A Delay-Optimized Adaptive Routing Method and System for Satellite Networks

Abstract

This invention relates to a latency-optimized adaptive routing method and system for satellite networks, belonging to the field of satellite communication technology. The method includes steps such as constructing multi-hop information age, multi-path routing, link state awareness, traffic segmentation adaptive routing, and dynamic routing decision-making. It employs a multi-hop information age model to consider the time-varying nature of link states, accurately estimating the real-time nature of path information and ensuring the timeliness of data updates. A multi-path search algorithm predicts link occupancy frequencies based on ephemeris information to construct multiple non-overlapping transmission paths, alleviating traffic congestion and optimizing link resource utilization. A traffic segmentation adaptive routing algorithm adjusts the path traffic allocation ratio, reducing data packet loss rate and information age, ensuring the stability of latency-sensitive data transmission. In highly dynamic satellite networks, this method balances resource utilization, effectively prevents link congestion, improves overall transmission efficiency and reliability, and provides a highly adaptable and efficient transmission solution for latency-sensitive services.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of satellite communication, and in particular to a time-delay-optimized satellite network adaptive routing method and system. BACKGROUND

[0002] With the continuous expansion of satellite networks in global communication, their potential in fields such as battlefield situation awareness, disaster relief and communication in remote areas is increasingly apparent. Due to the wide coverage and low time delay of satellite networks, they have unique advantages in dealing with emergencies and ensuring critical communication. At the same time, the dynamic and multi-hop nature of satellite networks also puts higher requirements on the routing mechanism of the network. In particular for time-sensitive services, how to ensure the real-time nature of data has become a pressing problem. In satellite networks, satellite nodes move at high speed, and the network topology changes frequently, and the availability and transmission delay of communication links fluctuate at all times. Traditional routing methods are mainly divided into static and dynamic types: static routing constructs paths based on preset orbital information, but it is difficult to respond to changes in network state in real time, often leading to link congestion and data loss; dynamic routing can update routing tables in a timely manner and has better adaptability, but it needs to continuously obtain real-time link information between satellites, increasing the communication overhead of the system. Although the existing dynamic and static combined routing strategy can balance the network state update and resource consumption to some extent, it still has great limitations in complex dynamic environments and cannot meet the high requirements of time-sensitive services for data real-time performance. In satellite networks, data needs to be transmitted through multi-hop transmission of multiple satellite nodes, and the frequent interruption of link status leads to dynamic changes in the path, and the selection of different paths directly affects the information age. The existing methods mainly focus on the optimization of single-path shortest time delay or maximum throughput, and do not effectively consider the information age in multi-hop transmission paths, which cannot guarantee the freshness of data, especially in the case of frequent link interruption, the existing scheme is difficult to realize the real-time update of data. The link status of satellite networks has a high degree of time variability. The link availability and delay in different time periods change with the change of satellite position, affecting the selection of paths and the control of information age. The existing methods usually do not consider the time variability of link status in path selection, which leads to congestion of service flow on high-load links, increases the transmission delay and packet loss rate, and affects the timeliness of data.

[0003] Therefore, in view of the deficiencies of the existing satellite network in information age control and resource utilization, an effective multi-hop information age optimization method is urgently needed to meet the needs of delay-sensitive services. Some existing research has alleviated the congestion problem of single-path transmission through multi-path routing and optimized the link utilization within a limited range. However, these methods often only optimize the path selection for a single transmission, ignoring the dynamic control of multi-hop information age, making it difficult to achieve comprehensive information age optimization. In addition, the existing multi-path method does not consider load balancing on different paths in path selection, which easily leads to frequent congestion of high-load links and cannot achieve effective updating of data.

[0004] In summary, the multi-hop routing mechanism of the existing satellite network has obvious deficiencies in meeting the needs of delay-sensitive services. SUMMARY

[0005] Therefore, the problem to be solved by the present application is: how to optimize the multi-hop routing mechanism and information age management in a satellite network to ensure real-time transmission of data for delay-sensitive services and meet multiple network constraints.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] A delay-optimized satellite network adaptive routing method, comprising the following steps:

[0008] S1: construction of multi-hop information age, specifically:

[0009] S11: in the low-orbit satellite constellation system model, assume that A fi (t) represents the instantaneous information age of the i-th satellite node of the data flow f at time slot t, and the instantaneous information age is calculated;

[0010] S12: introduce binary variable x fik (t) represents whether the i-th satellite node of the data flow f receives the k-th data packet at time slot t; if x fik (t) = 1, it means that at time slot t, the i-th satellite node of the data flow f successfully receives the k-th update packet, otherwise x fik (t) = 0;

[0011] S13: describe the relationship between the information age of the current satellite node and the last-hop satellite node as follows: if a data packet is received at t-1, the information age is represented as A f(i-1) (t-1) + 1; if no data packet is received, the information age is represented as A fi (t-1) + 1; the basic model of information age is constructed;

[0012] S14: Each satellite node is regarded as a service unit with limited service capacity and queue to store data packets waiting for processing; the queuing theory of first-come-first-served is used to model the waiting time of data packets in the queue;

[0013] S15: The congestion degree of the satellite network is measured by comparing the service rate of the satellite node and the data packet generation rate, and the network congestion degree is modeled:

[0014] S16: The connectivity constraint of the satellite network is modeled:

[0015] S17: A minimum information age model is constructed;

[0016] S2: Multi-path routing, through a path search algorithm based on multi-hop information age optimization to dynamically calculate multiple disjoint candidate paths, realizing traffic splitting;

[0017] S3: Link state awareness, each satellite periodically sends link state information to the control satellite, including the capacity of the satellite node to receive new data packets, the queue length of the current satellite node, the current network topology information, and the overall index of the network;

[0018] S4: Adaptive traffic splitting routing, based on path selection, through a deep reinforcement learning method, the optimized traffic distribution ratio and data generation rate are determined according to dynamic link parameters;

[0019] S5: Dynamic routing decision, according to the continuous iterative training of S4, the intelligent agent dynamically adjusts the traffic distribution ratio and data generation rate; the immediate reward is calculated using a reward function, and the policy network and Q value network are continuously optimized to minimize the information age; after training, the agent regularly collects network status and makes traffic splitting decisions according to the trained policy network.

