Resource allocation method for intensive tasks

Abstract

The invention relates to a resource allocation method for intensive tasks, and belongs to the technical field of wireless communication. Intensive tasks of ground terminal equipment are collected to a high-altitude platform to be processed or transmitted to a low-orbit satellite, and a system total energy consumption weighting and mathematical model is constructed; the method comprises the following steps: determining optimal path selection by constructing a bidirectional preference relationship between ground terminal equipment and a high-altitude platform and association between the high-altitude platform and a low-orbit satellite, and decoupling a system total energy consumption weighting and mathematical model into three sub-problems of a high-altitude platform selection association problem, task unloading and resource allocation for solving; joint iterative optimization is carried out on results of high-altitude platform selection association, task unloading decision and resource allocation, and a system total energy consumption weighted sum obtained by each iteration is used as a fitness evaluation standard until convergence is carried out to an optimal system total energy consumption weighted sum; the invention provides a hybrid unloading strategy with high energy efficiency, which not only can efficiently and flexibly allocate resources, but also can effectively improve the energy efficiency of the system.

Description

Technical Field

[0001] The invention belongs to the field of communication technology and relates to a resource allocation method for an air-ground-space integrated network oriented to intensive tasks. Background Art

[0002] With the rapid development of global informatization, the fifth generation communication system (5G) has gradually matured and has been widely used in various fields. At the same time, the research on the sixth generation communication system (6G) is also being actively promoted. Driven by the development of the 6G communication system, the communication and computing needs between terminal devices have increased dramatically, prompting the significant development of Mobile Edge Computing (MEC) technology. Compared with the single MEC server deployment method in the ground network, the deployment of MEC servers in the Space-Air-Ground Integrated Network (SAGIN) is more diversified. According to different network requirements, MEC servers can be deployed on low Earth Orbit (LEO) satellites or high altitude platforms (HAP) to provide more flexible and efficient computing services. Although SAGIN has the advantages of wide coverage, high throughput and strong elasticity, its highly dynamic and complex heterogeneous environment also brings significant technical challenges. For example, the frequent switching between nodes and users, the long-distance transmission of satellite communications, the dynamic connection between HAP and LEO, and the optimization of computing resources and energy consumption in multi-layer network structures are all key problems that need to be studied in depth. In addition, ground wireless networks are often limited by energy and computing resources, and it is difficult to meet the growing computing needs of IoT devices. Therefore, in order to fully utilize the advantages of SAGIN and cope with multiple challenges in intensive mission scenarios. The resource allocation method for space-ground integrated intensive tasks proposed in this paper can dynamically adjust the association relationship and reasonably allocate resources to minimize system energy consumption and ensure service quality, thereby effectively solving the resource allocation challenges of SAGIN in complex environments. This has important research significance for building an efficient, stable, and low-energy space-ground integrated network. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a resource allocation method for intensive tasks in an integrated air-ground-space network, aiming to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A resource allocation method for intensive tasks, comprising the following steps:

[0006] S1. The intensive tasks of the ground terminal equipment are collected by the High Altitude Platform (HAP) or forwarded by the HAP to the Low Earth Orbit (LEO) for processing, and a weighted and mathematical model of the total energy consumption of the system is constructed;

[0007] S2. Determine the optimal path selection by constructing the bidirectional preference relationship between the ground terminal equipment and HAP and the association between HAP and LEO; decouple the total energy consumption weighted and mathematical model of the system into three sub-problems: HAP selection association, task offloading decision and resource allocation, and solve them separately;

[0008] S3. The solution results of the three sub-problems of HAP selection association, task offloading decision and resource allocation are jointly iteratively optimized, and the weighted sum of the total energy consumption of the system obtained in each iteration is used as the fitness evaluation criterion until it converges to the optimal weighted sum of the total energy consumption of the system.

[0009] Furthermore, the HAP acts as a base station and a relay node. Under the maximum delay constraint, the weighted sum mathematical model of the total energy consumption of the construction system is expressed as:

[0010]

