Power transmission line-oriented space-air-ground integrated network multi-dimensional resource management method and device

Abstract

The invention discloses a power transmission line-oriented resource management method and device for a space-air-ground integrated network, and the method comprises the steps: building a target function which takes user scheduling, task unloading proportion, calculation resource distribution and unmanned plane deployment position as optimization variables through minimizing the task processing time delay of a terminal as a target; the terminal is clustered, the position of the clustering center is used as the initial deployment position of the unmanned aerial vehicle, the optimization problem of the target function is converted into three sub-problems to be solved, and then the deployment position of the unmanned aerial vehicle is optimized and solved; through continuous iteration, an optimal solution of a final unmanned aerial vehicle deployment position, an optimal solution of user scheduling, an optimal solution of a task unloading proportion and an optimal solution of computing resource allocation are obtained and serve as a resource management scheme of the space-air-ground integrated network, task processing time delay can be effectively reduced, and task processing continuity is guaranteed.

Description

Technical Field

[0001] This application provides embodiments in the field of power system technology, and particularly relates to a method and apparatus for multi-dimensional resource management of an integrated air-space-ground network for transmission lines. Background Technology

[0002] The Space-Air-Ground Integrated Network (SAGIN) can provide comprehensive, high-quality coverage and connectivity for users in remote areas. By deploying edge servers on drones and satellites, drones move over users, and ground tasks can be offloaded to mobile edge computing (MEC) devices carried by the drones or to satellite cloud servers. This network architecture can significantly shorten information transmission distances, quickly respond to computing needs, greatly reduce network latency, and significantly improve network reliability and anti-interference capabilities. When a node fails, tasks can be offloaded to other nodes, ensuring the continuity of network services. SAGIN technology can provide comprehensive communication coverage for power transmission lines in remote areas where terrestrial networks are difficult to deploy. However, the three-layer space-air-ground network is heterogeneous, with significant differences in the computing capabilities of nodes in different network layers. Given the functional characteristics and inherent constraints of each network segment, different networks need to collaboratively optimize unbalanced system resources, including computing and communication resources, to better adapt to diverse service needs. This involves selectively offloading computing tasks to appropriate network devices using ubiquitous communication connections to achieve efficient resource utilization.

[0003] The integrated air-space-ground network combines the wide-area coverage of satellite networks with the maneuverability of aerial platforms such as drones, effectively making up for the shortcomings of terrestrial networks. However, limited battery capacity can sometimes lead to gaps in the drone replacement process. After low-energy drones leave the mission area, the terminal can only rely on local or satellite computing, resulting in excessive mission processing latency. Summary of the Invention

[0004] This application provides a method and apparatus for multi-dimensional resource management of an integrated air-space-ground network for power transmission lines, which can effectively reduce task processing latency and ensure the continuity of task processing.

[0005] In a first aspect, embodiments of this application provide a resource management method for an integrated air-space-ground network for power transmission lines, comprising:

[0006] With the goal of minimizing the task processing latency of the terminal, an objective function is constructed with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network;

[0007] The terminals are clustered to obtain cluster centers, and the location of the cluster centers is used as the initial deployment location of the UAV. The optimization problem of the objective function is decomposed into user scheduling optimization sub-problem, task offloading ratio optimization sub-problem, and computing resource allocation optimization sub-problem.

[0008] The initial deployment location of the UAV is fixed as the fixed deployment location of the UAV. The user scheduling optimization subproblem, the task offloading ratio optimization subproblem, and the computing resource allocation optimization subproblem are solved under the corresponding constraints to obtain the optimal solutions for user scheduling, task offloading ratio, and computing resource allocation.

[0009] The optimal solution for user scheduling, the optimal solution for task unloading ratio, and the optimal solution for computing resource allocation are input into the objective function, and the optimal solution for the UAV deployment location is obtained under the corresponding constraints.

[0010] The optimal solution for the UAV deployment location is taken as the fixed deployment location. The steps of solving the user scheduling optimization subproblem, the task offloading ratio optimization subproblem, and the computing resource allocation optimization subproblem under the corresponding constraints are returned until the deviation between the optimal solutions of the UAV deployment locations in two adjacent iterations meets the iteration termination condition. The final optimal solutions for the UAV deployment location, the final optimal solutions for user scheduling, the final optimal solutions for task offloading ratio, and the final optimal solutions for computing resource allocation are obtained and used as the resource management scheme of the integrated air-space-ground network.

[0011] Secondly, embodiments of this application provide a resource management device for an integrated air-space-ground network for power transmission lines, comprising:

[0012] The module is used to construct an objective function with user scheduling, task offloading ratio, computing resource allocation and UAV deployment location as optimization variables, with the goal of minimizing the task processing latency of the terminal; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network;

[0013] The decomposition module is used to cluster the terminals to obtain cluster centers, and use the location of the cluster centers as the initial deployment location of the UAVs. It also decomposes the optimization problem of the objective function into user scheduling optimization sub-problems, task offloading ratio optimization sub-problems, and computing resource allocation optimization sub-problems.

[0014] The first solution module is used to fix the initial deployment position of the UAV as the fixed deployment position of the UAV, and solve the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints, respectively, to obtain the optimal solution for user scheduling, the optimal solution for task offloading ratio, and the optimal solution for computing resource allocation.

[0015] The second solution module is used to input the optimal solution of user scheduling, the optimal solution of task unloading ratio and the optimal solution of computing resource allocation into the objective function, and solve for the deployment location of the UAV under the corresponding constraints to obtain the optimal solution of the UAV deployment location.

[0016] The iterative solution module is used to take the optimal solution of the UAV deployment location as the fixed deployment location, and return the steps of solving the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints, until the deviation between the optimal solutions of the UAV deployment location in two adjacent iterations meets the iteration termination condition, and obtain the final optimal solution of the UAV deployment location, the final optimal solution of the user scheduling, the final optimal solution of the task offloading ratio, and the final optimal solution of the computing resource allocation, which are used as the resource management scheme of the integrated air-space-ground network.

[0017] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method provided in embodiments of this application.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the method provided in embodiments of this application.