[0020] Further, the S11 calculates the instantaneous information age, specifically:

[0021] The instantaneous information age is expressed as follows:

[0022]

[0023] Where g fk represents the generation time of the kth data packet of the data stream f;

[0024] The S13 reconstructs the basic model of information age, specifically:

[0025] The relationship between the information age of the current satellite node and the previous hop satellite node is described as follows: the basic model of information age is constructed as follows:

[0026]

[0027] Where K represents the total number of data streams: K f x represents the total number of data packets in data stream f; fik (t-1) represents a binary variable indicating whether data packet k was successfully received in time slot t-1;

[0028] Specifically, S14 is as follows:

[0029] When a data packet arrives at node j, the impact of congestion on information age is modeled as the queue waiting time W. i The increase of is expressed by the following formula:

[0030] A i (t)=tg i +W i (t)

[0031] Among them, A i (t) represents the information age of data packet i at time t; g i W represents the time when data packet i was generated at the source node; i (t) is the waiting time of data packet i in the queue;

[0032] Let the service rate of each satellite node i be μ. i Let λ represent the number of data packets a node can process per second; let the arrival rate be λ. i Let L represent the average rate at which data packets arrive at satellite node i; let the queue length be L. i Let (t) represent the number of data packets waiting to be processed in the queue of satellite node i at time t, expressed by the following formula:

[0033] L i (t+Δt)=L i (t)+λ i ·Δt-μ i Δt

[0034] Where Δt represents the time interval; λ represents the arrival rate;

[0035] Average service time is The waiting time W of the data packet in the queue i (t) can be expressed as the following formula:

[0036]

[0037] Specifically, S15 is as follows:

[0038] Define congestion index C congestion Let (t) represent the congestion level of satellite node i at time t, expressed as the following formula:

[0039]

[0040] Where, λ i (t) represents the arrival rate of satellite node i at time t; when C congestion When (t) > 0, it indicates that satellite node i is in a congested state; when C congestion When (t) < 0, it indicates that satellite node i has remaining processing capacity; when C congestion When (t) = 0, it means that satellite node i has just reached a balanced state;

[0041] The data packet generation rate λ is dynamically adjusted based on the current network congestion status and the remaining capacity of the satellite nodes; the adjustment formula is expressed as follows:

[0042] λ(t)=f(C congestion (t),C i (t))

[0043] Where λ(t) is the adjusted data packet generation rate, f(C) congestion (t),C i (t) is a function that adjusts the generation rate based on congestion indicators and the remaining capacity of nodes;

[0044] Specifically, S16 is as follows:

[0045] For each source node s and destination node d, there exists a path P s,d The collection makes:

[0046] Among them, e i Represents path P s,d The top edge, capacity(e) i ) represents edge e i The capacity, λ s This indicates the traffic demand of the source node;

[0047] Specifically, S17 is as follows:

[0048] The satellite network is modeled as a spatiotemporal diagram, represented by the following formula:

[0049] G = {(S,E,T,k)}

[0050] The set of satellites S is represented as S = {s1, s2, ..., s}. N}, where E is the set of links represented as E = {e1, e2, ..., e} L}, T is a vector consisting of k time slots, with a time interval τ = T / k; in the time interval τ i The number of existing links within is represented as The K paths between any nodes are defined as R, to achieve effective traffic splitting, and are expressed by the following formula:

[0051]

[0052] in, Indicates time slice τ i All path sets within; Indicates time slice τ i The first transmission path of the first data stream within the system; Indicates time slice τ i The Kth transmission path of the first data stream within the system; Indicates time slice τ i The Kth transmission path of the Mth data stream;

[0053] Assume the path space P = {P1, P2, ..., P} K The maximum flow rate C = {c1, c2, ..., c} within the observation time (0, T] is given by [c1, c2, ..., c]. K}; where P i The maximum flow is the maximum capacity of the link; based on the bandwidth resources of the relevant paths, each source node can split the flow into M sub-flows that simultaneously pass through K paths, letting... Let W represent the sub-flow ratio between the m-th node pair via the x-th path. The sub-flow splitting solution W between all sub-flow pairs between the source and destination nodes is expressed by the following formula:

[0054] W = {w 1,1 ,...,w 1,K ,...,w m,K}

[0055] Furthermore, S17 specifically includes:

[0056] By optimizing path selection R and traffic allocation strategy W, the average information age A of all paths in the network is minimized. avg This can be expressed as the following formula:

[0057]

[0058] Where K represents the total number of paths, τ i A represents a state in the network state set T; x (τ i ) indicates that in time slice τ i Within, the information age of the xth path;

[0059] Setting constraints includes:

[0060] 1) Packet generation rate constraint to ensure the generation rate of each satellite node. Not exceeding its maximum load capacity

[0061] 2) Satellite node processing capacity constraints to ensure the processing rate of each satellite node. Not exceeding its maximum processing capacity

[0062] 3) Network connectivity constraints ensure that the total link capacity of the selected path can meet the data traffic demand λ. s That is, the sum of the capacities of all links on the path, ∑capacity(e i The flow rate λ must be greater than or equal to the flow rate λ. s And there needs to be at least one feasible path p. i ∈P;

[0063] 4) Path constraints, ensuring that the path set P μ It only includes feasible paths that meet the link status and node capability conditions;

[0064] 5) Queue length constraint: The queue length L of all nodes must be constant. i Maintain maximum queue length The following measures are taken to avoid transmission delays caused by excessively long queues;

[0065] 6) Information age constraint: Ensure that the information age on the transmission path increases, i.e., the candidate paths between node pairs. satisfy Where 1≤x≤y≤K, m∈[1,M];

[0066] 7) Path-based flow conservation constraint: This requires that the sum of the sub-flows distributed across all paths from the total flow between any pair of nodes equals 1. m∈[1,M];

[0067] 8) Congestion index constraints, limiting the congestion index of each satellite node. Not exceeding the predefined threshold C threshold To avoid a decrease in the timeliness of information due to congestion;

[0068] 9) Bandwidth constraints to ensure traffic flow on each link in the path. Not exceeding the link bandwidth BW, of which And e j ∈E S .