[0011] Among them, E U,H and E H,L They represent the transmission energy consumption from the terminal device to HAP and from HAP to LEO, respectively. α1 and α2 are weight coefficients, which respectively measure the different contributions of the transmission energy consumption from the terminal device to HAP and from HAP to LEO to the optimization objective. E H represents the energy consumption of HAP calculation; E L represents LEO computing energy consumption; β and γ are constant positive weights of HAP and LEO transmission energy consumption, respectively; H represents H HAPs; h represents HAP index, from 1 to H; t represents time slot index, from 1 to T; o represents HAP offloading decision constraint: P represents the maximum power constraint of terminal equipment and HAP; Indicates whether the task is processed locally in HAP; Indicates whether the task is offloaded to LEO for processing; P represents the maximum power constraint set of the terminal device and HAP; Indicates that the power of path X should not exceed the maximum power constraint; U indicates that there are U terminal devices; u indicates the terminal device index, from 1 to U; S indicates that there are S LEOs; s indicates the LEO index, from 1 to S; Represented as the bit size of the current computing task; represents the computing resources of the current path X; F represents the maximum computing resource constraint of HAP and LEO; d u Indicates the actual completion delay of the task; Indicates the maximum tolerable delay of the task; C1 indicates that the terminal device can only choose one task offloading mode in a time slot; C2 indicates that the power of the terminal device and HAP does not exceed the maximum power limit; C3 indicates that the resources used by HAP and LEO do not exceed the maximum resource constraint; C4 indicates the maximum tolerable delay constraint for task processing;

[0012] The association constraint is expressed as:

[0013]

[0014] in, Indicates whether the terminal device u is associated with the HAP index at time slot t; Indicates whether the HAP index is associated with the LEO index at time slot t;

[0015] The maximum power constraint P of the terminal device and HAP is expressed as:

[0016]

[0017] in, Indicates the maximum power of the terminal equipment; Indicates the maximum power of the HAP;

[0018] o is represented by:

[0019]

[0020] in, when The task will be processed directly on the HAP. Then the task will be relayed to LEO via HAP for indirect processing;

[0021] F is represented by:

[0022]

[0023] in, Indicates the maximum available computing resources of HAP; represents the maximum available computing resources of LEO;

[0024] The path set X from the terminal device to the HAP and from the HAP to the LEO is expressed as:

[0025] X={(i,j)|i∈U∪H,j∈H∪S}.

[0026] Further, the It is expressed as:

[0027]

[0028] Said It is expressed as:

[0029]

[0030] Among them, u t represents the terminal device u that needs to process the task in time slot t; h t represents the HAP index participating in the task processing in time slot t; s t represents the LEO index participating in the processing task in time slot t;

[0031] Each terminal device can only be associated with one HAP, expressed as:

[0032]

[0033] Each HAP can only be associated with one satellite, expressed as:

[0034]

[0035] Each HAP has a limited load, which satisfies:

[0036]

[0037] Among them, N H is the maximum number of associations of HAP.

[0038] Further, the S2 is specifically:

[0039] Preference list P of terminal devices on HAP u,h , expressed as:

[0040]

[0041] Among them, φ u Expressed as task priority weight; It is expressed as the transmission rate from the terminal device to the HAP; It represents the available computing resources of HAP; t represents the time slot index, from 1 to T; U represents the number of terminal devices; u represents the terminal device index, from 1 to U; u t U represents the terminal device u that needs to process tasks in time slot t; t represents the set of all terminal devices that need to process tasks in time slot t;

[0042] HAP preference list P of the terminal device h,u , expressed as:

[0043]

[0044] in, Indicates the bit size of the current computing task; ht H represents the HAP index participating in the processing task in time slot t; t represents the set of all HAPs participating in processing tasks in time slot t;

[0045] The comprehensive preference objective is defined as:

[0046]

[0047] Where M is the matching set, λ and μ are weight coefficients; using the blocking pair theory, by determining whether the matching relationship between the terminal device and the HAP meets the blocking pair condition, M is gradually adjusted; in each iteration, if a blocking pair is found, resources are reallocated until all blocking pairs are eliminated, so that M reaches a stable state and meets the constraint conditions;

[0048] The matching problem between HAP and LEO is modeled as a many-to-many matching problem, which is solved by using the GUROBI optimizer to minimize the transmission energy consumption between HAP and LEO;

[0049] Based on the transmission and computing characteristics of tasks, a task offloading decision model is established, and the optimal offloading path is solved by particle swarm optimization algorithm;

[0050] Based on HAP selection association and task offloading decision, the problem of optimizing resource allocation is expressed as the problem of optimizing computing resources and power allocation. Under the condition of ensuring that the allocated computing resources do not exceed the maximum computing capacity, the whale optimization algorithm is adopted to deal with the constraints by introducing the penalty function, conduct local and global searches, explore the global optimum, and gradually approach the optimal solution through fitness evaluation. The penalty function Λ(p) is expressed as:

[0051]

[0052] Among them, δ1(t) and η1(t) represent the adaptive penalty factors of key constraints, which change dynamically with time slots; δ2(t) and η2(t) are used for the adaptive penalty factors of secondary constraints; Represents the indicator function, which is used to judge the constraints whether it has been violated; Represents the indicator function, which is used to judge the constraints Whether it is violated; r and z are parameters for adjusting the violation degree index, which are used to increase the penalty for solutions with greater violations; Indicates the degree of violation The logarithmic penalty is used for a smaller range of violations; Indicates the degree of violation The exponential penalty is used for rapidly growing violation levels; Q(p) represents the objective function, Represents the power weight factor.