[0019] The technical solution provided in this application aims to minimize the task processing latency of the terminal. It constructs an objective function with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables. By clustering the terminals and using the location of the cluster center as the initial deployment location of the UAV, the optimization problem of the objective function is transformed into three sub-problems for solving. Then, the deployment location of the UAV is optimized and solved. Through continuous iteration, the optimal solutions for the UAV deployment location, user scheduling, task offloading ratio, and computing resource allocation are obtained. As a resource management solution for an integrated air-space-ground network, it can effectively reduce task processing latency and ensure the continuity of task processing. Attached Figure Description

[0020] Figure 1 A flowchart illustrating a multi-dimensional resource management method for an integrated air-space-ground network for power transmission lines, provided in an embodiment of this application;

[0021] Figure 2 A schematic diagram of a wireless communication system for power transmission lines in an integrated air-space-ground network;

[0022] Figure 3 A graph showing the relationship between satellite computing power and latency;

[0023] Figure 4 A graph showing the relationship between the terminal's local computing power and latency;

[0024] Figure 5 A graph showing the relationship between the maximum computing power and latency of a drone;

[0025] Figure 6 A graph showing the relationship between the number of terminals and latency;

[0026] Figure 7 A graph showing the relationship between task data volume and unloading ratio;

[0027] Figure 8 A structural block diagram of a multi-dimensional resource management device for an integrated air-space-ground network for power transmission lines, provided in an embodiment of this application;

[0028] Figure 9 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0029] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] Figure 1This application provides a flowchart of a multi-dimensional resource management method for an integrated space-air-ground network for power transmission lines. The method can be executed by a multi-dimensional resource management device for an integrated space-air-ground network for power transmission lines. This device can be implemented using software and / or hardware, and can be configured in electronic devices such as computers. Figure 1 As shown, the technical solution provided in this application includes the following steps:

[0031] S110: To minimize the task processing latency of the terminal, an objective function is constructed with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network.

[0032] In this embodiment, there are multiple terminals and drones. The user scheduling refers to the association between drones and terminals; the task offloading ratio includes the proportion of tasks computed locally by the terminal, the proportion of tasks offloaded to the drone, and the proportion of tasks offloaded to the satellite; the computing resource allocation includes the number of computing resources allocated by the drone to the terminal and the number of computing resources allocated by the satellite to the drone.

[0033] In this embodiment, the objective function, which aims to minimize the terminal's task processing latency and uses user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables, includes: determining the task processing latency of the terminal's local computation as a first latency; wherein the first latency uses the proportion of tasks computed locally by the terminal as an optimization variable; determining the task processing latency of the terminal offloading tasks to the UAV as a second latency; wherein the second latency uses the proportion of tasks offloaded from the terminal to the UAV, the user scheduling, and the number of computing resources allocated to the terminal by the UAV as optimization variables; determining the task processing latency of the terminal offloading tasks to the satellite as a third latency; wherein the third latency uses the proportion of tasks offloaded from the terminal to the satellite and the number of computing resources allocated to the UAV by the satellite as optimization variables; taking the maximum latency among the first latency, second latency, and third latency as the terminal's total task processing latency, and minimizing the total task processing latency to obtain the objective function.

[0034] Specifically, the technical solution in this application embodiment is based on a wireless communication system for power transmission lines using an integrated air-space-ground network. Specifically, for example... Figure 2 As shown, the integrated air-space-ground network includes A set of status monitoring terminals can be represented as follows: Assuming all devices are at the same horizontal level, the first The location of each terminal can be represented by a two-dimensional vector. ; The first The coordinates of each terminal on the x and y axes; after the low-energy UAV leaves the mission area, there are... The remaining drones, their set can be represented as ;against The remaining drones, the first The position of the drone can be represented by a three-dimensional vector. ; They are the first The coordinates of the drone along the x, y, and z axes; a low-Earth orbit satellite (referred to as a satellite) with sufficient computing power is deployed in the integrated air-space-ground network to provide full-coverage service as a cloud server, denoted as... . No. The task parameters uploaded by each terminal include ,in, Indicates the first Data volume per terminal task This represents the number of CPU cycles required to compute 1 bit of data. Indicates the first The maximum tolerable latency for each terminal. Tasks generated by each terminal can be computed locally, offloaded to the drone, or via satellite. The relationship between the drone and the terminal can be represented by a binary variable:

[0035] (1);

[0036] when At that time, the representative will be the first The computing tasks generated by the terminal are offloaded to the terminal. Mount it on a drone, otherwise Since a terminal can only establish a communication link with one drone at a time, the following issues exist:

[0037] (2);

[0038] In resource management scenarios, when a backup drone fails to take over communication support tasks from a low-energy drone, the remaining drones must provide communication and computing services to the power terminals in the area after the low-energy drone leaves. Since the maximum tolerable latency varies for different types of terminals, drone locations need to be redeployed closer to terminals with lower latency tolerance to ensure all tasks can be completed within the tolerable latency. When a terminal is outside the drone's communication coverage area and chooses to perform calculations locally or offloaded to a satellite, the drone does not need to consider the maximum tolerable latency of that terminal's task. To prevent collisions between drones, a minimum distance must be set between them. Then there are constraints. Among the remaining drones, For the first The location of the drone.

[0039] In this embodiment, considering the communication scenario of power transmission lines in remote areas, the space is open with few tall buildings or trees obstructing the view, and the drone and terminal are far apart. Therefore, the communication link between the drone and the terminal is a line-of-sight (LoS) link. The drone and the first The channel gain between the terminals is :

[0040] (3);

[0041] in, This represents the channel gain at a reference distance of 1m. It is the path loss constant, under normal circumstances. . It indicates the first The drone and the first The distance between terminals.

[0042] To avoid co-channel interference during communication, frequency division multiplexing can be used for data transmission. The terminal and the first The information transmission rate of the drone is :

[0043] (4);

[0044] in, Indicates the first The drone was assigned to the first The transmission bandwidth of each terminal; Indicates the first The terminal and the first Transmission power between drones This represents the noise power spectral density of additive white Gaussian noise.

[0045] Because the computing power of terminals (condition monitoring terminals) on lightweight transmission lines is limited, performing all tasks locally would exceed the maximum tolerable latency. Therefore, when a task is generated, tasks beyond the required computing capacity can be offloaded to drones or satellites. Drones and satellites are equipped with servers with greater computing power, which can share the computing load with the terminals. Thus, there is a task offloading ratio. , respectively representing the first The proportion of tasks computed locally by each terminal, and the task unloaded to the terminal. The proportion of missions carried out by drones and the proportion of missions unloaded onto satellites. Local computing task processing latency of each terminal The first time delay, or first delay, can be expressed as:

[0046] (5);

[0047] in, It is the first The computing power of each terminal.

[0048] Among them, the The terminal will offload the task to the first terminal. The transmission and computation latency of the drone, i.e., the first The terminal will offload the task to the first terminal. Task processing latency of drones (i.e., the second delay) is:

[0049] (6);

[0050] in, It is the first The first drone allocation The number of computing resources per terminal is calculated based on the fact that the amount of data resulting from the completed task is much smaller than the original data, so the latency caused by the task return is ignored.

