Satellite communication network routing optimization method based on spaceborne multi-dimensional resource information

Abstract

The application provides a satellite communication network routing optimization method based on on-board multi-dimensional resource information. The method relates to a communication method, and steps include: obtaining multi-dimensional resource cost data and time delay cost data of a satellite communication network; using the multi-dimensional resource cost data and the time delay cost data, generating a multi-dimensional resource cost and time delay cost non-dominated Pareto approximation front elite solution set based on a shortest path method; based on the Pareto approximation front elite solution set, iterating a non-dominated sorting genetic algorithm until a preset termination condition is reached to obtain a Pareto optimal solution; and then mapping the Pareto optimal solution to a final transmission path as a satellite communication network routing considering on-board multi-dimensional resource information.

Description

Technical Field

[0001] This invention relates to communication methods, and more particularly to a satellite communication network routing optimization method based on onboard multidimensional resource information. Background Technology

[0002] In recent years, with the advancement of large-scale global satellite constellation construction, various satellite constellations have achieved networking through direct inter-satellite interconnection, which is crucial for achieving seamless global coverage and efficient data transmission. However, with the continuous expansion of satellite network scale and the increasing complexity of user demands and missions, large-scale satellite constellation networks face numerous challenges, including complex network structures, highly dynamic topology changes, and uneven distribution of service traffic. In particular, the processing capacity of individual satellite nodes is gradually approaching saturation, making it difficult to meet the ever-increasing mission demands. Therefore, how to achieve efficient on-board information processing and routing has become a core issue in ensuring the efficient, autonomous, and collaborative operation of satellite communication network systems.

[0003] Currently, existing inter-satellite routing primarily selects paths based on link state information and satellite communication network topology, forwarding data between two satellite nodes according to the principles of shortest path, minimum hop count, or nearest ground transmission. This approach focuses on the external link characteristics of the network while neglecting the dynamic changes and availability of multi-dimensional resources such as onboard computing and storage within the satellite communication network. It lacks consideration for the utilization of internal satellite resources, fails to fully leverage the potential of onboard computing and storage capabilities, and fails to adjust and optimize paths based on the resource conditions of different satellites, thus exhibiting certain limitations.

[0004] Meanwhile, with the improvement of onboard computing power, satellites can not only handle more local tasks, but also support other satellites when resources are idle, effectively alleviating the problem of low computing timeliness in the centralized computing mode of ground stations. However, existing routing methods fail to incorporate this internal resource information into the path selection consideration, resulting in the underutilization of the potential of onboard resources and limiting the further improvement of the overall performance of satellite communication networks. Summary of the Invention

[0005] To address the technical problems existing in the prior art, the present invention aims to provide a satellite communication network routing optimization method based on onboard multidimensional resource information, thereby achieving efficient optimization of satellite communication network routing and efficient utilization of onboard resources.

[0006] To achieve the above-mentioned objectives, this invention provides a satellite communication network routing optimization method based on onboard multi-dimensional resource information, comprising the following steps:

[0007] Obtain multi-dimensional resource overhead and latency overhead data of satellite communication networks;

[0008] The satellite communication network includes multiple satellites, ground stations, communication links between satellites, and communication links between satellites and ground stations;

[0009] The multidimensional resource overhead data includes the multidimensional resource overhead of all satellites, which includes computational overhead, storage resource overhead, and load status overhead; the latency overhead data includes the latency overhead of all communication links.

[0010] Using the aforementioned multidimensional resource overhead and latency overhead data, a Pareto approximate frontier elite solution set with non-dominated multidimensional resource overhead and latency overhead is generated based on the shortest path method.

[0011] Based on the Pareto approximate frontier elite solution set, the non-dominated sorting genetic algorithm is iterated until a preset termination condition is reached to obtain the Pareto optimal solution; then the Pareto optimal solution is mapped to the final transmission path as a satellite communication network route that considers onboard multidimensional resource information.

[0012] According to one technical solution of the present invention, before obtaining multi-dimensional resource overhead data and latency overhead data, the method further includes a step of constructing a weighted undirected graph of the satellite communication network, as follows:

[0013] The constellation operation cycle of the satellite communication network is divided into multiple time slices;

[0014] The topology of the satellite communication network within each time slice is represented by a weighted undirected graph;

[0015] The satellite communication network is a single layer, and the multiple satellites include ground-visible satellites and ground-invisible satellites;

[0016] In the weighted undirected graph, nodes represent satellites or ground stations; edges represent communication links between satellites or between a satellite and a ground station; and weights represent the distances between the nodes.

[0017] According to one technical solution of the present invention, the process of obtaining multi-dimensional resource overhead data of a satellite communication network is as follows:

[0018] Within the current time slice:

[0019] The available computing power of each satellite is calculated based on the floating-point computing power and floating-point computing power utilization of the onboard computer of each satellite; then the computing cost of each satellite is calculated using the available computing power.