[0069] Furthermore, S2 specifically includes the following steps:

[0070] S21: Input the source and destination satellite node pairs, ephemeris information, and a predefined random weight range, and initialize them, specifically as follows:

[0071] Input the satellite pair set (S,D) and ephemeris information of the source and target nodes, and the predefined random weight range (r1 < r2 < r3). ​​Initialize the link weight matrix M and the candidate path set R as empty sets, and set the link occupancy frequency. Let τ be 0, where e represents the link, and τ i Indicates a time segment;

[0072] S22: For each time slot τ i The shortest path within the current time slot is calculated using Dijkstra's shortest path algorithm based on information age. If the link is in the shortest path, then update the link occupancy frequency. To prevent the algorithm from getting trapped in local optima, a threshold setting method is used. If the link occupancy frequency... Less than a given threshold The ground control center randomly assigns an integer between r1 and r2 as the new link metric; otherwise, the link metric is set to a random integer between r2 and r3.

[0073] S23: Delete the link in the shortest transmission path of information age, and continue to select the second shortest path using the shortest path algorithm on the remaining links to make full use of idle links.

[0074] Furthermore, S3 specifically refers to:

[0075] The state S of each satellite node includes the following elements: C i L: The capacity of a satellite node to receive new data packets; i : The queue length of the current satellite node; G i : Current network topology information; C congestion,i : The overall index of the network.

[0076] Furthermore, S4 specifically includes:

[0077] S41: Let state S include C i and L i C i Indicates the capacity of a satellite node to receive new data packets; L i Indicates the queue length of the current satellite node; G i Indicates the current network topology information; C congestion,i The overall index of the network is represented by A; action A is the traffic share and data generation rate of candidate paths between all node pairs, expressed as: A={a1,a2,...,a... M}, each action a j Includes: path traffic allocation weights, packet generation rate The reward R is a weighted sum of the data age and the node congestion level, expressed as the following formula:

[0078] R(s,a)=-α·A avg (s,a)-β·C congestion (s)

[0079] Where R(s,a) represents the reward for performing action a in state s; A avg (s,a) represents the information age of taking action a in state s; C congestion (s) represents the link congestion level under state s; α and β are the weighting coefficients of information age and congestion level, respectively. The larger the reward value, the more effective the path selection and traffic allocation strategy is for optimizing information age.

[0080] S42: Input the path obtained by the path search algorithm based on multi-hop information age optimization in step S2 and the link state information obtained in step S3 into reinforcement learning, and train the agent to decide the traffic ratio and data transmission rate allocated to each transmission path; wherein the update of the optimal policy satisfies the following formula:

[0081]

[0082] Where E represents the expectation; α is the temperature coefficient (regularization coefficient), used to adjust the importance of entropy in the objective function; γ is the discount factor; r(s) t ,a t ) represents state s t Action a occurs under the following circumstances t The reward; H(π(·|s t )) is the entropy function, representing the state s t The entropy of the time-space policy is expressed by the following formula:

[0083] H(π(·|s t ))=E[-logπ(·|s t )]

[0084] Value function Q(s) t ,a t The definition is as follows:

[0085]

[0086] Among them, s t+1 V was obtained by sampling from the experience replay pool D; π (s t+1 ) represents the value function, indicating the state s. t+1 Below, the expected return when following strategy π;

[0087] State value function V(s) t The reward () represents the expected reward in a certain state, and is defined by the following formula:

[0088]

[0089] Furthermore, S4 specifically includes:

[0090] S41: Initialize reinforcement learning parameters. Initialize the network parameters in the Actor-Critic framework, including the initial weights of the policy network π and the value network Q, set the capacity of the experience pool M, and the weight parameters of the target network.

[0091] S42: Status observation. The system periodically collects the environmental status S of the satellite network, including: link utilization and latency information of the current path set R; and queue length L of satellite nodes. i Network link congestion index C congestion ;

[0092] S43: Execute the action, allocating a traffic proportion to each path according to the policy network π. and packet generation rate Apply these allocation parameters as action A to the current network environment;

[0093] S44: Calculate the reward function. For each time slice, calculate the reward value R(s,a).

[0094] R(s,a)=-α·A avg (s,a)-β·C congestion (s)

[0095] Among them, A avg (s,a) represents the average information age of the system after the action is performed, C congestion (s) represents the congestion level of the link, and α and β are weight parameters;

[0096] S45: Update the policy network and value network, and store the result of the current action into the experience pool M; randomly sample a small batch of data and use gradient descent to optimize the policy network and value network;

[0097] S46: Online learning and routing adjustment. The system updates the policy network in real time and adjusts the traffic allocation of each path according to the latest environmental status to ensure that high-priority data packets can pass through low-latency paths first, thereby reducing the overall information age; it outputs the trained optimal traffic allocation strategy and applies it to the data transmission process.

[0098] A latency-optimized adaptive routing system for satellite networks includes:

[0099] Path selection module: Based on known ephemeris information, a network topology snapshot map is constructed by measuring the real-time locations of satellite nodes; each time slice constitutes a topology map, and multiple candidate paths that meet feasible path constraints are calculated using a multi-path routing algorithm;

[0100] Packet generation rate adjustment module: Adjusts the generation rate based on the processing capacity of satellite nodes and network congestion, and feeds back the adjusted packet rate to the traffic allocation module; the adjustment of the packet generation rate affects subsequent path selection and traffic allocation;

[0101] Information Age Calculation Module: Calculates and updates the information age of each satellite node, providing data to the path selection module to help select the most suitable path; it also feeds back the information age to the traffic allocation module, affecting the optimized allocation of data streams; after updating the information age, it feeds back the calculation results to the reinforcement learning module, providing model learning and adjustment strategies.

[0102] Path selection module: Based on satellite ephemeris information and link status, select the optimal multi-hop path and feed the selection result back to the traffic allocation module;

[0103] Traffic allocation module: Based on information age, network congestion, and satellite node queue length, it optimizes traffic allocation and feeds back the traffic scheduling results to the data packet generation rate adjustment module and the path selection module; by optimizing traffic allocation, it affects the data forwarding rate and path selection strategy.

[0104] Reinforcement learning module: Based on network state, it adjusts path selection and traffic allocation strategies and optimizes long-term performance; this module interacts with the above modules to provide dynamic adjustment and adaptive optimization capabilities.

