Low earth orbit satellite system-oriented data compression and task scheduling joint optimization method

Abstract

The invention discloses a data compression and task scheduling joint optimization method for a low earth orbit satellite system, and belongs to the technical field of spatial information. The method comprises the steps that a low-orbit satellite system obtains an unloading task request and task information, a task to be unloaded is screened out, and then a feasible time window set is constructed; according to the task information, the feasible time window set and the compression ratio set of each low earth orbit satellite, constructing a multi-resource conflict graph to describe a communication and computing resource joint allocation conflict restriction relationship in a data unloading process; and finally, solving a maximum point weight independent set of the multi-resource conflict graph by using a classical algorithm, and generating a data compression and task scheduling joint allocation scheme. The method is simple in steps and convenient to use, the multi-resource conflict graph model is designed and used for representing conflicts in communication and computing resource joint optimization, the complexity of multi-dimensional resource joint allocation is greatly reduced, and efficient and accurate matching of massive space tasks and multi-dimensional resources is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of space information technology, and in particular relates to a data compression and task scheduling joint optimization method for a low-orbit satellite system, which is used for satellite data compression and task scheduling of a low-orbit satellite system. Background Art

[0002] With the surge in global information, satellite networks have become an important space component of 5G and its subsequent 6G networks, with their remarkable features such as seamless global coverage and high data transmission rates. Compared with medium-orbit and high-orbit satellites, low-orbit satellites can provide shorter communication latency, lower launch costs and higher networking flexibility. In view of these advantages, many commercial satellite companies, including SpaceX, Telesat and Eutelsat, have launched a large number of low-orbit satellites to establish low-orbit satellite systems. At present, low-orbit satellite systems are widely used in various fields such as wireless communications and earth observation, promoting the rapid expansion of space missions.

[0003] However, the communication time window between low-orbit satellites and ground stations is intermittent and very limited, which greatly reduces the data offloading efficiency of low-orbit satellite systems. Therefore, how to efficiently allocate these intermittent and limited communication time windows to adapt to the rapidly growing space missions has become a core challenge for improving the data offloading performance of low-orbit satellite systems. However, in previous studies on low-orbit satellite data offloading, the impact of computing resources on network offloading efficiency has been ignored, which often makes it difficult to fully decompose task transmission conflicts within the time window, thus greatly affecting the efficient use of low-orbit satellite system resources. Summary of the invention

[0004] In view of the shortcomings of the existing technology, a data compression and task scheduling joint optimization method for low-orbit satellite systems is provided. By introducing computing resources in low-orbit satellite task scheduling, the offloaded tasks are reasonably compressed. Then, a data compression and task scheduling joint optimization method for low-orbit satellite systems is proposed to improve the efficiency of low-orbit satellite system task planning.

[0005] To achieve the above technical objectives, the present invention discloses a data compression and task scheduling joint optimization method for a low-orbit satellite system. For a low-orbit satellite system constructed with S low-orbit satellites and H ground stations, each low-orbit satellite and each ground station is respectively equipped with a data transmitting antenna and a data receiving antenna; a multi-resource conflict graph model is constructed, and the multi-resource conflict graph model is used to characterize the conflict constraint relationship of the joint allocation of communication and computing resources that evolves complexly over time under a high dynamic network state during a typical data unloading process of the low-orbit satellite system; thereby, the problem of joint allocation of communication and computing resources for the low-orbit satellite system is converted into a maximum point weight independent set problem in the multi-resource conflict graph model, the complexity of the data unloading algorithm is simplified, and efficient and accurate matching of massive space tasks with multi-dimensional network resources is achieved;

[0006] The specific steps are as follows:

[0007] S1. Obtain S low-orbit satellites. According to the orbital parameters of the low-orbit satellites and the geographic coordinate information of H ground stations, input the orbital parameters and geographic coordinate information into the Satellite Tool Kit software, calculate the time window information between the low-orbit satellites and each ground station, and construct a time window set to obtain the data compression ratio set of each low-orbit satellite s for all ground stations to perform communication tasks;

[0008] S2. The low-orbit satellite receives the satellite data offloading request sent by each user through the ground station, obtains the offloading task information and constructs the offloading task set. J represents the total number of unloaded tasks; unloaded tasks include: the number of the low-orbit satellite that unloads the task s(j), the earliest unloading time a j , the latest unloading time of the task is d j , the amount of data for the task D j , the weight value of the task ω j , where j represents the unloading task number;

[0009] S3. The low-orbit satellite performs a preliminary screening of the obtained unloading task set. The preliminary screening process is as follows: For each unloading task in, first find the satellite number s(j) of the unloading task, and then obtain the time window set TW belonging to the low-orbit satellite s(j) s(j) Finally, according to the latest uninstallation time d of the uninstallation task j Determine whether there is a feasible time window for unloading, thereby eliminating unloaded tasks that cannot be completed, thereby constructing the tasks to be unloaded and forming a set of tasks to be unloaded

[0010] S4, preprocessing the tasks to be uninstalled that have been initially screened out: assigning a feasible time window set to each task to be uninstalled according to the task information and time window set of the tasks to be uninstalled that have been initially screened out;