[0051] Among them, the The terminal offloads the task to the satellite, reducing transmission and computation latency. The time required for terminals to offload tasks to satellites for task processing (i.e., the third delay). It can be represented as:

[0052] (7)

[0053] in, The satellite is allocated to the first The number of computing resources per terminal; This refers to the satellite-to-ground transmission rate.

[0054] Among them, the The total transmission and computation latency of each terminal, i.e., the total task processing latency, can be expressed as:

[0055] (8);

[0056] The relevant parameters and corresponding descriptions involved in the embodiments of this application can be found in Table 1.

[0057] Table 1

[0058]

[0059] In this embodiment, the goal of this application is to minimize the task processing latency of the terminal, and a user scheduling mechanism is constructed. Task uninstallation ratio Computing resource allocation Deployment location of drones To optimize the objective function of the variables, the objective function can be the function in formula (9) in Table 2.

[0060] Table 2

[0061]

[0062] in, It is the first The maximum remaining energy of the drone Indicates the first The maximum available computing resources for a drone No. The maximum available bandwidth resources for the drone. Constraints This indicates that the computational tasks on each terminal should be completed before the maximum tolerable deadline; constraints. This indicates the association constraints between the terminal and the drone; constraints It is a constraint on the task unloading ratio; constraint This indicates that the computing resources allocated to the terminal by the drone cannot exceed its maximum computing resources; constraint. This indicates that each terminal can be associated with at most one drone; constraints This indicates that to avoid collisions, the minimum distance between drones should be greater than [missing information]. ;constraint Limiting the flight altitude of drones; constraints All variables are restricted to positive integers. C1-C8 can be constraints corresponding to the objective function.

[0063] S120: Cluster the terminals to obtain cluster centers, and use the location of the cluster centers as the initial deployment location of the UAVs. Also, decompose the optimization problem of the objective function into user scheduling optimization sub-problem, task offloading ratio optimization sub-problem, and computing resource allocation optimization sub-problem.

[0064] In this embodiment, the optimization problem of UAV deployment location in the objective function is an NP problem, making it difficult to obtain a globally optimal solution. This complicates the solution and makes it difficult for general optimization algorithms to solve. However, heuristic algorithms can solve this non-convex problem without using gradient information. Since the UAV deployment location is coupled with the task offloading ratio and computational resource allocation, this embodiment first uses a weighted k-means clustering (Wk-means) method to determine the initial deployment location of the UAV before solving the optimization problem in the objective function. This facilitates finding a better solution for user scheduling, task offloading ratio, and computational resource allocation.

[0065] In this embodiment, optionally, clustering the terminals to obtain cluster centers includes: selecting a preset number of terminals as initial cluster centers; wherein the preset number is equal to the number of drones; determining the weighted distance between the terminal and each cluster center, and updating the cluster centers based on the weighted distance; if the deviation between the updated cluster centers and the unupdated cluster centers does not meet a first preset condition, randomly generating cluster centers, and returning to the step of determining the weighted distance between the terminal and each cluster center, until the deviation between the updated cluster centers and the unupdated cluster centers meets the first preset condition, thus obtaining the final cluster centers; wherein the weighted distance is determined based on the distance between the terminal and the cluster centers and a target weight; wherein the target weight is determined based on the maximum tolerable latency of the terminal.

[0066] K-means clustering can be used to cluster terminals, grouping data with similar characteristics. However, since each terminal has a different maximum tolerable latency, this embodiment uses an improved K-means clustering algorithm—Wk-means clustering—to cluster the terminals, using the cluster centers as the initial deployment locations for the UAVs. The specific Wk-means algorithm flow is shown in Table 3. Within each cluster, the sum of the weighted distances from all terminals to the cluster center is minimized; this embodiment prioritizes minimizing the maximum tolerable latency of the terminals. Determine the target weight, i.e., the target weight The weighted distance can be determined based on the following formula:

[0067] (10);

[0068] in, To obtain the weighted distance between the nth terminal and the mth cluster center, The coordinates of the m-th cluster center;

[0069] The cluster center update formula can be expressed as:

[0070] (11);

[0071] Terminals with lower maximum tolerable latency have a greater impact on the update of cluster centers; therefore, cluster centers will be closer to terminals with lower maximum tolerable latency. The assignment and update steps are repeated until the positions of the cluster centers no longer change significantly (i.e., convergence), or the preset maximum number of iterations is reached.

[0072] Table 3

[0073]

[0074] To solve the optimization problem (P1) in the objective function, auxiliary variables are introduced. At this point, (P1) can be transformed into the optimization problem (P2) shown in Table 4:

[0075] Table 4

[0076]

[0077] Therefore, (P2) remains a non-convex problem, making it difficult to obtain a globally optimal solution. Thus, the deployment location of the drones can first be fixed at a position determined by... The cluster centers obtained by the clustering algorithm are used to decompose user scheduling, task offloading ratio, and computing resource allocation optimization into three sub-problems through the block coordinate descent method. That is, the optimization problem in the objective function is decomposed into three sub-problems, and the successive convex approximation method is used to iteratively solve the three sub-problems.

[0078] S130: Fix the initial deployment position of the UAV as the fixed deployment position of the UAV, and solve the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints to obtain the optimal solution for user scheduling, the optimal solution for task offloading ratio, and the optimal solution for computing resource allocation.

[0079] In this embodiment, fixing the initial deployment location of the UAV as its fixed deployment location, and solving the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under corresponding constraints to obtain the optimal solutions for user scheduling, task offloading ratio, and computing resource allocation, includes: given the user scheduling, the task offloading ratio, and the computing resource allocation, respectively used as the current user scheduling, the current task carrying ratio, and the current computing resource allocation, and using the fixed deployment location as the current UAV deployment location; solving the user scheduling optimization sub-problem based on the current task offloading ratio, the current computing resource allocation, and the current UAV deployment location to obtain the current iteration's user scheduling optimization solution; and based on the current user scheduling, the current task carrying ratio, and the current computing resource allocation, solving the user scheduling optimization sub-problem .... The aforementioned computational resource allocation and the current UAV deployment location are used to solve the sub-problem of task unloading ratio optimization, resulting in the current iteration's optimized solution for task unloading ratio. Based on the current user scheduling, the current task unloading ratio, and the current UAV deployment location, the computational resource allocation optimization sub-problem is solved to obtain the current iteration's optimized solution for computational resource allocation. If the deviations between the current iteration's optimized solution for user scheduling, the current iteration's optimized solution for task unloading ratio, and the current iteration's optimized solution for computational resource allocation and the corresponding optimization solution from the previous iteration do not satisfy a second preset condition, the process returns to the steps given the user scheduling, the task unloading ratio, and the computational resource allocation, until the deviations between the corresponding optimization solutions from two adjacent iterations satisfy the second preset condition, thus obtaining the optimal solution for user scheduling, the optimal solution for task unloading ratio, and the optimal solution for computational resource allocation.