[0053] Furthermore, the task offloading decision model is specifically:

[0054]

[0055]

[0056]

[0057] Further, if Then the indicator function G h =1; otherwise G h =0.

[0058] Further, if Then H s =1; otherwise H s =0.

[0059] Furthermore, the objective function Q(p) is specifically:

[0060]

[0061] Furthermore, S3 is specifically as follows: the weighted sum of the total system energy consumption obtained by the iteration is used as a fitness evaluation criterion, and in each iteration, the optimal solution is gradually approached by optimizing the HAP association selection, task offloading path and resource allocation configuration; the iterative process updates the fitness evaluation criterion after each round of iteration until the weighted sum of the total system energy consumption reaches the convergence condition to obtain the optimal weighted sum of the total system energy consumption.

[0062] The beneficial effects of the present invention are:

[0063] In view of the problem of task offloading of intensive tasks, in view of the sharp increase in the number of terminal devices and computing, the complexity of communication infrastructure in some areas, etc.; considering the dynamic association relationship of HAP, and its large-scale access to act as a relay and air base station to provide edge computing services for intensive tasks, and the edge server on the satellite can also provide edge computing services; a high-energy-efficiency hybrid offloading strategy is proposed, which realizes the efficiency and flexibility of resource allocation by jointly optimizing the selection association of HAP, the task offloading path and reasonably allocating computing resources. The present invention effectively improves the energy efficiency of the system.

[0064] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0066] Figure 1 A network architecture diagram of a specific embodiment of the present invention;

[0067] Figure 2 The present invention is a flowchart of a method for allocating resources for intensive tasks according to a specific embodiment of the present invention. DETAILED DESCRIPTION

[0068] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in the specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the features in the following embodiments and embodiments can be combined with each other without conflict.

[0069] Among them, the figures are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the figures may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the figures may be omitted.

[0070] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0071] See also Figure 1 , which is a network architecture diagram of a specific embodiment of the present invention; please refer to Figure 2 , which is a flow chart of a resource allocation method for intensive tasks according to a specific embodiment of the present invention.

[0072] The resource allocation method for intensive tasks in this embodiment specifically includes the following steps:

[0073] S1: Intensive tasks of ground terminal equipment are collected and processed by the high altitude platform (HAP) or forwarded by the HAP to the low orbit satellite (LEO) for processing, and a weighted and mathematical model of the total energy consumption of the system is constructed;

[0074] S1.1: Intensive tasks of terminal equipment are performed through the link are collected to the HAP, which acts as a base station and relay node, processing them locally or through a link The task is passed to LEO; the task processing delay of terminal device u can be expressed as T u , where the delay for different offloading modes is mainly composed of task transmission delay and computing delay; the energy consumption of task processing is composed of the transmission energy consumption of the terminal device and the transmission energy consumption and computing energy consumption of different offloading modes, expressed as W E ;

[0075] S1.2: Under the maximum delay constraint, a weighted sum mathematical model of the total energy consumption of the system is constructed, which is expressed as:

[0076]

[0077] Among them, E U,H and E H,L They represent the transmission energy consumption from the terminal device to HAP and from HAP to LEO, respectively. α1 and α2 are weight coefficients, which respectively measure the different contributions of the transmission energy consumption from the terminal device to HAP and from HAP to LEO to the optimization objective. E H represents the energy consumption of HAP calculation; E L represents LEO computing energy consumption; β and γ are constant positive weights of HAP and LEO transmission energy consumption, respectively; H represents H HAPs; h represents HAP index, from 1 to H; t represents time slot index, from 1 to T; o represents HAP offloading decision constraint: P represents the maximum power constraint of terminal equipment and HAP; Indicates whether the task is processed locally in HAP; Indicates whether the task is offloaded to LEO for processing; P represents the maximum power constraint set of the terminal device and HAP; Indicates that the power of path X should not exceed the maximum power constraint; U indicates that there are U terminal devices; u indicates the terminal device index, from 1 to U; S indicates that there are S LEOs; s indicates the LEO index, from 1 to S; Represented as the bit size of the current computing task; represents the computing resources of the current path X; F represents the maximum computing resource constraint of HAP and LEO; d u Indicates the actual completion delay of the task; Indicates the maximum tolerable delay of the task; C1 means that the terminal device can only choose one task offloading method in a time slot; C2 means that the power of the terminal device and HAP does not exceed the maximum power limit; C3 means that the resources used by HAP and LEO do not exceed the maximum resource constraint; C4 indicates the maximum tolerable delay constraint for task processing.