[0011] S5. Construct a multi-resource conflict graph and construct point weights in the multi-resource conflict graph. The joint allocation of communication and computing resources for the satellite task scheduling problem is jointly represented by the conflict graph: According to the task information, the set of feasible time windows and the set of data compression ratios of each low-orbit satellite, the set V of multi-resource conflict graph points, the set E of multi-resource conflict graph edges and the set W of multi-resource conflict graph point weights are constructed respectively, so as to construct the multi-resource conflict graph G. conflict (V, E, W), thereby describing the conflict constraint relationship between the joint allocation of communication and computing resources between S low-orbit satellites and H ground stations during data unloading; wherein, the point weight of any point in the set V of multi-resource conflict graph points is calculated by using the Tchebycheff method, and the multi-resource conflict graph G is traversed. conflict All points in the point set V (V, E, W) can construct a multi-resource conflict graph G conflict The point weight value set W of (V, E, W);

[0012] S6. Solve the multi-resource conflict graph G conflict The maximum point weight independent set of (V, E, W) is used to make use of the multi-resource conflict graph G conflict The maximum point weight independent set of (V, E, W) realizes the temporal conflict decomposition of communication resources within the time window and generates data compression and task scheduling schemes for low-orbit satellites.

[0013] Furthermore, the time window set between all low-orbit satellites and all ground stations calculated by the Satellite Tool Kit software is Among them, TW k =(st k ,et k ,s,h) and represents the visible time window numbered k and the time window index set between the data receiving antenna of the low-orbit satellite s and the ground station h, respectively, k et k Respectively represent the start and end time of time window k, and further obtain the compression ratio set of any low-orbit satellite s for any task j

[0014] Further, exclude The tasks that are obviously unfinished in the task list are used to build a collection of tasks to be uninstalled. The process of eliminating unfinishable tasks is as follows:

[0015] 3a. Initialization: Task set to be uninstalled j represents the unloading task number;

[0016] 3b. Assume that the satellite number in the unloading task is fixed, and select the satellite number from the unloading task set Get the number j of the unloading task to be determined, and then get the low-orbit satellite number s(j), from the time window set TW = {TW k} to find the time window corresponding to the low-orbit satellite number s(j), and construct the time window set belonging to the low-orbit satellite s(j) Where h represents the ground station index corresponding to the data receiving antenna, Respectively represent the time window index set between the low-orbit satellite s(j) and the data receiving antenna of the ground station h, and then obtain the time window index set between the low-orbit satellite s(j) and the data receiving antenna of all ground stations

[0017] 3c. Initialize time window index number k=1;

[0018] 3d, from the time window set TW belonging to the low-orbit satellite s(j) s(j) Get the start time st of the time window with index number k k ;

[0019] 3e. Determine the starting time st of time window k k Is it less than the latest unloading time d of unloading task j? j :

[0020] If k <d j , then execute step 3f;

[0021] If k ≥d j , determine the time window set TW s(j) Is the traversal complete? If yes, update the task number j=j+1 and return to step 3b; otherwise, update the time window index number k=k+1 and make the following judgment:

[0022] If k belongs to the set Then return to step 3d;

[0023] Otherwise, end the current step;

[0024] 3f. Add the unloading task j that is determined to have a feasible time window to the set of unloading tasks without renumbering. In the process, update the task number j=j+1. If the task number j<J, return to step 3b, otherwise, end the current step.

[0025] Further, treat the offload task set Preprocessing is performed, that is, according to the task information to be unloaded and the time window set, for each task to be unloaded Allocate feasible time window set FW j ={FW jk},in, Indicates the time window with index k assigned to the task j to be offloaded. represents the start time of time window k, β jk represents the end time of time window k.

[0026] 4a. For any task to be uninstalled Initialize FW j is an empty set, that is, FW j ={}, and set j as the set of tasks to be uninstalled The smallest task number in ;

[0027] 4b. From the set of tasks to be uninstalled Get the task number j to be unloaded, and then get the low-orbit satellite number s(j), and get the time window set TW of the low-orbit satellite s(j) , and set the initialization time window index number k=1;

[0028] 4c. From the time window set TW s(j) Get the time window with index number k and get the start time st of the time window k and the end time et k ;

[0029] 4d. Construct a feasible time window in β jk =min(d j ,et k ), update the firmware j =FW j ∪FW jk , determine the time window set TW s(j) Whether the traversal is completed, if yes, execute step 4d1, otherwise, execute step 4d2;

[0030] 4d1. Update task number j=j+1 and make the following judgment:

[0031] If the task number j≤J and Then return to step 4b; suppose is {1,2,3,4,5}, is {1,2,3,5}, then J is still based on the set According to the above steps, first start from Select j=1, complete step 4d and go to 4d1. After updating task number j=2, it is necessary to determine whether j=2 is in Obviously, we go to 4b. And so on until j=4, we are no longer , then return 4d1 (according to: if the task number j≤J but Then return to step 4d1) and execute j=j+1 to get j=5, which means j is not in You can skip it directly.

[0032] If task number j ≤ J but Then return to step 4d1;

[0033] If task number j>J, terminate step 4;

[0034] 4d2 Update time window index k = k + 1, if k belongs to the set And return to step 4c.

