Satellite network dynamic domain division optimization method based on real-time flow potential energy perception

Abstract

The invention relates to a satellite network dynamic domain division optimization method based on real-time flow potential energy perception, and belongs to the technical field of satellite communication networks, and the method comprises the following steps: S1, constructing a satellite network domain division management and control system; s2, off-line static reference planning based on multi-objective evolution; s3, establishing a flow space-time potential energy sensing and domain master control autonomous model; and S4, online closed-loop fine adjustment based on the flow potential energy gradient. According to the method, quantitative measurement of data load accumulation degree and distribution gradient on a satellite network cross-domain link is described through flow potential energy, and the core logic is that real-time throughput pressure of a boundary link is mapped into logic potential energy; and when the load accumulation of the specific boundary exceeds a preset threshold value, the state is used as a driving signal for triggering the sub-domain boundary to carry out adaptive deformation towards the flow direction of the flow, so that the dynamic adaptation of the whole network topology and the flow distribution is realized.

Description

Technical Field

[0001] This invention belongs to the field of satellite communication network technology and relates to a dynamic domain optimization method for satellite networks based on real-time traffic potential perception. Background Technology

[0002] Currently, low-Earth orbit (LEO) satellite internet systems are facing the dual challenges of "ultra-large-scale networking" and "massive data transmission." To reduce the dimensionality explosion problem of network-wide routing calculations and control signaling overhead, multi-level domain partitioning has become the mainstream architectural choice. By dividing the constellation into multiple logical autonomous domains, local topology changes can be effectively isolated, improving network scalability.

[0003] However, in the actual operation of satellite networks, there is a significant adaptation bottleneck between the preset static domain topology and the spatiotemporal heterogeneous evolution characteristics of network data traffic: 1) Spatiotemporal Drift Characteristics of Traffic: Influenced by the Earth's rotation, time zone differences, and population distribution, data traffic generated on the ground exhibits significant "tidal effects" and "hotspot drift" in geographic space. For high-speed satellite networks, high-load traffic hotspots will undergo continuous logical displacement relative to the satellite constellation, constantly crossing physical domain boundaries.

[0004] 2) Limitations of Static Domain Segmentation: Traditional static domain segmentation methods based on orbital planes or fixed geographic grids lack the ability to detect traffic flow. When a high-intensity traffic center drifts to the boundary between two domains, the complete business flow is forcibly truncated by the rigid domain boundary, causing a large amount of data flow to have to pass through overloaded inter-domain links, resulting in severe boundary congestion and queuing packet loss.

[0005] 3) Cost of dynamic adjustment: Although the existing dynamic reconstruction schemes for the entire network can adapt to traffic changes, they usually require triggering global topology recalculation and flow table updates, which has extremely high computational complexity (NP-Hard). Furthermore, frequent global reconstructions can easily cause signaling storms in the control plane and network-wide routing oscillations.

[0006] Therefore, existing technologies face a dilemma when dealing with "dynamic changes in traffic distribution": either tolerate the congestion and throughput bottlenecks caused by static boundaries, or bear the high computational and signaling overhead caused by global reconstruction.

[0007] Currently, there is a lack of a mechanism that "coordinates the initial benchmark based on global static optimization with dynamic fine-tuning based on real-time traffic potential perception"—that is, after establishing a highly reliable network-wide basic topology through multi-objective optimization during the network initialization phase, no global reconstruction is performed during operation. Instead, low-overhead boundary migration (micro-adjustment) is performed only for traffic hotspots that emerge in real time, so that the domain structure can adaptively deform with the traffic center of gravity. Summary of the Invention

[0008] In view of this, the purpose of this invention is to provide a dynamic domain optimization method for satellite networks based on real-time traffic potential sensing.

[0009] To achieve the above objectives, the present invention provides the following technical solution: A dynamic domain optimization method for satellite networks based on real-time traffic potential sensing includes the following steps: S1: Construct a satellite network domain-based management and control system; S2: Offline static baseline planning based on multi-objective evolution; S3: Establish a spatiotemporal potential energy perception and domain-controlled autonomous model for traffic flow; S4: Online closed-loop fine-tuning based on flow potential gradient.

[0010] Furthermore, in step S1, a satellite network domain management and control system is constructed using a Software Defined Network (SDN) architecture, including an SDN control plane and a satellite data plane. The SDN control plane is equipped with a topology management module, an improved NSGA-II global planning engine, a traffic potential analyzer and a dynamic migration controller, a DHC election manager, and a multi-dimensional constraint checker. The satellite data plane is responsible for performing traffic forwarding, link load monitoring, and flow table updates, and integrates a DHC agent within the satellite nodes to periodically report Link State Information (LSA) to the control plane, providing a real-time topology view for global planning.

[0011] Furthermore, in the satellite data plane portion, each low-Earth orbit satellite node , It integrates traffic monitoring, flow table execution, and status reporting functions, specifically including: Traffic forwarding and monitoring: Satellite nodes act as execution units, responsible for forwarding data packets for the inter-satellite link (ISL); for any data flow, the node determines the next-hop node based on local flow table entries. The home domain status; if This is then marked as cross-domain traffic; the node's embedded counters provide real-time statistics on the throughput rate of each port. ; Node Status Awareness Module: Used to sense and report link load and topology status, and construct local status vectors. :

[0012] Among them: link utilization Characterizing the current load saturation of the link; set of neighboring nodes This represents the list of physically visible neighbors with established links at the current time; the current home domain identifier. The entire network is initially divided into K Each domain; Dynamic instruction execution engine: Satellite nodes receive domain migration instructions from the control plane via the southbound interface; upon receiving the instruction, the node immediately updates its local... It is then broadcast to neighbors, completing the logical domain switching.

[0013] Furthermore, the SDN control plane portion includes: Topology Management Module: Responsible for maintaining a global network view; collecting data reported by each node. Real-time construction of dynamic connection graphs And infer physical topological changes based on satellite ephemeris; The NSGA-II global planning engine is configured to run during system initialization or periodic refactoring phases. Internally, it encapsulates a dual optimization objective function that maximizes domain-average reliability and minimizes overall network control overhead. This engine is responsible for outputting a static baseline domain-specific scheme. It provides the basic framework for network operation; Traffic potential energy analysis and dynamic migration controller: includes a traffic matrix builder and a potential energy gradient analyzer; the controller calculates the potential energy difference of cross-domain traffic in real time during runtime slots, identifies heavily loaded links in the network, and generates fine-tuning instructions to be sent to the data plane.