[0020] Based on the storage capacity and storage capacity utilization rate of the onboard computer of each satellite, the available storage capacity of each satellite is calculated; then, the storage resource cost of each satellite is calculated using the available storage capacity.

[0021] The load status cost of each satellite is calculated based on the number of tasks to be completed for each satellite and the total number of tasks.

[0022] By utilizing the computational cost weight, storage resource cost weight, and load condition cost weight of each satellite, and combining them with the computational cost, storage resource cost, and load condition cost respectively, a weighted sum is obtained as the multidimensional resource cost of each satellite.

[0023] A multidimensional resource cost matrix is ​​constructed using the multidimensional resource costs of all satellites to represent the multidimensional resource cost data.

[0024] According to one technical solution of the present invention, the process of obtaining the latency overhead data of a satellite communication network is as follows:

[0025] The time delay between the two nodes is calculated by using the distance of the communication link between the two nodes and the propagation speed of the communication signal.

[0026] By constructing a matrix using all the delays, and scaling the matrix to a preset delay range using a linear transformation, a delay overhead matrix is ​​obtained to represent the delay overhead data.

[0027] According to one technical solution of the present invention, using the multidimensional resource overhead data and latency overhead data, an elite solution set for the Pareto approximate front is generated based on the shortest path method, as follows:

[0028] Candidate paths are defined based on a weighted undirected graph;

[0029] The candidate path includes a satellite node as the source node and a ground station node as the target node; or includes a satellite node as the source node, at least one satellite node as an intermediate node, and a ground station node as the target node.

[0030] Based on the multidimensional resource cost of each node in the candidate path, a multidimensional resource cost model is constructed.

[0031] The multidimensional resource cost model represents the sum of the multidimensional resource costs of each node in the candidate path;

[0032] Based on the latency cost of each node in the candidate path, a latency cost model is constructed.

[0033] The latency overhead model represents the sum of latency overhead of each node in the candidate path;

[0034] Minimizing the multidimensional resource overhead model and minimizing the latency overhead model are respectively taken as objectives, and a routing optimization objective function with a two-dimensional vector as the output result is constructed.

[0035] Based on the routing optimization objective function, multidimensional resource overhead data, and latency overhead data, a Pareto approximate frontier elite solution set that satisfies the preset optimization path constraints is generated using the shortest path method.

[0036] According to one technical solution of the present invention, a Pareto optimal solution is obtained as a satellite communication network route considering onboard multidimensional resource information. The process is as follows:

[0037] The Pareto approximate front elite solution set is mapped to the elite subset of the initial population in the non-dominated sorting genetic algorithm.

[0038] The initial population is obtained by using the union of randomly generated individuals and the elite subset.

[0039] Map the continuous decision variable vectors in the population to transmission paths;

[0040] The transmission path is subjected to multi-constraint checks; the multi-constraint checks include node range checks, uniqueness checks, connectivity checks, and maximum hop count checks.

[0041] The penalty value is calculated based on the results of the multi-constraint check, and a comprehensive fitness function is designed using the penalty value;

[0042] The population is adjusted using the comprehensive fitness function.

[0043] Based on the adjusted initial population, two parent individuals are selected for crossover and mutation operations to generate two offspring individuals;

[0044] The populations of parent individuals and offspring individuals are merged, and the Pareto approximate front solution is obtained by non-dominated sorting and crowding distance calculation.

[0045] The Pareto approximate front solution is used as a new population, and the non-dominated sorting genetic algorithm is repeatedly iterated until a preset termination condition is reached to obtain the Pareto optimal solution.

[0046] Map the continuous decision variable vector in the Pareto optimal solution to the final transmission path.

[0047] According to one technical solution of the present invention, the penalty value is the sum of the products of the degree to which the transmission path violates each constraint and the corresponding constraint penalty coefficient.

[0048] According to one technical solution of the present invention, the crossover operation is a single-point crossover operation, and the crossover point is random;

[0049] The mutation operation is a Gaussian mutation operation.

[0050] The present invention also provides an electronic device, comprising: one or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory to enable the electronic device to perform the above-described satellite communication network routing optimization method based on onboard multidimensional resource information.

[0051] The present invention also provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the above-described satellite communication network routing optimization method based on onboard multidimensional resource information.

[0052] Compared with the prior art, the present invention has the following advantages:

[0053] This invention discloses a satellite communication network routing optimization method based on onboard multidimensional resource information. By constructing a resource overhead model that includes multidimensional information such as computation, storage, and load, and comprehensively considering network topology and onboard resource constraints, the method optimizes the path selection process. It also injects elite solutions generated by the shortest path algorithm based on latency and resource overhead matrices into the initial population of a non-dominated sorting genetic algorithm. This method can comprehensively consider the internal resource status of the satellite and the link information of the network, efficiently optimize routing selection based on onboard multidimensional resource information, thereby optimizing the utilization efficiency of onboard resources and improving transmission performance.

