A satellite gateway station device resource dynamic priority allocation method based on graph theory

Abstract

The application relates to the field of operation and maintenance management of a satellite gateway station, and discloses a satellite gateway station equipment resource dynamic priority allocation method based on graph theory, which is suitable for a satellite communication gateway station and relates to the field of operation and maintenance management of the satellite gateway station, equipment resource dynamic allocation, multi-service link adaptive construction, graph theory modeling and a dynamic priority algorithm. The method mainly realizes closed-loop control of resource allocation and task execution state by constructing a resource modeling module, a priority evaluation engine, a graph theory algorithm engine and a link control module. The application has the characteristics of clarity, good continuity and strong scalability, and can especially solve the problems of unattended operation and automatic task execution of the current satellite communication gateway station.

Description

Technical Field

[0001] This invention applies to satellite communication gateway stations, specifically in the field of gateway station operation and maintenance management, focusing on dynamic allocation of equipment resources, adaptive construction of multi-service links, graph theory modeling, and dynamic priority algorithms. It primarily achieves closed-loop control of resource allocation and task execution status by constructing a resource modeling module, a priority evaluation engine, a graph theory algorithm engine, and a link control module. Background Technology

[0002] In areas such as oceans, lakes, deserts, deep mountains, and low altitudes, traditional terrestrial communication cannot provide coverage, yet the need for information exchange remains. Thus, satellite communication emerged. Initially used in the military field, satellite communication gradually expanded to the civilian sector, initially primarily in maritime transport and rescue operations, and in recent years has extended to internet applications.

[0003] As a branch of satellite mobile communication systems, emerging satellite internet constellations can provide broadband data services and realize internet transmission functions, attracting global attention and rising to the level of national strategy.

[0004] Satellite communication gateway stations, as key facilities in satellite internet constellations, are responsible for converting signals from satellites into data that can be processed by terrestrial networks, serving as crucial hubs between satellites and terrestrial networks. Satellite internet constellations typically deploy a comprehensive terrestrial control architecture, establishing a highly integrated and scalable terrestrial integrated control system to achieve integrated satellite-to-ground control, satellite-to-ground interconnection, and terrestrial network interconnection, meeting the global service needs of multiple users and multiple services.

[0005] Resource scheduling and management within the gateway station, along with high system resource utilization and low power consumption, have become important components of the ground-based integrated control system and key elements for achieving space-to-ground interconnection. Summary of the Invention

[0006] To address the problems of chaotic resource allocation, unbalanced service load, inability to automatically plan and adjust resources, and difficulty in achieving unattended operation of satellite gateway station equipment, this invention proposes a dynamic priority allocation method for satellite gateway station equipment resources based on graph theory.

[0007] The technical solution of this invention is as follows:

[0008] A graph theory-based method for dynamic priority allocation of satellite gateway station equipment resources includes the following steps:

[0009] Step 1: Based on the collected parameters and status of the gateway station equipment, the communication resources of the gateway station are abstracted into a connection topology graph. Nodes identify gateway station equipment, edges identify available links, and weights are assigned to the edges, including latency, bandwidth, and load rate.

[0010] Step 2: Traverse the nodes in the connection topology graph, obtain all reachable links in the graph as a candidate link set, and filter the candidate link set according to the constraints and the characteristics of the gateway station;

[0011] Step 3: Construct the relation matrix of the filtered link set, including the adjacency matrix, interference matrix and resource matrix, and generate a weighted adjacency matrix.

[0012] Step 4: Based on the matrix generated in Step 3, identify path conflict problems in resource allocation. For the identified conflicts, use the minimum cut set algorithm to resolve the conflicts and use the LSTM prediction model to predict the conflicts.

[0013] Step 5: Based on business needs, user type, and link status, perform dynamic priority evaluation and calculate dynamic priority weight and inherent link weight.

[0014] Step 6: Integrate dynamic priority weights and link-specific weights to form a comprehensive weight, and select the optimal path from the candidate paths based on the comprehensive weight;

[0015] Step 7: Perform adaptive link construction and maintenance based on the optimal path selected in Step 6.