[0014] Furthermore, in step S2, a topology-aware repair operator is innovatively introduced into the standard NSGA-II framework. After crossover and mutation operations, this operator acts as a mandatory constraint layer to perform real-time connectivity detection on chromosomes, and uses a heuristic neighborhood merging strategy to forcibly eliminate disconnected nodes, ensuring that the domain division schemes are all physically strongly connected subgraphs. With the goal of maximizing intra-domain reliability and minimizing control overhead, a Pareto optimal solution set is generated through evolutionary search, and the baseline topology for the initial operation of the network is selected from it.

[0015] Furthermore, step S2 specifically includes the following steps: S21: Population Initialization and Chromosome Encoding Chromosome encoding uses integer encoding, and the length of the encoding is equal to the total number of satellite nodes. chromosome vector Among them, genes Representing the The domain ID of each satellite; An initial population was generated using a hybrid strategy. .in: 80% of individuals are based on physical orbits: satellites are pre-divided according to their physical orbital planes, and satellites in the same orbital plane are preferentially assigned to the same domain to conform to the physical connection characteristics of inter-satellite links; 20% of individuals are randomly generated: nodes are randomly assigned to domains to increase population diversity; S22: Design of dual optimization objective functions, defining two conflicting optimization objective functions: Objective 1: Maximize domain average reliability The goal is to increase the link connectivity density within each domain, ensuring uninterrupted communication within the domain even when some nodes fail. The calculation formula is as follows:

[0016] in: This represents the total number of domains. For the first The number of physical links actually existing within each domain; For the first The total number of nodes in each domain; the denominator is the theoretical upper limit of the number of fully connected links under this size; Objective 2: Minimize overall network control overhead The goal is to reduce the signaling costs associated with cross-domain communication and ensure load balancing across domains. The calculation formula is as follows:

[0017] in: For indicator functions, when the link The two connected nodes belong to different domains, that is... The value is 1 if it is true, and 0 otherwise. This represents the communication and management cost coefficient for cross-domain links. For all The variance of the number of nodes in each domain is used to penalize solutions with extremely unbalanced domain sizes; This is a load balancing weighting factor; S23: Evolutionary Search and Topology-Aware Repair: Deeply integrating graph theory repair mechanisms based on standard genetic operations; S24: Baseline Topology Decision: After After generational evolution, a set of Pareto optimal solutions is output; a weighted scoring method is used to select the best solution from the frontier solution set.

[0018] Select rating The highest individual As the static reference topology of the system.

[0019] Furthermore, step S23 specifically includes the following steps: S231: Evolutionary Iteration Operator Design: Binary tournament selection: Crowding distance and non-dominant ranking rank are used as joint criteria; two individuals are randomly selected from the parent population each time, with priority given to individuals with higher ranking rank; if the ranks are the same, individuals with greater crowding distance are selected to maintain population diversity. Multi-point crossover: Set crossover probability For a selected pair of parent chromosomes, multiple loci are randomly selected to exchange gene fragments, generating offspring. Adaptive mutation: Set the base mutation probability Genes on chromosomes Randomly flipping satellites means randomly changing their positions. The domain of belonging; the mutation rate with the number of iterations The increase in exhibits a non-linear decay. S232: Analyze and repair topological integrity constraints and islanding phenomena; S233: Topology-aware repair operator execution flow: This operator, as a mandatory post-processing step, is embedded after each generation of the population. 1) Connected component detection; for each region in the chromosome Construct its corresponding induced subgraph ; Use breadth-first search algorithm to traverse Calculate the set of connected components within this domain. If m=1, it means the domain is physically connected and no repair is needed; if m>1, it means the domain is broken and repair logic needs to be triggered. 2) Main domain determination and fragment identification: In the set In the process, the connected component with the most nodes is selected as the principal domain kernel, and the remaining smaller components are marked as free fragments. 3) Greedy merge strategy: For each node in the free fragment traverse its set of physical neighbor nodes Analyze the distribution of neighboring nodes' home regions and calculate the merge cost function:

[0020] in For nodes and target domains The number of physical links between them, the latter being the deviation of the target domain size from its average value; Execute the following action: Forcibly redistribute the free debris to In the smallest neighboring domain.

[0021] Furthermore, step S3 specifically includes the following steps: S31: Construct a dynamic spatiotemporal flow matrix model, including a dynamic flow matrix incorporating time, space, and smoothing factors. ,definition Time Node With nodes Flow intensity between for:

[0022] in, To measure the confidence coefficient; The prediction model is decomposed into:

[0023] Time tidal factor: simulates the periodic fluctuations in network throughput caused by human work and rest schedules, among which... The Earth's rotation period, For nodes Phase offset corresponding to the current covered time zone; Spatial drift factor: A Gaussian mixture model is used to characterize the traffic pull effect of ground service hotspots on transit satellites, in which... For the first The latitude and longitude centers of each hotspot area The coordinates of the satellite's nadir point. The radius of the hotspot coverage; Matrix update: Historical data is stored using a sliding window mechanism, and the above model parameters are periodically updated to form a global view; S32: Potential energy monitoring and heavy load link identification, introducing an exponentially weighted moving average to calculate the effective potential energy of the link, including link smoothing load calculation and heavy load link set screening; S33: Design of the transfer gain function, including establishing a multi-dimensional gain evaluation function. Based on a greedy strategy, quantize nodes. Merge into the target domain Net comprehensive utility value after:

[0024] The specific definitions are as follows: Flow internalization gain :node After the migration, the original cross-domain traffic becomes the sum of intra-domain traffic:

[0025] Traffic overflow penalty The negative consequences of node v migrating from its original domain to cross-domain traffic:

[0026] Migration total cost Ψ: Includes signaling overhead and penalties for load balancing constraints.

