Inter-satellite link satellite data transmission task scheduling method and system and storage medium

Abstract

The application discloses an intersatellite link satellite data transmission task scheduling method and system and a storage medium, and comprises the following steps: acquiring a satellite set, a task target set and a ground station set; determining a network connection plan and a data transmission plan in each time slot; acquiring a task constraint condition, establishing an intersatellite link satellite data transmission task scheduling model, and taking the maximum complete task data value transmitted to the ground station after T time slots as the scheduling model; constructing a parallel algorithm group, acquiring an optimal algorithm and corresponding parameter configuration, and matching the optimal algorithm with a scene instance. The method constructs an integer programming model of the intersatellite link satellite data transmission task model, constructs a parallel algorithm group, searches for an optimal algorithm configuration in a machine learning mode, and overcomes the contradiction between the cost and quality of huge computing resources required by parallelism.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of satellite scheduling, and particularly relates to an inter-satellite link satellite data transmission task scheduling method and system and a storage medium. BACKGROUND

[0002] The inter-satellite link satellite data transmission task model focuses on realizing efficient data transmission in a satellite network, including communication between satellite constellations and between satellites and ground stations. The main purpose of the model is to optimize the communication process through a carefully designed scheduling scheme and resource management to improve the efficiency and reliability of data transmission. In the design and execution of the task, various operating constraints and performance indicators need to be considered to ensure that the communication task can stably run under different conditions. The inter-satellite link data transmission task is a task involving periodic cooperative transmission between different satellite constellations. The main goal of these tasks is to transmit communication instructions and imaging data through cooperative transmission, while considering satellite constraints and related objective functions. Among them, communication instruction transmission refers to the exchange of communication instructions between different satellite constellations on a regular basis to ensure the normal operation of the satellite system. These instructions may involve satellite orbit adjustment, communication parameter setting, task scheduling, etc., to optimize the performance of the entire satellite network. Imaging data transmission refers to the satellite constellation carrying imaging payloads to regularly acquire image data of the Earth or other celestial bodies. Through cross-domain periodic joint data transmission tasks of the constellation, these imaging data can be transmitted to other satellite constellations or ground stations to support Earth observation, environmental monitoring and other applications. The construction of the above two transmission models is the premise of the task model.

[0003] The patent with publication number CN 108804220 A discloses a method for researching satellite task planning algorithm based on parallel computing, which includes six steps: first, constructing a multi-satellite cooperative task database; obtaining an observation element task set and a receiving element task set according to the rules and constraints of satellite operation; performing model assumption and constraint variable definition according to the constraint conditions of the satellite, and establishing a task scheduling model based on the assumption of the model. The optimization objectives established include: the sum of the priorities of completed tasks is maximum; the number of completed tasks is maximum; the sum of the observation time lengths of completed tasks is maximum; different weights are set for the two objectives to obtain a new planning objective. However, this model includes three optimization objectives, and the calculation amount is large. SUMMARY

[0004] In view of the above technical problems, the purpose of the present application is to provide an inter-satellite link satellite data transmission task scheduling method, system and storage medium, which constructs an inter-satellite link satellite data transmission task model integer programming model, constructs a parallel algorithm group, and searches for the best algorithm configuration in a machine learning manner to overcome the contradiction between the cost and quality of the huge computing resources required by parallel computing.

[0005] The technical scheme of the present application is:

[0006] An inter-satellite link satellite data transmission task scheduling method, comprising the following steps:

[0007] S01: obtaining a satellite set, a task target set and a ground station set;

[0008] S02: determining a network connection plan and a data transmission plan in each time slot;

[0009] S03: obtaining a task constraint condition, establishing an inter-satellite link satellite data transmission task scheduling model, and the inter-satellite link satellite data transmission task scheduling model is to maximize the value of complete task data transmitted to the ground station after T time slots;

[0010] S04: constructing a parallel algorithm group, obtaining a best algorithm and a corresponding parameter configuration, and matching the best algorithm with a scene instance.

[0011] In the preferred technical solution, the step S02 comprises:

[0012] a binary integer variable is introduced to represent the connectivity of a link , which is 1 if connected and 0 otherwise, is a node; a variable is introduced to represent the transmission amount of a transmission task f in a link , and a variable I(f) is introduced to represent the completion degree of the task f, which is 1 if all data is transmitted to the ground station, and 0 otherwise.