[0016] Furthermore, step 1 specifically involves:

[0017] Step 101: Abstract the communication resources of the gateway station into a connection topology graph. Nodes identify gateway station devices, edges identify available links, and weights are assigned to the edges, including latency, bandwidth, and load rate.

[0018] Step 102: Collect parameters and status of the gateway station equipment through the monitoring protocol;

[0019] Step 103: Based on the collected parameters and status of the gateway station equipment, determine whether the links between the devices are reachable and dynamically construct the connection topology.

[0020] Furthermore, step 2 specifically involves:

[0021] Step 201: Using a depth-first approach, traverse the nodes in the connection topology graph, record the paths and backtrack to obtain all reachable links in the graph as a set of candidate links.

[0022] Step 202: Set the rules that need to be constrained, and filter available links from the candidate link set; and pre-set the device components in the link as needed to filter available links from the candidate links.

[0023] Furthermore, step 3 specifically involves:

[0024] Step 301: Construct a relationship matrix, including an adjacency matrix, for the filtered available links. Interference matrix and resource matrix Adjacency matrix Indicates the direct connection relationship between nodes, interference matrix The resource matrix represents the degree of interference between different paths passing through the same node. Record parameters between nodes, including latency, bandwidth, and load rate;

[0025] Step 302, adjust the adjacency matrix. Interference matrix and resource matrix Dynamic fusion generates a weighted adjacency matrix, the formula of which is:

[0026]

[0027] in, , and These are the weighting coefficients.

[0028] Furthermore, step 4 specifically involves:

[0029] Step 401: Based on the weighted adjacency matrix and resource matrix generated in step 3, use graph theory methods to identify resource allocation conflicts, specifically including determining whether there are multiple candidate paths sharing the same node or link resource, exceeding its capacity.

[0030] Step 402, Conflict resolution: For the identified conflicts, the minimum cut set algorithm is used to divide the resource allocation area, or the path selection is adjusted according to the preset conflict resolution strategy to make the resource allocation conflict-free.

[0031] Conflict prediction: Based on the LSTM prediction model, the resource utilization rate of each node in the future is predicted using historical resource utilization time series data, and the resource contention probability of each node is calculated to generate a conflict probability matrix. When the conflict probability exceeds the preset threshold, the path replanning process is triggered, that is, returning to step 4 or returning to step 2.

[0032] Furthermore, in step 5, the dynamic priority weight is calculated as follows:

[0033] Comprehensive business type Q 业务类型 User type Q 用户类型 and real-time demand Q 实时类型 Generate dynamic priority weights The formula is:

[0034]

[0035] In the formula, , ,and The priority weight coefficient is dynamically adjusted through online learning and feedback optimization.

[0036] Link Inherent Weight The calculation is as follows:

[0037]

[0038] In the formula, These are the device nodes that pass through the link. This represents the total number of link nodes.

[0039] Furthermore, in step 5, the process of dynamically adjusting the dynamic priority weight coefficient is as follows:

[0040] Step 501, construct the state space: business type vector, user type vector, and real-time demand vector;

[0041] Constructing the action space: priority weight coefficients , and The adjustment amount;

[0042] Design a reward function:

[0043]

[0044] In the formula, This represents the percentage of links successfully established within the current period. For resource utilization rate; The number of times a conflict occurs; , and This is the adjustment coefficient;

[0045] Step 502: Collect the status in each decision cycle t. It calculates the current reward and outputs the corresponding action through the DQN network. And execute the corresponding actions to adjust the priority weight coefficients. , and .

[0046] Furthermore, step 6 specifically involves:

[0047] Step 601: Convert the dynamic priority weight and the inherent link weight into a comprehensive weight. The formula is as follows:

[0048]

[0049] In the formula, and To set coefficients;

[0050] Step 602: Based on the fusion and the obtained path comprehensive weights, solve for the optimal solution path under the system constraints.

[0051] Step 603: During the search process, monitor the link status in real time, dynamically prune the candidate path set, and exclude links that are unusable due to status changes.

[0052] Furthermore, step 7 specifically involves:

[0053] Step 701: Based on the optimal path and corresponding comprehensive weight calculated in Step 6, configure the parameters of the relevant devices on the path to complete the establishment of the communication link;

[0054] Step 702: During link operation, continuously monitor the channel status;

[0055] When the detected link quality change exceeds the dynamic threshold, the relationship matrix is ​​dynamically updated, and the process returns to step 4 to reselect the optimal path.