[0105] The beneficial effects of this invention are as follows:

[0106] This invention significantly improves the timeliness of data transmission and network resource utilization for latency-sensitive services by constructing a multi-hop information age optimization model for satellite networks, proposing a multi-path search algorithm, and an information age-optimized traffic segmentation adaptive routing algorithm. The multi-hop information age model considers the time-varying nature of link states, accurately estimates the real-time nature of path information, and ensures the timeliness of data updates. The multi-path search algorithm predicts link occupancy frequencies based on ephemeris information, constructs multiple non-overlapping transmission paths, effectively alleviates traffic congestion, and optimizes link resource utilization. The traffic segmentation adaptive routing algorithm adaptively adjusts the path traffic allocation ratio through deep reinforcement learning, significantly reducing data packet loss rate and information age, and ensuring the stability of latency-sensitive data transmission. Through these optimizations, this invention achieves a comprehensive improvement in real-time transmission, link resource utilization efficiency, and transmission reliability in satellite networks.

[0107] 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

[0108] 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:

[0109] Fig. 1 This is a diagram illustrating a satellite network data transmission application architecture according to an embodiment of the present invention.

[0110] Fig. 2 This is an interaction diagram of the modules in a latency-optimized adaptive routing system for satellite networks according to an embodiment of the present invention.

[0111] Fig. 3 This is a diagram illustrating the training process of dynamic routing decision based on the Soft Actor-Critic (SAC) algorithm according to an embodiment of the present invention. Detailed Implementation

[0112] 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.

[0113] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0114] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0115] The purpose of this invention is to provide a latency-optimized adaptive routing method for satellite networks. By dynamically optimizing information age, path selection, and traffic allocation in multi-hop paths, it ensures data freshness and transmission timeliness. Combining packet generation rate, satellite node processing capacity, bandwidth, queue length, and energy constraints, this invention can balance resource utilization in highly dynamic satellite networks, effectively prevent link congestion, improve overall transmission efficiency and reliability, and provide a highly adaptable and efficient transmission solution for latency-sensitive services. The specific steps of this method include the following:

[0116] S1: Constructing the Age of Multi-hop Information

[0117] S11: In the low Earth orbit satellite constellation system model, assuming A fi (t) represents the instantaneous information age of the i-th satellite node in time slot t of data stream f. The instantaneous information age can be expressed by the following formula:

[0118]

[0119] Among them, g fk This indicates the generation time of the kth data packet in data stream f;

[0120] S12: Introducing binary variable x fik Let (t) represent whether the i-th hop satellite node of data stream f receives the k-th data packet in time slot t. If x fik If (t) = 1, it means that in time slot t, the i-th hop satellite node on data stream f successfully received the k-th update packet; otherwise, x fik (t) = 0.

[0121] S13: The relationship between the information age of the current satellite node and the previous hop satellite node is described as follows: If a data packet is received at time t-1, the information age is represented as A. f(i-1) (t-1)+1; If no data packet is received, the information age is represented as A. fi (t-1)+1; The basic model for constructing the information age is as follows:

[0122]

[0123] Where K represents the total number of data streams: K f x represents the total number of data packets in data stream f; fik (t-1) represents a binary variable indicating whether data packet k was successfully received in time slot t-1;

[0124] S14: In satellite networks, information age is dynamically updated as data packets are transmitted and network conditions change. When a data packet arrives at node j, it may be delayed due to link interruptions and queuing, thus increasing the information age of the transmitted data. Therefore, the impact of congestion on information age is modeled as the queue waiting time W. i The increase of is expressed by the following formula:

[0125] A i (t)=tg i +W i (t)

[0126] Among them, A i (t) represents the information age of data packet i at time t; g i W represents the time when data packet i was generated at the source node; i (t) is the waiting time of data packet i in the queue, which depends on the network congestion level and the processing capacity of the satellite node.

[0127] The following section will analyze the waiting time in the queue.

[0128] In a satellite network, each satellite node can be considered a service unit with limited service capacity and a queue to store data packets waiting to be processed. First-come, first-served queuing theory is used to model the waiting time of data packets in the queue.

[0129] Service rate: Each satellite node i has a service rate μ i This indicates the number of data packets a node can process per second; the service rate depends on the satellite's hardware performance and current workload.

[0130] Packet generation rate: arrival rate λ i This refers to the average rate at which data packets arrive at satellite node i.

[0131] Queue length: Queue length L i (t) represents the number of data packets waiting to be processed in the queue of satellite node i at time t. It is expressed by the following formula:

[0132] L i (t+Δt)=Li (t)+λ i ·Δt-μ i Δt

[0133] Where Δt represents the time interval; λ represents the arrival rate;

[0134] Waiting time: The time W for a data packet to wait in the queue. i (t) can be expressed as the following formula:

[0135]

[0136] Service time follows an exponential distribution, meaning that the processing time of data packets is random, but the average service time is constant.

[0137] S15: The congestion level of a satellite network is measured by comparing the service rate and packet generation rate of satellite nodes. This modeling of network congestion is specifically expressed as follows: Network congestion can be measured by comparing the service rate and packet generation rate of satellite nodes. When the arrival rate exceeds the service rate, the network becomes congested. Therefore, a congestion metric C is defined. congestion Let (t) represent the congestion level of satellite node i at time t, expressed as the following formula:

[0138]

[0139] Where, λ i (t) represents the arrival rate of satellite node i at time t; when C congestion When (t) > 0, it indicates that satellite node i is in a congested state; when C congestion When (t) < 0, it indicates that satellite node i has remaining processing capacity; when C congestion When (t) = 0, it indicates that satellite node i has reached a balanced state. The data packet generation rate λ can be dynamically adjusted based on the current network congestion status and the remaining capacity of the satellite nodes. The adjustment formula is as follows:

[0140] λ(t)=f(C congestion (t),C i (t))

[0141] Where λ(t) is the adjusted data packet generation rate, f(C) congestion (t),C i (t) is a function that adjusts the generation rate based on congestion indicators and the remaining capacity of nodes.

[0142] S16: Model the connectivity constraints of the satellite network. In a satellite network, data packets must be transmitted from the source node to the destination node through at least one feasible path. For each source node s and destination node d, there exists a path P. s,d The collection makes:

[0143]

[0144] Among them, e i Represents path P s,d The top edge, capacity(e) i ) represents edge e i The capacity, λ s This indicates the traffic demand of the source node.