[0013] In the preferred technical solution, the task constraint condition comprises a node data flow conservation constraint, and the node data flow conservation constraint comprises:

[0014] a capacity of link transmission

[0015] a task data amount injected into a satellite system:

[0016]

[0017] for any one task data flow, the following constraint is met,

[0018]

[0019] the transmission amount of each link should be less than the maximum load of the link:

[0020]

[0021] wherein, is the transmission rate of the link, Δτ is the time length of the time slot, E l is a network graph of the link, r fFor the available amount of tasks, v i ′, v j ′ is the satellite node to which data is to be injected, E so is the link of the satellite with the task, s(f) is the source target of the task f, f is a single task, is the task set, is the cache of the task f owned by the i-th satellite in the t-th time slot, is the cache of the task f owned by the i-th satellite in the t-1-th time slot.

[0022] In the preferred technical solution, the task constraint condition comprises an inter-satellite connection quantity constraint, and the inter-satellite connection quantity constraint comprises:

[0023]

[0024] For any satellite, at most one satellite connection can be established:

[0025]

[0026] For any satellite, at most one downlink connection can be established with a ground station:

[0027]

[0028] For any ground station, at most one satellite connection can be established:

[0029]

[0030] wherein E ss represents an inter-satellite link, E sg represents a satellite-ground station link, V s is all nodes of the satellite s, is all nodes of the ground station g.

[0031] In the preferred technical solution, the task constraint condition comprises a transmission capacity constraint of a link, and the transmission capacity constraint of the link comprises:

[0032] For each transmission link, the amount of task data of a single transmission cannot exceed the upper limit of the transmission capacity of the link:

[0033]

[0034] The total amount of data cached by each satellite is not greater than the fixed storage capacity of the satellite:

[0035]

[0036] wherein, the capacity of the link transmission, B is the buffer of the task f owned by the i-th satellite in the t-th time slot, imax T is the maximum storage capacity of the satellite, and T is the total number of time slots, is the node set of all links, is the satellite node set without buffered data.

[0037] In the preferred technical solution, the task constraint condition includes the energy constraint of the satellite, and the energy constraint of the satellite includes:

[0038] The energy consumed by the satellite i for forwarding data is:

[0039]

[0040] Wherein, P is the energy consumed by the satellite i for forwarding data in the t-th time slot, ss P represents the power of forwarding data between satellites, sg P represents the power of forwarding data from the satellite to the ground station, is the capacity of the link transmission, and Δτ is the time length of the time slot;

[0041] The energy consumed by the satellite i for forwarding data in the t-th time slot is:

[0042]

[0043] Wherein, P r P is the power of receiving data by the satellite, P represents the data volume of the task f transmitted by the link, and the following constraint is established:

[0044]

[0045] Wherein, P represents the energy consumed by the satellite i for obtaining imaging data, so P is the transmission power of the satellite;

[0046] The basic energy consumed by the satellite i for maintaining normal operation is:

[0047]

[0048] Wherein, P0 represents the basic power of the satellite;

[0049]

[0050] P represents the solar energy that can be obtained by the satellite, P represents the power of the satellite c, P represents the time length of the shadow in the time slot t;

[0051] remaining energy of satellite i at the beginning of time slot t

[0052]

[0053] remaining energy at the beginning of time slot t-1, energy consumed at time slot t;

[0054] the completion degree I(f) of task f is:

[0055]

[0056] wherein, cache of task f of the i-th satellite in the t-th time slot, G is a set of all satellites, r (f) total amount of transmission data of task f.

[0057] In the preferred technical solution, the optimization objective of the inter-satellite link satellite data transmission task scheduling model is:

[0058]

[0059] wherein, w(f) is the imaging task value set by the task control center, f is a single task, task set, optimization objective of the task.

[0060] In the preferred technical solution, step S04 specifically comprises:

[0061] S41: constructing a parallel algorithm group, using algorithm performance as a feature to cluster scenario instances of inter-satellite link satellite data transmission scheduling, to obtain a scenario instance set;

[0062] S42: taking the scenario instance set as a generator of a generative adversarial network, which aims to generate scenario instances that cannot be solved by the current algorithm group with high quality, and taking the algorithm group as a discriminator, which aims to evolve algorithm configuration to solve the scenario instance set including the new scenario instance;

[0063] S43: searching for the best algorithm configuration according to the scenario instance group, performing algorithm evolution, clustering scenario instances again, and after multiple evolution iterations, locking the best algorithm under the current scenario instance set, obtaining the parameter configuration corresponding to the best algorithm, and matching the best algorithm with the scenario instance.