[0027] In the first item, For single-stream entry update overhead, The node degree reflects the signaling cost of notifying neighbors. The second item is the overload protection penalty; if the target domain load rate exceeds the safety threshold... This increases costs, thereby inhibiting migration behavior to the overloaded domain; S34: Dynamic election of the domain master satellite based on multi-dimensional weights, including: introducing the domain master satellite role in each logical domain, periodically calculating the comprehensive weight based on the real-time status of the nodes and executing the election mechanism, the calculation process is as follows: Node comprehensive weight calculation: for the current domain any satellite node in Define its voting weight for:

[0028] in: Degree centrality of nodes; It is a normalized index of the remaining computing resources and storage space of a node; This refers to the orbital priority coefficient. , These are non-negative weighting coefficients; Election and proxy responsibilities: Compare the performance of all nodes within the domain. Value, select The largest node is designated as the DHC agent node for the domain. The DHC is responsible for aggregating and compressing the link status information of ordinary nodes within the domain and reporting it. It also serves as the anchor point for issuing flow table update instructions from the SDN control plane and is responsible for broadcast distribution within the domain. The elected DHC node will take over the dynamic instruction distribution responsibility in step S1 and act as a signaling relay between the control system and ordinary satellites.

[0029] Furthermore, in step S32, the link smoothing load calculation includes: for each cross-domain link Its smooth load The updated formula is:

[0030] in As a smoothing factor, it effectively filters high-frequency jitter; The heavy-load link set filtering includes: defining a high-load threshold. Real-time filtering of link sets that meet the following conditions :

[0031] This means that not only is the load limit required, but also the duration. Exceeding the minimum time window ;gather The links in the diagram are marked as high-voltage lines, indicating that the domain boundaries must expand in the direction of outflow.

[0032] Furthermore, step S4 specifically includes the following steps: S41: Discrete-Time Slice State Monitoring and Triggering Model System on Discrete-Time Axis Run on; at every moment Define the state trigger indicator function Based on the set of heavy-load links output in step S32 The logic that triggers the judgment is as follows:

[0033] in Represents the cardinality of a set; For emergency trigger threshold; only when When =1, proceed to the next stage of fine-tuning; otherwise, the system maintains the static baseline configuration. ; S42: Constructing a candidate migration set and gradient sorting for each boundary node in the high-load link set. Calculate its direction to the target domain Migration gain Construct a candidate migration set It only includes nodes with positive returns:

[0034] For sets Perform a descending sort operation based on the gain gradient. Generate an ordered sequence :

[0035] This represents the steepest gradient direction for the release of traffic potential energy in the current network; S43: Multidimensional Topological Constraint Verification: Defining Boolean Constraint Functions For sequences Candidate nodes in the dataset must simultaneously satisfy the following three strict mathematical constraints: Connectivity preservation constraint: Definition To make nodes The induced subgraph of the original domain after removal. Using the cut vertex criterion in graph theory, we can determine the number of connected components after the original domain is split. Keep it at 1:

[0036] Domain capacity limit constraint: To prevent the controller from overloaded due to the expansion of the number of nodes in the target domain, the limit is set... The current number of nodes in the target domain. Maximum allowable deviation:

[0037] Ping-Pong Effect Suppression Constraints: Recording Nodes The last time the migration occurred The time difference between the current moment and the last migration must exceed the cooldown period. :

[0038] Only when both conditions are met Only nodes that meet the above criteria are marked as qualified. S44: Optimal Subset Execution Based on Braking Mechanism: A braking mechanism is introduced to model the fine-tuning problem as a cardinality-constrained maximum gain subset selection problem; the system executes from an ordered sequence... In the process, qualified nodes that meet the constraints are extracted sequentially in descending order of gain gradient until the maximum allowable number of migrations in a single step is reached. This completes the final execution set. :

[0039] The constraints are:

[0040] in The maximum number of migrations allowed per step; finally, the control system issues a command to update the entire network topology status:

[0041] This process ensures that the network achieves maximum traffic load balancing benefits with minimal topology disturbance in each time slice, and simultaneously triggers updates to the member list of DHC agents within the relevant domain.

[0042] The beneficial effects of this invention are as follows: It solves the problem of traffic misalignment in time and space: overcomes the shortcomings of static domain division in adapting to the "tidal effect" and "geographic hotspot drift" of ground business traffic, and eliminates cross-domain link congestion caused by boundary solidification; It breaks through the performance bottleneck of online computing: avoiding high-complexity global reconstruction (such as online genetic algorithms) during network operation, and achieving millisecond-level traffic hotspot response and congestion mitigation; It ensures long-term network topology stability: by "minimally invasive" boundary adjustments, it can absorb sudden traffic while controlling the range of network topology fluctuations locally, thereby reducing the overall network control signaling overhead. Increased overall network throughput: By converting cross-domain elephant flows into intra-domain traffic, bandwidth consumption of inter-domain interfaces was significantly reduced, improving resource utilization. A highly reliable self-healing architecture was constructed: forming a closed-loop system of "offline global planning as a safety net + online dynamic fine-tuning and adaptation" to ensure that the network connectivity and integrity are not compromised during any adjustment process.

[0043] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0044] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 A flowchart of a dynamic domain optimization method for satellite networks based on real-time traffic potential sensing; Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0045] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0046] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0047] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0048] Example 1: like Figure 1-2 As shown, this invention provides a dynamic domain optimization method for satellite networks based on real-time traffic potential sensing, comprising the following steps: S1: Construct a satellite network domain-based management and control system. The system adopts a Software-Defined Networking (SDN) architecture, divided into two main parts: the "SDN control plane" and the "satellite data plane." 1) The SDN control plane deploys a topology management module, an improved NSGA-II global planning engine, a traffic potential analyzer and dynamic migration controller, a DHC election manager, and a multi-dimensional constraint checker. 2) The satellite data plane is responsible for performing traffic forwarding, link load monitoring, and flow table updates. It also integrates a DHC agent within the satellite nodes and periodically reports link state information (LSA) to the control plane, providing a real-time topology view for global planning.

[0049] S2: Offline Static Benchmark Planning Based on Multi-Objective Evolution. Addressing the weakness of traditional evolutionary algorithms in generating disconnected "island solutions" during graph partitioning, this invention innovatively introduces a topology-aware repair operator within the standard NSGA-II framework. After crossover and mutation operations, this operator acts as a mandatory constraint layer to perform real-time connectivity checks on chromosomes, using a heuristic neighborhood merging strategy to forcibly eliminate disconnected nodes, ensuring that all partitioning schemes are physically strongly connected subgraphs. The algorithm aims to maximize intra-domain reliability and minimize control overhead, generating a Pareto optimal solution set through evolutionary search, and selecting the best solution from this set to determine the initial benchmark topology for network operation.