[0145] S17: Construct a model that minimizes the information age. The satellite network is modeled as a spatiotemporal graph, expressed by the following formula:

[0146] G = {(S,E,T,k)}

[0147] The set of satellites S is represented as S = {s1, s2, ..., s}. N}, where E is the set of links represented as E = {e1, e2, ..., e} L}, where T is a vector consisting of k time slots, with a time interval τ = T / k. During the time interval τ i The number of existing links within is represented as Since there are multiple transmission paths between any source satellite and destination satellite pair, K paths between any nodes, defined as R, can be obtained during each topology snapshot to achieve effective traffic offloading, as expressed by the following formula:

[0148]

[0149] in, Indicates time slice τ i All path sets within; Indicates time slice τ i The first transmission path of the first data stream within the system; Indicates time slice τ i The Kth transmission path of the first data stream within the system; Indicates time slice τ i The Kth transmission path of the Mth data stream.

[0150] Assume the path space P = {P1, P2, ..., P} K The maximum flow rate C = {c1, c2, ..., c} within the observation time (0, T] is given by [c1, c2, ..., c]. K}. Where P iThe maximum traffic is equal to the maximum capacity of the link. Based on the bandwidth resources of the relevant paths, each source node can split the flow into M sub-flows that simultaneously traverse K paths, letting... Let represent the sub-flow ratio between the m-th node pair via the x-th path. The sub-flow splitting solution for all sub-flows between the source and destination node pairs can be expressed by the following formula:

[0151] W = {w 1,1 ,...,w 1,K ,...,w m,K}

[0152] This invention addresses the adaptive routing problem in satellite networks by proposing an optimization model that minimizes information age. The details are as follows:

[0153] Objective function: Minimize the average information age A of all paths in the network by optimizing path selection R and traffic allocation strategy W. avg ,Right now:

[0154]

[0155] Where K represents the total number of paths, τ i A represents a state in the network state set T; x (τ i ) indicates that in time slice τ i Within, the information age of the xth path.

[0156] The main constraints include:

[0157] (1) Data packet generation rate constraint to ensure the generation rate of each satellite node. Not exceeding its maximum load capacity

[0158] (2) Satellite node processing capacity constraints to ensure the processing rate of each satellite node. Not exceeding its maximum processing capacity

[0159] (3) Network connectivity constraints to ensure that the total link capacity of the selected path can meet the data traffic demand λ. s That is, the sum of the capacities of all links on the path, ∑capacity(e i The flow rate λ must be greater than or equal to the flow rate λ. s And there needs to be at least one feasible path p. i ∈P.

[0160] (4) Path constraints ensure that the path set P μ It only includes feasible paths that meet the link status and node capability conditions.

[0161] (5) Queue length constraint, requiring the queue length L of all nodes to be L. i Maintain maximum queue length The following steps aim to avoid transmission delays caused by excessively long queues.

[0162] (6) Information age constraint, ensuring that the information age on the transmission path increases, i.e., the candidate paths between node pairs satisfy Where 1≤x≤y≤K, m∈[1,M].

[0163] (7) Path flow conservation constraint: The total flow between any pair of nodes must be distributed to the sum of the sub-flows of all paths equal to 1. m∈[1,M].

[0164] (8) Congestion index constraint, limiting the congestion index of each satellite node. Not exceeding the predefined threshold C threshold This is to avoid a decrease in the timeliness of information due to congestion.

[0165] (9) Bandwidth constraints to ensure traffic flow on each link in the path. Not exceeding the link bandwidth BW, of which And e j ∈E S .

[0166] To address the aforementioned objective optimization model, this invention relates to a path search algorithm based on multi-hop information age optimization, used to optimize the transmission efficiency of latency-sensitive data in a network. This algorithm optimizes the information age of the entire network by calculating multiple transmission paths with low latency. Specifically, this invention achieves effective traffic splitting by dynamically calculating multiple disjoint candidate paths, thereby improving path utilization while avoiding network congestion. The following are the specific implementation steps of the multi-path search algorithm of this invention:

[0167] S2: Multi-path routing, the specific steps are as follows:

[0168] S21: Input the satellite pair set (S,D) of the source and target nodes, ephemeris information (i.e., satellite positions and topology information within each time segment), and a predefined random weight range (r1 < r2 < r3). ​​Initialize the link's weight matrix M and candidate path set R as empty sets, and set the link occupancy frequency. Let τ be 0, where e represents the link, and τ i Indicates a time segment.

[0169] S22: For each time segment τ iInternally, based on satellite ephemeris information, the Dijkstra's Shortest Path algorithm (DSP) is used to calculate the initial set of paths. The calculated paths are then incorporated into the candidate path set. Simultaneously, other candidate paths are calculated to increase path diversity. For each satellite pair of source node S and target node D, the link occupancy frequency is iteratively updated across all paths. Specifically, for each shortest path... Above, for each edge (a,b) in the path, increase its link occupancy frequency, i.e., execute... To prevent the algorithm from getting trapped in local optima, this invention sets a threshold based on the link frequency calculation. To dynamically adjust the link weights. If the link occupancy frequency... Below the threshold Then the weights of the link will be reassigned randomly. Its value is within a random range (r1, r2). Otherwise, it is randomly assigned within a higher weight range (r2, r3).

[0170] S23: Remove links included in the already calculated shortest path and re-run the shortest path algorithm on the remaining set of links to calculate the suboptimal path. This maximizes the use of unused link resources, ensuring the system still has backup paths during network congestion. Repeat the above steps until K candidate paths are found between each pair of source and destination nodes, thus satisfying the multi-path traffic splitting requirement. Through this path selection process, the present invention can reserve redundant channels on the transmission path to improve the robustness and stability of data transmission.

[0171] S3: Link Status Awareness, the specific steps of which are as follows:

[0172] S31: Each satellite periodically sends link status information to the control satellite, including the satellite node's capacity to receive new data packets, the current satellite node's queue length, the current network topology information, and the overall network index. The state S of each satellite node includes the following elements: C i The capacity of a satellite node to receive new data packets. L i : The current queue length of the satellite node. G i Current network topology information. (C) congestion,i : The overall index of the network.