[0054] Simulation results show that the proposed method maintains a higher hypervolume throughout the entire iteration process, demonstrating that its Pareto front outperforms the benchmark algorithm in both convergence and diversity, and that the solution set covers a wider range and is closer to the ideal solution.

[0055] Furthermore, the overall cost is reduced by 10.15% and 7.21% compared to Dijkstra's algorithm and the non-dominated sorting genetic algorithm that only includes inter-satellite link delay information, respectively. Compared to Dijkstra's algorithm and the non-dominated sorting genetic algorithm that only includes inter-satellite link delay information, the lowest cost is consistently achieved on all nodes. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating a satellite communication network routing optimization method based on onboard multidimensional resource information according to an embodiment of the present invention is shown.

[0058] Figure 2 The illustration shows a scenario of a satellite communication network in a satellite communication network routing optimization method based on onboard multidimensional resource information according to an embodiment of the present invention.

[0059] Figure 3 This illustration shows a comparison of Pareto solution overvolume verified by simulation in a satellite communication network routing optimization method based on onboard multidimensional resource information according to an embodiment of the present invention.

[0060] Figure 4 This illustration shows a comparison of path delay overhead and multidimensional resource overhead in a satellite communication network routing optimization method based on onboard multidimensional resource information according to an embodiment of the present invention, as verified by simulation. Detailed Implementation

[0061] The description of the embodiments in this specification should be taken in conjunction with the accompanying drawings, which should form part of the complete specification. In the drawings, the shape or thickness of the embodiments may be exaggerated and may be indicated in a simplified or convenient manner. Furthermore, parts of the various structures in the drawings will be described separately; it is worth noting that elements not shown in the figures or not described in words are in a form known to those skilled in the art.

[0062] The descriptions of the embodiments herein, including any references to directions and orientations, are for ease of description only and should not be construed as limiting the scope of the invention. The following description of preferred embodiments involves combinations of features, which may exist independently or in combination; the invention is not particularly limited to the preferred embodiments. The scope of the invention is defined by the claims.

[0063] like Figures 1-4 As shown, the satellite communication network routing optimization method based on joint multidimensional resource information (IMG-MRSP, improved multi-objective genetic algorithm based on joint multidimensional resource information and shortest path) of the present invention comprises the following steps:

[0064] S1. Obtain multi-dimensional resource overhead data and latency overhead data of the satellite communication network;

[0065] Satellite communication networks consist of multiple satellites, ground stations, communication links between satellites, and communication links between satellites and ground stations.

[0066] Multidimensional resource overhead data includes the multidimensional resource overhead of all satellites, which includes computational overhead, storage resource overhead, and load status overhead; latency overhead data includes the latency overhead of all communication links.

[0067] S2. Using multidimensional resource overhead data and latency overhead data, generate Pareto approximate frontier elite solution sets with non-dominated multidimensional resource overhead and latency overhead based on the shortest path method.

[0068] S3. Based on the Pareto approximation frontier elite solution set, the non-dominated sorting genetic algorithm is iterated until the preset termination condition is reached to obtain the Pareto optimal solution; then the Pareto optimal solution is mapped to the final transmission path as the satellite communication network route that considers the multi-dimensional resource information on the satellite.

[0069] In this embodiment, the satellite communication network routing optimization method includes:

[0070] Construct a low-Earth orbit satellite communication network, and based on its topology G(S,E) and resource information, build a path total delay cost model T(P) and a multi-dimensional resource cost model cost passing through node i. i ;

[0071] Using the shortest path algorithm, based on the delay cost matrix T (delay cost data) and the resource cost matrix C (C comp C stor C load (Resource overhead data) Generate multidimensional elite solution P d And inject it into the initial population P of the non-dominated sorting genetic algorithm. opt ;

[0072] Perform path decoding and node update checks to ensure the validity and feasibility of paths in the population;

[0073] Design a fitness function fitness(X) to comprehensively evaluate the latency and resource overhead of a path, and handle constraint violations through a penalty mechanism;

[0074] New offspring individuals are generated by crossover and mutation operations. Through multiple iterations using non-dominated sorting and crowding distance calculation, the Pareto optimal solution is finally obtained.

[0075] Before obtaining multidimensional resource overhead data and latency overhead data, the process also includes constructing a weighted undirected graph of the satellite communication network, as detailed below:

[0076] S01. Divide the constellation operation cycle of the satellite communication network into multiple time slices;

[0077] S02. Represent the topology of the satellite communication network within each time slice using a weighted undirected graph;

[0078] The satellite communication network is a single layer, consisting of multiple satellites, including those visible to the ground and those invisible to the ground.

