Satellite network reconfiguration scheme design method based on DRDES

Abstract

The application discloses a satellite network reconstruction scheme design method based on DRDES, which models an actual problem by using DRDES, and then combines a simulated annealing algorithm with the DRDES to design a reconstruction scheme of a low-orbit satellite communication network. The application is a network reconstruction method before network node interference, failure and attack, which transforms a structure into a target topology while ensuring that the invulnerability in the reconstruction process is high enough. The method can effectively find a reconstruction scheme which can meet the time consumption requirement and the average invulnerability, and can also reach a preset value within an acceptable time range according to actual needs.

Description

Technical Field

[0001] This invention belongs to the field of network reconfiguration technology for satellite communication network systems, and relates to a design method for satellite network reconfiguration schemes based on DRDES. Background Technology

[0002] With the development of aerospace technology, people's requirements for the performance and services of satellite communication systems are gradually increasing. A considerable portion of the services provided by satellite communication network systems involve national security issues, making the resilience of the network system particularly important. In the reconstruction of low-Earth orbit satellite communication networks, global resilience is a key concern. The network must maintain high resilience throughout the reconstruction process, even when subjected to a series of unknown intermediate configurations. This ensures that the entire system can stably provide the intended service in the face of periodic topology reconfigurations, sporadic attacks, or satellite / link failures. Therefore, designing an algorithm that can reconstruct the network to the target topology while ensuring high resilience during the reconstruction process is highly valuable.

[0003] Dynamic reconfigurability refers to the ability of a system to adjust its configuration online in real time according to changes in the internal and external execution environment to ensure that the system continuously meets control parameters. We define a system with dynamic characteristics where discrete events play a decisive role in the system process as a Dynamically Reconfigurable Discrete Event System (DRDES). When considering only the topology, logical behavior, and related performance of a low-Earth orbit (LEO) satellite communication network system, it can be abstracted as a DRDES. Modeling LEO satellite communication networks as DRDES for network reconfiguration is a previously unexplored approach. Given the large scale of LEO satellite communication networks, heuristic algorithms are needed for the reconfiguration of DRDES systems.

[0004] Currently, there is no design method for a reconfiguration scheme that takes a given target network as the reconfiguration target and requires average resilience during the reconfiguration process. Some researchers (Fuad NAA, Azlin AA N, Zahari MF D, et al. Power distribution loss reduction using cuckoo search[C] / / 2018 7th International Conference on Computer and Communication Engineering(ICCCE).IEEE,2018:333-337.) proposed to use heuristic algorithms to reconfigure the distribution network and reduce the power loss in the target configuration system by changing the state of the reconfiguration switch. However, the scale of reconfiguration that occurs during the operation of the power grid is relatively small, and this method is not suitable for large-scale satellite communication networks. Zhao Zhigang (Zhao Zhigang, Liu Jun, Yu Hua, et al. Reconfiguration and self-recovery of LEO satellite networks[J]. Journal of Northeastern University,2004(10):931-933.) and others detected faulty links through dynamic routing strategies, used power control to reconfigure the links, and switched back to the original topology after the fault was recovered. They focused on the link fault problem without involving the combination of resilience and search algorithms. The dynamic reconstruction technology of the space-ground integrated information network proposed by Dong Ruonan (Dong Ruonan, Zhang Guangjie, Liu Yuan, et al. Dynamic reconstruction technology and simulation method of space-ground integrated information network[J]. Small and Microcomputer Systems, 2020, 41(05):1065-1070.) enables automatic path switching to ensure normal network service when satellite nodes fail. However, this research focuses more on the real-time detection and reconstruction of the network itself, rather than reconstructing to the target topology. Summary of the Invention

[0005] The purpose of this invention is to provide a satellite network reconfiguration scheme design method based on Dynamically Reconfigurable Discrete Event System (DRDES). This method can effectively find a satellite network reconfiguration scheme that meets both time consumption requirements and average resilience.

[0006] The technical solution adopted in this invention is a satellite network reconstruction scheme design method based on DRDES, which is implemented according to the following steps:

[0007] Step 1: Perform DRDES modeling on the actual problem;

[0008] Step 2: Combine the simulated annealing algorithm with DRDES to design a reconstruction scheme for low-Earth orbit satellite communication networks.