[0035] Further, construct a multi-resource conflict graph G conflict The (V,E,W) process is as follows:

[0036] 5a. Construct multi-resource conflict graph G conflict The point set V of (V,E,W) is:

[0037] 5a1. Initialization: Set j to the set of tasks to be uninstalled The smallest task number in , k = 1, and initialize the point set V of the task j to be offloaded j ={};

[0038] 5a2. Obtain the feasible time window set FW according to the task number j to be offloaded j The feasible time window with index k in And divide it into M time slots with equal intervals, that is Where τ is the length of the time slot, symbol Indicates rounding down;

[0039] 5a3. Get the data volume D of task j to be offloaded j , from the compression ratio set Select the compression ratio c from the table and calculate the time required for the unloading task j to be completed in the feasible time window. The transmission time required Among them, R jk represents the low-orbit satellite s(j) in the feasible time window FW jk The transmission rate within, where c = 1 means that the offload task j is not compressed, and then by traversing the set For all compression ratios, construct offloading task j in the feasible time window The processing time set P in jk ={p jkc};

[0040] 5a4. From the processing time set P jk ={p jkc} to extract the processing time p corresponding to the selected compression ratio c jkc , execute steps 5a4-1 and 5a4-2 in sequence to construct the point set V of offloading task j under compression ratio c in time window k jkc , and traverse the compression ratio set All compressions in the time window k are obtained to obtain the point set V of the task j to be offloaded in the time window k. jk = {V jkc};

[0041] 5a4-1, with For the start time and Constructing a time block for the end moment In the time window Slide backwards according to each time slot, thus forming For the start and end times, and will correspond to the processing time p jkc The start and end time set of Where σ represents the start and end time The number corresponding to the processing time p is obtained. jkc The number set Λ={σ};

[0042] 5a4-2. For any pair of start and end times numbered σ∈Λ Build as G conflict A point in (V,E,W) in Indicates the start time, Indicates the end time, traverses all the numbers in the number set Λ, and then constructs the set V jkc = {v jkσ ,σ∈Λ};

[0043] 5a5. Update set V j =V j ∪V jk , update time window index k = k + 1, if k belongs to the set Then execute step 5a3, otherwise proceed to step 5a6;

[0044] 5a6. Update the set V = V ∪ V j and task number j = j + 1, if task number j ≤ J and Then return to step 5a2, otherwise the construction of the point set V is completed; any task j obtains different transmission times in different time windows, and each transmission time forms a pair of start and end times in its window. Each start and end time corresponds to a point, so multiple points are derived from a task, and all tasks are traversed in turn to construct a set V of multi-resource conflict graph points;

[0045] 5b. Determine whether there is a conflict edge between any two points u and v in the set V of multi-resource conflict graph points. If so, add a conflict edge between the two points u and v. Specifically, first obtain the index numbers i and j of the unloaded tasks of points u and v, the start time of point u The end time of point u The ground station index h(u) of point u, the starting time of point v Point v end time and the ground station index h(v) of point v, and make the following judgment:

[0046] If i=j, add a conflict edge between two points u and v; if i≠j, determine the time interval and Is there an intersection? If not, there is no need to add a conflict edge between the two points u and v. Otherwise, it is necessary to determine whether the condition s(i) = s(j) or h(u) = h(v) holds: If one of the conditions s(i) = s(j) or h(u) = h(v) holds, add a conflict edge between the two points u and v. Otherwise, do not add it. Repeat the above steps. After traversing all any two points in the set V of multi-resource conflict graph points, the set E of multi-resource conflict graph edges can be constructed.

[0047] 5c. Constructing multi-resource conflict graph G conflict The point weight set W of (V, E, W);

[0048] 5c1. Take point u from the set V of multi-resource conflict graph points in turn, and obtain the task number j and time window index k of the point, starting at time Ending time

[0049] 5c2. Get the data volume D of unloaded task j according to task number j j and weight value ω j ;

[0050] 5c3. According to the task number j and the time window index k, obtain the transmission rate R of the low-orbit satellite s(j) in the time window k jk , use the following formula to calculate the data compression ratio c of the task corresponding to point u:

[0051]

[0052] 5c4. Use the Tchebycheff method to calculate the weight value w corresponding to point u:

[0053]

[0054] Among them, J represents the total number of tasks, θ is a constant between 0 and 1, here it is 0.5, ω j represents the weight value of task j, ω max represents the maximum weight value of the task, c represents the data compression ratio of point u corresponding to task j, D j represents the amount of data for task j, D max Represents the maximum data volume of the task. According to the above steps, all points in the set V of multi-resource conflict graph points can be traversed to construct the point weight set W of the multi-resource conflict graph.

[0055] A computer device, characterized in that it includes a processor and a memory, wherein the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used to execute a joint optimization method for data compression and task scheduling for a low-orbit satellite system.

[0056] Compared with the prior art, the present invention has the following advantages:

[0057] 1) The present invention introduces computing resources in the process of low-orbit satellite data unloading, thereby reasonably compressing the mission data to improve the utilization efficiency of low-orbit satellite system communication resources, which will help to adapt to massive mission unloading needs in the future.