[0173] Based on path selection, an adaptive traffic segmentation algorithm is proposed, transforming the information age minimization problem into a Markov decision process. Deep reinforcement learning is used to optimize the traffic allocation ratio and data generation rate based on dynamic link parameters. The specific method is as follows:

[0174] S4: Adaptive routing for traffic splitting, the specific steps are as follows:

[0175] S41: State S includes C i The capacity of satellite nodes to receive new data packets and L i : The current queue length of the satellite node. G i Current network topology information. (C) congestion,i The overall index of the network. Action A represents the traffic share and data generation rate of candidate paths between all node pairs, expressed as: A = {a1, a2, ..., a...} M}, each action a j Includes: path traffic allocation weights, packet generation rate The reward R is a weighted sum of the data age and the node congestion level, expressed as the following formula:

[0176] R(s,a)=-α·A avg (s,a)-β·C congestion (s)

[0177] Where R(s,a) represents the reward for performing action a in state s; A avg (s,a) represents the information age of taking action a in state s; C congestion (s) represents the link congestion level under state s; α and β are the weighting coefficients of information age and congestion level, respectively. The larger the reward value, the more effective the path selection and traffic allocation strategy is for optimizing information age.

[0178] S42: The paths obtained by the path search algorithm based on multi-hop information age optimization in step S2 and the link state information obtained in step S3 are input into reinforcement learning. The agent is trained to decide the traffic ratio and data transmission rate allocated to each transmission path. The update of the optimal policy follows the following formula:

[0179]

[0180] Where E represents the expectation; α is the temperature coefficient (regularization coefficient), used to adjust the importance of entropy in the objective function; γ is the discount factor; r(s) t ,a t ) represents state s t Action a occurs under the following circumstances t The reward; H(π(·|s t )) is the entropy function, representing the state s t The entropy of the time-space policy is expressed by the following formula:

[0181] H(π(·|s t ))=E[-logπ(·|s t )]

[0182] Value function Q(s) t ,a t The definition is as follows:

[0183]

[0184] Among them, s t+1 V was obtained by sampling from the experience replay pool D; π (s t+1 ) represents the value function, indicating the state s. t+1 Below, the expected return when following strategy π.

[0185] State value function V(s) t The reward () represents the expected reward in a certain state, and is defined by the following formula:

[0186]

[0187] Furthermore, the policy network π in the SoftActor-Critic (SAC) algorithm φ (a t ,s t ), Soft-state value network, Target-state value network And two softQ networks Each is composed of different parameter sets φ, ψ, θ parameterization, where φ represents the parameters of the policy network, which is responsible for generating the policy; ψ represents the parameters of the state-value network, which estimates the value of the state; The parameters of the target state value network are represented by θ, which are delayed in updating to provide a more stable target value; θ represents the parameter set of the two Q networks, which estimate the Q value of each state-action pair; the objective function of the network is optimized by applying stochastic gradient descent.

[0188] S5: Dynamic routing decision, the specific steps are as follows:

[0189] S51: Based on continuous iterative training from S4, the intelligent agent dynamically adjusts the traffic allocation ratio and data generation rate. It calculates immediate rewards using a reward function, continuously optimizing the policy network and Q-value network to ensure information age is minimized. After training is complete, the agent periodically collects network state data and makes traffic splitting decisions based on the trained policy network.

[0190] The following is in conjunction with the appendix Figs. 1-3 The specific embodiments of the present invention will be described in detail below.

[0191] Please see Fig. 1This is a satellite network data transmission application architecture diagram according to an embodiment of the present invention. After the source satellite node s obtains the freshness information, it transmits the data to the destination node d via paths p1 and p2, taking into account the connectivity issues between satellites. The data is then received by the Ground Receiving Equipment (GRE) and sent to the Ground Control Center (GCC). The Data Center (DC) updates the real-time monitoring data. It is assumed that each satellite node has information transmission, processing, and storage capabilities, and can forward data via multi-hop paths. The system should be able to dynamically collect information from each node, including information age, queue length, network load, etc. Each satellite node updates its information age based on the time of the received data packet. If a data packet is successfully received and removed from the queue, the node's information age is updated.

[0192] Please see Fig. 2 This is an interaction diagram of the modules in a latency-optimized adaptive routing system for satellite networks according to an embodiment of the present invention. The system includes:

[0193] The path selection module constructs a network topology snapshot based on known ephemeris information and the real-time locations of satellite nodes. Each time slice constitutes a topology map, and a multi-path routing algorithm is used to calculate multiple candidate paths that meet the feasible path constraints.

[0194] The packet generation rate adjustment module adjusts the generation rate based on the processing capacity of the satellite nodes and network congestion, and feeds the adjusted packet rate back to the traffic allocation module. Adjustments to the packet generation rate affect subsequent path selection and traffic allocation.

[0195] Information Age Calculation Module: This module calculates and updates the information age (information age) of each satellite node, providing data to the path selection module to help choose the most suitable path. It also feeds the information age back to the traffic allocation module, influencing the optimized allocation of data streams. After updating the information age, the calculation results are fed back to the reinforcement learning module for model learning and strategy adjustment.

[0196] Path selection module: Based on satellite ephemeris information and link status, it selects the optimal multi-hop path and feeds the selection result back to the traffic allocation module. Path selection optimization is affected by information age and queue status, and therefore interacts closely with the information age calculation module.

[0197] Traffic allocation module: Based on information age, network congestion, and satellite node queue length, it optimizes traffic allocation and feeds back the traffic scheduling results to the packet generation rate adjustment module and the path selection module. By optimizing traffic allocation, it influences the data forwarding rate and path selection strategy.

[0198] The reinforcement learning module adjusts path selection and traffic allocation strategies based on network status (including feedback information such as information age, path selection, and traffic allocation), and optimizes long-term performance. This module interacts with all other modules, providing dynamic adjustment and adaptive optimization capabilities.

[0199] The specific workflow of this system is as follows:

[0200] (1) Initialize path weights and link occupancy frequency

[0201] The system first initializes the path set R and the link occupancy frequency matrix. The initial value is set to zero. Simultaneously, based on satellite ephemeris information, a network topology graph G is constructed, containing all nodes V and links E.