[0009] The invention is further characterized by:

[0010] The DRDES model in step 1 is as follows:

[0011] DRDES consists of triples Definition, where It is a finite set of declared configurations. It is a set of reconstruction schemes. This is the initial configuration; the refactoring scheme is... Definition, where Indicates in Implement the previously declared configuration. Indicates in The configuration declared after implementation; G o =(V o E o ) and G d =(V d E d The initial and target configurations of the reconfiguration are represented by ) which both belong to a finite set of declared configurations.

[0012] A declared configuration The system is described by an undirected graph G = (V, E), where the vertex set V of G represents the satellites in configuration, the edge E represents the available communication links between satellites in V, and I represents the global performance index of the system. The adjacency matrix M of the graph G = (V, E) is a |V|×|V| matrix representing the adjacency relationship between vertices.

[0013] The refactoring event defined in the DRDES model is:

[0014] Make the declaration configuration finite set Initial configuration G o and target configuration G d The same dimension, i.e., V o =V d =N; if a vertex is not connected to any other vertex, it is considered to be zero, and then M is calculated. r =M d -M o , to obtain G o and G d The difference matrix between the two is used to define and reconstruct the event set R based on the information in the difference matrix. od ;R od Represents the reconstruction of G o To G d The required set of refactoring events; ultimately, ensuring the refactoring events After execution, the low-orbit satellite network will be reconstructed.

[0015] The optimization problem in the DRDES model is: Given a set of components that can complete the reconstruction task G... o →G d Reconstruction scheme The option with the highest average survivability was selected from among them;

[0016] The objective function is:

[0017]

[0018] Among them: I during the reconstruction process i (i∈[1,n]) is the i-th digit obtained by using r. th The resilience of each intermediate configuration is calculated; the resilience of the intermediate configurations is summed, and the average resilience I of these reconfiguration configurations is used to evaluate the reconfiguration scheme r.

[0019] The resilience was calculated using either the natural connectivity method or the structural difference method.

[0020] Step 2 is as follows:

[0021] Step 2.1: Initialization, setting parameters: initial temperature T0, termination temperature T end The number of cycles L at each temperature k Temperature decay coefficient α and optimal value matrix result;

[0022] Step 2.2: Randomly generate a reconstruction scheme r based on the DRDES model obtained in Step 1. c ;

[0023] Step 2.3: Calculate the reconstruction scheme r generated in step 2.2. c Average damage resistance I c ;

[0024] Step 2.4: The optimal reconstruction scheme r optimal renew;

[0025] Step 2.5: Current temperature T < Termination temperature T end When the external loop is activated, the loop exits when the initially set termination temperature is reached. After the external loop ends, the optimal reconstruction scheme r is obtained. optimal and maximum average damage resistance I max ;

[0026] Step 2.6, Output: Optimal Reconstruction Scheme r optimal and maximum average damage resistance I max .

[0027] The initial scheme in step 2.2 is a series of distinct consecutive natural numbers generated in MATLAB. According to the rules of the reconstruction event, |R od After a series of distinct natural numbers occur, the satellite network will complete the transition from G... o To G d The reconstruction.

[0028] In step 2.4: initially r optimal =r c The current maximum average damage resistance I max The optimal solution at the current temperature is recorded in the optimal value matrix result, and initially I max =I c .

[0029] Step 2.5 specifically involves:

[0030] Step 2.5.1: Current cycle number i < cycle number L set at each temperature k When the inner loop is started, the condition for exiting the loop is to reach the initially set maximum number of inner loops. After the inner loop ends, the optimal reconstruction scheme and average tamper resistance at that temperature are obtained and recorded in the result.

[0031] After the inner loops of steps 2.5.2 and 2.5.1 are completed, the optimal reconstruction scheme and its corresponding average survivability at this temperature are obtained and recorded in the optimal value matrix result.

[0032] Step 2.5.3: Perform temperature decay T = T × α;

[0033] Step 2.5.4: Current temperature T < Termination temperature T end If the condition is met, proceed to step 2.5.1; otherwise, the outer loop ends and proceed to step 2.6.