[0050] S3: Establish a traffic spatiotemporal potential energy perception and domain master autonomous model. Its core objective is to capture in real time the traffic center of gravity shift caused by the Earth's rotation and service tides in the network, and establish the physical basis for domain autonomous management. Definitions: 1) Traffic matrix model: A dynamic load matrix that integrates time tide factors and spatial drift factors (GMM), abandoning static assumptions; 2) Traffic potential energy monitoring mechanism: Based on the aforementioned definition of traffic potential energy, the EWMA (Exponential Weighted Moving Average) smoothing algorithm is used to quantify the real-time load accumulation degree of cross-domain links, thereby identifying "heavy load links (Heavy Hitters)" whose potential energy intensity exceeds a preset threshold; 3) Migration gain function: Based on a greedy strategy, the difference between the "traffic internalization benefit" and the "signaling migration cost" after moving boundary nodes into adjacent domains is calculated; 4) Domain master election mechanism: The DHC election manager in the control plane dynamically elects satellite nodes with DHC agent functions based on node degree centrality, remaining resources, and orbital priority to achieve local aggregation of signaling within the domain and "micro-autonomy" management.

[0051] S4: Online Closed-Loop Fine-Tuning Based on Traffic Potential Gradient. The main design involves sequentially completing the following steps within each discrete time slice: 1) Status Monitoring: Collecting cross-domain link traffic data; 2) Threshold Decision: Using a traffic potential analyzer to filter out traffic exceeding a threshold. The key cross-domain flows; 3) Topology fine-tuning: cross-domain migration is performed on boundary nodes with high migration gain values, and the multi-dimensional constraint checker performs constraint checks such as connectivity, domain size, and ping-pong effect; 4) Braking mechanism: the maximum number of migrations in a single step is limited to prevent route oscillation. This section describes the online adjustment process at runtime, which forms a "dynamic and static combination" logical closed loop with the offline baseline planning in S2.

[0052] In one optional implementation, step S1 includes: S11: Construct satellite data plane node modules. Each low-Earth orbit satellite node. (in In this embodiment =288) integrates traffic monitoring, flow table execution, and status reporting functions.

[0053] 1. Service Traffic Forwarding and Monitoring: Satellite nodes, acting as execution units, are responsible for handling packet forwarding on the inter-satellite link (ISL). For any data flow, the node determines the next-hop node based on its local flow table entries. The domain status. If This is then marked as cross-domain traffic. The node's embedded counters provide real-time statistics on the throughput rate of each port. .

[0054] 2. Node Status Awareness Module: Used to sense and report link load and topology status, and construct local status vectors. :

[0055] in: Link utilization : Represents the current load saturation of the link; Neighbor node set : Represents the list of physically visible neighbors with established links at the current moment; Current home domain identifier In this embodiment, the entire network is initially divided into: K =12 fields.

[0056] 3. Dynamic Command Execution Engine: Satellite nodes receive "domain migration commands" from the control plane via the southbound interface. When a command containing... When the instruction is given, the node immediately updates its local data. It is then broadcast to neighbors, completing the logical domain switching within milliseconds without the need for complex route recalculation.

[0057] S12: SDN Domain-Specific Control Center Design. The control center (deployed at a ground station or backbone satellite) is the "brain" of the system, responsible for global resource management. To achieve a combined static and dynamic control strategy, the center is designed with a modular architecture: Topology Management Module: Responsible for maintaining the global network view. This is achieved by collecting data reported by each node. Real-time construction of dynamic connection graphs And based on the satellite ephemeris, physical topological changes are deduced.

[0058] NSGA-II Global Planning Engine: This is the system's offline decision-making core. Configured to run during system initialization or periodic refactoring, this engine encapsulates a dual optimization objective function: maximizing domain-average reliability and minimizing overall network control overhead. This engine is responsible for outputting the static baseline domain-specific scheme. It provides a highly reliable basic framework for network operation.

[0059] Traffic Potential Analysis and Dynamic Migration Controller: This is the core of the system's online execution. This module includes a "Traffic Matrix Builder" and a "Potential Gradient Analyzer." It calculates the potential energy difference of cross-domain traffic in real time during runtime slots, identifies "Heavy Hitters" in the network, and generates fine-tuning instructions to be sent to the data plane.

[0060] S13: Construct a highly reliable communication interface. To ensure efficient interaction between the control plane and the data plane, two types of dedicated channels are designed: Telemetry Channel: Employs a lightweight UDP / gRPC protocol. Satellite nodes periodically report Link Status Information (LSA) and traffic statistics to ensure the timeliness of the control layer view.

[0061] Control Channel: Employs a highly reliable TCP long-connection protocol. Used for transmitting "domain update messages" and "flow table change commands" from the control center. This interface introduces an application-layer acknowledgment (ACK) mechanism to ensure that commands are not lost in long-latency satellite-to-ground links. Furthermore, the interface integrates a heartbeat keep-alive mechanism. If a node reaches a set threshold (e.g., ...), the heartbeat keeps-alive mechanism will be activated. If no central heartbeat is received within a certain period, the network will automatically revert to the baseline configuration of the previous version to prevent the network from being paralyzed due to control failure.

[0062] In one optional implementation, step S2 specifically includes: S21: Population Initialization and Chromosome Encoding. To map the satellite network domain division problem into a form that can be handled by evolutionary algorithms, encoding and population initialization are performed first: Chromosome encoding: Integer encoding is used, and the constructed length is... Chromosome vector of (total number of satellite nodes) Among them, genes Representing the The domain ID of each satellite.

[0063] Hybrid Initialization Strategy: To improve the algorithm's convergence speed and ensure the physical feasibility of the initial solution, a hybrid strategy of "orbit priority + random perturbation" is adopted to generate the initial population. .in: 80% Individual (based on physical orbit): Pre-divided according to the physical orbital plane of the satellite, satellites in the same orbital plane are preferentially assigned to the same domain to conform to the physical connection characteristics of inter-satellite links.

[0064] 20% Individuals (randomly generated): Randomly assign domain affiliation to nodes to increase population diversity and prevent the algorithm from getting trapped in local optima too early.