[0080] In this embodiment, the objective function of the user scheduling optimization sub-problem is determined based on the following method:

[0081] The objective function for the user scheduling optimization subproblem is obtained by taking the task offloading ratio, the computing resource allocation, and the UAV deployment location as given quantities in the objective function, and the user scheduling as the optimization variable in the objective function; wherein, the UAV deployment location is given as the fixed deployment location.

[0082] The constraints of the user scheduling optimization sub-problem include the following conditions:

[0083] The first delay, the second delay, and the third delay are all less than or equal to the total task processing delay;

[0084] The second delay is less than or equal to the maximum tolerable delay of the terminal;

[0085] The range of user scheduling is [0,1].

[0086] Specifically, given At this point, user scheduling is a binary variable, which needs to be relaxed into a continuous variable. The user scheduling optimization subproblem can be referred to Table 5. As shown in Table 5, the objective function of the user scheduling optimization subproblem can be formula (13), and the corresponding constraints can be referred to the constraints in Table 5. The user scheduling optimization subproblem can be converted into an optimization problem (P3).

[0087] Table 5

[0088]

[0089] At this point, (P3) is a standard convex optimization problem, which can be solved using standard convex optimization toolboxes such as CVX.

[0090] In this embodiment, the objective function of the task offloading ratio optimization sub-problem is determined based on the following method:

[0091] The objective function for the task unloading ratio optimization subproblem is obtained by taking the user scheduling, the computing resource allocation, and the UAV deployment location as given quantities in the objective optimization function, and taking the task unloading ratio as the optimization variable in the objective optimization function; wherein, the UAV deployment location is given as the fixed deployment location of the UAV.

[0092] The constraints of the task offloading ratio optimization problem include the following conditions:

[0093] The first delay, the second delay, and the third delay are all less than or equal to the total task processing delay;

[0094] The first delay, the second delay, and the third delay are all less than or equal to the terminal's maximum tolerable delay;

[0095] The sum of the task ratio calculated locally by the terminal, the task ratio offloaded to the drone, and the task ratio offloaded to the satellite is 1;

[0096] The proportion of tasks computed locally by the terminal, the proportion of tasks offloaded to the drone, the proportion of tasks offloaded to the satellite, and the number of computing resources allocated to the terminal by the drone are all greater than or equal to 0.

[0097] Specifically, given The task unloading ratio optimization subproblem can be found in Table 6 and can be converted into problem (P4). Formula (14) can be the objective function of the task unloading ratio subproblem, and the corresponding constraints can be found in Table 6. Figure 6 The constraints in the text.

[0098] Table 6

[0099]

[0100] The problem at this point (P4) is a linear programming problem, which can be solved using standard toolkits such as CVX.

[0101] In this embodiment, the objective function of the computational resource allocation optimization sub-problem is determined based on the following method:

[0102] The objective function for the task unloading ratio optimization subproblem is obtained by taking the user scheduling, the task unloading ratio, and the drone deployment location as given quantities in the objective function, and taking the task unloading ratio as the optimization variable in the objective function; wherein, the drone deployment location is given as the fixed deployment location;

[0103] The constraints of the computational resource allocation optimization subproblem include the following constraints:

[0104] The proportion of tasks computed locally by the terminal, the proportion of tasks offloaded to the drone, the proportion of tasks offloaded to the satellite, and the number of computing resources allocated to the terminal by the drone are all greater than or equal to 0.

[0105] The first delay, the second delay, and the third delay are all greater than or equal to the total task processing delay;

[0106] The second delay is less than or equal to the maximum tolerable delay of the terminal;

[0107] The number of computing resources allocated to the terminal by the drone cannot exceed the maximum number of computing resources of the drone.

[0108] Specifically, in the given In the case of , the computational resource allocation optimization subproblem can be expressed as the problem in Table 7 (P5), where the objective function of the computational resource allocation optimization subproblem can be formula (15) in Table 7, and the corresponding constraints can be found in Table 7.

[0109] Table 7

[0110]

[0111] because It is a convex function, so the constraint and It is a convex constraint. Therefore, problem (P5) is a convex optimization problem, which can be solved using toolboxes such as CVX.

[0112] In this embodiment, the specific solution process based on the above three sub-problems can be referred to the method shown in Table 8.

[0113] Table 8

[0114]

[0115] in, These are the optimal total task processing latency, the optimal solution for user scheduling, the optimal solution for task offloading ratio, and the optimal solution for computing resource allocation. , , They are the first The next iteration provides optimized solutions for user scheduling, the proportion of unloaded tasks, and the allocation of computing resources.

[0116] S140: Input the optimal solution of user scheduling, the optimal solution of task unloading ratio, and the optimal solution of computing resource allocation into the objective function, and solve for the deployment location of the UAV under the corresponding constraints to obtain the optimal solution of the UAV deployment location.

[0117] This application embodiment can use an improved particle swarm optimization algorithm to solve for the deployment location of the UAV, If the optimal solutions for user scheduling, task offloading ratio, and computational resource allocation are fixed, that is, the results obtained through problems (P3)-(P5), then the UAV deployment location optimization problem can be expressed as problem (P6) in Table 9.

[0118] Table 9

[0119]

[0120] By leveraging an improved particle swarm optimization algorithm, particles are used to explore the three-dimensional search space for the optimal UAV position. Based on the traditional Particle Swarm Optimization (PSO) algorithm, a Genetic Algorithm (GA) is introduced, using crossover and mutation probabilities from GA to improve the particle update strategy of the traditional PSO algorithm. The GA-PSO algorithm not only retains the advantages of the traditional PSO algorithm—ease of solution and fast convergence—but also incorporates the excellent global search capabilities of the GA algorithm, thus overcoming the traditional PSO algorithm's tendency to get trapped in local optima and achieving a better solution.