[0034] Step 2.5.1 specifically refers to:

[0035] Step 2.5.1.1, in r c Generate a new solution r nearby;

[0036] Step 2.5.1.2: Calculate the average resilience I of the reconstruction scheme r;

[0037] Step 2.5.1.3: Apply the Metropolis criterion of the simulated annealing algorithm to determine whether to accept the new solution or the original solution. After the determination, obtain the reconstruction scheme r accepted under this determination at this temperature. c ;

[0038] Step 2.5.1.4: Compare the average survivability I of the obtained scheme with the maximum average survivability I of the current cycle. max If I>I max Update I with the value of I max The value of I if I ≤ I max Keep I max The value remains unchanged;

[0039] Step 2.5.1.5: Current cycle number i < cycle number L set at each temperature k If the condition is met, proceed to step 2.5.1.1; otherwise, the inner loop ends and proceed to step 2.5.2.

[0040] The beneficial effects of this invention are:

[0041] Existing satellite network reconfiguration methods focus on how to quickly perform local reconfiguration of low-Earth orbit satellite communication networks after link failures or node damage to ensure the stability of the topology and the reliability of communication. They do not involve reconfiguring the initial network topology to the target network topology. The reconfiguration achieved by these methods involves quickly searching for the optimal node to establish a connection after a failure occurs.

[0042] This invention proposes a reconstruction scheme design method for modeling low-Earth orbit (LEO) satellite communication networks as dynamically reconfigurable discrete event systems, based on simulated annealing. Such systems with dynamic reconfiguration characteristics can adjust their configuration online in real time according to changes in the internal and external execution environment, such as network node failures, interference, attacks, and changes in demand, to ensure the system continuously meets control indicators. The task of LEO satellite communication network reconstruction in this invention is to transform the initial topology into a known topology, which is a near-global reconstruction of the entire network. The dynamic reconstruction from the current topology to the target topology is achieved gradually, not immediately. Many reconstruction schemes may be used to reconstruct the satellite communication network from one topology to another, and the resilience of the network varies depending on the topology traversed. This invention provides a network reconstruction performed before network node interference, failures, and attacks occur, transforming the structure to the target topology while ensuring sufficient resilience during the reconstruction process. This method can effectively find reconstruction schemes that meet both time consumption requirements and average resilience, and can also achieve preset values ​​within an acceptable time range according to actual needs.

[0043] For low-Earth orbit (LEO) satellite communication networks, more evaluation criteria, such as network service quality, can be incorporated in the future. Furthermore, the method of this invention can not only achieve network reconfiguration for LEO satellite communication networks, but also provide reconfiguration ideas for other dynamically reconfigurable systems. For example, in computer network communication systems, network QoS during reconfiguration can be used as a relevant global performance indicator; in power distribution network systems, power supply reliability during reconfiguration can be used as a global performance indicator. Attached Figure Description

[0044] Figure 1 This is a state space diagram whose configuration is not declared during the reconstruction process of the method of this invention;

[0045] Figure 2 This is a small-scale satellite communication network topology diagram of Embodiment 1 of the present invention;

[0046] Figure 3 This is a large-scale satellite communication network topology diagram of Embodiment 2 of the present invention;

[0047] Figure 4 This is a line graph showing the average resilience calculation results of the acceptable reconfiguration scheme in Embodiment 2 of the present invention. Detailed Implementation

[0048] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0049] This invention presents a satellite network reconfiguration scheme design method based on Dynamically Reconfigurable Discrete Event Systems (DRDES). It combines simulated annealing algorithm with DRDES to design a reconfiguration scheme for low-Earth orbit satellite communication networks, specifically implemented according to the following steps:

[0050] Step 1: Perform DRDES modeling on the actual problem:

[0051] DRDES consists of triples Definition, where It is a finite set of declared configurations. It is a set of reconstruction schemes. This is the initial configuration. The refactoring scheme is... Definition, where Indicates in Implement the configuration declared before (or after). A declared configuration. The system is described by an undirected graph G = (V, E). In this invention, the vertex set V of G represents the satellites in configuration, the edges E represent the available communication links between satellites in V, I represents the global performance index of the system, and G is used as the basis for the description. o =(V o E o ) and G d =(V d E d Let ) represent the initial and target configurations of the reconfiguration, both of which belong to a finite set of declared configurations. The adjacency matrix M of graph G = (V, E) is a |V| × |V| matrix representing the adjacency relationships between vertices.

[0052] Therefore, from G o To G d The reconstruction is achieved by gradually changing M o The entries in, up to M o Change to M d Finally, we define a series of actions that set the corresponding elements in the adjacency matrix as a reconstruction event r. i , If it is a set of reconstruction schemes, then the reconstruction process is defined as: initial configuration G o Restructuring scheme Obtain the target configuration G d .