[0079] In a weighted undirected graph, nodes represent satellites or ground stations; edges represent communication links between satellites or between satellites and ground stations; and weights represent the distances between nodes.

[0080] In this embodiment, the topology G(S,E) of the low-Earth orbit satellite communication network is constructed as follows:

[0081] like Figure 2 As shown, a single-layer low Earth orbit satellite constellation network is constructed, consisting of N satellites. pl Each of the N low-Earth orbit satellite orbital planes consists of N orbital planes. sat The constellation consists of several low-Earth orbit satellites. The operational cycle of the satellite constellation is divided into multiple time slices. Within each time slice, the network topology is regarded as a fixed virtual topology, which is represented by a weighted undirected graph G(S,E).

[0082] Where S = S m ∪S vis A set of satellite nodes; This is a set of inter-satellite links.

[0083] In this implementation, satellite nodes are divided into three categories:

[0084] Ordinary satellite node S m ={s1,s2,…,s i};

[0085] Ground-visible satellite node S vis ={s i+1 ,s i+2 ,…,s N};

[0086] Ground station node G ground ={g1,g2,…,g q};

[0087] Where N is the total number of satellites, i and j are the ordinal numbers of the satellites, and i≠j, and q is the ordinal number of the ground station.

[0088] like Figure 2As shown, each satellite maintains four inter-satellite links, connecting to two satellites in the same orbit and two satellites in adjacent orbits. Each satellite node has a specific resource cost value, reflecting the resource consumption required to reach that node. A satellite node is randomly selected as the source node to initiate a computation task. The computation task needs to be transmitted from the source node to the ground station node. In the process of selecting the transmission path, firstly, based on the satellite's satellite-to-ground visibility information, a set of satellite nodes that can serve as ground access points is determined; then, considering the connectivity of the inter-satellite links, intermediate nodes with better resource conditions and ground-visible nodes are prioritized to construct the transmission path for the computation task.

[0089] In step S1, the process of obtaining multi-dimensional resource overhead data of the satellite communication network is as follows:

[0090] Within the current time slice:

[0091] The available computing power of each satellite is calculated based on the floating-point computing power and floating-point computing power utilization of the onboard computer of each satellite; then the computing cost of each satellite is calculated using the available computing power.

[0092] Based on the storage capacity and storage capacity utilization of the onboard computer of each satellite, the available storage capacity of each satellite is calculated; then the storage resource cost of each satellite is calculated using the available storage capacity.

[0093] The load status cost of each satellite is calculated based on the number of tasks to be completed for each satellite and the total number of tasks.

[0094] By combining the computational cost weight, storage resource cost weight, and load condition cost weight of each satellite, and then combining these with the computational cost, storage resource cost, and load condition cost, a weighted sum is obtained as the multidimensional resource cost of each satellite.

[0095] A multidimensional resource cost matrix is ​​constructed using the multidimensional resource costs of all satellites to represent the multidimensional resource cost data.

[0096] In this embodiment, the calculation method for each resource expense in the multidimensional resource expense data is as follows:

[0097] 1. Satellite node computing resource utilization based on the floating-point computing power of the onboard computer and its usage rate An assessment was conducted to determine the onboard computing power of each satellite.

[0098] Based on the available computing power on the satellite The onboard computing overhead generated by this node can be expressed as follows:

[0099]

[0100] in, The normalization scaling factor for node computing power. M represents the difference in the quantization range of node computing power, defined as:

[0101] 2. Satellite node storage resource usage is based on the onboard computer's storage capacity. and its usage rate An evaluation is conducted, and the available storage capacity of node i is defined as follows:

[0102] Modeling the heterogeneity of onboard storage capacity and the stochasticity of its utilization, the storage resource overhead of satellite node i is expressed as:

[0103]

[0104] Where, χ sup and χ inf Let V represent the upper and lower bounds of the target interval for normalized available storage capacity, respectively. Define V as:

[0105] 3. The number of tasks that satellite node i has not yet completed in time slice t. (Number of tasks to be completed) and total number of tasks to be completed (Total number of tasks) Map the load condition calculation function to a uniform distribution. The load condition cost of satellite node i is expressed as follows:

[0106]

[0107] Where θ represents the cost baseline of all nodes in the satellite communication network (i.e., the minimum load cost of all nodes), Λ δ This represents a linear operator based on the overhead scaling factor δ.

[0108] Multidimensional resource overhead is calculated using a multidimensional resource overhead model.

[0109] The multidimensional resource overhead model integrates multiple parameters such as the computational overhead, storage overhead, and load status overhead of satellite nodes to construct a comprehensive cost assessment framework, providing a more efficient task allocation and data transmission scheme for resource-constrained satellite communication networks. The multidimensional resource overhead model in this implementation is represented as follows:

[0110]

[0111] Wherein, α, β, and γ are weight values ​​set according to the influence of onboard computing power, storage capacity, and load on node participation in routing selection and resource overhead quantification, satisfying α+β+γ=1.