[0121] Because the PSO algorithm, as a population-based swarm intelligence search algorithm, directly affects its convergence efficiency and optimization performance through its particle encoding strategy, and traditional binary and integer encoding methods suffer from dimensional constraints and discretization characteristics, failing to effectively represent the feasible solution space of problem (P6). The drone (i.e., the remaining drones) deployment problem in this embodiment is its three-dimensional coordinate optimization problem. Based on this, the following encoding mechanism is designed: each particle represents a deployment scheme for a set of drones, mathematically defined as containing... A set of 3D coordinate vectors, where each vector corresponds to the 3D coordinate position of a UAV within the system. In the algorithm's... During the iteration, the first A particle can be represented as:

[0122] (17);

[0123] in, In the first In the nth iteration The position of each particle; Indicates the first In the nth iteration The three-dimensional coordinates of the drone can be specifically represented as:

[0124] (18);

[0125] in, They represent the first The first particle The deployment location of the UAV is determined by three dimensions: horizontal position, vertical position, and height. The maximum latency of the terminal is used as the fitness function of the G-PSO algorithm to evaluate the quality of the obtained solution. Substituting the UAV deployment location obtained by the GA-PSO algorithm and the task unloading ratio and computing resource allocation scheme obtained by the BCD algorithm into equation (18) yields the fitness of a particle.

[0126] In the traditional PSO algorithm framework, each particle possesses two attributes: position and velocity. Its core optimization mechanism involves particles independently exploring the optimal solution space, with their motion patterns dynamically adjusting position and velocity guided by both individual and swarm optimal solutions. The update rules for these two particle attributes can be expressed as follows:

[0127] (19);

[0128] (20);

[0129] in, and They represent the first time. In the nth iteration The latest and current velocity of each particle; and They represent the first The particle in the first The second iteration and the... The position of +1 iteration; and They represent the process. After the nth iteration The historical best position of an individual particle and the historical optimal position of a group of examples; It is the inertia factor, which can control the intensity of the influence of historical velocity on the current state; and These represent the individual learning coefficient and the population learning coefficient, respectively, quantifying the degree to which a particle responds to the historical optimal solution of an individual and a group. and exist Random distribution within the interval is used to maintain the stochastic exploration characteristic of the search process.

[0130] The GA-PSO algorithm proposed in this application improves the update process of the original particle position and velocity by using crossover and mutation probabilities in GA, based on the traditional PSO algorithm. Therefore, the particle iterative update method in the GA-PSO algorithm can be expressed as:

[0131] (twenty one)

[0132] in, It is the mutation probability in the GA-PSO algorithm. and It is the crossover probability in the GA-PSO algorithm.

[0133] In the GA-PSO algorithm optimization provided in this application embodiment, the dynamic enhancement of exploration capability is achieved by introducing the mutation mechanism of GA into the inertia update stage of the PSO algorithm. The mathematical expression of the improved inertia term update rule is as follows:

[0134] (twenty two);

[0135] in, After the particle is subjected to inertia and random mutation, the first The position of +1 iteration; This represents the coordinates of a randomly selected particle, i.e., any drone. The improved inertial term can randomly select a dimension index from the particle position vector and resample the coordinates of that dimension within a preset threshold range. In the GA-PSO algorithm optimization provided in this application embodiment, combining the crossover mechanism of GA and the individual and group learning stages of the PSO algorithm can result in a new update method:

[0136] (twenty three);

[0137] (twenty four);

[0138] in, Indicates in The updated position is obtained by intersecting the particle's historical best position. Indicates in Based on this, the updated position is obtained by intersecting with the global optimal position of the particle; and Same in The crossover probability is randomly distributed within the interval. Two quantiles are randomly selected on the particle, and the values ​​between the quantiles are used... and replace.

[0139] In the traditional PSO algorithm, the inertia factor As a key control parameter, its value directly determines the convergence speed and search capability of the algorithm. In the GA-PSO algorithm used in this application embodiment, when the inertia factor... When the value is large, the probability of particle mutation increases significantly, allowing the algorithm to explore a wider problem space, significantly expanding the coverage of the search space, and possessing strong global search capabilities, which helps to discover potential optimal solutions to the problem. And when... When the value is small, the probability of particle mutation in the GA-PSO algorithm is significantly reduced. At this time, by suppressing the intensity of random perturbations, the algorithm focuses on the details of the current search region, thus possessing good local search capabilities. This is beneficial for fine-grained searches near already discovered possible solutions, further optimizing the quality of the solution. Different stages of the algorithm execution have different search requirements. In the early stages of algorithm execution, to fully understand the distribution of the problem space, it is necessary to emphasize the diversity of the search and explore as many regions as possible. As the search continues to deepen, the approximate range of the optimal solution gradually becomes clear. At this point, more emphasis should be placed on local search capabilities to accurately find the optimal solution. Given that the traditional PSO algorithm uses a fixed inertia weight factor, which cannot well adapt to the search requirements of different stages of the algorithm, this invention adopts a linear inertia weight factor adjustment strategy. Under this strategy, the inertia factor... It decreases linearly with the number of iterations. This adjustment method allows the algorithm to utilize a larger amount of data in the early stages. Perform a global search on the value, and later utilize the smaller value. The algorithm employs a local search to make its search process more scientific and rational. The specific adjustment strategy is as follows:

[0140] (25);

[0141] in, and They represent the inertia factors, respectively. The upper and lower bounds, This is the maximum number of iterations allowed. Besides the inertia factor, for the individual learning factor in formula (21)... Social learning factors The embodiments of this application also employ a similar linear adjustment strategy, aiming to further optimize the search performance of the algorithm and achieve a more effective balance between global search and local search, specifically as follows:

[0142] (26);

[0143] (27);

[0144] in, and Representing parameters respectively The initial value before the iteration begins and the final value during the iteration process. and Representing parameters respectively The initial value before the iteration begins and the final value during the iteration process.

[0145] The UAV deployment location optimization method based on the GA-PSO algorithm is shown in Table 10. First, the population is initialized, including the population size. Maximum number of iterations and the particles Initialized to pass The clustering algorithm obtains the deployment locations of drones, and also identifies the local optima for each population. and the global optimal solution Initialization is performed. Under the condition that the algorithm execution conditions are met, the particle population is updated by performing crossover and mutation operations on the particles. The BCD algorithm is called to obtain the task unloading ratio, computational resource allocation, and user scheduling scheme. The fitness of each particle is calculated according to formula (9). Based on the calculated fitness value, the local optimum is determined. and the global optimal solution Update.

[0146] Table 10

[0147]

[0148] S150: Using the optimal solution for the UAV deployment location as the fixed deployment location, return the steps of solving the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints, until the deviation between the optimal solutions for the UAV deployment location in two adjacent iterations satisfies the iteration termination condition, and obtain the optimal solutions for the UAV deployment location, user scheduling, task offloading ratio, and computing resource allocation, and use them as the resource management scheme for the integrated air-space-ground network.