[0202] (2) Calculate the shortest path

[0203] Based on each time slice τ i Given the topology information, use Dijkstra's algorithm to calculate the shortest path between the source node and the target node. Store the shortest path in a path set R and update the link occupancy frequency on that path.

[0204] (3) Dynamically allocate link weights

[0205] To avoid path selection getting trapped in local optima, link weights are dynamically adjusted: if the link occupancy frequency... Less than the threshold The weights are randomly assigned between the interval [r1, r2]; if the link occupancy frequency is greater than the threshold, the weights are assigned between the interval [r2, r3].

[0206] (4) Iteratively find candidate paths

[0207] After each shortest path is selected, the links used in the current path are removed from the graph, and the shortest paths on the remaining links are recalculated. This process is repeated until K candidate paths that meet the criteria are found.

[0208] (5) Output candidate paths

[0209] The output consists of multiple disjoint paths, which can be used for subsequent data traffic segmentation and transmission to ensure load balancing and improve link utilization and data transmission efficiency.

[0210] Please see Fig. 3 This is a diagram illustrating the dynamic routing decision training process based on the Soft Actor-Critic (SAC) algorithm, according to an embodiment of the present invention. This dynamic routing decision training process is... Fig. 2The reinforcement learning module executes the algorithm, selecting multiple candidate paths from the path selection module and training a model based on the satellite network status to minimize information age in multi-path transmission. To address dynamic changes in transmission links and variations in resource distribution, the algorithm models the routing decision problem as a Markov Decision Process (MDP) and utilizes the Soft Actor-Critic (SAC) algorithm from deep reinforcement learning to dynamically optimize the routing strategy. The specific steps are as follows:

[0211] (1) Initialize reinforcement learning parameters

[0212] Initialize the network parameters in the Actor-Critic framework, including the initial weights of the policy network π and the value network Q, setting the capacity of the experience pool M, and the weight parameters of the target network.

[0213] (2) State observation

[0214] The system periodically collects the environmental status S of the satellite network, including: link utilization and latency information of the current path set R; and the queue length L of the satellite nodes. i Network link congestion index C congestion .

[0215] (3) Perform the action

[0216] Based on the policy network π, allocate a traffic proportion to each path. and packet generation rate These allocation parameters are applied as action A to the current network environment.

[0217] (4) Calculate the reward function

[0218] Calculate the reward value R(s,a) for each time slice:

[0219] R(s,a)=-α·A avg (s,a)-β·C congestion (s)

[0220] Among them, A avg (s,a) represents the average information age of the system after the action is performed, C congestion (s) represents the congestion level of the link, and α and β are weight parameters.

[0221] (5) Update the policy network and value network

[0222] The result of the current action (state, reward, next state) is stored in the experience pool M. Randomly sample mini-batch data and use gradient descent to optimize the policy network and value network.

[0223] (6) Online learning and routing adjustment

[0224] The system updates the policy network in real time, adjusting traffic allocation for each path based on the latest environmental conditions to ensure that high-priority packets are prioritized for low-latency paths, thereby reducing the overall data age. It outputs the trained optimal traffic allocation policy and applies it to the data transmission process.

[0225] This invention proposes a latency-optimized adaptive routing method and system for satellite networks. By leveraging the dynamic topology of satellite networks and optimizing information age across multi-hop paths, it ensures data timeliness. Furthermore, by incorporating link time-varying characteristics into the routing strategy, this invention provides a highly adaptable solution aimed at improving the transmission efficiency of satellite networks and meeting the real-time data requirements of latency-sensitive services.

[0226] 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 latency-optimized satellite network adaptive routing method, characterized in that, Comprising the following steps: S1: Construction of multi-hop information age, specifically: S11: In the low earth orbit satellite constellation system model, assuming that the data flow is transmitted from the first satellite node to the second satellite node, and the data flow is transmitted from the second satellite node to the third satellite node, the instantaneous information age of the first satellite node in the time slot is calculated as follows: ​ S12: Introduce binary variables representing data flow of the first hop satellite node in time slot whether the first data packet is received by the first hop satellite node; when the first update packet is successfully received by the first hop satellite node in time slot =1; otherwise =0; S13: the relationship between the information age of the current satellite node and the information age of the last-hop satellite node is described as: if a data packet is received at time, the information age is expressed as ; if no data packet is received, the information age is expressed as ; a basic model of the information age is constructed; S14: Each satellite node is regarded as a service unit with limited service capacity and a queue to store data packets waiting for processing; the waiting time of data packets in the queue is modeled using the queuing theory of first-come-first-served; S15: The congestion level of the satellite network is measured by comparing the service rate of the satellite node with the data packet generation rate, and the network congestion level is modeled; S16: The connectivity constraints of the satellite network are modeled; S17: Construction of the minimum information age model; S2: Multi-path routing, dynamic calculation of multiple disjoint candidate paths through path search algorithm based on multi-hop information age optimization, to achieve traffic splitting; S3: Link state awareness, each satellite periodically sends link state information to the control satellite, including the capacity of the satellite node to receive new data packets, the queue length of the current satellite node, the current network topology information, and the overall index of the network; S4: Adaptive traffic splitting routing, based on path selection, the traffic allocation ratio and data generation rate are optimized according to dynamic link parameters through deep reinforcement learning method; S5: Dynamic routing decision, continuously iterated training according to S4, intelligent agent dynamically adjusts traffic allocation ratio and data packet generation rate; uses reward function to calculate immediate reward, continuously optimizes policy network and Q value network to minimize information age; after training, the agent regularly collects network status and makes traffic splitting decisions based on the trained policy network.