[0064] The application further discloses an inter-satellite link satellite data transmission task scheduling system, comprising:

[0065] an acquisition module, which acquires a satellite set, a task target set and a ground station set;

[0066] a determining module, configured to determine a network connection plan and a data transmission plan in each time slot;

[0067] a inter-satellite link satellite data transmission task scheduling model construction module, configured to obtain task constraint conditions, and establish an inter-satellite link satellite data transmission task scheduling model, wherein the inter-satellite link satellite data transmission task scheduling model is used to maximize the value of complete task data transmitted to a ground station after T time slots;

[0068] a parallel solving module, configured to construct a parallel algorithm group, obtain a best algorithm and a corresponding parameter configuration, and match the best algorithm with a scene instance.

[0069] The application further discloses a computer storage medium, which stores a computer program, and the computer program is executed to realize the inter-satellite link satellite data transmission task scheduling method.

[0070] Compared with the prior art, the application has the following beneficial effects:

[0071] 1. The method constructs an inter-satellite link satellite data transmission task model integer programming model, and maximizes communication efficiency under the constraint of limited resources (such as communication bandwidth, storage capacity and energy), and reasonable construction and implementation of the model are crucial to maintaining an efficient and reliable space communication network, especially in the case of a large amount of data transmission demand (for example, earth observation, planetary exploration, deep space communication, etc.). By optimizing the execution of periodic joint communication tasks, the operation efficiency and data processing capacity of the satellite network can be significantly improved, thereby supporting complex space tasks and scientific research activities.

[0072] 2. The parallel algorithm group is constructed, different algorithms are used to test the performance of each algorithm in multiple inter-satellite link satellite data transmission scenes, the best algorithm and the corresponding parameter configuration corresponding to different instance groups are searched from the algorithm configuration space, and the best algorithm is matched with the scene instance. By utilizing the complementarity between optimization algorithms, the algorithm group can achieve stronger overall performance than any component solver. With the aid of a parallel computer architecture mechanism, the contradiction between cost and quality of the huge computing resources required by parallel computing is overcome. A group of feasible solutions can be quickly solved. BRIEF DESCRIPTION OF DRAWINGS

[0073] The application will be further described below in combination with the drawings and embodiments:

[0074] Figure 1 The flowchart of the inter-satellite link satellite data transmission task scheduling method of the embodiment is shown in the figure;

[0075] Figure 2 The principle block diagram of the inter-satellite link satellite data transmission task scheduling system of the embodiment is shown in the figure;

[0076] Figure 3 This is a schematic diagram illustrating the evolution of the scenario instance clustering and parallel algorithm in this embodiment;

[0077] Figure 4 This is a diagram showing the co-evolution of the algorithm and instances based on generative adversarial networks in this embodiment. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0079] Example

[0080] like Figure 1 As shown, an inter-satellite link satellite data transmission task scheduling method includes the following steps:

[0081] S01: Acquire the satellite set, mission target set, and ground station set;

[0082] S02: Determine the network connection plan and data transmission plan for each time slot;

[0083] S03: Obtain the task constraints and establish an inter-satellite link satellite data transmission task scheduling model. The inter-satellite link satellite data transmission task scheduling model maximizes the value of the complete task data transmitted to the ground station after T time slots.

[0084] S04: Construct a parallel algorithm group, obtain the best algorithm and its corresponding parameter configuration, and match the best algorithm with the scenario instance.

[0085] In a preferred embodiment, step S02 includes:

[0086] Introducing binary integer variables To represent the link The connectivity status is determined by a value of 1 if the network is connected, and 0 otherwise. For nodes; introduce To represent a chain The amount of data transmitted in the transmission task f is introduced, and I(f) is introduced to represent the completion degree of task f. If all data is transmitted to the ground station, it is 1, otherwise it is 0.