[0065] S22: Design of a Dual Optimization Objective Function To balance network robustness and management efficiency, this step defines two conflicting optimization objective functions: Objective 1: Maximize domain average reliability ( The goal is to increase the link connectivity density within each domain, ensuring uninterrupted communication within the domain even when some nodes fail. The calculation formula is as follows:

[0066] in: This represents the total number of domains. For the first The number of physical links actually existing within each domain; For the first The total number of nodes within a domain; the denominator is the theoretical upper limit of the number of fully connected links under this scale.

[0067] Objective 2: Minimize the overall network control overhead ( This objective aims to reduce signaling costs associated with cross-domain communication and ensure load balancing across domains. The calculation formula is as follows:

[0068] in: For indicator functions, when the link The two nodes being connected belong to different domains (i.e.) The value is 1 when the condition is met, and 0 otherwise. This represents the communication and management cost coefficient for cross-domain links. For all The variance of the number of nodes in each domain is used to penalize solutions with extremely unbalanced domain sizes; This is the load balancing weighting factor.

[0069] S23: Evolutionary Search and Topology-Aware Repair To evolve a better partitioning scheme from the initial population, the algorithm enters an iterative loop. Considering the discreteness of satellite network topology and the strong constraints of graph structure, this step deeply integrates a graph theory repair mechanism on top of standard genetic operations. The specific process is as follows: 1. Evolutionary Iteration Operator Design: Binary Tournament Selection: This method uses crowding distance and non-dominant rank as a joint criterion. Each time, two individuals are randomly selected from the parent population. Priority is given to individuals with higher rank (i.e., closer to the Pareto front); if the ranks are the same, the individual with a larger crowding distance (i.e., sparser solution distribution) is selected to maintain population diversity.

[0070] Multi-point Crossover: Setting crossover probabilities (e.g., 0.8). For a selected pair of parent chromosomes, multiple loci are randomly selected to exchange gene fragments, generating offspring, thereby exploring new solution spaces while preserving superior gene patterns.

[0071] Adaptive Mutation: Sets the base mutation probability. (e.g., 0.1). For genes on chromosomes. Perform random flipping (i.e., randomly change the satellite) (The domain of the region to which the mutation rate belongs). To prevent later oscillations, the mutation rate increases with the number of iterations. The increase in exhibits a non-linear decay.

[0072] 2. Analysis of Topological Integrity Constraints and the "Island" Phenomenon: Due to the randomness of crossover and mutation operations, the physical continuity of domains is easily disrupted, leading to the creation of "island nodes." That is, a certain domain... The set of nodes within the domain is split into multiple unconnected sub-clusters on the physical topology graph. This disconnected structure prevents the domain controller from managing all nodes within the domain via inter-satellite links, making the solution impractical in engineering. Therefore, a remediation mechanism must be introduced.

[0073] 3. Topology-aware Repair Operator Execution Flow: This operator, as a mandatory post-processing step, is embedded after each generation of the population. Step 1): Component Detection. For each region in the chromosome... Construct its corresponding induced subgraph The breadth-first search (BFS) algorithm is used to traverse the traversal. (An equivalent algorithm, such as disjoint-set data structure, can also be used) to calculate the set of connected components within the domain. .

[0074] If m=1, it means that the domain is physically connected and no repair is needed; If m>1, it indicates that there is a break in the domain, and the repair logic needs to be triggered.

[0075] Step 2): Main domain determination and fragment identification. In the set In this process, the connected component with the most nodes is selected as the "Main Core", and the remaining smaller components are marked as "Fragments".

[0076] Step 3): Greedy Merging Strategy. For each node in the free fragment... The system traverses its set of physical neighbor nodes. Calculate the distribution of neighboring nodes' home regions. Calculate the merge cost function:

[0077] in For nodes and target domains The number of physical links between them (connection strength), and the next item is the deviation of the target domain size from its average value (load balancing item).

[0078] Execute the following action: Force the free debris to be redistributed to In the smallest (i.e., the most physically connected and least loaded) neighboring domain.

[0079] Through the above repairs, it is ensured that all individuals in each generation of the population satisfy 100% connectivity constraints in terms of physical topology, avoiding the waste of search resources by invalid solutions and significantly improving the convergence efficiency of the algorithm.

[0080] The algorithm pseudocode is shown below:

[0081] S24: Baseline Topology Decision go through After generational evolution, the algorithm outputs a set of Pareto optimal solutions. To select the baseline solution for final deployment, the system uses a weighted scoring method to select the best solution from the frontier solution set:

[0082] Select rating The highest individual This serves as the system's static baseline topology. This scheme will be distributed to all satellite nodes as the default operating configuration.

[0083] In one optional implementation, step S3 specifically includes: S31: Constructing a dynamic spatiotemporal traffic model To accurately describe the complex patterns of low-Earth orbit satellite network traffic evolution over time and spatial coverage changes, the system abandons static assumptions and constructs a dynamic traffic matrix that incorporates temporal, spatial, and smoothing factors. Definition Time Node With nodes Flow intensity between for:

[0084] in, To measure the confidence coefficient. The prediction model is further decomposed into:

[0085] Time Tidal Factor: Simulates the periodic fluctuations in network throughput caused by human activity patterns. The Earth's rotation period, For nodes Phase offset corresponding to the current covered time zone.

[0086] Spatial drift factor: Gaussian mixture model (GMM) is used to characterize ground business hotspots (such as large urban clusters). The flow pull effect on transiting satellites. For the first The latitude and longitude centers of each hotspot area The coordinates of the satellite's nadir point. This represents the hotspot coverage radius. This parameter accurately reflects the dramatic difference in payload when the satellite flies over cities and oceans.

[0087] Matrix Update: The control center uses a sliding window mechanism to store historical data and periodically updates the above model parameters to form a global view.

[0088] S32: Potential Monitoring Strategy The "traffic potential energy" in the network is defined as the degree of load accumulation on cross-domain links. To avoid misjudgments caused by instantaneous traffic spikes, this system introduces an exponentially weighted moving average (EWMA) mechanism to calculate the "effective potential energy" of the link.

[0089] 1. Link Smoothing Load Calculation: For each cross-domain link... Its smooth load The updated formula is:

[0090] in This is a smoothing factor (with a value of 0.1). 0.3), effectively filtering high-frequency jitter.