[0058] 2) The present invention constructs a multi-resource conflict graph to characterize the conflict constraint relationship between the compression ratio selection of low-orbit satellites in the low-orbit satellite system and the communication resource demand of the unloading task scheduling within the time window, and transforms the complex data unloading problem with time window constraints into the maximum point weight independent set problem in the multi-resource conflict graph, which greatly reduces the complexity of solving the problem and effectively improves the utilization efficiency of the resources of the low-orbit satellite system.

[0059] 3) The present invention uses the Tchebycheff method to transform the two optimization objectives of maximizing the task weight and minimizing the total data loss into a single optimization objective, which can ensure the task weight and maximize the minimum data loss at the same time, and is beneficial to reduce unnecessary data loss caused by the introduction of computing resources to task data while ensuring the system data unloading efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of the data compression and task scheduling joint optimization method for low-orbit satellite systems of the present invention;

[0061] Figure 2 A schematic diagram for constructing a multi-resource conflict graph in the present invention;

[0062] Figure 3 A simulation comparison curve diagram of target values ​​obtained by using the embodiment of the present invention and four existing data unloading solutions;

[0063] Figure 4 A simulation comparison curve diagram of the sum of weights of tasks obtained by using the embodiment of the present invention and four existing data offloading schemes;

[0064] Figure 5 It is a simulation comparison curve diagram of total data loss obtained by using the embodiment of the present invention and four existing data unloading solutions. DETAILED DESCRIPTION

[0065] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0066] The present invention discloses a data compression and task scheduling joint optimization method for a low-orbit satellite system.

[0067] This example starts from the unloading of satellite data from a low-orbit satellite to a ground station in a low-orbit satellite system to illustrate the implementation process of the present invention. The low-orbit satellite system of the present invention is composed of S low-orbit satellites and H ground stations. and Respectively represent the low-orbit satellite set and the ground station set, wherein each low-orbit satellite and each ground station is equipped with a data transmitting antenna and a data receiving antenna.

[0068] See also Figure 1 , the implementation steps are as follows:

[0069] Step 1: The mission planning and dispatching center of the low-orbit satellite system calculates the time window information based on the orbital parameters of the low-orbit satellite and the geographical location of the ground station, constructs a time window set, and obtains the compression ratio set of each low-orbit satellite.

[0070] The Mission Planning and Scheduling Center imports the latitude and longitude of each ground station and the orbital parameter information of the low-orbit satellite into the Satellite Tool Kit (STK) software to calculate the time window set between the low-orbit satellite and each ground station. Where TWk=(stk,etk,s,h) and represents the visible time window numbered k and the time window index set between the data receiving antenna of the low-orbit satellite s and the ground station h, respectively, k et k Respectively represent the start and end time of time window k, and further obtain the compression ratio set of any low-orbit satellite s for any task j

[0071] Step 2: The mission planning and dispatching center of the low-orbit satellite system receives satellite data offloading task requests sent by each user, obtains offloading task information, and constructs an offloading task set. Where J is the total number of offload tasks.

[0072] The format of uninstall task information is: Task j =(s(j),a j ,d j ,D j ,ω j ), where j represents the unloading task number, s(j) represents the low-orbit satellite number of unloading task j, and a j represents the earliest unloading time of unloading task j, d j represents the latest unloading time of unloading task j, D j represents the amount of data offloaded from task j, ω j represents the weight of offloading task j.

[0073] Step 3: Preliminary screening of the requested offload tasks, exclusion of unfinished tasks, and construction of a set of tasks to be offloaded

[0074] (3a) Initialization:

[0075] (3b) Obtain the task number j from the unloaded task set J, and then obtain the low-orbit satellite number s(j), and from the time window set TW = {TW k} to find the time window corresponding to the low-orbit satellite number s(j) and construct a time window set Where h represents the ground station index corresponding to the data receiving antenna, They represent the time window index set between the data receiving antenna of the low-orbit satellite s(j) and the ground station h respectively;

[0076] (3c) Initialize time window index k=1;

[0077] (3d) From the time window set TW s(j) Get the start time st of the time window with index number k k ;

[0078] (3e) Determine the starting time st k Is it less than the latest unloading time d of task j? j :

[0079] If k <d j , then execute step (3f);

[0080] If k ≥dj , determine the time window set TW s(j) Is the traversal complete?

[0081] If yes, update the task number j=j+1 and return to step (3b);

[0082] Otherwise, update the time window index k=k+1 and return to step (3d);

[0083] (3f) Add task j to the set of tasks to be uninstalled , update the task number j=j+1. If the task number j<J, return to step (3b), otherwise, end step 3.

[0084] Step 4: Treat the uninstall task set Preprocessing is performed, that is, according to the task information to be unloaded and the time window set, for each task to be unloaded Allocate feasible time window set FW j ={FW jk},in, Indicates the time window with index k assigned to the task j to be offloaded. represents the start time of time window k and β jk represents the end time of time window k.