[0087] In a preferred embodiment, the task constraints include node data flow conservation constraints, which include:

[0088] Link transmission capacity

[0089] Task data volume of an injection satellite system:

[0090]

[0091] For any one task data stream, the following constraints are satisfied,

[0092]

[0093] The transmission volume of each link should be less than the maximum load of the link:

[0094]

[0095] where, is the transmission rate of the link, Δτ is the time length of a time slot, E l is the network graph of the link, r f is the available task volume, v′ i , v′ j is the satellite node to which the data is injected, E so is the link of the satellite with the task, s(f) is the source target of the task f, f is a single task, and F is a task set, is the buffer of the task f of the i-th satellite in the t-th time slot, is the buffer of the task f of the i-th satellite in the t-1-th time slot.

[0096] In a preferred embodiment, the task constraint conditions include the inter-satellite connection number constraint, which includes:

[0097]

[0098] For any one satellite, at most one satellite connection can be established:

[0099]

[0100] For any one satellite, at most one downlink connection with a ground station can be established:

[0101]

[0102] For any one ground station, at most one satellite connection can be established:

[0103]

[0104] where, ss represents the inter-satellite link, E sg represents the link between the satellite and the ground station, V s is all nodes of the satellite s, is all nodes of the ground station g.

[0105] In a preferred embodiment, the task constraints include link transmission capacity constraints, which include:

[0106] For each transmission link, the amount of task data transmitted in a single transmission cannot exceed the upper limit of the link transmission capacity:

[0107]

[0108] The total amount of data stored in each satellite cannot exceed the satellite's storage capacity:

[0109]

[0110] wherein, is the capacity of the link transmission, is the cache of task f owned by satellite i in the tth time slot, B imax is the maximum storage capacity of the satellite, T is the total number of time slots, is the node set of all links, is the satellite node set without cached data.

[0111] In a preferred embodiment, the task constraints include satellite energy constraints, which include:

[0112] The energy consumed by satellite i to forward data is:

[0113]

[0114] wherein, is the energy consumed by satellite i to forward data in the tth time slot, P ss is the power of inter-satellite data forwarding, P sg represents the power of satellite data forwarding to the ground station, is the capacity of the link transmission, and Δτ is the time length of the time slot;

[0115] The energy consumed by satellite i to acquire data for forwarding in the tth time slot is:

[0116]

[0117] wherein, P r represents the power of inter-satellite data forwarding, represents the amount of data of task f transmitted by the link, and the following constraints are established:

[0118]

[0119] wherein, represents the energy consumed by satellite i to acquire imaging data, Pso Satellite transmission power;

[0120] The basic energy consumed by the satellite i to maintain normal operation is:

[0121]

[0122] Wherein, P0 represents the basic power of the satellite;

[0123]

[0124] represents the solar energy that the satellite can obtain, represents the power of the satellite c, represents the length of time in the time slot t that is in the shadow;

[0125] The remaining energy of the satellite i at the beginning of the time slot t

[0126]

[0127] is the remaining energy at the beginning of the time slot t-1, is the energy consumed at the time slot t;

[0128] The completion degree I(f) of the task f is:

[0129]

[0130] Wherein, is the cache of the task f owned by the i-th satellite in the t-th time slot, G is the set of all satellites, r (f) is the total amount of transmission data of the task f.

[0131] In a preferred embodiment, the optimization objective of the inter-satellite link satellite data transmission task scheduling model is:

[0132]

[0133] Wherein, w(f) is the imaging task value set by the task control center, f is a single task, is the task set, is the optimization objective of the task.

[0134] In a preferred embodiment, step S04 specifically comprises:

[0135] S41: Construct a parallel algorithm group, use algorithm performance as a feature to cluster the scene instances of the inter-satellite link satellite data transmission scheduling, and obtain a scene instance set;

[0136] S42: The scene instance set is used as the generator of the adversarial generative network. Its goal is to generate scene instances that the current algorithm group cannot solve with high quality. The algorithm group is the discriminator. Its goal is to evolve the algorithm configuration to solve the scene instance set, including new scene instances.

[0137] S43: Search for the best algorithm configuration based on the scene instance group, perform algorithm evolution, cluster the scene instances again, and after multiple evolution iterations, lock the best algorithm under the current scene instance set, obtain the parameter configuration corresponding to the best algorithm, and match the best algorithm with the scene instance.

[0138] In another embodiment, a computer storage medium stores a computer program that, when executed, implements the inter-satellite link satellite data transmission task scheduling method described above. The specific method is consistent with the inter-satellite link satellite data transmission task scheduling method described above, and will not be repeated here.