[0053] Before specifying the refactoring event, we ensure that the original configuration and the target configuration have the same dimension, i.e., V. o =V d =N. If a vertex is not connected to any other vertex, it is considered to be zero. Then, by calculating M... r =M d -M o , to obtain G o and G d The difference matrix between them. For all i,j∈N, M r =1 means that in v i and v j A new edge (v) has been added between them. i ,v j M r =0 means edge (v i ,v j No change, M r =-1 means that in v i and v j The edge (v) was deleted. i ,v j This change can be considered a refactoring event, i.e., adding or deleting an edge. od Represents the reconstruction of G o toG d The required set of reconstruction events is numbered sequentially from 1, 2, ..., n-1, n (|R) od |=n). Assume that only one refactoring event can be implemented in a step, and each refactoring event occurs only once. Then, after all refactoring events have occurred during the refactoring process, it is possible to complete the process from G. o and G d System restructuring. Clearly, there exists R... od Possible refactoring schemes. From G o To G d During a complete system refactoring process, the system will undergo a finite set of undeclared configurations, the specific topologies of which depend on the actual sequence of refactoring events. Using Ω... o (|Ω o |=n) represents applying the reconstruction scheme After execution, from G o To G d The set of configurations not declared in the middle. The complete configuration in this refactoring process is Ω. o =G o,G0,G1,G2,…,G n-1 G d These intermediate configurations can be obtained from Figure 1 To see it intuitively.

[0054] Finally, an objective function is defined, and the global performance index during the reconstruction process is defined as the average resilience during the reconstruction process. Two methods for evaluating resilience are exemplified in this embodiment: the structural difference method and the natural connectivity method.

[0055] The structural dissimilarity method compares the structural differences between fully connected and partially connected networks. The structural dissimilarity method calculates resilience as shown in formula (1):

[0056]

[0057] In equation (1), G = (V, E) is a partially connected network, where |V| = N. For any two distinct nodes v in G... i and v j v i and v j The number of paths between them (with a length no greater than t) is denoted by u(t), and Inv(G) is the ratio of the sum of the number of equivalent shortest paths to all pairs of nodes in G. Therefore, the summation in the formula is given. Combinations In G, v i and v j The number of shortest paths of length t between them.

[0058] In the natural connectivity method, the resilience of a network is measured by the redundancy of alternative paths between nodes, i.e., the total number of closed paths. The resilience of a network is calculated using formula (2):

[0059]

[0060] In equation (2), in the network G=(V,E), |V|=N, and the number of closed paths of length k is n. k S′ represents the formula for calculating resilience before modification. For ease of calculation, the eigenvalues ​​λ of the graph adjacency matrix will be used. i Calculate S′ using (i = 1, 2, ..., N) to obtain By combining |V| with logarithmic operations, the revised survivability formula is obtained.

[0061]

[0062] In equation (3), I during the reconstruction process i (i∈[1,n]) is the i-th digit obtained by using r. thThe resilience of each intermediate configuration is summed, and the average resilience I of these reconfiguration configurations is used to evaluate the reconfiguration scheme r.

[0063] Step 2: Design a reconstruction scheme for low-Earth orbit satellite communication networks by combining simulated annealing algorithm with DRDES:

[0064] Step 2.1: Initialization, setting parameters: initial temperature T0, termination temperature T end The number of cycles L at each temperature k Temperature decay coefficient α and optimal value matrix result;

[0065] Step 2.2: Randomly generate a reconstruction scheme r based on the DRDES model obtained in Step 1. c The initial scheme consists of a series of distinct consecutive natural numbers generated in MATLAB. According to the rules of the reconstruction event, |R| od After a series of distinct natural numbers occur, the satellite network will complete the transition from G... o To G d Reconstruction;

[0066] Step 2.3: Calculate the reconstruction scheme r generated in step 2.2. c Average damage resistance I c The two evaluation methods for resilience calculation are written into different sub-functions (structural difference method or natural connectivity method), which can be called according to different evaluation requirements when solving the objective function I;

[0067] Step 2.4: The optimal reconstruction scheme r optimal Update (initially r) optimal =r c ), will the current maximum average damage resistance I max (Initial time I) max =I c The optimal solution at the current temperature is recorded in the optimal value matrix reault;