[0056] When a link interruption is detected, the fault recovery mechanism is activated, and the system switches to the optimal backup link based on the currently available redundant links and their comprehensive weights; or it returns to step 2 to find the latest optimal path.

[0057] Furthermore, the dynamic threshold in step 702 The calculation formula is as follows:

[0058]

[0059] In the formula, This represents the total number of connected device nodes on the current communication link. Let be the comprehensive weight of the i-th node on the path; The real-time quality knowledge of the i-th node on the path includes the weighted values ​​of signal-to-noise ratio, latency jitter, and bit error rate; This represents the maximum capacity of critical nodes along the path.

[0060] Compared with the prior art, the beneficial effects achieved by this invention are as follows:

[0061] (1) This invention combines dynamic priority with graph theory, embeds business priority weights into graph model optimization, and solves the problem of resource waste caused by traditional static allocation.

[0062] (2) This invention adopts multidimensional conflict resolution, combined with minimum cut set algorithm and simulated annealing optimization, to realize the joint scheduling of bandwidth, power and computing resources.

[0063] (3) This invention improves communication stability under severe weather conditions and completes the closed loop through real-time channel sensing.

[0064] (4) The method of the present invention has the characteristics of being clear and easy to understand, having good continuity and strong scalability. It can dynamically adjust the resource allocation according to user needs, realize the resource allocation, load sharing and fault backup of multiple links in the satellite communication process, ensure the stable operation of the system under high-intensity working conditions, improve the system resource utilization rate, and thus improve the overall performance of the system. Attached Figure Description

[0065] Figure 1 This is a flowchart provided in an embodiment of the present invention.

[0066] Figure 2 This is a device resource connection diagram after model construction provided in an embodiment of the present invention.

[0067] Figure 3 This is a schematic diagram of the optimal link and the second-best link provided in the embodiments of the present invention.

[0068] Figure 4 This is a block diagram of the enhanced dynamic weight adjustment module of the present invention.

[0069] Figure 5 This is a flowchart of the LSTM conflict prediction model of the present invention.

[0070] Figure 6 This is the logic diagram for dynamic fault threshold calculation and decision-making in this invention. Detailed Implementation

[0071] This invention designs a graph theory-based dynamic priority allocation method for satellite gateway equipment resources. It can dynamically adjust resource allocation according to user needs, realize resource allocation, load sharing, and fault backup for multiple links in satellite communication, ensure stable operation of the system under high-intensity working conditions, improve system resource utilization, and thus enhance the overall performance of the system.

[0072] The present invention will be further described below with reference to the accompanying drawings.

[0073] This invention illustrates a graph theory-based dynamic priority allocation method for satellite gateway station equipment resources. The method mainly involves steps such as abstracting and modeling equipment resources within the gateway station, traversing network paths in the graph and filtering available paths, constructing a link relationship matrix, conflict resolution based on graph theory, dynamic priority evaluation, optimal path selection, and adaptive link construction. This achieves closed-loop control of resource allocation and task execution status. The main process is based on four parts: a resource modeling module, a priority evaluation engine, a graph theory algorithm engine, and a link control module.

[0074] This invention is used in gateway station development and integration projects when it is necessary to independently allocate communication resources within the station. Taking a certain gateway station integration project as an example, the equipment consists of two baseband devices, one matrix device, two frequency converters, two power amplifiers, and two sets of antenna servos, such as... Figure 1 As shown, the specific implementation method is as follows:

[0075] Step 1: Abstraction and model building of equipment resources within the gateway station;

[0076] Step 1 specifically includes:

[0077] Step 101: Draw a connection topology diagram of a gateway station using software, extracting the gateway station's communication resources into elements on the connection topology diagram. In the connection topology diagram, nodes represent communication equipment within the gateway station, such as antennas, power amplifiers, frequency converters, matrices, optical transceivers, basebands, etc.; the lines (edges) between nodes represent available links, i.e., the signal flow between nodes, and assign attributes such as delay, bandwidth, and load rate to the lines; for example... Figure 2 As shown;

[0078] Step 102: According to the equipment monitoring protocol, communicate with the equipment node devices to collect the parameters and status of the node devices in real time, such as the azimuth and elevation angle of the antenna, the signal-to-noise ratio and carrier-to-noise ratio of the baseband, etc.