[0139] In another embodiment, such as Figure 2 As shown, an inter-satellite link satellite data transmission mission scheduling system includes:

[0140] The acquisition module acquires the satellite set, mission target set, and ground station set;

[0141] The module determines the network connection plan and data transmission plan for each time slot;

[0142] The inter-satellite link satellite data transmission task scheduling model construction module obtains task constraints and establishes an inter-satellite link satellite data transmission task scheduling model. The inter-satellite link satellite data transmission task scheduling model maximizes the value of the complete task data transmitted to the ground station after T time slots.

[0143] The parallel solution module constructs a group of parallel algorithms, obtains the best algorithm and its corresponding parameter configuration, and matches the best algorithm with the scenario instance.

[0144] The following example illustrates in detail the workflow of an inter-satellite link satellite data transmission task scheduling system:

[0145] To address the lack of standardized mathematical models and parallel optimization algorithms for inter-satellite link satellite data transmission missions, 1) an inter-satellite link satellite data transmission mission scheduling model was established; 2) a self-learning technology based on parallel algorithm groups was provided.

[0146] I. Establishment of Inter-Satellite Link Satellite Data Transmission Task Scheduling Model:

[0147] 1. Task Input

[0148] The variables and symbols involved in the task input are shown in Table 1.

[0149] Table 1 Task Input

[0150]

[0151] 2. Task Output

[0152] This problem requires determining the network connection plan and data transmission plan within each time slot. Therefore, to represent the connectivity of links within each time slot, this project introduces binary integer variables. To represent the link The connectivity status, A value of 1 indicates connectivity, otherwise a value of 0; (Introducing...) To represent a chain The amount of data transmitted in the transmission task f is introduced, and I(f) is introduced to represent the degree of completion of task f. If all data is transmitted to the ground station, it is 1, otherwise it is 0.

[0153] 3. Task Constraints

[0154] This section introduces the relevant constraints of the model to represent the limited resources of a small satellite system. Due to the limited resources of small satellites, energy is consumed in observation, storage, data relay, and maintaining normal operation. Additionally, solar panels acquire energy during orbital operation. To ensure the normal operation and sustainable use of the satellites, there is a lower energy limit for each satellite, and a maximum link capacity exists between each link. Indicates link From node To the node The maximum transmission capacity is given by B, where B represents the satellite's buffer data size, and Bimax represents the satellite's maximum buffer capacity. Assuming the ground station has unlimited resources and can receive data indefinitely, the constraints in the model are set as follows:

[0155] (1) Node data flow conservation constraint

[0156] Mission data is transmitted in the form of data streams within a small satellite network. The formulas represent the following: the capacity of a link between two nodes equals the link's transmission rate multiplied by the time slot length; data injection; and the out-degree of a satellite node's data volume equals its in-degree. For each mission data stream, flow conservation is followed; within a time slot t, the total transmission volume of inter-satellite and satellite-to-ground links is less than the maximum transmission value of that link. Assume that the amount of mission data acquired by the satellite in each time slot is r. f (i.e., the initial amount of data injected into the network) This represents the amount of data transmitted between satellites. This invention defines the importance of each mission, where w(f) represents the standardized value factor for each mission set by the Mission Management Committee (MMC) to indicate the value of different observation missions.

[0157]

[0158] r f Let T be the total amount of tasks, i.e. the amount of task data to be injected into the satellite system:

[0159]

[0160] v i ′, v j ′ be the satellite node to which the data is to be injected, E so be the links of the satellite with tasks.

[0161] For each task data flow, the following constraints are satisfied,

[0162]

[0163] The transmission capacity of each link should be less than the maximum capacity of the link,

[0164]

[0165] (2) Inter-satellite connection quantity constraint

[0166] The present invention assumes that the ground stations have unlimited resources, data storage is not limited, and each ground station can be connected to each other. Since the resources of small satellites are limited, the number of transceivers carried by each satellite is small, and each satellite can only establish a connection with at most one satellite and one ground station, so not all potential connections can be activated. The present invention defines a binary integer variable δ to represent the state of the potential connection, which is 1 when active, and 0 otherwise. Each ground station can only establish a connection with one satellite at the same time.