[0091] 2. Filtering of Heavy Hitters: Define high load thresholds. Real-time filtering to select a set of links that meet the following conditions.

[0092]

[0093] This means that not only is the load limit required, but also the duration. Exceeding the minimum time window .gather The links in the diagram are marked as “potential energy high-voltage lines,” indicating that the domain boundaries must expand in the direction of outflow to release the accumulated potential energy.

[0094] S33: Migration Gain Function Design To scientifically determine whether boundary nodes should migrate across domains, this invention establishes a multi-dimensional gain evaluation function. This function uses a greedy strategy to quantize the nodes. Merge into the target domain The net value of overall utility after the calculation.

[0095]

[0096] The specific definitions are as follows: 1. Flow internalization gain :node After the migration, the original cross-domain traffic becomes the sum of intra-domain traffic:

[0097] This directly reduces the expensive overhead of cross-domain control.

[0098] 2. Traffic overflow penalties The negative consequences of node v migrating, where traffic within the original domain becomes cross-domain traffic:

[0099] 3. Overall migration cost Ψ: Includes signaling overhead and penalties for load balancing constraints.

[0100] First item: For single-stream entry update overhead, The node degree reflects the signaling cost of notifying neighbors.

[0101] The second item is an overload protection penalty. If the target domain load rate exceeds the safety threshold... This would drastically increase costs, thereby inhibiting migration behavior to the overloaded domain. S34: Dynamic DHC Election of Domain Master Satellite Based on Multidimensional Weights To reduce latency overhead caused by frequent satellite-to-ground interactions and to achieve "micro-autonomy" management within each domain, this invention introduces a "Domain Head Controller (DHC)" role within each logical domain. The system periodically calculates the overall weight and executes an election mechanism based on the real-time status of the nodes.

[0102] 1. Node comprehensive weight calculation: For the current domain any satellite node in Define its voting weight for:

[0103] in: Degree centrality of nodes; The normalization index of the node's remaining computing resources and storage space (range 0~1). This refers to the orbital priority coefficient. , Non-negative weighting coefficients 2. Election and Proxy Responsibilities: The control center or distributed protocol compares the performance of all nodes within the domain. Value, select The largest node is designated as the DHC proxy node for that domain. The DHC is responsible for aggregating and compressing the Link State Information (LSA) of ordinary nodes within the domain before reporting it. It also serves as the "anchor point" for issuing flow table update commands from the SDN control plane, handling broadcast distribution within the domain, thereby significantly reducing the frequency of signaling interactions between satellite and ground links. The elected DHC node will assume the dynamic command distribution responsibility described in S11, acting as a signaling relay between the control center and ordinary satellites.

[0104] In one optional implementation, step S4 specifically includes: S41: Discrete-time Monitoring and Triggering Model The system operates on a discrete-time axis. Running on. At every moment. Define the state trigger indicator function Based on the set of heavy-load links output in step S32 The logic that triggers the judgment is as follows:

[0105] in: Represents the cardinality of a set; Emergency trigger threshold Only when When =1, proceed to the next stage of fine-tuning; otherwise, the system maintains the static baseline configuration. .

[0106] S42: Constructing the Candidate Migration Set and Gradient Ranking (Candidate Set Construction) for each boundary node in the high-load link set. Calculate its direction to the target domain Migration gain (As defined by S33). Construct a candidate migration set. It only includes nodes with positive returns:

[0107] For sets Perform a descending sort operation based on the gain gradient. Generate an ordered sequence :

[0108] This represents the "steepest gradient direction" of the current network's potential energy release.

[0109] S43: Topology Constraint Validation. To ensure the physical feasibility of fine-tuning operations, Boolean constraint functions are defined. For sequences Candidate nodes in the dataset must simultaneously satisfy the following three strict mathematical constraints: Connectivity Preservation Constraint: Definition To make nodes The induced subgraph of the original domain after removal. Using the cut-vertex check in graph theory, we can determine the number of connected components after the original domain is split. Keep it at 1:

[0110] Domain capacity constraint: Prevents controller overload caused by the excessive number of nodes in the target domain. Let... The current number of nodes in the target domain. Maximum allowable deviation:

[0111] Ping-Pong Effect Suppression Constraint: Recording Nodes The last time the migration occurred The time difference between the current moment and the last migration must exceed the cooldown period. :

[0112] Only when both conditions are met Only nodes that meet the criteria are marked as "qualified".

[0113] S44: Braking & Execution Based on Braking Mechanism. To prevent routing oscillations, this step introduces a "braking mechanism," modeling the fine-tuning problem as a cardinality-constrained maximum gain subset selection problem. The system starts from an ordered sequence... In the process, qualified nodes that meet the constraints are extracted sequentially in descending order of gain gradient until the maximum allowable number of migrations in a single step is reached. This completes the final execution set. :

[0114] The constraints are:

[0115] in This represents the maximum number of migrations allowed per step. Finally, the control center issues a command to update the entire network topology status:

[0116] This process ensures that the network achieves maximum traffic load balancing benefits with minimal topology disturbance in each time slice, and simultaneously triggers updates to the member list of DHC agents within the relevant domain.

[0117] The algorithm pseudocode is shown below:

[0118] Example 2: An electronic device, comprising a memory and a processor; The memory is used to store computer programs; The processor is configured to implement the method described in Embodiment 1 when executing the computer program.

[0119] Example 3: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in Embodiment 1.

[0120] Example 4: A computer program product includes a computer program that, when executed by a processor, implements the method described in Example 1.

[0121] In the above embodiments, the reference to "this embodiment" in the specification indicates that a specific feature, structure, or characteristic described in connection with the embodiment is included in at least some embodiments, but not necessarily all embodiments. Multiple appearances of "this embodiment" do not necessarily refer to the same embodiment.

[0122] In the above embodiments, although the invention has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory structures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed. The embodiments of the invention are intended to cover all such substitutions, modifications, and variations falling within the broad scope of the appended claims.

[0123] As will be understood by those skilled in the art, the computer-readable storage medium described in this embodiment allows for the implementation of all or part of the steps in the above method embodiments by computer program-related hardware. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0124] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication between them. The memory is used to store computer programs, the communication interface is used to perform communication, and the processor and the transceiver are used to run the computer programs, so that the electronic terminal performs the steps of the above method.