[0112] In step S1, the process of obtaining the latency overhead data of the satellite communication network is as follows:

[0113] The time delay between the two nodes is calculated by using the distance of the communication link between them and the propagation speed of the communication signal.

[0114] By constructing a matrix using all the delays, and scaling the matrix to a preset delay range using a linear transformation, a delay overhead matrix is ​​obtained, representing the delay overhead data.

[0115] In step S2, using multidimensional resource cost data and latency cost data, an elite solution set for the Pareto approximate front is generated based on the shortest path method. The process is as follows:

[0116] Candidate paths are defined based on a weighted undirected graph;

[0117] The candidate path includes a satellite node as the source node and a ground station node as the target node; or includes a satellite node as the source node, at least one satellite node as an intermediate node, and a ground station node as the target node.

[0118] A multidimensional resource cost model is constructed based on the multidimensional resource cost of each node in the candidate path.

[0119] The multidimensional resource cost model represents the sum of multidimensional resource costs of each node in the candidate path;

[0120] Based on the latency cost of each node in the candidate path, a latency cost model is constructed.

[0121] The latency overhead model represents the sum of latency overhead of each node in the candidate path;

[0122] Minimizing the multidimensional resource overhead model and the latency overhead model are taken as objectives respectively, and a routing optimization objective function with a two-dimensional vector output is constructed.

[0123] Based on the routing optimization objective function, multidimensional resource overhead data, and latency overhead data, a Pareto approximate frontier elite solution set that satisfies the preset optimization path constraints is generated using the shortest path method.

[0124] In this embodiment, the candidate path is defined as a node sequence P = (x1, x2, ..., x...). n ), where x k ∈V P This indicates that in candidate path P, the path lies within the node set V. PThe node number of the k-th satellite, |E P |=m-1 represents the total number of hops in the candidate path, and n represents the maximum number of satellite nodes in the candidate path.

[0125] A delay overhead model is constructed based on the inter-satellite link distance matrix, and then the total delay overhead of any transmission path is calculated. To facilitate quantitative analysis of delay performance, the inter-satellite distance matrix is ​​scaled to a preset delay range through a linear transformation to generate a delay overhead matrix. For a transmission path P from the source node to the ground station, its total delay overhead T(P) is the sum of the delay overheads between all adjacent nodes on the path, expressed as:

[0126] T(P)=Δ t ·Ψ P (D P ,d min ,Δ d )

[0127] Where, d min Δ is the actual minimum value of the inter-satellite link distance. t Δ d D represents the range difference between quantization delay and distance, respectively. P Ψ is the sum of the distances between all adjacent nodes on path P. P T(P) is the path normalization cumulative factor. The smaller the value of T(P), the lower the propagation delay and the better the propagation performance of the path.

[0128] The candidate path resource cost function measures the comprehensive consumption of path P in terms of computation, storage, and load. To achieve resource allocation optimization, the resource cost R(P) is defined as the sum of the multi-dimensional resource costs of each node on the path, expressed as:

[0129]

[0130] Satellite routing optimization must balance transmission efficiency and resource efficiency, thus establishing two competing objective functions: minimizing path delay cost T(P) and minimizing resource cost R(P). Furthermore, to ensure the feasibility and practicality of the routing path within the satellite communication network, the routing process should satisfy the following constraints:

[0131] To ensure successful data transmission to the ground, the path endpoint node s is limited by the satellite's orbital motion. n It must be located in the set of currently visible satellites S at each ground station vis In, that is: s n ∈S vis .

[0132] To limit path complexity and control latency, and avoid excessive routing detours, the path hop count |E is determined based on the satellite constellation configuration. P| Must meet the maximum hop count threshold H max The limitation is: |E P |≤H max .

[0133] To ensure the physical existence of the path, and to ensure that the path nodes consist only of legal nodes conforming to the network topology definition, each node in the path must belong to the total set of nodes in the satellite communication network, i.e.: x k ∈S,

[0134] Adjacent nodes in the path need to be physically connected via inter-satellite links to ensure communication feasibility at each hop. The design constraints based on the link establishment strategy are expressed as: l k,k+1 ∈E,

[0135] To prevent node overload, ensure service stability, and avoid performance degradation due to resource overload, the resource consumption of each node in the path must not exceed the maximum threshold for computing resources, storage resources, and load status overhead. and

[0136] Based on the above constraints, considering Ω as the feasible solution space and F(P) as a two-dimensional objective vector containing path delay cost and resource cost, the final objective function for multi-objective route optimization is established as follows:

[0137]

[0138] sts n ∈S vis

[0139] |E P |≤H max

[0140]

[0141] To address the shortcomings of existing non-dominated sorting genetic algorithms, which are prone to getting trapped in local optima and have limited convergence speed when randomly initialized with sparse topologies of satellite communication networks.