[0068] Step 2.5: Current temperature T < Termination temperature T end When the external loop is activated, the loop exits when the initially set termination temperature is reached. After the external loop ends, the optimal reconstruction scheme r is obtained. optimal and maximum average damage resistance I max Specifically:

[0069] Step 2.5.1: Current cycle number i < cycle number L set at each temperature k At that time, the inner loop is started, and the condition for exiting the loop is reaching the initially set maximum number of inner loop iterations. After the inner loop ends, the optimal reconstruction scheme and average survivability at that temperature are obtained and recorded in the result. Specifically:

[0070] Step 2.5.1.1: Use the `gen new path function` in r c A new scheme r is generated nearby. Here, a new reconstruction scheme r is obtained by exchanging, shifting, and reversing the order of reconstruction events in the original reconstruction scheme;

[0071] Step 2.5.1.2: Calculate the average resilience I of the reconstruction scheme r (using the previously defined sub-function, i.e., using the structural difference method or the natural connectivity method);

[0072] Step 2.5.1.3: Apply the Metropolis criterion of the simulated annealing algorithm to determine whether to accept the new solution or the original solution. After the determination, obtain the reconstruction scheme r accepted under this determination at this temperature. c ;

[0073] Step 2.5.1.4: Compare the average survivability I of the obtained scheme with the maximum average survivability I of the current cycle. max If I>I max Update I with the value of I max The value of I if I ≤ I max Keep I max The value remains unchanged;

[0074] Step 2.5.1.5: Current cycle number i < cycle number L set at each temperature k If the condition is met, proceed to step 2.5.1.1; otherwise, the inner loop ends and proceed to step 2.5.2.

[0075] Step 2.5.2: Update the optimal value matrix result at the current temperature. After the inner loop of step 2.5.1 is completed, the optimal reconstruction scheme and its corresponding average resilience at this temperature are obtained and recorded in the optimal value matrix result;

[0076] Step 2.5.3: Perform temperature decay T = T × α, and then perform the inner loop at the next temperature;

[0077] Step 2.5.4: Current temperature T < Termination temperature T end If the condition is met, proceed to step 2.5.1; otherwise, end the outer loop and proceed to step 2.6.

[0078] Step 2.6, Output: Optimal Reconstruction Scheme r optimal and maximum average damage resistance I max .

[0079] This invention combines simulated annealing with DRDES to design a reconfiguration scheme search algorithm for low-Earth orbit satellite communication networks. This scheme is suitable for solving large-scale combinatorial optimization problems. When facing small-scale network reconfiguration problems, traditional exact algorithms, such as exhaustive search, dynamic programming, and integer linear programming, can be used to find the optimal reconfiguration scheme. The reconfiguration algorithm of this invention is as follows:

[0080]

[0081]

[0082] Example 1

[0083] This embodiment takes a small-scale network reconstruction as an example. The specific network structure is as follows: Figure 2 As shown, two of the networks are simple ring network topologies. Using the reconstruction scheme design method of this invention, the initial adjacency matrix M is compared... o and the target adjacency matrix M d The corresponding refactoring event set R was specified. od Based on the corresponding global performance metric of robustness I, the network is reconstructed according to the proposed reconstruction steps. At this scale, the defined set of reconstruction events R... od If there are 8 events, then there are 8! solutions in the solution space. The optimal reconstruction solution is to be found from these solutions.

[0084] First, generate an initial solution and obtain the initial objective function value: the solution of the objective function value is divided into two sub-functions. One sub-function uses the structural difference method to calculate the resilience, and the other sub-function uses the natural connectivity method to calculate the resilience. Different methods can be applied to evaluate the reconstruction scheme according to the requirements.

[0085] Secondly, new solutions are generated near the initial solution, and new objective function values ​​are obtained: An initial scheme is randomly generated by calling a built-in function in MATLAB. New reconstruction schemes are obtained near the initial scheme by swapping, shifting, and reversing the order of reconstruction events in the original scheme. After each inner loop iteration, the optimal solution generated at that temperature is compared with the known optimal solution to obtain a better solution. At the end of the final iteration, a suitable reconstruction scheme is obtained.

