Technology

[0002] With the development of intelligent and systematic mission execution, mission execution capabilities have gradually evolved from single-equipment mission execution to system-wide confrontation. In this system-based mission execution model, mission execution capability originates from a complex system composed of multiple functional units such as reconnaissance, command and control, firepower, and support. These system elements have multi-level and multi-link relationships. When the system is subjected to enemy attacks or communication disruptions, the damage to local nodes or links will cause a decrease in mission capability, and in severe cases, lead to a decline in the overall mission execution efficiency of the system. Therefore, the system's mission maintenance and recovery capabilities under disturbance conditions have become important indicators for the design and evaluation of intelligent systems.

[0003] Existing research primarily focuses on resilience assessment from the perspective of system topology or performance degradation. For example, hypernetwork-based resilience assessment methods describe the coupling relationships between capability, task, and elements to quantitatively calculate the system's capability recovery time and recovery performance ratio; percolation-based assessment methods analyze changes in structural connectivity under different perturbations from the perspective of link failure, quantitatively characterizing the overall robustness and recoverability of the system. While these methods have some reference value in static structural analysis, they suffer from the following limitations:

[0004] (1) Lack of dynamic reconstruction mechanism. Existing methods generally regard the system as a static network structure, which cannot characterize the dynamic evolution process of the system to restore its function through autonomous adjustment or collaborative reorganization after disturbance.

[0005] (2) Ignoring the autonomy and collaboration of intelligent agents. With the development of artificial intelligence technology, system nodes are gradually acquiring perception, decision-making and collaboration capabilities. The resilience of the system is no longer determined solely by structural robustness, but depends more on the collaborative reconstruction behavior among intelligent agents. Traditional topology or performance analysis is difficult to reflect this intelligent characteristic.

[0006] (3) The evaluation indicators are singular and lack task orientation. Most studies only focus on the performance recovery curve or node connectivity, and fail to comprehensively consider multi-dimensional factors such as task completion rate, agent response speed, and structural reconstruction efficiency, making it difficult to fully reflect the behavioral characteristics of the system's resilience.

[0007] Therefore, there is an urgent need for a new system resilience assessment method that can integrate intelligent agent modeling, autonomous reconstruction and task performance evaluation, in order to characterize the absorption, recovery and adaptation process of intelligent systems after being disturbed, thereby providing technical support for system design optimization and practical performance evaluation. Summary of the Invention

[0008] The purpose of this invention is to address the problems of existing resilience assessment methods and the fact that assessment indicators are singular and cannot reflect the dynamic evolution process of functional recovery, resulting in low assessment accuracy. This invention provides a dynamic assessment method for system resilience based on multi-agent collaborative reconstruction.

[0009] A dynamic assessment method for system resilience based on multi-agent collaborative reconfiguration is proposed, wherein the system is a whole composed of multiple functional units, each of which includes one or more agents, and each agent is a node; the dynamic assessment method is based on the dynamic evolution process of the system recovering its functions through multi-agent collaborative reconfiguration after a disturbance; this dynamic assessment method includes:

[0010] Obtain the absorption capacity index of the system after disturbance ;

[0011] Acquire the system's recovery capability indicators after collaborative reconstruction ;

[0012] Acquisition of structural reconstruction indicators of the system during collaborative reconstruction process and functional restructuring indicators ;

[0013] according to , , and Constructing system resilience assessment indicators .

[0014] Preferably, the disturbances include node disturbances, link disturbances, and task disturbances.

[0015] Preferably, ,in,

[0016] , These are the weighting coefficients for the system's absorption capacity, recovery capacity, structural reconstruction capacity, and functional reconstruction capacity, respectively.

[0017] Preferably, the absorption capacity index is obtained. and recovery capacity indicators The implementation methods include:

[0018] S11, Obtain Task At each moment Subsystem task completion rate ;

[0019] S12. Based on the completion rate of all tasks at each time point, obtain the time. Overall performance of the lower system ;

[0020] S13. Plot the overall system performance curve as a function of time based on the overall system performance at all times, and obtain... , and ;in,

[0021] Steady-state moment before disturbance The overall performance of the system under the following conditions For the time of disturbance Overall performance of the system The time to recover to steady state after the disturbance The overall performance of the system is as follows;

[0022] S14, according to and Determine absorption capacity indicators ;

[0023] according to , and Determine recovery capacity indicators .