[0142] This implementation proposes a multi-dimensional elite solution injection mechanism. This mechanism generates an elite solution set with a Pareto approximate front through a shortest path algorithm based on latency or node resource cost matrices, constructing an initial population with prior knowledge. The optimal path set obtained through shortest path search is represented as:

[0143]

[0144] Among them, V s and S visLet P represent the source node set and the target node set, respectively. Dijkstra This represents the optimal path from source node s to target node d, calculated using Dijkstra's algorithm based on a multidimensional resource cost matrix.

[0145] In step S3, the Pareto optimal solution is obtained and used as the satellite communication network route considering onboard multidimensional resource information. The process is as follows:

[0146] S31. Map the Pareto approximate front elite solution set to the elite subset of the initial population in the non-dominated sorting genetic algorithm.

[0147] S32. Use the union of randomly generated individuals and the elite subset as the initial population;

[0148] S33. Map the vector of continuous decision variables in the population to transmission paths;

[0149] S34. Perform path multi-constraint checks on the transmission path; path multi-constraint checks include node range checks, uniqueness checks, connectivity checks, and maximum hop count checks;

[0150] S35. Calculate the penalty value based on the results of the multi-constraint check, and design a comprehensive fitness function using the penalty value;

[0151] S36. Adjust the population using a comprehensive fitness function;

[0152] S37. Based on the adjusted initial population, select two parent individuals and perform crossover and mutation operations to generate two offspring individuals;

[0153] S38. Merge the population where the parent individual is located with the population where the offspring individual is located, and obtain the Pareto approximate front solution by non-dominated sorting and crowding distance calculation.

[0154] S39. Using the Pareto approximate front solution as a new population, repeatedly iterate the non-dominated sorting genetic algorithm until the preset termination condition is met to obtain the Pareto optimal solution.

[0155] S310. Map the continuous decision variable vector in the Pareto optimal solution to the final transmission path.

[0156] In this embodiment, the population size is preset to N. P ,Will Mapped to the initial population of a non-dominated sorting genetic algorithm The elite subset:

[0157]

[0158] This invention maps the continuous decision variable vectors representing individuals in a non-dominated sorting genetic algorithm to discrete transmission path sequences, and its decoding and construction method is as follows:

[0159]

[0160] in This is the set of visited nodes, thus avoiding loops in the path. Path decoding starts from the source node, determining subsequent nodes based on each dimension of the decision vector X. To ensure path validity, each node update requires checks on node range, uniqueness, connectivity, and maximum hop count constraints, i.e., satisfying:

[0161]

[0162] A comprehensive fitness function is designed to evaluate the performance of decision variables in terms of time delay overhead and resource utilization efficiency. A penalty mechanism (penalty factor λ) is used to effectively handle various constraints, ensuring a balance between solution feasibility and the optimization objective. The fitness function is defined as follows:

[0163]

[0164] Where T(P(X)) and R(P(X)) represent the time delay and resource cost of path P(X), respectively, and λ is the penalty factor. To ensure that the path meets multiple constraints, this implementation dynamically adjusts the penalty value according to the severity of constraint violations, defining a tiered penalty mechanism as a weighted sum of multiple constraint violation degrees, i.e.:

[0165]

[0166] Where K is the total number of constraints, ω m c is the penalty coefficient for the m-th constraint (default is 1). m (P(X)) is a function of the degree to which the path violates the m-th constraint (if the constraint-related quantity exceeds a preset threshold, the difference between it and the preset threshold is multiplied by a preset multiplier).

[0167] This implementation uses a single-point crossover method, randomly selecting a crossover point and swapping the segments of the two parent individuals X1 and X2 at that point, thereby generating the offspring Y1 = [x 1,1 ,…,x 1,k ,x 2,k+1 ,…,x 2,n ] and Y2 = [x 2,1 ,…,x 2,k ,x 1,k+1 ,…,x 1,n Gaussian mutation is used for each decision variable x. i With the probability of mutation Perform mutation, where n is the total number of decision variables. The mutated value is calculated as follows:

[0168] In step S35, the penalty value is the sum of the products of the degree to which the transmission path violates each constraint and the corresponding constraint penalty coefficient.

[0169] In step S39, the crossover operation is a single-point crossover operation, and the crossover point is random;

[0170] The mutation operation is a Gaussian mutation operation.

[0171] According to one aspect of the present invention, an electronic device is provided, comprising: one or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, and the one or more computer programs are stored in the memory; when the electronic device is running, the processor executes the one or more computer programs stored in the memory to cause the electronic device to perform the satellite communication network routing optimization method based on onboard multidimensional resource information as described above.

[0172] According to one aspect of the present invention, a computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the satellite communication network routing optimization method based on onboard multidimensional resource information as described above.