2. The delay-optimized satellite network adaptive routing method according to claim 1, characterized in that: The S11 calculates the instantaneous information age, specifically: The instantaneous information age is represented by the following formula: wherein, indicates a generation time of a first data packet of a data stream of the data stream . The basic model of the S13 construction information age Specifically: The basic model of information age is constructed based on the relationship between the information age of the current satellite node and the information age of the previous hop satellite node, represented by the following formula: wherein, denotes the total number of data streams: denotes the data stream in which the total number of data packets is counted; denotes a binary variable which indicates whether a data packet was successfully received in the time slot ; The S14, specifically: When a data packet arrives at a node The impact of congestion on the information age is modeled as an increase in the queue's waiting time , represented as follows: wherein, representing a data packet k at the time of generation of the data packet at the source node; representing a data packet k at the time of generation of the data packet at the source node; is the waiting time of the data packet k in the queue; Let the service rate of each satellite node be , which represents the number of packets that the node can process per second; let the arrival rate be , which represents the average rate at which packets arrive at the satellite node ; let the queue length be , which represents the number of packets waiting to be processed in the queue of the satellite node at time , which is expressed as follows: wherein denotes a time interval; denotes an arrival rate; The average service time is The waiting time of a data packet in the queue Can be expressed as follows: The S15, specifically: Defining congestion metrics to represent the satellite node at a time the congestion level at the time instant is represented by the following equation: wherein, represents a satellite node in the arrival rate of the time; when represents a satellite node is in a congested state; when represents a satellite node has residual processing capacity; when represents a satellite node is exactly in a balanced state; The data packet generation rate is dynamically adjusted according to the current network congestion state and the residual capacity of the satellite node The adjustment formula is expressed as follows: wherein, is the adjusted data packet generation rate, is a function of adjusting the data packet generation rate according to the congestion index and the node residual capacity; The S16, specifically: For each source node and destination node there exists a set of paths such that: wherein, represents an edge on a path, represents a capacity of an edge, represents a source node traffic demand amount;​​ The S17, specifically: The satellite network is modeled as a space-time graph, represented by the following formula: wherein, The set of satellites is denoted as , The set of links is denoted as , is a vector consisting of time slots with a time interval of ; within a time slice , the number of existing links is denoted as ; the paths between any two nodes are defined as , to achieve effective traffic splitting, denoted as the following equation: wherein, represents a time slice all paths set within the time slice; represents a time slice the first transmission path of the first data stream within the time slice; represents a time slice the second transmission path of the first data stream within the time slice; represents a time slice the third transmission path of the first data stream within the time slice; represents a time slice the fourth transmission path of the first data stream within the time slice; represents a time slice Assumed path space At observation time Maximum flow ; wherein The maximum flow of a link is the maximum capacity of the link; according to the bandwidth resources of the relevant path, each source node can split the flow into K sub-streams that pass through M paths at the same time.

3. The delay-optimized satellite network adaptive routing method according to claim 1, characterized in that: The S2, specifically comprising the following steps: S21: Input source satellite node and destination satellite node pair, ephemeris information and pre-defined random weight range, and initialize, specifically: a set of satellite pairs of input source nodes and target nodes , ephemeris information, and a predefined random weight range , a weight matrix of the initialized links and a set of candidate paths is set to 0, the link occupation frequency is set to 0, where denotes a link, denotes a time slice; S22: For each time slice , the shortest path in the current time slice is calculated by using Dijkstra shortest path algorithm based on information age , if the link is in the shortest path, the link occupation frequency is updated, in order to prevent the algorithm from falling into local optimum, the threshold setting method is adopted, if the occupation frequency of the link is less than the given threshold , the ground control center randomly allocates an integer between and as the new link metric; otherwise, the metric of the link is set to a random integer between and ; S23: Remove the link in the shortest transmission path of information age, continue to select the second shortest path using the shortest path algorithm on the remaining links, and fully utilize the idle links.

4. The delay-optimized satellite network adaptive routing method according to claim 1, characterized in that: The S3, specifically: the state of each satellite node comprises the following elements: : the capacity of the satellite node to receive new data packets; : the queue length of the current satellite node; : the current network topology information; : the overall index of the network.

5. The delay-optimized satellite network adaptive routing method according to claim 1, characterized in that: The S4, specifically: S41: Set state S includes and ; represents the capacity of the satellite node to receive new data packets; represents the queue length of the current satellite node; represents the current network topology information; represents the overall index of the network; the action A is the traffic proportion of the candidate path between all node pairs and the data packet generation rate, represented as: , each action contains: path traffic allocation weight, data packet generation rate , the reward R is the weighted sum of data information age and node congestion degree, represented as the following formula: wherein, represents the state takes action ; represents the state takes action ; represents the link congestion degree in the state ; and are weight coefficients of the information age and the congestion degree, respectively, and the greater the reward value, the more effective the path selection and the traffic distribution strategy are in optimizing the information age. S42: input the path obtained by the path search algorithm based on multi-hop information age optimization in step S2 and the link state information obtained in step S3 into reinforcement learning, and train the agent to decide the traffic ratio and data transmission rate allocated to each transmission path; wherein the update of the optimal strategy satisfies the following formula: where, represents a desire; is a temperature coefficient, used to adjust the importance of entropy in the objective function; is a discount factor; represents a state action occurs a reward; is an entropy function, representing the entropy of the policy at state is represented as the following equation: Value function is defined as the following equation: wherein, from an experience replay pool D sampled from; represents a value function, representing the expected return in state following policy . State-Value Function represents the expected reward obtained in a certain state, which is defined as follows: 。 6. A latency-optimized satellite network adaptive routing system, characterized in that, Comprise: Path selection module: based on known ephemeris information, construct a network topology snapshot graph by real-time position of satellite nodes; each time slice constitutes a topology graph, and a multi-path routing algorithm is used to calculate a plurality of candidate paths satisfying the feasible path constraint; Data packet generation rate adjustment module: adjust the data packet generation rate according to the processing capacity of the satellite node and the network congestion degree, and feed back the adjusted data packet generation rate to the traffic allocation module; the adjustment of the data packet generation rate affects the subsequent path selection and traffic allocation; Information age calculation module: calculate and update the information age of each satellite node, provide data for the path selection module, and help select the most suitable path; The information age is also fed back to the traffic allocation module to affect the optimization of data flow; After updating the information age, the calculation result is fed back to the reinforcement learning module to provide model learning and adjustment strategy; Path selection module: select the optimal multi-hop path according to the satellite ephemeris information and link state, and feed back the selection result to the traffic allocation module; Traffic allocation module: optimize traffic allocation according to information age, network congestion and satellite node queue length, and feed back the traffic scheduling result to the data packet generation rate adjustment module and the path selection module; by optimizing traffic allocation, the data forwarding rate and path selection strategy are affected; Reinforcement learning module: based on the network state, adjust the path selection and traffic allocation strategy, and optimize the long-term performance; this module interacts with other modules in the system to dynamically adjust and optimize the adaptation.

Images