[0149] In this embodiment, the iteration termination condition can be that the deviation between two optimized solutions is less than a preset deviation. By continuously iterating until the iteration termination condition is met, the optimal solution for the drone deployment location, the optimal solution for user scheduling, the optimal solution for task offloading ratio, and the optimal solution for computing resource allocation corresponding to the optimal iteration are taken as the final optimal solutions for drone deployment location, user scheduling, task offloading ratio, and computing resource allocation.

[0150] In this embodiment, the algorithm for solving the optimization problem in the objective function can be called the GP-B algorithm. The optimization problem of the objective function is optimization problem (P1), as shown in Table 11. This embodiment uses the GP-B algorithm to alternately optimize four optimization variables. When the difference between two adjacent iterations is less than the set precision... In this way, the optimal solutions for user scheduling, task offloading ratio, computing resource allocation, and drone deployment location can be obtained, ultimately minimizing the maximum task processing latency.

[0151] Table 11

[0152]

[0153] in, These are the optimal solutions for final user scheduling, task offloading ratio, computing resource allocation, and drone deployment location.

[0154] The technical solution provided in this application aims to minimize the task processing latency of the terminal. It constructs an objective function with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables. By clustering the terminals and using the location of the cluster center as the initial deployment location of the UAV, the optimization problem of the objective function is transformed into three sub-problems for solving. Then, the deployment location of the UAV is optimized and solved. Through continuous iteration, the optimal solutions for the deployment location of the UAV, user scheduling, task offloading ratio, and computing resource allocation are obtained. As a resource management solution for integrated air-space-ground management, it can effectively reduce task processing latency and ensure the continuity of task processing.

[0155] To verify the technical solutions provided in the embodiments of this application, the optimization algorithm provided in the embodiments of this application is simulated using Matlab to verify the effectiveness of the algorithm. The embodiments of this application provide parameter configurations in the simulation, analyze the impact of different situations on the maximum processing latency of the task, and compare them with other solutions.

[0156] The simulation scene is set to (800×800)m. 2 The simulation parameters are configured as shown in Table 12, with M=3 drones and 1 low-orbit satellite providing communication services to terminals in the power transmission line within the region.

[0157] Table 12

[0158]

[0159] This application introduces two comparison algorithms to compare with the optimization algorithm provided in this application. Since the problem in this application is a mixed nonlinear optimization problem, the optimal solution cannot be obtained by conventional methods. Therefore, two benchmark algorithms commonly used in related research are used.

[0160] 1) Comparison with Algorithm 1: The UAV remains stationary, and the tasks of terminals within the original coverage area of ​​the low-energy UAV are calculated locally or by low-Earth orbit satellites. Specifically: the UAV's position is fixed, and then the optimal solutions for user scheduling, task offloading ratio, and computing resource allocation are obtained according to the optimization methods for the three sub-problems provided in the embodiments of this application, which are then used as the final optimized solutions.

[0161] 2) Comparison Algorithm 2: The location of the UAV is deployed according to the GA-PSO algorithm provided in the embodiment of this application. A random algorithm is used to associate the terminal with the UAV and the task is randomly offloaded to the UAV and satellite for calculation.

[0162] Figure 3This demonstrates the impact of satellite computing power on the maximum processing latency of a mission. Figure 4 The diagram illustrates the impact of the terminal's local computing power on the maximum processing latency of a task. Figure 5 The impact of the UAV's maximum computing power on the maximum processing latency of a task is demonstrated. The results show that the maximum processing latency decreases with increasing computing power. Due to the high uncertainty of the random algorithm in Comparison Algorithm 2, 300 simulations were performed on this algorithm, and the average value was taken as the result. It can be seen that the optimized algorithm proposed in this application is significantly superior to the two comparative algorithms.

[0163] like Figure 4 As shown, with the increase of local computing power, the maximum processing latency of the task compared to Algorithm 1 is reduced more significantly than that of the proposed algorithm. This is because when the local computing power is small, more tasks are offloaded to satellite computing. Although the satellite computing power is powerful, there is a large latency compared to the computing resources offloaded to the drone. Therefore, when there are no drones available to provide services, the transmission latency in this algorithm is greater. However, when the local computing power is large, the task can be computed locally, and both the computation latency and the transmission latency are reduced.

[0164] like Figure 5 As shown, when the maximum computing power of the UAV increases, the maximum processing latency of the algorithm provided in this embodiment shows a more significant downward trend compared to the comparative algorithm 1. This is because the increase in UAV computing power in comparative algorithm 1 has no impact on the original coverage area of ​​the low-energy UAV, and only reduces the processing latency in the area originally covered by the remaining UAVs. The GP-B algorithm proposed in this embodiment is significantly better than comparative algorithm 2. This is because an effective UAV deployment method allows the terminal to offload tasks to the UAV for execution, making full use of the UAV's resources. In contrast, in the offloading scheme using a random algorithm, the resources of the UAV cannot be effectively utilized, and some terminals cannot connect to the UAV, thereby reducing the utilization rate of computing and communication resources.

[0165] Besides the computing power of satellites, local systems, and drones affecting latency, the number of terminals deployed along power transmission lines also significantly impacts task processing latency. This application introduces the Differential Evolution (DE) algorithm as the DE-B algorithm for drone location deployment, while the task offloading and resource allocation methods remain as per the algorithm in this application. Figure 6As can be seen, the GP-B algorithm proposed in this application generally outperforms the DE-B algorithm. When the number of terminals is small, the GP-B algorithm is significantly better than the DE-B algorithm. However, when the number of terminals is large, the DE-B algorithm shows better performance than the GP-B algorithm. This is because the DE algorithm is superior to the GA-PSO algorithm in terms of global search capability and is more likely to escape local optima, thus having an advantage when the number of terminals is large. In contrast, the GA-PSO algorithm is prone to reaching local optima and has a fast convergence speed, making it suitable for situations with a small number of terminals.

[0166] like Figure 7 As shown, as the amount of task data increases, more tasks are offloaded to drones and satellites. This is because when the amount of task data increases, local computing cannot meet the maximum tolerable latency constraint of the terminal. Therefore, the terminal will offload the tasks to drones and satellites with more powerful computing capabilities for computation.

[0167] Therefore, addressing the lack of backup drones in communication networks based on integrated air-space-ground power transmission lines, the technical solution provided in this application can jointly optimize user scheduling, task offloading ratio, computing resource allocation, and drone deployment location to minimize the maximum task processing latency. Since drone deployment location and resource allocation are closely coupled, the optimization problem is difficult to solve. This application proposes a hierarchical GP-B algorithm to address this issue. This algorithm uses the GA-PSO algorithm at the upper layer for drone deployment and the BCD algorithm at the lower layer for optimizing task offloading ratio and resource allocation. Simulation results show that the technical solution provided in this application can effectively reduce the task processing latency of power transmission line terminals.