[0024] Preferably, ;

[0025] in, For the task The weights satisfy , For the number of tasks, For the task At any moment The completion rate of tasks in the system.

[0026] Preferably, structural reconstruction indicators are obtained. The implementation method is as follows:

[0027] Based on the number of edges in the architecture when it recovers to a steady state after collaborative reconfiguration Number of edges connected to the pre-disturbance architecture Determine structural reconstruction indicators .

[0028] Preferably, obtain functional refactoring metrics. The implementation method is as follows:

[0029] ;

[0030] in, For the task The time to recover to steady state after the disturbance The completion rate of the system tasks under the following conditions For the task Steady-state time before disturbance The completion rate of the system tasks under the following conditions The number of tasks.

[0031] Preferably,

[0032] ;

[0033] ;

[0034] ;

[0035] ;

[0036] in, Indicates task At any moment completion rate Indicates time Next node For the task Contribution weight, For a moment Next node For the task Contribution weight, For time intervals, The total number of nodes. For nodes For the task The proportion of takeover, For the task The increase in demand, For nodes At any moment The health level below, For nodes At any moment The health level below, For nodes The disturbance intensity coefficient, For nodes The effective connectivity factor, For a moment The system exhibits cooperative relationships and also functions as an adjacency matrix. These are the first and second empirical parameters, respectively. For nodes The degree of the nodes, and , , For a moment Next node With nodes The collaborative relationship between nodes, if nodes With nodes If there is task collaboration, then ,otherwise , This represents the maximum node degree among all nodes.

[0037] Preferably, .

[0038] Preferably, resource constraints are set. .

[0039] The beneficial effects of this invention are:

[0040] This invention establishes a system resilience assessment index from four dimensions: absorption capacity characteristics, recovery capacity characteristics, structural reconstruction, and functional reconstruction. It considers all influencing factors in the dynamic evolution process of functional recovery through multi-agent collaborative reconstruction after perturbation, enabling a quantitative assessment of the task performance maintenance and recovery capability of complex systems under perturbation conditions. The system resilience assessment results are accurate. The method of this invention can simulate the damage, reconstruction, and task migration processes of system elements, revealing the absorption, recovery, and adaptation characteristics of system resilience from the perspective of agent behavior. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the system performance curves before and after interference injection;

[0042] Figure 2 This is a schematic diagram showing the changes in the system architecture before and after interference injection;

[0043] Figure 3 This is a comparison chart of changes in the system's collaborative restructuring;

[0044] Figure 4 This is a diagram showing the changes due to the system architecture refactoring;

[0045] Figure 5 It is a graph showing the overall performance of the system.

[0046] Figure 6 This is a schematic diagram illustrating the principle of the system resilience dynamic assessment method based on multi-agent collaborative reconstruction described in this invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0049] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0050] System resilience refers to the ability of a system to maintain its critical functions and recover to an acceptable performance level within a certain period of time when subjected to external disturbances or internal failures. Unlike traditional reliability, which focuses solely on the "probability of failure," it emphasizes the dynamic adaptation and continuous operation characteristics of the system throughout the entire process of "disturbance-degradation-recovery." The core characteristics of system resilience can be summarized into three stages: absorption, recovery, and adaptation.

[0051] In complex systems, these three characteristics are not only reflected in changes in task performance but also in the dynamic reconfiguration of network connectivity and functional allocation. Based on this, a quantitative assessment of the task performance maintenance and recovery capabilities of complex systems under perturbation conditions is achieved through a multi-agent collaboration and dynamic reconfiguration mechanism. It can simulate the damage, reconstruction, and task migration processes of system elements, revealing the absorption, recovery, and adaptation characteristics of system resilience from the perspective of agent behavior. This invention establishes system resilience assessment indicators from four dimensions: absorption capacity characteristics, recovery capacity characteristics, structural reconfiguration, and functional reconfiguration, to comprehensively characterize the system's resistance to impact, recovery, and collaborative performance under perturbation conditions.