[0173] Computer-readable storage media can include any medium capable of storing or transmitting information. Examples of computer-readable storage media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and so on. Code segments can be downloaded via computer networks such as the Internet and intranets.

[0174] The present invention provides a satellite communication network routing optimization method based on onboard multidimensional resource information. The method, applicable to the field of information retrieval, comprises the following steps: Constructing a logical operation expression for satellite telemetry parameters based on multiple telemetry parameters and operators; obtaining a set of effective time periods for each telemetry parameter according to the filtering conditions of each telemetry parameter; the set of effective time periods for each parameter being the set of time periods in which the telemetry parameter meets the corresponding filtering conditions; substituting the set of effective time periods for each telemetry parameter into the logical operation expression for the satellite telemetry parameters and solving it to obtain the effective time periods for multiple telemetry parameters.

[0175] This invention employs the Walker-Delta constellation for simulation analysis, constructing a single-layer low Earth orbit network with 72 orbital planes at an altitude of 550 km. Each orbit contains 22 satellites evenly distributed, and the constellation tilt angle is 53°. Nine strategically deployed ground stations are used, and high-precision simulations are performed using satellite orbit simulation software. The inter-satellite link distance matrix and satellite-to-ground visibility matrix are obtained by combining satellite trajectories and the effects of Earth's curvature. Onboard multidimensional resource information is modeled using statistical distribution to reflect the heterogeneous characteristics of the nodes, where computing resources (unit: TFLOPS) follow a Gaussian distribution. Storage resources (unit: TB) follow a Gaussian distribution Load condition compliance Based on experience, the maximum hop count threshold for the path is set to 70, and the population size is set to 1600. Furthermore, based on the severity of constraint violations, the range distribution is set to 10. 3 ~10 5 The penalty factor is used. To comprehensively evaluate the performance of the algorithm, this invention introduces Dijkstra's algorithm and a non-dominated sorting genetic algorithm that does not include multidimensional resource information matching, i.e., only includes inter-satellite link delay information, as a comparison benchmark.

[0176] Hypervolume measures the convergence and diversity of the solution set by calculating the union volume of the hyperrectangle formed by the non-dominated solution and the reference point. Figure 3 The proposed IMG-MRSP method is compared with the non-dominated sorting genetic algorithm using only time-delay initialization in terms of hypervolume metrics over 150 iterations. Simulation results show that the proposed method maintains a higher hypervolume throughout the entire iteration process, demonstrating that its Pareto front outperforms the benchmark algorithm in both convergence and diversity, and that its solution set covers a wider range and is closer to the ideal solution.

[0177] Figure 4 A comparative analysis of path delay and overall node resource overhead is conducted. Simulation experiments were repeated 150 times on nodes initiating different tasks. Considering the priority of resource optimization for computational tasks, the total path cost is a weighted combination of delay (weight 0.3) and resource consumption (weight 0.7). Simulation results show that, compared with Dijkstra's algorithm and a non-dominated sorting genetic algorithm containing only inter-satellite link delay information, the proposed method consistently achieves the lowest cost on all nodes. For example, at node B, the overall cost of the proposed IMG-MRSP algorithm is reduced by 10.15% and 7.21% compared to Dijkstra's algorithm and the non-dominated sorting genetic algorithm containing only inter-satellite link delay information, respectively. This is because the elite solution injection strategy of the proposed method significantly enhances the multi-objective optimization process, guiding the evolutionary search towards a balance and efficiency between delay and resource utilization. Therefore, the embodiments of this invention achieve efficient optimization of satellite communication network routing and efficient utilization of onboard resources.

[0178] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code.