[0086] In this embodiment, the parameters are set as follows: initial temperature 1000, ending temperature 0.5, and the number of inner loops and decay coefficient are set as follows: when using the natural connectivity method to calculate resilience, the temperature decay coefficient α is set to 0.95, and the number of inner loops L... k Set to 50. When using the structural differential method to calculate tamper resistance, the temperature decay coefficient α is set to 0.95, and the number of internal cycles L is set to... kThe value is 10. The simulated annealing algorithm for small-scale DRDES applications was coded in MATLAB 2021 and executed on an Intel i5 processor and 16GB RAM system, as shown in step 2 above. Table 1 lists the results of the simulated annealing method for small-scale DRDES applications under different evaluation methods. It can be seen that in small-scale network reconstruction, the application of this algorithm takes very little time to obtain the reconstruction scheme and can quickly find the optimal reconstruction scheme.

[0087] Table 1 Results of small-scale DRDES application of simulated annealing under different evaluation methods

[0088]

[0089] Example 2

[0090] This embodiment takes large-scale network reconstruction as an example:

[0091] In practical low-Earth orbit satellite communication network reconfiguration problems, the network size is often very large. The satellite network topology in this embodiment is as follows: Figure 1 The diagram shown illustrates the topology of a large-scale satellite network, with two networks each having 100 nodes. The diagram is generated using MATLAB.

[0092] In this embodiment, 75 reconstruction events are declared, meaning there are 75! solutions in the solution space. We know that specific satellite network resilience assessment algorithms are related to the network topology. Sometimes, high redundancy in alternative paths between some nodes can lead to different reconstruction sequences producing the same result, easily trapping the system in a local optimum during the search process. Simulated annealing algorithms accept a solution worse than the current one with a certain probability, thus potentially escaping this local optimum and reaching the global optimum. In large-scale satellite network reconstruction, the specific performance of this algorithm depends on the annealing parameter settings, specifically the number of inner loops L. k And the attenuation coefficient α. This means that multiple experiments are needed to find suitable parameters. The specific parameter combinations are selected based on the experimental results in Table 2.

[0093] When dealing with large-scale problems, it is essential not only to find good solutions but also to minimize time consumption. Therefore, the algorithm should ideally select a parameter set that simultaneously satisfies low time consumption and high average resilience. The initial temperature was set to 1000, and the final temperature to 0.5. Experiments yielded the optimal parameter set.

[0094] Table 2. Influence of different parameters on the calculation results in the simulated annealing algorithm.

[0095]

[0096] Table 2 records the performance of the simulated annealing algorithm under different cooling parameters. Ultimately, a suitable parameter combination needs to be selected by considering both time cost and average resilience. The experimental parameters and data above represent the optimal results obtained from multiple experiments. As shown in Table 2, the optimal parameter combination is: when using the natural connectivity method to calculate resilience, the temperature decay coefficient α is set to 0.95, and the number of inner loops L... k Set to 50. When using the structural differential method to calculate tamper resistance, the temperature decay coefficient α is set to 0.95, and the number of internal cycles L is set to... k The value is 10. Table 3 lists the results of applying simulated annealing to large-scale DRDES under different evaluation methods. It can be seen that even when the scale of the research object increases, the time cost of using the simulated annealing algorithm to search for reconstruction schemes is acceptable.

[0097] Table 3 Results of large-scale DRDES application using simulated annealing under different evaluation methods

[0098]

[0099] As shown in Table 3, the design method of this invention can provide a reconstruction scheme, survivability, and time consumption for a large-scale low-Earth orbit satellite communication network system, and the time consumption is relatively short.

[0100] In practical problems, the resilience of satellite networks has fixed metrics. Sometimes a reconstruction scheme that can achieve preset values ​​is needed, requiring the search to end when the corresponding metrics are reached. Taking the natural connectivity method as an example, experimental results show that when the number of inner loops L... k When the temperature decay coefficient α is 0.95 and the value is 50, the algorithm can find a good solution within an acceptable time range. Its average resilience fluctuates between 1.085 and 1.110. Therefore, 1.090 is used as the preset average resilience for testing.

[0101] Table 4 shows the calculation results of acceptable reconfiguration schemes within a given time period.