[0085] (4a) For any task to be offloaded Initialize FW j is an empty set, that is, FW j ={}, and set j=1;

[0086] (4b) From the set of tasks to be offloaded Get the task number j to be unloaded, and then get the low-orbit satellite number s(j), and then get the time window set TW s(j) , and set the initialization time window index number k=1;

[0087] (4c) From the time window set TW s(j) Get the time window with index number k and get the start time st of the time window k and the end time et k ;

[0088] (4d) Constructing a feasible time window in β jk =min(d j ,et k ), update the firmware j =FW j ∪FW jk , determine the time window set TW s(j)Is the traversal complete? If yes, execute step (4d1); otherwise, execute step (4d2);

[0089] (4d1) Update task number j=j+1 and make the following judgment:

[0090] If the task number j≤J and Then return to step (4b);

[0091] If task number j ≤ J but Then return to step (4d1);

[0092] If task number j>J, terminate step 4;

[0093] (4d2) Update the time window index k=k+1 and return to step (4c).

[0094] Step 5: Construct a multi-resource conflict graph G according to the mission information, the feasible time window set and the optional compression ratio set of each low-orbit satellite. conflict (V, E, W) to characterize the conflict constraints between the joint allocation of communication and computing resources during data offloading, where V represents the multi-resource conflict graph G conflict The point set (V, E, W), E represents the multi-resource conflict graph G conflict The edge set of (V, E, W), W represents the multi-resource conflict graph G conflict The point weight set (V,E,W).

[0095] (5a) Constructing the multi-resource conflict graph G conflict The point set V of (V,E,W) is:

[0096] (5a1) Initialization: j=1, k=1;

[0097] (5a2) From the feasible time window set FW j Get the feasible time window with index k in And divide it into M time slots with equal intervals, that is Where τ is the length of the time slot, symbol Indicates rounding down;

[0098] (5a2) Obtain the data volume D of task j to be offloaded j , from the compression ratio set Select the compression ratio c from the table and calculate the time required for the unloading task j to be completed in the feasible time window. The transmission time required Among them, R jk represents the low-orbit satellite s(j) in the feasible time window FW jk The transmission rate within, where c = 1 means that the offload task j is not compressed, and then by traversing the set For all compression ratios, construct offloading task j in the feasible time window The processing time set P in jk ={p jkc};

[0099] (5a3) From the processing time set P jk ={p jkc} to extract the processing time p jkc , to start For the start time and Construct a time block for the end time, in the time window Slide backwards according to each time slot, thus forming The start and end times are numbered, and then a number set Λ is constructed;

[0100] (5a4) Construct a pair of start and end times of any number σ∈Λ into G conflict A point in (V,E,W) in Indicates the start time, Indicates the end time, traverses all the numbers in the number set Λ, and constructs the point set corresponding to the task j to be unloaded. Similarly, for other tasks to be unloaded, the point set V can be constructed;

[0101] See also Figure 2 In this example, two tasks to be unloaded are taken as examples. Task 1 to be unloaded can choose between uncompressed and compressed. In the former case, the compression ratio is 1, and the data block can slide three times in the time window k = 1, forming a point set {v 111 ,v 112 ,v 113}, in the case of a compression ratio of 3 / 4, the data block can slide four times in the time window k = 1 to form a point set {v 114 ,v 115 ,v 116 ,v 116}, similarly, in the two cases of no compression and compression, the point sets formed by the unloaded task 2 in the time window k = 2 are {v 221 ,v 222} and {v 223 ,v 224 ,v 225}, finally in this example, we get the point set V = {v 111 ,v 112 ,v 113 ,v 114 ,v 115 ,v 116 ,v 116 ,v 221 ,v 222 ,v223 ,v 224 ,v 225};

[0102] (5b) Traverse any two points u and v in the point set V (the subscripts of the points are omitted for ease of representation), obtain the index numbers i and j of the tasks to be offloaded at these two points, and the starting time of point u The end time of point u The ground station index h(u) of point u, the starting time of point v Point v end time and the ground station index h(v) of point v, and make the following judgment:

[0103] If i=j, add a conflict edge between two points u and v; if i≠j, determine the time interval and Is there an intersection? If not, there is no need to add a conflict edge between the two points u and v. Otherwise, it is necessary to determine whether the condition s(i) = s(j) or h(u) = h(v) holds. If one of the two holds, add a conflict edge between the two points u and v. Otherwise, do not add one. Repeat the above steps. After traversing all two points in the point set V, the edge set E can be constructed.

[0104] From step (5b), we can see that conflicting edges mainly exist in two situations: First, the task index numbers are the same, such as Figure 2 As shown, the point set generated by the unloading task 1 is {v 111 ,v 112 ,v 113 ,v 114 ,v 115 ,v 116 ,v 116}, the task index numbers of all points in this set are equal, so there are conflicting edges between any two points. The second is that the task numbers corresponding to the two points are different, but the two pairs of start and end times corresponding to the same low-orbit satellite or the same ground station have an intersection, such as Figure 2 Midpoint v 115 and v 223 ;

[0105] (5c) Constructing the multi-resource conflict graph G conflict The point weight set W of (V, E, W);

[0106] (5c1) Take point u from point set V in turn, and obtain the task number j and time window index k of the point, starting at time Ending time

[0107] (5c2) Get the data volume D according to task number j j and weight value ω j ;

[0108] (5c3) According to the task number j and the time window index k, the transmission rate R can be obtained jk , and calculate the compression ratio of the task corresponding to this point as follows:

[0109]

[0110] Among them, the start time and end time In (5c1), we get the data volume D j Obtained in (5c2);

[0111] (5c4) According to the compression ratio c obtained in (5c3), the weight value corresponding to the point is calculated using the Tchebycheff method as follows:

[0112]

[0113] Among them, J represents the total number of tasks, ω max represents the maximum weight of the task, D max Represents the maximum amount of data for the task, θ is a constant between 0 and 1, which can be set to 0.5 here. The point weight set W can be constructed by traversing all points in the point set V.