[0125] In this embodiment, the memory may include random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device.

[0126] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0127] This invention can be used in a wide range of general-purpose or special-purpose computing system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0128] This invention can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A dynamic domain optimization method for satellite networks based on real-time traffic potential sensing, characterized in that: Includes the following steps: S1: Construct a satellite network domain-based management and control system; S2: Offline static baseline planning based on multi-objective evolution; S3: Establish a spatiotemporal potential energy perception and domain-controlled autonomous model for traffic flow; S4: Online closed-loop fine-tuning based on flow potential gradient.

2. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 1, characterized in that: In step S1, a satellite network domain management and control system is constructed using a Software Defined Network (SDN) architecture, including an SDN control plane and a satellite data plane. The SDN control plane is equipped with a topology management module, an improved NSGA-II global planning engine, a traffic potential analyzer, a dynamic migration controller, a DHC election manager, and a multi-dimensional constraint checker. The satellite data plane is responsible for performing traffic forwarding, link load monitoring, and flow table updates, and integrates a DHC agent within the satellite nodes to periodically report Link State Information (LSA) to the control plane, providing a real-time topology view for global planning.

3. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 2, characterized in that: In the satellite data plane section, each low-Earth orbit satellite node , It integrates traffic monitoring, flow table execution, and status reporting functions, specifically including: Traffic forwarding and monitoring: Satellite nodes act as execution units, responsible for forwarding data packets for the inter-satellite link (ISL); for any data flow, the node determines the next-hop node based on local flow table entries. The home domain status; if This is then marked as cross-domain traffic; the node's embedded counters provide real-time statistics on the throughput rate of each port. ; Node Status Awareness Module: Used to sense and report link load and topology status, and construct local status vectors. : Among them: link utilization Characterizing the current load saturation of the link; set of neighboring nodes This represents the list of physically visible neighbors with established links at the current time; the current home domain identifier. The entire network is initially divided into K Each domain; Dynamic instruction execution engine: Satellite nodes receive domain migration instructions from the control plane via the southbound interface; upon receiving the instruction, the node immediately updates its local... It is then broadcast to neighbors, completing the logical domain switching.

4. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 3, characterized in that: The SDN control plane portion includes: Topology Management Module: Responsible for maintaining a global network view; collecting data reported by each node. Real-time construction of dynamic connection graphs And infer physical topological changes based on satellite ephemeris; The NSGA-II global planning engine is configured to run during system initialization or periodic refactoring phases. Internally, it encapsulates a dual optimization objective function that maximizes domain-average reliability and minimizes overall network control overhead. This engine is responsible for outputting a static baseline domain-specific scheme. It provides the basic framework for network operation; Traffic potential energy analysis and dynamic migration controller: includes a traffic matrix builder and a potential energy gradient analyzer; the controller calculates the potential energy difference of cross-domain traffic in real time during runtime slots, identifies heavily loaded links in the network, and generates fine-tuning instructions to be sent to the data plane.

5. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 1, characterized in that: In step S2, a topology-aware repair operator is innovatively introduced into the standard NSGA-II framework. After crossover and mutation operations, this operator acts as a mandatory constraint layer to perform real-time connectivity detection on chromosomes and uses a heuristic neighborhood merging strategy to forcibly eliminate disconnected nodes, ensuring that all domain division schemes are physically strongly connected subgraphs. With the goal of maximizing intra-domain reliability and minimizing control overhead, a Pareto optimal solution set is generated through evolutionary search, and the baseline topology for the initial operation of the network is selected from this set.

6. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 5, characterized in that: Step S2 specifically includes the following steps: S21: Population Initialization and Chromosome Encoding Chromosome encoding uses integer encoding, and the length of the encoding is equal to the total number of satellite nodes. chromosome vector Among them, genes Representing the The domain ID of each satellite; An initial population was generated using a hybrid strategy. ,in: 80% of individuals are based on physical orbits: satellites are pre-divided according to their physical orbital planes, and satellites in the same orbital plane are preferentially assigned to the same domain to conform to the physical connection characteristics of inter-satellite links; 20% of individuals are randomly generated: nodes are randomly assigned to domains to increase population diversity; S22: Design of dual optimization objective functions, defining two conflicting optimization objective functions: Objective 1: Maximize domain average reliability The goal is to increase the link connectivity density within each domain, ensuring uninterrupted communication within the domain even when some nodes fail. The calculation formula is as follows: in: This represents the total number of domains. For the first The number of physical links actually existing within each domain; For the first The total number of nodes in each domain; the denominator is the theoretical upper limit of the number of fully connected links under this size; Objective 2: Minimize overall network control overhead The goal is to reduce the signaling costs associated with cross-domain communication and ensure load balancing across domains. The calculation formula is as follows: in: For indicator functions, when the link The two connected nodes belong to different domains, that is... The value is 1 if it is true, and 0 otherwise. This represents the communication and management cost coefficient for cross-domain links. For all The variance of the number of nodes in each domain is used to penalize solutions with extremely unbalanced domain sizes; This is a load balancing weighting factor; S23: Evolutionary Search and Topology-Aware Repair: Deeply integrating graph theory repair mechanisms based on standard genetic operations; S24: Baseline Topology Decision: After After generational evolution, a set of Pareto optimal solutions is output; a weighted scoring method is used to select the best solution from the frontier solution set. Select rating The highest individual As the static reference topology of the system.

7. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 6, characterized in that: Step S23 specifically includes the following steps: S231: Evolutionary Iteration Operator Design: Binary tournament selection: Crowding distance and non-dominant ranking rank are used as joint criteria; two individuals are randomly selected from the parent population each time, with priority given to individuals with higher ranking rank; if the ranks are the same, individuals with greater crowding distance are selected to maintain population diversity. Multi-point crossover: Set crossover probability For a selected pair of parent chromosomes, multiple loci are randomly selected to exchange gene fragments, generating offspring. Adaptive mutation: Set the base mutation probability Genes on chromosomes Randomly flipping satellites means randomly changing their positions. The domain of belonging; the mutation rate with the number of iterations The increase in exhibits a non-linear decay. S232: Analyze and repair topological integrity constraints and islanding phenomena; S233: Topology-aware repair operator execution flow: This operator, as a mandatory post-processing step, is embedded after each generation of the population. 1) Connected component detection; for each region in the chromosome Construct its corresponding induced subgraph ; Use breadth-first search algorithm to traverse Calculate the set of connected components within this domain. If m=1, it means the domain is physically connected and no repair is needed; if m>1, it means the domain is broken and repair logic needs to be triggered. 2) Main domain determination and fragment identification: In the set In the process, the connected component with the most nodes is selected as the principal domain kernel, and the remaining smaller components are marked as free fragments. 3) Greedy merge strategy: For each node in the free fragment traverse its set of physical neighbor nodes Analyze the distribution of neighboring nodes' home regions and calculate the merge cost function: in For nodes and target domains The number of physical links between them, the latter being the deviation of the target domain size from its average value; Execute the following action: Forcibly redistribute the free debris to In the smallest neighboring domain.

8. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 1, characterized in that: Step S3 specifically includes the following steps: S31: Construct a dynamic spatiotemporal flow matrix model, including a dynamic flow matrix incorporating time, space, and smoothing factors. ,definition Time Node With nodes Flow intensity between for: in, To measure the confidence coefficient; The prediction model is decomposed into: Time tidal factor: simulates the periodic fluctuations in network throughput caused by human work and rest schedules, among which... The Earth's rotation period, For nodes Phase offset corresponding to the current covered time zone; Spatial drift factor: A Gaussian mixture model is used to characterize the traffic pull effect of ground service hotspots on transit satellites, in which... For the first The latitude and longitude centers of each hotspot area The coordinates of the satellite's nadir point. The radius of the hotspot coverage; Matrix update: Historical data is stored using a sliding window mechanism, and the above model parameters are periodically updated to form a global view; S32: Potential energy monitoring and heavy load link identification, introducing an exponentially weighted moving average to calculate the effective potential energy of the link, including link smoothing load calculation and heavy load link set screening; S33: Design of the transfer gain function, including establishing a multi-dimensional gain evaluation function. Based on a greedy strategy, quantize nodes. Merge into the target domain Net comprehensive utility value after: The specific definitions are as follows: Flow internalization gain :node After the migration, the original cross-domain traffic becomes the sum of intra-domain traffic: Traffic overflow penalty The negative consequences of node v migrating from its original domain to cross-domain traffic: Migration total cost Ψ: Includes signaling overhead and penalties for load balancing constraints. In the first item, For single-stream entry update overhead, The node degree reflects the signaling cost of notifying neighbors. The second item is the overload protection penalty; if the target domain load rate exceeds the safety threshold... This increases costs, thereby inhibiting migration behavior to the overloaded domain; S34: Dynamic election of the domain master satellite based on multi-dimensional weights, including: introducing the domain master satellite role in each logical domain, periodically calculating the comprehensive weight based on the real-time status of the nodes and executing the election mechanism, the calculation process is as follows: Node comprehensive weight calculation: for the current domain any satellite node in Define its voting weight for: in: Degree centrality of nodes; It is a normalized index of the remaining computing resources and storage space of a node; This refers to the orbital priority coefficient. , These are non-negative weighting coefficients; Election and proxy responsibilities: Compare the performance of all nodes within the domain. Value, select The largest node is designated as the DHC agent node for the domain. The DHC is responsible for aggregating and compressing the link status information of ordinary nodes within the domain and reporting it. It also serves as the anchor point for issuing flow table update instructions from the SDN control plane and is responsible for broadcast distribution within the domain. The elected DHC node will take over the dynamic instruction distribution responsibility in step S1 and act as a signaling relay between the control system and ordinary satellites.

9. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 8, characterized in that: In step S32, the link smoothing load calculation includes: for each cross-domain link Its smooth load The updated formula is: in As a smoothing factor, it effectively filters high-frequency jitter; The heavy-load link set filtering includes: defining a high-load threshold. Real-time filtering of link sets that meet the following conditions : This means that not only is the load limit required, but also the duration. Exceeding the minimum time window ;gather The links in the diagram are marked as high-voltage lines, indicating that the domain boundaries must expand in the direction of outflow.

10. The satellite network dynamic domain optimization method based on real-time traffic potential sensing according to claim 1, characterized in that: Step S4 Specifically, the following steps are included: S41: Discrete-Time Slice State Monitoring and Triggering Model System on Discrete-Time Axis Run on; at every moment Define the state trigger indicator function Based on the set of heavy-load links output in step S32 The logic that triggers the judgment is as follows: in Represents the cardinality of a set; For emergency trigger threshold; only when When =1, proceed to the next stage of fine-tuning; otherwise, the system maintains the static baseline configuration. ; S42: Constructing a candidate migration set and gradient sorting for each boundary node in the high-load link set. Calculate its direction to the target domain Migration gain Construct a candidate migration set It only includes nodes with positive returns: For sets Perform a descending sort operation based on the gain gradient. Generate an ordered sequence : This represents the steepest gradient direction for the release of traffic potential energy in the current network; S43: Multidimensional Topological Constraint Verification: Defining Boolean Constraint Functions For sequences Candidate nodes in the dataset must simultaneously satisfy the following three strict mathematical constraints: Connectivity preservation constraint: Definition To make nodes The induced subgraph of the original domain after removal. Using the cut vertex criterion in graph theory, we can determine the number of connected components after the original domain is split. Keep it at 1: Domain capacity limit constraint: To prevent the controller from overloaded due to the expansion of the number of nodes in the target domain, the limit is set... The current number of nodes in the target domain. Maximum allowable deviation: Ping-Pong Effect Suppression Constraints: Recording Nodes The last time the migration occurred The time difference between the current moment and the last migration must exceed the cooldown period. : Only when both conditions are met Only nodes that meet the above criteria are marked as qualified. S44: Optimal Subset Execution Based on Braking Mechanism: A braking mechanism is introduced to model the fine-tuning problem as a cardinality-constrained maximum gain subset selection problem; the system executes from an ordered sequence... In the process, qualified nodes that meet the constraints are extracted sequentially in descending order of gain gradient until the maximum allowable number of migrations in a single step is reached. This completes the final execution set. : The constraints are: in The maximum number of migrations allowed per step; finally, the control system issues a command to update the entire network topology status: This process ensures that the network achieves maximum traffic load balancing benefits with minimal topology disturbance in each time slice, and simultaneously triggers updates to the member list of DHC agents within the relevant domain.

Images