[0168] Figure 8 A structural block diagram of a resource management device for an integrated air-space-ground network for power transmission lines, as provided in this application embodiment, is shown below. Figure 8 As shown, the device includes:

[0169] The construction module 810 is used to construct an objective function with user scheduling, task offloading ratio, computing resource allocation and UAV deployment location as optimization variables, with the goal of minimizing the task processing latency of the terminal; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network.

[0170] The decomposition module 820 is used to cluster the terminals to obtain cluster centers, and use the location of the cluster centers as the initial deployment location of the UAV, and decompose the optimization problem of the objective function into user scheduling optimization sub-problem, task offloading ratio optimization sub-problem and computing resource allocation optimization sub-problem;

[0171] The first solution module 830 is used to fix the initial deployment position of the UAV as the fixed deployment position of the UAV, and solve the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under corresponding constraints to obtain the optimal solution for user scheduling, the optimal solution for task offloading ratio, and the optimal solution for computing resource allocation.

[0172] The second solution module 840 is used to input the optimal solution of user scheduling, the optimal solution of task unloading ratio and the optimal solution of computing resource allocation into the objective function, and solve for the deployment location of the UAV under the corresponding constraints to obtain the optimal solution of the UAV deployment location.

[0173] The iterative solution module 850 is used to take the optimal solution of the UAV deployment location as the fixed deployment location, and return the steps of solving the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints, until the deviation between the optimal solutions of the UAV deployment location in two adjacent iterations meets the iteration termination condition, and obtain the optimal solutions of the UAV deployment location, user scheduling, task offloading ratio, and computing resource allocation, and use them as the resource management scheme of the integrated air-space-ground network.

[0174] like Figure 9 As shown in the figure, this application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0175] Memory 113 is used to store computer programs;

[0176] In one embodiment of this application, when the processor 111 executes a program stored in the memory 113, it implements the method provided in any of the foregoing method embodiments, including:

[0177] With the goal of minimizing the task processing latency of the terminal, an objective function is constructed with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network;

[0178] The terminals are clustered to obtain cluster centers, and the location of the cluster centers is used as the initial deployment location of the UAV. The optimization problem of the objective function is decomposed into user scheduling optimization sub-problem, task offloading ratio optimization sub-problem, and computing resource allocation optimization sub-problem.

[0179] The initial deployment location of the UAV is fixed as the fixed deployment location of the UAV. The user scheduling optimization subproblem, the task offloading ratio optimization subproblem, and the computing resource allocation optimization subproblem are solved under the corresponding constraints to obtain the optimal solutions for user scheduling, task offloading ratio, and computing resource allocation.

[0180] The optimal solution for user scheduling, the optimal solution for task unloading ratio, and the optimal solution for computing resource allocation are input into the objective function, and the optimal solution for the UAV deployment location is obtained under the corresponding constraints.

[0181] The optimal solution for the UAV deployment location is taken as the fixed deployment location. The steps of solving the user scheduling optimization subproblem, the task offloading ratio optimization subproblem, and the computing resource allocation optimization subproblem under the corresponding constraints are returned until the deviation between the optimal solutions of the UAV deployment location in two adjacent iterations meets the iteration termination condition. The optimal solutions for the UAV deployment location, user scheduling, task offloading ratio, and computing resource allocation are obtained and used as the resource management scheme of the integrated air-space-ground network.

[0182] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method provided in any of the foregoing method embodiments.

[0183] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0184] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0185] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this application.

Claims

1. A resource management method for an integrated air-space-ground network for power transmission lines, characterized in that, include: With the goal of minimizing the task processing latency of the terminal, an objective function is constructed with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network; The terminals are clustered to obtain cluster centers, and the location of the cluster centers is used as the initial deployment location of the UAV. The optimization problem of the objective function is decomposed into user scheduling optimization sub-problem, task offloading ratio optimization sub-problem, and computing resource allocation optimization sub-problem. The initial deployment location of the UAV is fixed as the fixed deployment location of the UAV. The user scheduling optimization subproblem, the task offloading ratio optimization subproblem, and the computing resource allocation optimization subproblem are solved under the corresponding constraints to obtain the optimal solutions for user scheduling, task offloading ratio, and computing resource allocation. The optimal solution for user scheduling, the optimal solution for task unloading ratio, and the optimal solution for computing resource allocation are input into the objective function, and the optimal solution for the UAV deployment location is obtained under the corresponding constraints. The optimal solution for the UAV deployment location is taken as the fixed deployment location. The steps of solving the user scheduling optimization subproblem, the task offloading ratio optimization subproblem, and the computing resource allocation optimization subproblem under the corresponding constraints are returned until the deviation between the optimal solutions of the UAV deployment locations in two adjacent iterations meets the iteration termination condition. The final optimal solutions for the UAV deployment location, the final optimal solutions for user scheduling, the final optimal solutions for task offloading ratio, and the final optimal solutions for computing resource allocation are obtained and used as the resource management scheme of the integrated air-space-ground network.

2. The method according to claim 1, characterized in that, The user scheduling refers to the association between the drone and the terminal; the task offloading ratio includes the proportion of tasks computed locally by the terminal, the proportion of tasks offloaded to the drone, and the proportion of tasks offloaded to the satellite; the computing resource allocation includes the number of computing resources allocated by the drone to the terminal and the number of computing resources allocated by the satellite to the drone. The objective function, which aims to minimize the terminal's task processing latency, is constructed with user scheduling, task offloading ratio, computing resource allocation, and UAV deployment location as optimization variables. This includes: The task processing latency of the terminal computed locally is determined as the first latency; wherein, the first latency is optimized by the proportion of tasks computed locally by the terminal. The task processing delay at which the terminal offloads tasks to the drone is determined as the second delay; wherein the second delay is optimized by the proportion of tasks offloaded from the terminal to the drone, the user scheduling, and the number of computing resources allocated by the drone to the terminal. The task processing delay at which the terminal offloads tasks to the satellite is determined as the third delay; wherein, the third delay is optimized by the proportion of tasks offloaded by the terminal to the satellite and the amount of computing resources allocated by the satellite to the UAV. The maximum delay among the first delay, the second delay, and the third delay is taken as the total task processing delay of the terminal. The total task processing delay is minimized to obtain the objective function.