[0052] Specific Implementation Method 1: Combination Figure 6 This embodiment describes a system resilience dynamic assessment method based on multi-agent collaborative reconfiguration. The system is a whole composed of multiple functional units, each including one or more agents, with each agent acting as a node. The dynamic assessment method is based on the dynamic evolution process of the system recovering its function through multi-agent collaborative reconfiguration after a disturbance. This dynamic assessment method includes:

[0053] Obtain the absorption capacity index of the system after disturbance ;

[0054] Acquire the system's recovery capability indicators after collaborative reconstruction ;

[0055] Acquisition of structural reconstruction indicators of the system during collaborative reconstruction process and functional restructuring indicators ;

[0056] according to , , and Constructing system resilience assessment indicators .in, , , These are the weighting coefficients for the system's absorption capacity, recovery capacity, structural reconstruction capacity, and functional reconstruction capacity, respectively. These four coefficients can be set according to the system type or task type.

[0057] This invention establishes a system resilience assessment index from four dimensions: absorption capacity characteristics, recovery capacity characteristics, structural reconstruction, and functional reconstruction, enabling a quantitative assessment of the task performance maintenance and recovery capabilities of complex systems under perturbation conditions. The method of this invention can simulate the damage, reconstruction, and task migration processes of system elements, revealing the absorption, recovery, and adaptation characteristics of system resilience from the perspective of agent behavior. Furthermore, the dynamic evolution process of functional recovery achieved by the system through multi-agent collaborative reconstruction after perturbation can be achieved using existing technologies.

[0058] Furthermore, to characterize the system's dynamic response under external attacks or internal failures, a disturbance injection mechanism is introduced. Disturbances can be categorized into three types: node disturbances, link disturbances, and task disturbances.

[0059] Node-type perturbations are used to describe the damage or performance degradation of functional units, and their effects can be represented as follows: .in, For nodes The perturbation intensity coefficient characterizes the node. The percentage decrease in health;

[0060] Link-type disturbances weaken the adjacency matrix through random or targeted methods. This is achieved through several connection elements to simulate communication interruption or loss of cooperation.

[0061] Task-related perturbations are manifested as task importance weights. The mutation or redistribution is used to simulate changes in external task priorities.

[0062] After a disturbance occurs, the overall performance curve of the system will show varying degrees of degradation.

[0063] In practical applications, in order to characterize the dynamic behavior of intelligent systems under disturbance conditions, a disturbance injection mechanism is defined based on the functional structure model of the system to achieve a computable characterization of the system's performance.

[0064] The system consists of several functional units, which may include elements such as reconnaissance, command and control, firepower, and support. Based on the system capability-task-entity (C-T-E) relationship, a three-layer abstract model of the system architecture is performed.

[0065] In the element layer, each functional unit is abstracted as an intelligent agent node with perception, decision-making, and execution capabilities. Its status can be determined by its health level. Resource utilization rate and capability vector .in, Represents a node For the task Contribution weight, The number of tasks.

[0066] The cooperative relationships within the system are established through an adjacency matrix. Description, if node With nodes If there is a task collaboration relationship, then ,otherwise Based on this Nodes can The effective connectivity factor is expressed as :

[0067] At the task level, the system task completion rate is defined as... , Indicates task At any moment Completion level;

[0068] In the system capability layer, the overall system performance is defined as... .

[0069] The system restructuring is divided into two levels: structural restructuring and functional restructuring.

[0070] Structural reconfiguration refers to the process by which intelligent agents restore network connectivity by re-establishing communication links or adjusting cooperative relationships. Structural reconfiguration can be modeled as an adjacency matrix. The dynamic adjustment process: .in, This indicates a new or replacement collaborative relationship, which can be generated based on distance, task similarity, or communication reachability constraints.

[0071] Functional reconfiguration: This refers to maintaining the continuity of system functionality through task migration, function takeover, or resource reallocation mechanisms after some nodes fail. The task contribution update for a node can be defined as follows: .in, For the task The increase in demand, Represents a node For the task The proportion of takeover. And this should be achieved by setting resource constraints. This ensures that node capacity is not overloaded during the reconstruction process.