[0179] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0180] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0181] It should also be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0182] Finally, it should be noted that the above description represents a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A satellite communication network routing optimization method based on onboard multidimensional resource information, characterized in that, The steps are as follows: Obtain multi-dimensional resource overhead and latency overhead data of satellite communication networks; The satellite communication network includes multiple satellites, ground stations, communication links between satellites, and communication links between satellites and ground stations; The multidimensional resource overhead data includes the multidimensional resource overhead of all satellites, which includes computational overhead, storage resource overhead, and load status overhead; the latency overhead data includes the latency overhead of all communication links. Using the aforementioned multidimensional resource overhead and latency overhead data, a Pareto approximate frontier elite solution set with non-dominated multidimensional resource overhead and latency overhead is generated based on the shortest path method. Based on the Pareto approximate front elite solution set, the non-dominated sorting genetic algorithm is iterated until a preset termination condition is reached to obtain the Pareto optimal solution; then the Pareto optimal solution is mapped to the final transmission path as a satellite communication network route that considers onboard multidimensional resource information. Using the aforementioned multidimensional resource overhead data and latency overhead data, an elite solution set with approximate Pareto front is generated based on the shortest path method, as follows: Candidate paths are defined based on a weighted undirected graph; The candidate path includes a satellite node as the source node and a ground station node as the target node; or includes a satellite node as the source node, at least one satellite node as an intermediate node, and a ground station node as the target node. Based on the multidimensional resource cost of each node in the candidate path, a multidimensional resource cost model is constructed. The multidimensional resource cost model represents the sum of the multidimensional resource costs of each node in the candidate path; Based on the latency cost of each node in the candidate path, a latency cost model is constructed. The latency overhead model represents the sum of latency overhead of each node in the candidate path; Minimizing the multidimensional resource overhead model and minimizing the latency overhead model are respectively taken as objectives, and a routing optimization objective function with a two-dimensional vector as the output result is constructed. Based on the route optimization objective function, multidimensional resource overhead data, and delay overhead data, a Pareto approximate frontier elite solution set that satisfies the preset optimization path constraints is generated using the shortest path method. The Pareto optimal solution is obtained as the routing for a satellite communication network that considers onboard multidimensional resource information, and the process is as follows: The Pareto approximate front elite solution set is mapped to the elite subset of the initial population in the non-dominated sorting genetic algorithm. The initial population is obtained by using the union of randomly generated individuals and the elite subset. Map the continuous decision variable vectors in the population to transmission paths; The transmission path is subjected to multi-constraint checks; the multi-constraint checks include node range checks, uniqueness checks, connectivity checks, and maximum hop count checks. The penalty value is calculated based on the results of the multi-constraint check, and a comprehensive fitness function is designed using the penalty value; The population is adjusted using the comprehensive fitness function. Based on the adjusted initial population, two parent individuals are selected for crossover and mutation operations to generate two offspring individuals; The populations of parent individuals and offspring individuals are merged, and the Pareto approximate front solution is obtained by non-dominated sorting and crowding distance calculation. The Pareto approximate front solution is used as a new population, and the non-dominated sorting genetic algorithm is repeatedly iterated until a preset termination condition is reached to obtain the Pareto optimal solution. Map the continuous decision variable vector in the Pareto optimal solution to the final transmission path; The penalty value is the sum of the products of the degree to which the transmission path violates each constraint and the corresponding constraint penalty coefficient; The crossover operation is a single-point crossover operation, and the crossover point is random; The mutation operation is a Gaussian mutation operation.

2. The satellite communication network routing optimization method based on onboard multidimensional resource information according to claim 1, characterized in that, Before obtaining multidimensional resource overhead data and latency overhead data, the process also includes the step of constructing a weighted undirected graph of the satellite communication network, as follows: The constellation operation cycle of the satellite communication network is divided into multiple time slices; The topology of the satellite communication network within each time slice is represented by a weighted undirected graph; The satellite communication network is a single layer, and the multiple satellites include ground-visible satellites and ground-invisible satellites; In the weighted undirected graph, nodes represent satellites or ground stations; edges represent communication links between satellites or between a satellite and a ground station; and weights represent the distances between the nodes.

3. The satellite communication network routing optimization method based on onboard multidimensional resource information according to claim 2, characterized in that, The process of obtaining multidimensional resource overhead data for satellite communication networks is as follows: Within the current time slice: The available computing power of each satellite is calculated based on the floating-point computing power and floating-point computing power utilization of the onboard computer of each satellite; then the computing cost of each satellite is calculated using the available computing power. The available storage capacity of each satellite is calculated based on the storage capacity and storage capacity utilization rate of the onboard computer of each satellite. The storage resource cost for each satellite is then calculated using the available storage capacity. The load status cost of each satellite is calculated based on the number of tasks to be completed for each satellite and the total number of tasks. By utilizing the computational cost weight, storage resource cost weight, and load condition cost weight of each satellite, and combining them with the computational cost, storage resource cost, and load condition cost respectively, a weighted sum is obtained as the multidimensional resource cost of each satellite. A multidimensional resource cost matrix is ​​constructed using the multidimensional resource costs of all satellites to represent the multidimensional resource cost data.

4. The satellite communication network routing optimization method based on onboard multidimensional resource information according to claim 2, characterized in that, The process of obtaining latency overhead data for satellite communication networks is as follows: The time delay between the two nodes is calculated by using the distance of the communication link between the two nodes and the propagation speed of the communication signal. By constructing a matrix using all the delays, and scaling the matrix to a preset delay range using a linear transformation, a delay overhead matrix is ​​obtained to represent the delay overhead data.

5. An electronic device, characterized in that, include: One or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory to cause the electronic device to perform the satellite communication network routing optimization method based on onboard multidimensional resource information as described in any one of claims 1 to 4.

6. A computer-readable storage medium, characterized in that, Used to store computer instructions, which, when executed by a processor, implement the satellite communication network routing optimization method based on onboard multidimensional resource information as described in any one of claims 1 to 4.

Images