[0102]

[0103] Table 4 shows the calculation results of acceptable reconfiguration schemes within a given time period. It can be seen that satellite network reconfiguration based on the simulated annealing algorithm can achieve corresponding results according to different requirements. Multiple experiments were conducted using preset values ​​as a baseline, and the experimental results were plotted as a line graph and recorded. Figure 4 , Figure 4 A line graph showing the calculated average resilience of acceptable reconfiguration schemes within a given time range (s) [500, 1200]. (From Table 4 and...) Figure 4 It can be seen that the method of the present invention can not only complete the task of finding a better reconstruction scheme within a certain time, but also find a reconstruction scheme that reaches the preset value within an acceptable time range.

[0104] Comparative Example 1

[0105] Examples 1 and 2 use the method of the present invention to solve the reconstruction schemes of two pairs of low-Earth orbit satellite communication networks of different sizes. This comparison uses the traditional algorithm on satellite communication networks of different sizes, and uses the obtained results as a benchmark to compare with the schemes obtained by using the simulated annealing algorithm on the corresponding size network.

[0106] For small-scale low-Earth orbit satellite communication networks, an exhaustive search method is used to find the sequence with the highest average resilience as the optimal solution. This is because the corresponding reconstruction event set R for this scale... od The value is not large and can be calculated precisely, so we use Python to solve for the reconstructed event set R. od The xlwt library is used to write data to an Excel spreadsheet for use in the exhaustive search algorithm. The exhaustive search algorithm searches for the best refactoring solution under identical hardware conditions.

[0107] Table 5 Comparison of results of two reconstruction scheme search methods in small-scale networks under different evaluation methods.

[0108]

[0109] Table 5 lists the optimal reconstruction schemes for small-scale networks, the average resilience of the reconstruction schemes, and the time consumption of the reconstruction schemes found by exhaustive search and simulated annealing under two different resilience evaluation criteria. This is because the exhaustive search method traverses the set of reconstruction schemes R. od If all possible solutions are considered, then the optimal solution is obtained. The table shows that in small-scale satellite network reconstruction, simulated annealing can find the optimal reconstruction solution and requires less time to calculate resilience. Therefore, it can be concluded that when the scale of the network reconstruction problem is the same, the simulated annealing algorithm used in this invention takes less time than the exhaustive search method.

[0110] Large-scale low-Earth orbit satellite communication networks, due to their own reconfiguration scheme set R od The current solution is too large to be traversed one by one, thus failing to obtain the globally optimal solution. Therefore, a better reconstruction scheme can only be found by performing a random search on a larger set of solutions, and then compared with the result of the simulated annealing method. The optimal reconstruction scheme found by this method is not the globally optimal solution, but a better solution can be found by increasing the size of the solution and the number of searches, and then used for comparison with the simulated annealing algorithm.

[0111] Table 6 Comparison of results of two reconstruction scheme search methods in large-scale networks under different evaluation methods.

[0112]

[0113] Table 6 lists the optimal reconstruction schemes for large-scale networks found by simulated annealing and random search under two different resilience evaluation criteria, along with the average resilience and time consumption of the reconstruction schemes. The table clearly shows that, under different resilience evaluation criteria, simulated annealing is superior in searching for reconstruction schemes, exhibiting low time consumption and high average resilience. Therefore, in large-scale reconstruction problems, the reconstruction scheme search algorithm designed in this invention yields results with high average resilience.

[0114] This invention is not limited to the given embodiments. For systems similar to low-Earth orbit satellite communication networks, it can proactively modify some components of the current configuration, the connections between components, and the data, enabling the current configuration to be transformed online into another configuration, thereby ensuring that the system can continuously meet control requirements. All of these can be modeled as DRDES (Device Restructuring and Execution System), such as safety-critical systems being restructured to higher robustness while maintaining high reliability during the restructuring process, and communication networks being restructured to a more robust network to improve confidentiality while ensuring normal service.