[0167]

[0168] For each satellite, at most one satellite connection can be established,

[0169]

[0170] For each satellite, at most one ground station can establish a downlink connection,

[0171]

[0172] For each ground station, at most one satellite connection can be established.

[0173]

[0174] (3) Link transmission capacity constraint

[0175] The total amount of task data transmitted in a single transmission cannot exceed the upper limit of the transmission capacity of the link,

[0176]

[0177] The total amount of data stored in each satellite cannot exceed the storage capacity of the satellite,

[0178]

[0179] (4) Energy constraints of satellites

[0180] The satellite consumes energy in maintaining basic operation, establishing communication with other satellites, and forwarding and receiving data. The solar panels on the satellite can obtain part of the energy. The formulas represent the energy consumed by the satellite in collecting data, forwarding data, receiving data, and maintaining normal operation, as well as the solar energy obtained by the satellite. Realize, where S is the length of time in the shadow, i.e. the length of time not exposed to sunlight. represents the lower bound of the amount of power the satellite has to maintain its normal operating state. θ is the battery depth set in advance.

[0181]

[0182] Esi represents the energy consumed by satellite i in forwarding data, where Pss represents the power of inter-satellite data forwarding, and Psg represents the power of satellite data forwarding to ground stations. represents the amount of data transmitted by link transmission task f.

[0183]

[0184] represents the energy consumed by satellite i in acquiring forwarding data in the tth time slot, where P r represents the power of satellite data reception, represents the amount of data transmitted by link transmission task f, and the following constraints are established.

[0185]

[0186] represents the energy consumed by satellite i in acquiring imaging data.

[0187]

[0188] represents the basic energy consumed by satellite i in maintaining normal operation, where P0 represents the basic power of the satellite.

[0189]

[0190] Si(t) represents the solar energy that satellite i can acquire at time slot t.

[0191]

[0192] Ri(t) represents the residual energy of satellite i at the beginning of time slot t, which is equal to the residual energy at the beginning of the previous time slot minus the consumed energy, plus the acquired solar energy.

[0193]

[0194] θ is an adaptive parameter, which can be set to 0.5, is the set of points in the time diagram composed of satellites s.

[0195] I(f) is the data integrity variable, which is 1 if the data is transmitted completely, and 0 otherwise, where Ri(t) represents the amount of data buffer contained by satellite i at time slot t,

[0196]

[0197] 4. Optimization objective

[0198]

[0199] The optimization objective is to maximize the value of the complete mission data transmitted to the ground station after T time slots, where w(f) is the imaging mission value set by the mission management center (MMC).

[0200] II. Self-learning method based on parallel algorithm group

[0201] In order to solve the above optimization objective, the algorithm performance is used as a feature to cluster scenario instances, and machine learning is used to search for the best algorithm configuration according to the scenario instance group, that is, algorithm evolution, and then the instances are clustered again. After multiple evolution iterations, the best algorithm under the current scenario instance set is locked. Figure 3 The process of this framework is shown. Among them, A1, A2, A3 represent different algorithms, such as transmission algorithm, hill climbing algorithm, tabu search algorithm and other algorithms, and A'1 is the algorithm after evolution of algorithm A1, and the final A"1 is the algorithm after evolution for many times. S1, S2, S3, S4, S5 represent different inter-satellite link satellite data transmission scheduling scenario instances. The scenario instances here can be data transmission requirements (such as earth observation, planetary exploration, deep space communication, etc.), and different colors represent different scenario instance sets, that is, clustering these instances into an instance set. Taking Figure 3For example, from the beginning of matching A1 with S1, S2, as the current best algorithm, to improving the algorithm configuration of A1 with the method of machine learning, evolving algorithm A'1, it is found through testing that algorithm A'1 surpasses the current best algorithm matched with it in solving S4, so it is added to the green instance set. At the same time, algorithm A2 also realizes an evolution. After multiple iterations of evolution, the best algorithm and the corresponding scene instance are matched.

[0202] Specifically, the application adds a self-learning framework to the generative adversarial network framework. The generative adversarial network (GAN) is a machine learning model composed of a generator and a discriminator. Its goal is to enable the generator to generate realistic data samples by having the generator and the discriminator learn from each other through mutual competition. In the basic framework, the task of the generator is to learn to generate new samples similar to real data samples, while the discriminator determines whether the sample to be discriminated is a real sample by classifying it. The generator and the discriminator learn through repeated iterative training.