[0114] Step 6: Solve the multi-resource conflict graph G conflict The maximum point weight independent set of (V, E, W) is used to decompose the time conflicts of communication resources within the time window and generate data compression and task scheduling schemes for low-orbit satellites.

[0115] Solving the maximum independent set of vertex weights is a classic problem in graph theory, which can be solved efficiently using heuristic algorithms such as simulated annealing and genetic algorithms.

[0116] The technical effect of the present invention can be further illustrated by the following simulation.

[0117] 1. Simulation conditions

[0118] A low-orbit satellite system consisting of three low-orbit satellites and three data receiving antennas is taken as an example to verify and illustrate the advantages of the present invention.

[0119] The simulation duration is set to two hours, and the length of each time slot is three minutes. The dataset in Swate Rojanasoonthon's doctoral thesis is used to generate a set of time windows and the transmission duration of each task within its time window. The weight of each task is generated using a uniform distribution function between the interval [1,5]. The compression ratio set of each low-orbit satellite is set to {0.3, 0.5, 0.8}.

[0120] There are four comparison schemes used in the simulation: the first is the GRASP (Greedy Randomized Adaptive Search Procedure) algorithm proposed in Swate Rojanasoonthon's doctoral thesis, that is, the GRASP algorithm is used to schedule the tasks without compressing them; the second is the GRASP-RC algorithm, that is, the compression ratio is randomly selected from the compression ratio set to compress the tasks, and then the GRASP algorithm is used to schedule the tasks; the third is to use the TPTS (Two-Phase Task Scheduling) in the paper "Two-phase taskscheduling in data relay satellite systems" to schedule tasks without compressing them; the fourth is the TPTS-RC algorithm, that is, all tasks are compressed with a randomly selected compression ratio and then the TPTS algorithm is used to schedule the tasks.

[0121] 2. Simulation content and results

[0122] Simulation 1, using the above four comparison schemes and the present invention to simulate and compare the optimization target value in the low-orbit satellite system, the results are as follows Figure 3 .like Figure 3 As shown in Figure 2, the target value obtained by the present invention is significantly greater than that of the comparative schemes TPTS-RC and GRASP-RC. This is because the present invention allocates appropriate compression ratios to offload tasks according to the time window and the number of tasks. Figure 3 As shown in the figure, as the number of tasks increases, the present invention obtains higher target values ​​than TPTS and GRASP, except when the number of tasks is exactly 30. This is because under this simulation scenario and parameter settings, when the number of tasks is 30 or less, the communication resources of the low-orbit satellite system are sufficient, so there is no need to introduce computing resources to compress the tasks. However, when the number of tasks is greater than 30, the communication resources in the network are difficult to meet the task requirements, so the network performance can be improved by introducing computing resources to reasonably compress the tasks.

[0123] Simulation 2, using the above four comparison schemes and the present invention to simulate and compare the weights and total data loss of the tasks in the low-orbit satellite system, the results are as follows: Figure 4 and Figure 5 .like Figure 4 As shown in FIG. 1 , the weights and values ​​of the tasks obtained by the present invention are significantly better than those of the comparison schemes TPTS-RC and GRASP-RC. Figure 5As shown in Figure 1, the total data loss caused by the present invention is significantly lower than that of the comparative schemes TPTS-RC and GRASP-RC. This is because the present invention improves network performance by jointly optimizing data compression and task scheduling. This also emphasizes the importance of jointly optimizing data compression and task scheduling in low-orbit satellite systems.