[0072] Therefore, we further provide an index for the absorption capacity of the system after it has been disturbed. and the system's resilience indicators after collaborative reconstruction The implementation methods include:

[0073] S11, Obtain Task At each moment Subsystem task completion rate ;

[0074] S12. Based on the completion rate of all tasks at each time point, obtain the time. Overall performance of the lower system Specifically,

[0075] ;

[0076] ;

[0077] ;

[0078] ;

[0079] ;

[0080] ;

[0081] in, For the task The weights satisfy , For the number of tasks, For the task At any moment Subsystem task completion rate The total number of nodes;

[0082] Indicates task At any moment completion rate Indicates time Next node For the task Contribution weight, For a moment Next node For the task Contribution weight, For time intervals, For nodes For the task The proportion of takeover, For the task The increase in demand, For nodes At any moment The health level below, For nodes At any moment The health level below, For nodes The disturbance intensity coefficient, For nodes The effective connectivity factor, For a moment The system exhibits cooperative relationships and also functions as an adjacency matrix. These are the first and second empirical parameters, respectively. For nodes The degree of the nodes, and , , For a moment Next node With nodes The collaborative relationship between nodes, if nodes With nodes If there is task collaboration, then ,otherwise , This represents the maximum node degree among all nodes; as an example, Typically, values ​​of 0.8 and 0.2 are used to reflect the efficiency correction of nodes in system coordination;

[0083] S13. Plot the overall system performance curve as a function of time based on the overall system performance at all times, and obtain... , and ;in,

[0084] Steady-state moment before disturbance The overall performance of the system under the following conditions For the time of disturbance Overall performance of the system The time to recover to steady state after the disturbance The overall performance of the system is as follows;

[0085] S14, according to and Determine absorption capacity indicators ;

[0086] according to , and Determine recovery capacity indicators .

[0087] In this preferred embodiment, the absorption phase reflects the system's initial ability to resist disturbances, i.e., the extent of performance degradation when subjected to external shocks. Therefore, an absorption capacity index is defined. , The closer the value is to 1, the stronger the system's resistance to disturbances and the smaller the performance degradation.

[0088] The recovery phase reflects the system's ability and speed to recover its performance through reconfiguration mechanisms after instability. A recovery capability index is defined. .in, The larger the value, the faster the system recovers and the greater the recovery rate.

[0089] In addition to overall performance, system resilience is also reflected in the reconstruction indicators at the structural and functional reconstruction levels during the system's collaborative reconstruction process.

[0090] Furthermore, structural reconstruction indicators of the acquisition system during the collaborative reconstruction process are provided. The implementation method is as follows:

[0091] Based on the number of edges in the architecture when it recovers to a steady state after collaborative reconfiguration Number of edges connected to the pre-disturbance architecture Determine structural reconstruction indicators The connecting edge is the line connecting the nodes.

[0092] In this preferred embodiment, the structural reconstruction index The degree to which the structural connectivity of the system is restored after reconstruction. The larger the value, the more complete the system's topological connections, and the stronger its foundation for information transmission and collaboration.

[0093] Furthermore, functional reconstruction indicators for the acquisition system during the collaborative reconstruction process are provided. The implementation method is as follows:

[0094] ;

[0095] in, For the task The time to recover to steady state after the disturbance The completion rate of the system tasks under the following conditions For the task Steady-state time before disturbance The completion rate of the system tasks under the following conditions The number of tasks.

[0096] In this preferred embodiment, the functional reconstruction index To measure the degree of recovery of the system's ability to perform tasks. The higher the value, the more effective the system has in terms of functional takeover and resource reallocation at the task level.

[0097] Example:

[0098] Preliminary preparation stage:

[0099] To characterize the dynamic behavior of intelligent systems under disturbance conditions, we first construct a functional structure model of the system, and then define a disturbance injection mechanism based on this model to achieve a computable representation of the system's performance.

[0100] The system consists of several functional units, which may include elements such as reconnaissance, command and control, firepower, and support. Based on the system capability-task-entity (C-T-E) relationship, a three-layer abstract model of the system architecture is performed.

[0101] In the element layer, each functional unit is abstracted as an intelligent agent node with perception, decision-making, and execution capabilities. Its status can be determined by its health level. Resource utilization rate and capability vector .in, Represents a node For the task Contribution weight, The number of tasks.

[0102] The cooperative relationships within the system are established through an adjacency matrix. Description, if node With nodes If there is a communication or task collaboration relationship between them, then ,otherwise Based on this, the node Effective connectivity factor It can be represented as: .in, Represents a node The degree of the node, These are the first and second empirical parameters, typically set to 0.8 and 0.2, respectively, used to reflect the efficiency correction of nodes in system coordination.

[0103] At the task level, the system task completion rate can be defined as: ;in, Indicates task At any moment completion rate The total number of nodes;

[0104] In the system capability layer, the overall system performance is represented as: ,in, For the task The weights satisfy .