[0203] During training, the generator tries to increasingly improve the quality of the generated samples to deceive the discriminator, while the discriminator tries to improve its discrimination accuracy to distinguish between the samples generated by the generator and the real samples. By integrating the generative adversarial network into the parallel evolution framework, the scene instance set is the generator, and its goal is to generate scene instances that cannot be solved by the current algorithm group with high quality. The algorithm group is the discriminator, and its goal is to evolve the algorithm configuration to more efficiently solve the scene instance set including the new scene instance. Under this framework, algorithms and scene instances realize co-evolution. Figure 4 A round of iteration under this framework is shown. Among them, A1, A2, and A3 represent different algorithms, S1, S2, S3, and S4 represent different scene instances, S4 is the newly generated scene instance of this round, and red represents a new algorithm or scene instance. For example, in this figure, the current algorithm group cannot solve the instance set with high quality, and S4 is generated as the target. The algorithm group is to solve the new scene instance set with high quality, and algorithms A1 and A2 evolve into A'1 and A'2, and the new scene instance is matched. After multiple iterations, a large amount of scene data will be generated, and a machine learning model based on algorithm performance will also find a better algorithm configuration.

[0204] The advantages of this method are:

[0205] A parallel algorithm group is constructed, and different algorithms are used to test the performance of each algorithm in multiple inter-satellite link satellite data transmission scenarios. The best algorithm and its corresponding parameter configuration for different instance groups are found from the algorithm configuration space, and the best algorithm is matched with the scene instance.

[0206] By exploiting the complementarity between the optimization algorithms, the algorithmic ensemble can achieve a stronger overall performance than any of its component solvers. With the aid of parallel computer architecture mechanisms, the cost versus quality tradeoff of the massive computational resources required for parallelism can be overcome. A set of feasible solutions can be quickly solved.

[0207] It should be understood that the above specific embodiments of the present application are merely exemplary or illustrative for explaining the principles of the present application, and not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made without departing from the spirit and scope of the present application should be included in the protection scope of the present application. In addition, the appended claims of the present application are intended to cover all variations and modifications falling within the scope and boundary of the appended claims, or the equivalent forms of such scope and boundary.

Claims

1. An inter-satellite link satellite data transmission task scheduling method, characterized in that, The method comprises the following steps: S01: acquiring a satellite set, a task target set and a ground station set; S02: determining a network connection plan and a data transmission plan in each time slot; S03: acquiring a task constraint condition, and establishing an inter-satellite link satellite data transmission task scheduling model, the inter-satellite link satellite data transmission task scheduling model being to maximize the value of complete task data transmitted to the ground station after T time slots; The task constraint condition comprises a node data flow conservation constraint, and the node data flow conservation constraint comprises: Capacity of link transmission ; representing a link, is a node; The amount of task data injected into the satellite system: representing chain transmission task transmission amount; For any one task data flow, the following constraint is met, The transmission amount of each link should be less than the maximum load of the link: wherein, is the rate of the link transmission, is the duration of the time slot, is the network graph of the links, is the amount of tasks available, is the satellite node into which the data is to be injected, is the link of the satellite with the task, is the task source target, is the single task, is the set of tasks, is the cache of tasks f owned by the satellite number in the time slot number , is the cache of tasks f owned by the satellite number in the time slot number . S04: constructing a parallel algorithm group, acquiring a best algorithm and a corresponding parameter configuration, and matching the best algorithm with a scene instance. 2.The method of Claim 1, wherein, The step S02 comprises: Introduce binary integer variable to represent the connectivity of link , 1 if connected, 0 otherwise, is a node; introduce to represent the transmission volume of the transmission task in the chain ; introduce to represent the completion degree of the task , 1 if all data are transmitted to the ground station, 0 otherwise. 3.The method of Claim 2, wherein, The task constraint condition comprises an inter-satellite connection quantity constraint, and the inter-satellite connection quantity constraint comprises: For any one satellite, at most, a connection can be established with one satellite: For any one satellite, at most, a downlink connection can be established with one ground station: For any one ground station, at most, a connection can be established with one satellite: wherein, denotes an inter-satellite link, denotes a satellite-to-ground station link, is the set of all nodes of satellite s, is the set of all nodes of ground station g.