Claims

1. A data compression and task scheduling joint optimization method for low-orbit satellite systems, characterized in that: A low-orbit satellite system is constructed for S low-orbit satellites and H ground stations, each of which is equipped with a data transmitting antenna and a data receiving antenna. A multi-resource conflict graph model is constructed, and the multi-resource conflict graph model is used to characterize the conflict constraints of the joint allocation of communication and computing resources that evolves complexly over time under a highly dynamic network state during a typical data unloading process of the low-orbit satellite system. Thus, the problem of joint allocation of communication and computing resources for the low-orbit satellite system is transformed into the problem of maximum point-weighted independent sets in the multi-resource conflict graph model, which simplifies the complexity of the data unloading algorithm and realizes efficient and accurate matching of massive space tasks with multi-dimensional network resources. The specific steps are as follows: S1. Obtain S low-orbit satellites. According to the orbital parameters of the low-orbit satellites and the geographic coordinate information of H ground stations, input the orbital parameters and geographic coordinate information into the Satellite Tool Kit software, calculate the time window information between the low-orbit satellites and each ground station, and construct a time window set to obtain the data compression ratio set of each low-orbit satellite s for all ground stations to perform communication tasks; S2. The low-orbit satellite receives the satellite data offloading request sent by each user through the ground station, obtains the offloading task information and constructs the offloading task set. J represents the total number of unloaded tasks; unloaded tasks include: the number of the low-orbit satellite that unloads the task s(j), the earliest unloading time a j , the latest unloading time of the task is d j , the amount of data for the task D j , the weight value of the task ω j , where j represents the unloading task number; S3. The low-orbit satellite performs a preliminary screening of the obtained unloading task set. The preliminary screening process is as follows: For each unloading task in, first find the satellite number s(j) of the unloading task, and then obtain the time window set TW belonging to the low-orbit satellite s(j) s(j) Finally, according to the latest uninstallation time d of the uninstallation task j Determine whether there is a feasible time window for unloading, thereby eliminating unloaded tasks that cannot be completed, thereby constructing the tasks to be unloaded and forming a set of tasks to be unloaded S4, preprocessing the tasks to be uninstalled that have been initially screened out: assigning a feasible time window set to each task to be uninstalled according to the task information and time window set of the tasks to be uninstalled that have been initially screened out; S5. Construct a multi-resource conflict graph and construct point weights in the multi-resource conflict graph. The joint allocation of communication and computing resources for the satellite task scheduling problem is jointly represented by the conflict graph: According to the task information, the set of feasible time windows and the set of data compression ratios of each low-orbit satellite, the set V of multi-resource conflict graph points, the set E of multi-resource conflict graph edges and the set W of multi-resource conflict graph point weights are constructed respectively, so as to construct the multi-resource conflict graph G. conflict (V, E, W), thereby describing the conflict constraint relationship between the joint allocation of communication and computing resources between S low-orbit satellites and H ground stations during data unloading; wherein, the point weight of any point in the multi-resource conflict graph point set V is calculated using the Tchebycheff method, and the multi-resource conflict graph G is traversed. conflict All points in the point set V (V, E, W) can construct a multi-resource conflict graph G conflict The point weight set W of (V, E, W); S6. Solve the multi-resource conflict graph G conflict The maximum point weight independent set of (V, E, W) is used to make use of the multi-resource conflict graph G conflict The maximum point weight independent set of (V, E, W) realizes the temporal conflict decomposition of communication resources within the time window and generates data compression and task scheduling schemes for low-orbit satellites.

2. The data compression and task scheduling joint optimization method for low-orbit satellite systems according to claim 1, characterized in that: The set of time windows between all low-orbit satellites and all ground stations calculated using the Satellite Tool Kit software Among them, TW k =(st k ,et k ,s,h) and represents the visible time window numbered k and the time window index set between the data receiving antenna of the low-orbit satellite s and the ground station h, respectively. k et k Represent the start and end time of time window k respectively, and further obtain the compression ratio set C of any low-orbit satellite s for any task j s(j) .

3. The data compression and task scheduling joint optimization method for low-orbit satellite systems according to claim 2 is characterized in that: Eliminate the tasks that are obviously unfinished in J and build a set of tasks to be uninstalled The process of eliminating unfinishable tasks is as follows: 3a. Initialization: Task set to be uninstalled j=1, j represents the unloading task number; 3b. Assume that the satellite number in the unloading task is fixed, and select the satellite number from the unloading task set Get the number j of the unloading task to be determined, and then get the low-orbit satellite number s(j), from the time window set TW = {TW k } to find the time window corresponding to the low-orbit satellite number s(j), and construct the time window set belonging to the low-orbit satellite s(j) Where h represents the ground station index corresponding to the data receiving antenna, Respectively represent the time window index set between the low-orbit satellite s(j) and the data receiving antenna of the ground station h, and then obtain the time window index set between the low-orbit satellite s(j) and the data receiving antenna of all ground stations 3c. Initialize time window index number k=1; 3d, from the time window set TW belonging to the low-orbit satellite s(j) s(j) Get the start time st of the time window with index number k k ; 3e. Determine the starting time st of time window k k Is it less than the latest unloading time d of unloading task j? j : If k <d j , then execute step 3f; If k ≥d j , determine the time window set TW s(j) Is the traversal complete? If yes, update the task number j=j+1 and return to step 3b; otherwise, update the time window index number k=k+1 and make the following judgment: If k belongs to the set Then return to step 3d; Otherwise, end the current step; 3f. Add the unloading task j that is determined to have a feasible time window to the set of unloading tasks without renumbering. In the process, update the task number j=j+1. If the task number j<J, return to step 3b, otherwise, end the current step.

4. The data compression and task scheduling joint optimization method for low-orbit satellite systems according to claim 1, characterized in that: To be uninstalled task collection Preprocessing is performed, that is, according to the task information to be unloaded and the time window set, for each task to be unloaded Allocate feasible time window set FW j ={FW jk },in, Indicates the time window with index k assigned to the task j to be offloaded. represents the start time of time window k, β jk represents the end time of time window k. 4a. For any task to be uninstalled Initialize FW j is an empty set, that is, FW j ={}, and set j as the set of tasks to be uninstalled The smallest task number in ; 4b. From the set of tasks to be uninstalled Get the task number j to be unloaded, and then get the low-orbit satellite number s(j), and get the time window set TW of the low-orbit satellite s(j) , and set the initialization time window index number k=1; 4c. From the time window set TW s(j) Get the time window with index number k and get the start time st of the time window k and the end time et k ; 4d. Construct a feasible time window in β jk =min(d j ,et k ), update the firmware j =FW j ∪FW jk , determine the time window set TW s(j) Whether the traversal is completed, if yes, execute step 4d1, otherwise, execute step 4d2; 4d1. Update task number j=j+1 and make the following judgment: If the task number j≤J and Then return to step 4b; If task number j ≤ J but Then return to step 4d1; If task number j>J, terminate step 4; 4d2 Update time window index k = k + 1, if k belongs to the set And return to step 4c.