[0079] Step 103: Based on the parameters and status of the node devices collected through the monitoring protocol, determine whether the links / paths between devices are reachable, and mark them on the connection topology map as lines; calculate additional attributes of the connections based on the data, such as latency, bandwidth, load rate, etc., and mark them on the connection topology map with line colors. When device parameters and status change, dynamically update the topology information.

[0080] Step 2: Traversing the network paths in the diagram and filtering available paths;

[0081] Step 2 specifically involves:

[0082] Step 201: Using a depth-first approach, traverse the nodes in the connection topology graph, record the paths and backtrack to obtain all reachable links in the graph as a set of candidate links.

[0083] Step 202: Define the rules that need to be constrained, encapsulate the constraints into different plugins, and have the software call them as needed to filter available links from the candidate link set. For example: the path needs to include a Ka-band power amplifier, the path needs to include a Ka-band antenna with circular polarization, and the baseband equipment resource margin in the path needs to be greater than 2Mbps, etc.

[0084] Based on the characteristics of the gateway station, the links are preset. For example, if the uplink of this station needs to be a combination of baseband + matrix + upconverter + power amplifier + antenna, then the candidate links are screened to ensure that each path has a specified start point, intermediate points and end point.

[0085] Table 1. List of Available Paths

[0086] Serial Number baseband matrix frequency conversion power amplifier antenna 1 1 1 1 1 1 2 1 1 1 2 2 3 1 1 2 1 1 4 1 1 2 2 2 5 2 1 1 1 1 6 2 1 1 2 2 7 2 1 2 1 1 8 2 1 2 2 2

[0087] In step 3, a relationship matrix is ​​constructed for the filtered link set;

[0088] Step 3 specifically includes:

[0089] Step 301: Construct a relationship matrix, including an adjacency matrix, for the filtered available links. Interference matrix and resource matrix ;

[0090] Adjacency matrix This indicates a direct connection between nodes, and the element... Indicates the existence of a link. Indicates a non-existent link;

[0091]

[0092] Interference matrix Quantify the degree of interference between different paths passing through the same node. This indicates that there is strong interference in the link between node i and node j;

[0093]

[0094] Resource Matrix Record parameters such as bandwidth and load rate between each node;

[0095]

[0096] Step 302: Dynamically fuse the adjacency matrix, interference matrix, and resource matrix to generate a weighted adjacency matrix. Its formula is:

[0097]

[0098] in, , and Initialize to 0.6, -0.2, 0.6;

[0099] According to the formula, we can obtain:

[0100]

[0101] In step 4, conflict resolution is based on graph theory;

[0102] Step 4 specifically involves:

[0103] Step 401: Based on the weighted adjacency matrix and resource matrix generated in step 3, use graph theory methods to identify resource allocation conflicts, specifically including determining whether there are multiple candidate paths sharing the same node or link resource, exceeding its capacity.

[0104] Step 402, Conflict resolution: For the identified conflicts, the minimum cut set algorithm is used to divide the resource allocation area, or the path selection is adjusted according to the preset conflict resolution strategy to make the resource allocation conflict-free.

[0105] Conflict Prediction: To address the real-time requirements of high-load scenarios, this example introduces a Long Short-Term Memory (LSTM) prediction model to achieve early conflict avoidance. For example... Figure 5 As shown, the specific implementation is as follows:

[0106] (1) Data acquisition and preprocessing

[0107] Collect historical resource occupancy rate time-series data of key equipment nodes, select data from the past T minutes (e.g., T=30 minutes) as input serial port, and perform normalization processing on the data;

[0108] (2) Construction of LSTM prediction model

[0109] Construct an LSTM neural network consisting of an input layer, an LSTM hidden layer, and a fully connected output layer; use a forget gate. Input gate and output gate Controlling the flow of information can effectively capture long-term relationships in occupancy rates. The model's input is a normalized time-series data matrix. Output for the future Time (e.g.) Predicted resource utilization of each node within 5 minutes. ;

[0110] (3) Calculation of conflict probability

[0111] Based on the predicted value Calculate the resource contention probability of each node. :

[0112]

[0113] in, It is the Sigmoid activation function. This represents the resource bottleneck threshold.