4. The method of Claim 2, wherein, The task constraint condition comprises a transmission capacity constraint of a link, and the transmission capacity constraint of the link comprises: For each transmission link, the amount of task data of single transmission cannot exceed the upper limit of the transmission capacity of the link: The total amount of data stored in each satellite is not greater than the fixed storage capacity of the satellite: wherein, is the capacity of the link transmission, is the number of time slots, is the number of satellites, is the cache of task f owned by the i-th satellite, is the maximum storage capacity of the satellite, T is the total number of time slots, is the set of nodes for all links, is the set of satellite nodes without cached data.

5. The method of Claim 2, wherein, The task constraint condition comprises a satellite energy constraint, and the satellite energy constraint comprises: Satellite The energy consumed for forwarding data is: in, For the first Satellites in each time slot The energy consumed in forwarding data, The power of data relay between satellites, The power at which the satellite relays data to the ground station. For the capacity of link transmission, The duration of the time slot; Satellite In a first The energy consumed in acquiring the forwarded data within a time slot is: wherein, P is the power for satellite receiving data, is the data volume of the link transmission task , the following constraint is established: wherein represents a satellite energy consumed at the time of observation, is the satellite transmission power; Satellite The basic energy consumed to maintain normal operation is: wherein, represents the base power of the satellite; In a time slot Energy consumed by the time slot is: denotes the solar energy available to the satellite, denotes the power of the satellite c, denotes the length of time in which the satellite is in the shadow during the time slot t; at the start of the time slot at the start of the time slot of the remaining energy : remaining energy at the start of the time slot, remaining energy at the start of the time slot, remaining energy at the start of the time slot, energy consumed during the time slot; Completion of task f Is: wherein, is the number of tasks f owned by the satellite, is the number of time slots in the time period, is the cache of task f owned by the satellite, is the set of ground stations, is the total amount of transmission data for task f.

6. The method of Claim 1, wherein, The optimization target of the inter-satellite link satellite data transmission task scheduling model is: wherein, an imaging task value for a mission control center, for a single task, for a collection of tasks, an optimization goal for a task.

7. The method of Claim 1, wherein, The step S04 specifically comprises: S41: constructing a parallel algorithm group, using algorithm performance as a feature to cluster scene instances of inter-satellite link satellite data transmission scheduling, and obtaining a scene instance set; S42: taking the scene instance set as a generator of a generative adversarial network, the target of the generator being to generate a scene instance that cannot be solved by a current algorithm group, and taking the algorithm group as a discriminator, the target of the discriminator being to evolve algorithm configuration to solve the scene instance set including the new scene instance; S43: searching for a best algorithm configuration according to the scene instance group, evolving the algorithm, clustering the scene instance again, locking the best algorithm under the current scene instance set through multiple evolution iterations, acquiring a parameter configuration corresponding to the best algorithm, and matching the best algorithm with the scene instance.

8. An inter-satellite link satellite data transmission task scheduling system, characterized in that, It comprises: An acquisition module that acquires a satellite set, a task target set and a ground station set; A determination module that determines a network connection plan and a data transmission plan in each time slot; An inter-satellite link satellite data transmission task scheduling model construction module that acquires a task constraint condition, and establishes an inter-satellite link satellite data transmission task scheduling model, the inter-satellite link satellite data transmission task scheduling model being to maximize the value of complete task data transmitted to the ground station after T time slots; The task constraint condition comprises a node data flow conservation constraint, and the node data flow conservation constraint comprises: Capacity of link transmission ; representing a link, is a node; The amount of task data injected into the satellite system: representative chain transmission task transmission amount; For any one task data flow, the following constraint is met, The transmission amount of each link should be less than the maximum load of the link: wherein, is the rate of the link transmission, is the duration of the time slot, is the network graph of the links, is the amount of tasks available, is the satellite node into which the data is to be injected, is the link of the satellite with the task, is the task source target, is the single task, is the set of tasks, is the cache of tasks f owned by the satellite number in the time slot number , is the cache of tasks f owned by the satellite number in the time slot number , A parallel solution module is configured to build a parallel algorithm group, acquire an optimal algorithm and a corresponding parameter configuration, and match the optimal algorithm with a scenario instance.

9. A computer storage medium having stored thereon a computer program, characterized in that The computer program, when executed, implements the inter-satellite link satellite data transmission task scheduling method of any one of claims 1-7.

Images