3. The method according to claim 1, characterized in that, The step of clustering the terminals to obtain cluster centers includes: A preset number of terminals are selected as initial cluster centers; wherein, the preset number is equal to the number of drones. Determine the weighted distance between the terminal and each cluster center, and update the cluster centers based on the weighted distance; If the deviation between the updated cluster centers and the original cluster centers does not meet the first preset condition, cluster centers are randomly generated, and the step of determining the weighted distance between the terminal and each cluster center is returned until the deviation between the updated cluster centers and the original cluster centers meets the first preset condition, and the final cluster centers are obtained. The weighted distance is determined based on the distance between the terminal and the cluster center and the target weight; wherein the target weight is determined based on the maximum tolerable latency of the terminal.

4. The method according to claim 1, characterized in that, The initial deployment location of the UAV is fixed as its fixed deployment location. Then, under corresponding constraints, the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem are solved to obtain the optimal solutions for user scheduling, task offloading ratio, and computing resource allocation. This includes: Given the user scheduling, the task offloading ratio, and the computing resource allocation, respectively, these are taken as the current user scheduling, the current task carrying ratio, and the current computing resource allocation, and the fixed deployment location is taken as the current UAV deployment location; Based on the current task unloading ratio, the current computing resource allocation, and the current UAV deployment location, the user scheduling optimization sub-problem is solved to obtain the user scheduling optimization solution for the current iteration; Based on the current user scheduling, the current computing resource allocation, and the current UAV deployment location, the task unloading ratio optimization subproblem is solved to obtain the task unloading ratio optimization solution for the current iteration; The computational resource allocation optimization subproblem is solved based on the current user scheduling, the current task unloading ratio, and the current UAV deployment location to obtain the computational resource allocation optimization solution for the current iteration. If the deviations between the current iteration's optimized user scheduling solution, the current iteration's optimized task unloading ratio solution, and the current iteration's optimized computational resource allocation solution and the optimization solution corresponding to the previous iteration do not meet the second preset condition, return to the steps of giving the user scheduling, the task unloading ratio, and the computational resource allocation, until the deviations between the corresponding optimization solutions of two adjacent iterations meet the second preset condition, and obtain the optimal solution for user scheduling, the optimal solution for task unloading ratio, and the optimal solution for computational resource allocation.

5. The method according to claim 2, characterized in that, The objective function of the user scheduling optimization sub-problem is determined based on the following method: The objective function for the user scheduling optimization subproblem is obtained by taking the task offloading ratio, the computing resource allocation, and the UAV deployment location as given quantities in the objective function, and the user scheduling as the optimization variable in the objective function; wherein, the UAV deployment location is given as the fixed deployment location. The constraints of the user scheduling optimization sub-problem include the following conditions: The first delay, the second delay, and the third delay are all less than or equal to the total task processing delay; The second delay is less than or equal to the maximum tolerable delay of the terminal; The range of user scheduling is [0,1].

6. The method according to claim 2, characterized in that, The objective function for the task unloading ratio optimization sub-problem is determined based on the following method: The objective function for the task offloading ratio optimization subproblem is obtained by taking the user scheduling, the computing resource allocation, and the UAV deployment location as given quantities in the objective function, and taking the task offloading ratio as the optimization variable in the objective function; wherein, the UAV deployment location is given as the fixed deployment location; The constraints of the task offloading ratio optimization problem include the following conditions: The first delay, the second delay, and the third delay are all less than or equal to the total task processing delay; The first delay, the second delay, and the third delay are all less than or equal to the maximum tolerable delay of the terminal; The sum of the task ratio calculated locally by the terminal, the task ratio offloaded to the drone, and the task ratio offloaded to the satellite is 1; The proportion of tasks computed locally by the terminal, the proportion of tasks offloaded to the drone, the proportion of tasks offloaded to the satellite, and the number of computing resources allocated to the terminal by the drone are all greater than or equal to 0.

7. The method according to claim 2, characterized in that, The objective function of the computational resource allocation optimization subproblem is determined based on the following method: The objective function for the task unloading ratio optimization subproblem is obtained by taking the user scheduling, the task unloading ratio, and the drone deployment location as given quantities in the objective function, and taking the task unloading ratio as the optimization variable in the objective function; wherein, the drone deployment location is given as the fixed deployment location; The constraints of the computational resource allocation optimization subproblem include the following constraints: The proportion of tasks computed locally by the terminal, the proportion of tasks offloaded to the drone, the proportion of tasks offloaded to the satellite, and the number of computing resources allocated to the terminal by the drone are all greater than or equal to 0. The first delay, the second delay, and the third delay are all greater than or equal to the total task processing delay; The second delay is less than or equal to the maximum tolerable delay of the terminal; The number of computing resources allocated to the terminal by the drone cannot exceed the maximum number of computing resources of the drone.

8. A resource management device for an integrated air-space-ground network for power transmission lines, characterized in that, include: The module is used to construct an objective function with user scheduling, task offloading ratio, computing resource allocation and UAV deployment location as optimization variables, with the goal of minimizing the task processing latency of the terminal; wherein, the terminal is a status monitoring terminal on the transmission line in the integrated air-space-ground network; The decomposition module is used to cluster the terminals to obtain cluster centers, and use the location of the cluster centers as the initial deployment location of the UAVs. It also decomposes the optimization problem of the objective function into user scheduling optimization sub-problems, task offloading ratio optimization sub-problems, and computing resource allocation optimization sub-problems. The first solution module is used to fix the initial deployment position of the UAV as the fixed deployment position of the UAV, and solve the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints, respectively, to obtain the optimal solution for user scheduling, the optimal solution for task offloading ratio, and the optimal solution for computing resource allocation. The second solution module is used to input the optimal solution of user scheduling, the optimal solution of task unloading ratio and the optimal solution of computing resource allocation into the objective function, and solve for the deployment location of the UAV under the corresponding constraints to obtain the optimal solution of the UAV deployment location. The iterative solution module is used to take the optimized solution of the UAV deployment location as the fixed deployment location, and return the steps of solving the user scheduling optimization sub-problem, the task offloading ratio optimization sub-problem, and the computing resource allocation optimization sub-problem under the corresponding constraints, until the deviation between the optimal solutions of the UAV deployment location in two adjacent iterations meets the iteration termination condition, and obtain the final optimal solution of the UAV deployment location, the final optimal solution of the user scheduling, the final optimal solution of the task offloading ratio, and the final optimal solution of the computing resource allocation, which are used as the resource management scheme of the integrated air-space-ground network.

9. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed in a computer, causes the computer to perform the method described in any one of claims 1-7.

Images