Claims

1. A satellite network reconfiguration scheme design method based on DRDES, characterized in that, The specific steps are as follows: Step 1: Perform DRDES modeling on the actual problem; Step 2: Design a reconstruction scheme for low-Earth orbit satellite communication networks by combining the simulated annealing algorithm with DRDES; The DRDES model in step 1 is as follows: DRDES consists of triples Definition, where It is a finite set of declared configurations. It is a set of reconstruction schemes. This is the initial configuration; the refactoring scheme is... =(• , •) Defined, where Indicates in Implement the previously declared configuration. • Indicates in The configuration declared after implementation; and The initial and target configurations used to represent the reconfiguration both belong to a finite set of declared configurations. ; A declared configuration From undirected graph To describe, among which vertex set Indicates the configuration of satellites, side express Available communication links between satellites, using To represent the global performance indicators of a satellite network system; Figure adjacency matrix It represents the adjacency relationship between vertices. matrix; The defined reconstruction event in the DRDES model is: Make the declaration configuration finite set initial configuration and target configuration Same dimension, that is If a vertex is not connected to any other vertex, it is considered to be zero, and then calculated... ,get and The difference matrix between them is used to define and reconstruct the event set based on the information in the difference matrix. ; Represents refactoring arrive The required set of refactoring events; ultimately, ensuring the refactoring events After execution, the low-Earth orbit satellite network will be reconstructed; The optimization problem in the DRDES model is: to find the optimal solution from the set point that can complete the reconstruction task. Reconstruction scheme The option with the highest average survivability was selected from among them; The objective function is: Maximize: (3) Among them: during the reconstruction process It is through the use of The obtained number The resilience of each intermediate configuration is calculated; the resilience of these intermediate configurations is summed, and the final average resilience using these reconfigured configurations is obtained. To evaluate the reconstruction plan .

2. The satellite network reconfiguration scheme design method based on DRDES according to claim 1, characterized in that, The resilience was calculated using either the natural connectivity method or the structural difference method.

3. The satellite network reconfiguration scheme design method based on DRDES according to claim 1, characterized in that, Step 2 is as follows: Step 2.1, Initialization, setting parameters: Initial temperature Termination temperature Number of cycles at each temperature Temperature decay coefficient α and optimal value matrix ; Step 2.2: Randomly generate a reconstruction scheme based on the DRDES model obtained in Step 1. ; Step 2.3: Calculate the reconstruction scheme generated in Step 2.

2. Average damage resistance ; Step 2.4: Optimal Reconstruction Scheme renew; Step 2.5, Current Temperature < When the external circulation is activated, the loop exits when the initially set termination temperature is reached. After the external circulation ends, the optimal reconstruction scheme is obtained. and maximum average damage resistance ; Step 2.6, Output: Optimal Reconstruction Scheme and maximum average damage resistance .

4. The satellite network reconfiguration scheme design method based on DRDES according to claim 3, characterized in that, The initial scheme in step 2.2 is a series of distinct consecutive natural numbers generated in MATLAB. According to the rules of the reconstruction event, After a series of distinct natural numbers occur, the satellite network will complete the transition from... arrive The reconstruction.

5. The satellite network reconfiguration scheme design method based on DRDES according to claim 3, characterized in that, In step 2.4: Initially The current maximum average damage resistance The optimal solution at the current temperature is recorded in the optimal value matrix. Initially .

6. The satellite network reconfiguration scheme design method based on DRDES according to claim 3, characterized in that, Step 2.5 specifically involves: Step 2.5.1, Current loop count < When the inner loop is started, the condition for exiting the loop is to reach the initially set maximum number of inner loops. After the inner loop ends, the optimal reconstruction scheme and average damage resistance at that temperature are obtained and recorded in the result. After the inner loops of steps 2.5.2 and 2.5.1 are completed, an optimal reconstruction scheme and its corresponding average resilience at that temperature are obtained and recorded in the optimal value matrix. Step 2.5.3: Perform temperature decay. ; Step 2.5.4, Current Temperature < If the condition is met, proceed to step 2.5.1; otherwise, the outer loop ends and proceed to step 2.

6.

7. The satellite network reconfiguration scheme design method based on DRDES according to claim 6, characterized in that, Step 2.5.1 specifically refers to: Step 2.5.1.1, in Generate a new solution nearby ; Step 2.5.1.2: Calculate the reconstruction scheme. Average damage resistance ; Step 2.5.1.3: Apply the Metropolis criterion of the simulated annealing algorithm to determine whether to accept the new solution or the original solution. After the determination, obtain the reconstruction scheme accepted under this determination at this temperature. ; Step 2.5.1.4: Compare the average survivability of the obtained schemes. Maximum average damage resistance in the current cycle ,if > ,use Value update The value, if ≤ ,Keep The value remains unchanged; Step 2.5.1.5, Current loop count < If the condition is met, proceed to step 2.5.1.1; otherwise, the inner loop ends and proceed to step 2.5.2.

Images