[0105] Disturbance injection phase:

[0106] To characterize the system's dynamic response under external attacks or internal failures, a disturbance injection mechanism is introduced. Disturbances can be categorized into three types: node disturbances, link disturbances, and task disturbances.

[0107] Node-type perturbations are used to describe the damage or performance degradation of functional units, and their effects can be represented as follows: .in, For nodes The disturbance intensity coefficient represents the node. The percentage decrease in health;

[0108] Link-type disturbances weaken the adjacency matrix through random or targeted methods. This is achieved through several connection elements to simulate communication interruption or loss of cooperation.

[0109] Task-related perturbations are manifested as task importance weights. The mutation or redistribution is used to simulate changes in external task priorities.

[0110] After a disturbance occurs, the overall performance curve of the system will decay to varying degrees. The magnitude of the decay is closely related to the disturbance intensity, the importance of the damaged nodes, and the connectivity of the system network. Through multi-step simulations, the dynamic trajectory of the system performance over time can be obtained, providing data support for subsequent multi-agent collaborative reconfiguration modeling and system resilience assessment.

[0111] The performance curves and structural changes of the system before and after interference injection are as follows: Figure 1 and Figure 2 As shown.

[0112] Multi-agent collaborative reconstruction phase:

[0113] Under disturbances, the system's performance will degrade to varying degrees. To characterize the system's adaptive recovery capability, multi-agent collaborative reconstruction is required, describing the structural reconstruction and functional reallocation process from two levels: individual decision-making behavior and group collaboration mechanisms. This collaborative reconstruction process includes four main stages: reconstruction trigger determination, reconstruction strategy generation, agent interaction decision-making, and system performance recovery.

[0114] 1) Reconstruction Trigger Determination

[0115] System reconfiguration is typically triggered by performance metrics. When the overall system performance... Drop to a certain threshold When the threshold is reached, the system is considered to have lost its original steady-state capability and needs to enter the reconstruction phase. The triggering condition can be determined based on the importance of the task or the failure rate of critical nodes in the system. Triggering conditions can be driven not only by performance degradation but also by the node's self-aware local state (such as health status). Below the threshold This triggers a distributed, bottom-up refactoring pattern.

[0116] 2) Modeling the Restructuring Mechanism

[0117] The restructuring of the system is divided into two levels: structural restructuring and functional restructuring.

[0118] Structural reconfiguration refers to the process by which intelligent agents restore network connectivity by re-establishing communication links or adjusting cooperative relationships. Structural reconfiguration can be modeled as an adjacency matrix. The dynamic adjustment process: .in, This indicates a new or replacement collaborative relationship, which can be generated based on distance, task similarity, or communication reachability constraints.

[0119] Functional reconfiguration: This refers to maintaining the continuity of system functionality through task migration, function takeover, or resource reallocation mechanisms after some nodes fail. The task contribution update for a node can be defined as follows: .in, For the task The increase in demand, Represents a node For the task The takeover ratio. This is achieved by setting resource constraints. This ensures that node capacity is not overloaded during the reconstruction process.

[0120] 3) Agent cooperative decision-making modeling

[0121] During the reconstruction process, each node acts as an independent intelligent agent, making decisions based on its local state (such as health, task load, and neighbor connectivity). A rule-based collaborative mechanism is employed.

[0122] ① Healthy nodes take over tasks according to their remaining capacity ratio;

[0123] ② Neighboring nodes are connected by probability Establish or enhance communication links.

[0124] This mechanism enables distributed, self-organizing collaborative evolution, allowing the system to restore its overall functionality without global control.

[0125] 4) System performance recovery

[0126] The dynamic changes in system performance recovery are contributed by structural remodeling. and contributions to functional refactoring Driven by each other, defining .in, This represents the total number of edges connected to the system (a structural connectivity index). Weighted values ​​(functional metrics) are assigned to task completion rates. Indicates the moment of the disturbance. This represents the steady-state moment before the disturbance.

[0127] The evolution of system performance recovery can be represented as: .in, These are weighting coefficients, reflecting the contributions of structural restoration and functional takeover to the overall performance improvement, respectively.