[0114] (4) Triggering mechanism

[0115] Generate the conflict probability matrix When any element of the matrix Greater than the preset warning threshold (For example When the probability of a conflict is 0.7, and a high-probability conflict is determined to occur in the future, the system immediately triggers the path replanning process (returning to step 4 or step 2) to release or adjust resources on high-risk paths in advance.

[0116] Step 5, Dynamic Priority Evaluation;

[0117] Step 5 specifically includes:

[0118] The dynamic priority weight is calculated as follows:

[0119]

[0120] In the formula, , ,and The priority weight coefficient is dynamically adjusted through online learning and feedback optimization.

[0121] To further enhance the adaptive capability of weight adjustment, this example introduces a deep Q-network as a reinforcement learning optimization submodule. For example... Figure 4 As shown, the specific implementation process is as follows:

[0122] (1) Constructing the state space

[0123] Define the state vector S as the feature set of the current system environment, specifically as follows:

[0124]

[0125] in, Identifier business type vector, User type vector, Identify the real-time demand vector.

[0126] (2) Define the action space

[0127] Define action vector A as priority weight coefficient , and The fine-tuning operation, namely:

[0128]

[0129] The action set includes preset step sizes for increasing or decreasing each weight coefficient, such as ±0.5.

[0130] (3) Design the reward function

[0131] Define a reward function R to simultaneously optimize resource utilization and business success rate:

[0132]

[0133] in, This represents the percentage of links successfully established within the current period. For resource utilization rate; The number of times a conflict occurs; , and This is the adjustment coefficient.

[0134] (4) Online learning process

[0135] The system collects the status in each decision cycle t. Actions are output through the DQN network. This action adjusts the weighting coefficients. The reward is calculated based on the environmental feedback after execution. and will experience ( , , , Store the data in the experience buffer. Update the DQN network parameters by minimizing the loss function L.

[0136] Through the reinforcement learning mechanism described above, the weight coefficients , and It can dynamically adjust according to real-time business scenarios, thereby replacing traditional static configuration or black-box feedback optimization and significantly improving the scientific nature of weight allocation.

[0137] Furthermore, the inherent weight of the link The calculation is as follows:

[0138]

[0139] in, These are the device nodes that pass through the link. This refers to the total number of link nodes; in this project, it is 5.

[0140] The available path weight table can be obtained based on the formula.

[0141] Table 2 Available Path Weights

[0142] 1 1.16 2 1.04 3 0.92 4 0.92 5 1.16 6 1.04 7 0.92 8 0.92

[0143] Step 6, Optimal Path Selection;

[0144] Step 6 specifically includes:

[0145] Step 601: Convert the dynamic priority weights (service type, user level, real-time requirements) and the inherent weights of the link (interference, resources) into a comprehensive weight. The formula is as follows:

[0146]

[0147] In the formula, and To set coefficients;

[0148] Step 602: Based on the fusion and the obtained path comprehensive weights, solve for the optimal solution path under the system constraints, such as... Figure 3 As shown;

[0149] Step 603: During the search process, monitor the link status in real time, dynamically prune the candidate path set, and exclude links that are unusable due to status changes.

[0150] Step 7, adaptive link construction;

[0151] like Figure 6 As shown, step 7 specifically involves:

[0152] Step 701: Based on the optimal path and corresponding comprehensive weight calculated in Step 6, configure the parameters of the relevant devices on the path to complete the establishment of the communication link;

[0153] Step 702: During link operation, continuously monitor the channel status; when the detected link quality change exceeds the dynamic threshold, dynamically update the relationship matrix and return to step 4 to reselect the optimal path;

[0154] To achieve more precise fault switching, this example defines a dynamic threshold. This threshold is used to determine whether a link has deteriorated to the point where a switchover is necessary. This threshold is no longer a fixed value, but rather a value calculated in real-time based on the overall weight of the current path and the device's capabilities. Specifically:

[0155] (1) Dynamic threshold formula

[0156]

[0157] in,

[0158] n: The total number of device nodes connected in series on the current communication link;

[0159] The comprehensive weight of the i-th device node in the current optimal path is as follows:

[0160]

[0161] This reflects the importance of this node to the current business;

[0162] The comprehensive real-time quality index value of the i-th node. The specific calculation is as follows:

[0163]

[0164] in:

[0165] The normalized signal-to-noise ratio;

[0166] For normalized latency jitter;

[0167] The normalized bit error rate;

[0168] , ,and These are the corresponding weighting coefficients.

[0169] A comprehensive metric for the maximum capacity (e.g., maximum throughput or maximum output power) of critical nodes on the path (which can be taken from all nodes on the path). (Minimum value or weighted sum).

[0170] (2) Judgment Logic

[0171] The system monitors the overall quality indicators of the current path in real time. The fault recovery mechanism is triggered when any of the following conditions are met:

[0172] 1) Hard handover: The current quality is below the dynamic threshold.

[0173] 2) Soft handover: #imgpt135#, the rate of quality change exceeds the dynamic threshold.

[0174] By introducing dynamic thresholds, the system can adapt its sensitivity to the importance of the business and the real-time status of the equipment, avoiding the problems of "over-switching" or "post-switching" caused by fixed thresholds.

[0175] Furthermore, when a link interruption is detected, a fault recovery mechanism is activated, switching to the optimal backup link based on the currently available redundant links and their comprehensive weights; or returning to step 2 to find the latest optimal path.

[0176] By designing and implementing the above content, the allocation of satellite gateway station resources can be completed.

[0177] The embodiments of the present invention present an original graph-based dynamic priority allocation method for satellite gateway station equipment resources, which solves problems such as chaotic resource allocation, unbalanced service load, inability to automatically plan and adjust resources, and difficulty in achieving unattended operation. This ensures stable system operation under high-intensity working conditions, improves system resource utilization, and thus enhances the overall system performance.

[0178] Those skilled in the art will understand that a graph theory-based method for dynamic priority allocation of satellite gateway station equipment resources can be implemented using file editing software, auxiliary software, and related hardware. When executed, this method performs the steps described in the above-described method embodiments.

Claims

1. A method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory, characterized in that, Includes the following steps: Step 1: Based on the collected parameters and status of the gateway station equipment, the communication resources of the gateway station are abstracted into a connection topology graph. Nodes identify gateway station equipment, edges identify available links, and weights are assigned to the edges, including latency, bandwidth, and load rate. Step 2: Traverse the nodes in the connection topology graph, obtain all reachable links in the graph as a candidate link set, and filter the candidate link set according to the constraints and the characteristics of the gateway station; Step 3: Construct the relation matrix of the filtered link set, including the adjacency matrix, interference matrix and resource matrix, and generate a weighted adjacency matrix. Step 4: Based on the matrix generated in Step 3, identify path conflict problems in resource allocation. For the identified conflicts, use the minimum cut set algorithm to resolve the conflicts and use the LSTM prediction model to predict the conflicts. Step 5: Based on business needs, user type, and link status, perform dynamic priority evaluation and calculate dynamic priority weight and inherent link weight. Step 6: Integrate dynamic priority weights and link-specific weights to form a comprehensive weight, and select the optimal path from the candidate paths based on the comprehensive weight; Step 7: Perform adaptive link construction and maintenance based on the optimal path selected in Step 6.

2. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 1, characterized in that, Step 1 is as follows: Step 101: Abstract the communication resources of the gateway station into a connection topology graph. Nodes identify gateway station devices, edges identify available links, and weights are assigned to the edges, including latency, bandwidth, and load rate. Step 102: Collect parameters and status of the gateway station equipment through the monitoring protocol; Step 103: Based on the collected parameters and status of the gateway station equipment, determine whether the links between the devices are reachable and dynamically construct the connection topology.

3. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 1, characterized in that, Step 2 is as follows: Step 201: Using a depth-first approach, traverse the nodes in the connection topology graph, record the paths and backtrack to obtain all reachable links in the graph as a set of candidate links. Step 202: Set the rules that need to be constrained and filter available links from the candidate link set; And as needed, the components of the devices in the link are pre-defined to filter available links from the candidate links.

4. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 3, characterized in that, Step 3 specifically involves: Step 301: Construct a relationship matrix, including an adjacency matrix, for the filtered available links. Interference matrix and resource matrix Adjacency matrix Indicates the direct connection relationship between nodes, interference matrix The resource matrix represents the degree of interference between different paths passing through the same node. Record parameters between nodes, including latency, bandwidth, and load rate; Step 302, adjust the adjacency matrix. Interference matrix and resource matrix Dynamic fusion generates a weighted adjacency matrix, the formula of which is: ； in, , and These are the weighting coefficients.

5. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 1, characterized in that, Step 4 is as follows: Step 401: Based on the weighted adjacency matrix and resource matrix generated in step 3, use graph theory methods to identify resource allocation conflicts, specifically including determining whether there are multiple candidate paths sharing the same node or link resource, exceeding its capacity. Step 402, Conflict resolution: For the identified conflicts, the minimum cut set algorithm is used to divide the resource allocation area, or the path selection is adjusted according to the preset conflict resolution strategy to make the resource allocation conflict-free. Conflict prediction: Based on the LSTM prediction model, the resource utilization rate of each node in the future is predicted using historical resource utilization time series data, and the resource contention probability of each node is calculated to generate a conflict probability matrix. When the conflict probability exceeds the preset threshold, the path replanning process is triggered, that is, returning to step 4 or returning to step 2.

6. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 1, characterized in that, In step 5, the dynamic priority weight is calculated as follows: Comprehensive business type Q 业务类型 User type Q 用户类型 and real-time demand Q 实时类型 Generate dynamic priority weights The formula is: ； In the formula, , ,and The priority weight coefficient is dynamically adjusted through online learning and feedback optimization. Link Inherent Weight The calculation is as follows: ； In the formula, These are the device nodes that pass through the link. This represents the total number of link nodes.

7. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 6, characterized in that, In step 5, the process of dynamically adjusting the dynamic priority weight coefficient is as follows: Step 501, construct the state space: business type vector, user type vector, and real-time demand vector; Constructing the action space: priority weight coefficients , and The adjustment amount; Design a reward function: ； In the formula, This represents the percentage of links successfully established within the current period. For resource utilization rate; The number of times a conflict occurs; , and This is the adjustment coefficient; Step 502: Collect the status in each decision cycle t. It calculates the current reward and outputs the corresponding action through the DQN network. And execute the corresponding actions to adjust the priority weight coefficients. , and .

8. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 1, characterized in that, Step 6 specifically involves: Step 601: Convert the dynamic priority weight and the inherent link weight into a comprehensive weight. The formula is as follows: ； In the formula, and To set coefficients; Step 602: Based on the fusion and the obtained path comprehensive weights, solve for the optimal solution path under the system constraints. Step 603: During the search process, monitor the link status in real time, dynamically prune the candidate path set, and exclude links that are unusable due to status changes.

9. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 1, characterized in that, Step 7 specifically includes: Step 701: Based on the optimal path and corresponding comprehensive weight calculated in Step 6, configure the parameters of the relevant devices on the path to complete the establishment of the communication link; Step 702: During link operation, continuously monitor the channel status; When the detected link quality change exceeds the dynamic threshold, the relationship matrix is ​​dynamically updated, and the process returns to step 4 to reselect the optimal path. When a link interruption is detected, the fault recovery mechanism is activated, and the system switches to the optimal backup link based on the currently available redundant links and their comprehensive weights; or it returns to step 2 to find the latest optimal path.

10. The method for dynamic priority allocation of satellite gateway station equipment resources based on graph theory as described in claim 9, characterized in that, Dynamic threshold in step 702 The calculation formula is as follows: ； In the formula, This represents the total number of connected device nodes on the current communication link. Let be the comprehensive weight of the i-th node on the path; The real-time quality knowledge of the i-th node on the path includes the weighted values ​​of signal-to-noise ratio, latency jitter, and bit error rate; This represents the maximum capacity of critical nodes along the path.

Images