5. The data compression and task scheduling joint optimization method for low-orbit satellite systems according to claim 4, characterized in that: Constructing multi-resource conflict graph G conflict The (V,E,W) process is as follows: 5a. Construct multi-resource conflict graph G conflict The point set V of (V,E,W) is: 5a1. Initialization: Set j to the set of tasks to be uninstalled The smallest task number in , k = 1, and initialize the point set V of the task j to be offloaded j ={}; 5a2. Obtain the feasible time window set FW according to the task number j to be offloaded j The feasible time window with index k in And divide it into M time slots with equal intervals, that is Where τ is the length of the time slot, symbol Indicates rounding down; 5a3. Get the data volume D of task j to be offloaded j , from the compression ratio set Select the compression ratio c from the table and calculate the time required for the unloading task j to be completed in the feasible time window. The transmission time required Among them, R jk represents the low-orbit satellite s(j) in the feasible time window FW jk The transmission rate within, where c = 1 means that the offload task j is not compressed, and then by traversing the set For all compression ratios, construct offloading task j in the feasible time window The processing time set P in jk ={p jkc }; 5a4. From the processing time set P jk ={p jkc } to extract the processing time p corresponding to the selected compression ratio c jkc , execute steps 5a4-1 and 5a4-2 in sequence to construct the point set V of offloading task j under compression ratio c in time window k jkc , and traverse the compression ratio set All compressions in the time window k are obtained to obtain the point set V of the task j to be offloaded in the time window k jk = {V jkc }; 5a4-1, with For the start time and Constructing a time block for the end moment In the time window Slide backwards according to each time slot, thus forming For the start and end times, and will correspond to the processing time p jkc The start and end time set of Where σ represents the start and end time The number corresponding to the processing time p is obtained. jkc The number set Λ={σ}; 5a4-2. For any pair of start and end times numbered σ∈Λ Build as G conflict A point in (V,E,W) in Indicates the start time, Indicates the end time, traverses all the numbers in the number set Λ, and then constructs the set V jkc = {v jkσ ,σ∈Λ}; 5a5 Update Set V j =V j ∪V jk , update time window index k = k + 1, if k belongs to the set Then execute step 5a3, otherwise proceed to step 5a6; 5a6 Update set V = V ∪ V j and task number j = j + 1, if task number j ≤ J and Then return to step 5a2, otherwise the construction of the point set V is completed; any task j obtains different transmission times in different time windows, and each transmission time forms a pair of start and end times in its window. Each start and end time corresponds to a point, so multiple points are derived from a task, and all tasks are traversed in turn to construct a set V of multi-resource conflict graph points; 5b. Determine whether there is a conflict edge between any two points u and v in the set V of multi-resource conflict graph points. If so, add a conflict edge between the two points u and v. Specifically, first obtain the index numbers i and j of the unloaded tasks of points u and v, the start time of point u End time of point u The ground station index h(u) of point u, the starting time of point v Point v end time and the ground station index h(v) of point v, and make the following judgment: If i=j, add a conflict edge between two points u and v; if i≠j, determine the time interval and Is there an intersection? If not, there is no need to add a conflict edge between the two points u and v. Otherwise, it is necessary to determine whether the condition s(i) = s(j) or h(u) = h(v) holds: If one of the conditions s(i) = s(j) or h(u) = h(v) holds, add a conflict edge between the two points u and v. Otherwise, do not add it. Repeat the above steps. After traversing all any two points in the set V of multi-resource conflict graph points, the set E of multi-resource conflict graph edges can be constructed. 5c. Constructing multi-resource conflict graph G conflict The point weight set W of (V, E, W); 5c1. Take point u from the set V of multi-resource conflict graph points in turn, and obtain the task number j and time window index k of the point, starting at time Ending time 5c2. Get the data volume D of unloaded task j according to task number j j and weight value ω j ; 5c3. According to the task number j and the time window index k, obtain the transmission rate R of the low-orbit satellite s(j) in the time window k jk , use the following formula to calculate the data compression ratio c of the task corresponding to point u: 5c4. Use the Tchebycheff method to calculate the weight value w corresponding to point u: Among them, J represents the total number of tasks, θ is a constant between 0 and 1, here it is 0.5, ω j represents the weight value of task j, ω max represents the maximum weight value of the task, c represents the data compression ratio of point u corresponding to task j, D j represents the amount of data for task j, D max Represents the maximum data volume of the task. According to the above steps, all points in the set V of multi-resource conflict graph points can be traversed to construct the point weight set W of the multi-resource conflict graph.

6. A computer device, characterized in that: It includes a processor and a memory, the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used to execute the data compression and task scheduling joint optimization method for low-orbit satellite systems described in any one of claims 1-5.

Images