[0128] like Figure 3 As shown, when the system is not reconfigured, its performance remains low for a long time after the disturbance; however, after introducing a collaborative reconfiguration mechanism, the structural connectivity is quickly restored, the task completion rate gradually improves, and the system performance... The system recovered and stabilized quickly, demonstrating strong adaptive recovery capabilities. System architecture refactoring, such as... Figure 4 As shown.

[0129] System resilience assessment phase:

[0130] System resilience is a comprehensive reflection of multidimensional characteristics; therefore, it is necessary to construct a system resilience assessment index. .in, The four weights respectively reflect the relative importance of the system's resistance to attack (absorption), self-organizing recovery (recovery), structural robustness (structural layer), and functional adaptability (functional layer).

[0131] Pick The results of the calculation system resilience assessment indicators, such as Figure 5 As shown, the steady-state time before the disturbance is obtained. Overall system performance , Overall performance of the lower system (i.e., the lowest performance value after the disturbance) and the time when the performance recovers to steady state after the disturbance. Overall system performance .

[0132] The calculated toughness assessment results are shown in Table 1.

[0133] Table 1. Results of System Toughness Assessment

[0134]

[0135] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration, characterized in that, The system is a whole composed of multiple functional units working together, each functional unit including one or more agents, and each agent as a node; the dynamic evaluation method is based on the dynamic evolution process of the system recovering its function through multi-agent collaborative reconstruction after a disturbance; the dynamic evaluation method includes: Obtain the absorption capacity index of the system after disturbance ; Acquire the system's recovery capability indicators after collaborative reconstruction ; Acquisition of structural reconstruction indicators of the system during collaborative reconstruction process and functional restructuring indicators ; according to , , and Constructing system resilience assessment indicators .

2. The system resilience dynamic assessment method based on multi-agent collaborative reconfiguration according to claim 1, characterized in that, Disturbances include node disturbances, link disturbances, and task disturbances.

3. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 1, characterized in that, ,in, , These are the weighting coefficients for the system's absorption capacity, recovery capacity, structural reconstruction capacity, and functional reconstruction capacity, respectively.

4. The system resilience dynamic assessment method based on multi-agent collaborative reconfiguration according to claim 1, characterized in that, Obtaining absorption capacity indicators and recovery capacity indicators The implementation methods include: S11, Obtain Task At each moment Subsystem task completion rate ; S12. Based on the completion rate of all tasks at each time point, obtain the time. Overall performance of the lower system ; S13. Plot the overall system performance curve as a function of time based on the overall system performance at all times, and obtain... , and ;in, Steady-state moment before disturbance The overall performance of the system under the following conditions For the time of disturbance Overall performance of the system The time to recover to steady state after the disturbance The overall performance of the system is as follows; S14, according to and Determine absorption capacity indicators ; according to , and Determine recovery capacity indicators .

5. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 4, characterized in that, ； in, For the task The weights satisfy , For the number of tasks, For the task At any moment The completion rate of tasks in the system.

6. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 1, characterized in that, Obtain structural reconstruction indicators The implementation method is as follows: Based on the number of edges in the architecture when it recovers to a steady state after collaborative reconfiguration Number of edges connected to the pre-disturbance architecture Determine structural reconstruction indicators .

7. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 1, characterized in that, Get Function Refactoring Metrics The implementation method is as follows: ； in, For the task The time to recover to steady state after the disturbance The completion rate of the system tasks under the following system For the task Steady-state time before disturbance The completion rate of the system tasks under the following conditions The number of tasks.

8. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 4 or 7, characterized in that, ； ； ； ； in, Indicates task At any moment completion rate Indicates time Next node For the task Contribution weight, For a moment Next node For the task Contribution weight, For time intervals, The total number of nodes. For nodes For the task The proportion of takeover, For the task The increase in demand, For nodes At any moment The health level below, For nodes At any moment The health level below, For nodes The disturbance intensity coefficient, For nodes The effective connectivity factor, For a moment The system exhibits cooperative relationships and also functions as an adjacency matrix. These are the first and second empirical parameters, respectively. For nodes The degree of the nodes, and , , For a moment Next node With nodes The collaborative relationship between nodes, if nodes With nodes If there is task collaboration, then ,otherwise , This represents the maximum node degree among all nodes.

9. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 8, characterized in that, 。 10. The method for dynamic evaluation of system resilience based on multi-agent collaborative reconfiguration according to claim 8, characterized in that, Set resource constraints .

Images