[0078] The associated set is represented as:

[0079]

[0080] in, Indicates whether the terminal device u is associated with the HAP index at time slot t; Indicates whether the HAP index is associated with the LEO index at time slot t;

[0081]

[0082]

[0083] Among them, u t represents the terminal device u that needs to process the task in time slot t; h t represents the HAP index participating in the task processing in time slot t; s t represents the LEO index participating in the processing task in time slot t;

[0084] It is stipulated that each terminal device can only be associated with one HAP, which is expressed as:

[0085]

[0086] It is stipulated that each HAP can only be associated with one satellite, which is expressed as:

[0087]

[0088] The load of each HAP is limited and meets the following requirements:

[0089]

[0090] Among them, N H is the maximum association number of HAP;

[0091] The maximum power constraint P of the terminal device and HAP is expressed as:

[0092]

[0093] in, Indicates the maximum power of the terminal equipment; Indicates the maximum power of the HAP;

[0094] o is represented by:

[0095]

[0096] in, when The task will be processed directly on the HAP. Then the task will be relayed to LEO via HAP for indirect processing;

[0097] F is represented by:

[0098]

[0099] in, Indicates the maximum available computing resources of HAP; represents the maximum available computing resources of LEO;

[0100] The path set X from the terminal device to the HAP and from the HAP to the LEO is expressed as:

[0101] X={(i,j)|i∈U∪H,j∈H∪S}

[0102] S2: By constructing the bidirectional preference relationship between ground terminal equipment and HAP and the association between HAP and LEO, the optimal path selection is determined, and the weighted and mathematical model of the total energy consumption of the system is decoupled into three sub-problems: HAP selection association problem, task offloading, and resource allocation, which are solved separately;

[0103] S2.1: The task makes an offloading decision at the HAP. The two association problems of the terminal device and the HAP and the HAP and the LEO satellite are independent. The terminal device and the HAP association problem can be modeled as a many-to-one matching game problem. The preference list P of the terminal device on the HAP is u,h , expressed as:

[0104]

[0105] Among them, φ u Expressed as task priority weight; It is expressed as the transmission rate from the terminal device to the HAP; It represents the available computing resources of HAP; t represents the time slot index, from 1 to T; U represents the number of terminal devices; u represents the terminal device index, from 1 to U; u t U represents the terminal device u that needs to process tasks in time slot t; t represents the set of all terminal devices that need to process tasks in time slot t;

[0106] HAP preference list P of the terminal device h,u , expressed as:

[0107]

[0108] in, Indicates the bit size of the current computing task; h t H represents the HAP index participating in the processing task in time slot t; t represents the set of all HAPs participating in processing tasks in time slot t;

[0109] Taking into account the preference of the terminal device for the HAP and the preference of the HAP for the terminal device, the comprehensive preference target is defined as:

[0110]

[0111] Where M is the matching set, λ and μ are weight coefficients; using the blocking pair theory, by determining whether the matching relationship between the terminal device and the HAP meets the blocking pair condition, M is gradually adjusted; in each iteration, if a blocking pair is found, resources are reallocated until all blocking pairs are eliminated, so that M reaches a stable state and meets the constraint conditions;

[0112] The matching problem between HAP and LEO is modeled as a many-to-many matching problem and solved by using the GUROBI optimizer. The goal is to minimize the transmission energy consumption between HAP and LEO while meeting the computing task offloading requirements of HAP so that the matching relationship reaches a stable state.

[0113] S2.2: Based on the transmission and computing characteristics of the task, each task only selects one processing method, which is represented by binary decision variables to establish a task offloading decision model, which is expressed as:

[0114]

[0115]

[0116]

[0117] Through the particle swarm optimization algorithm, the optimal offloading decision is solved to minimize the system energy consumption under the constraint of meeting the task delay;

[0118] S2.3: Optimize resource allocation based on HAP selection association and task offloading decision; First, dynamically select the most suitable HAP according to the task characteristics of the terminal device and the resource status of the HAP. Then, decide whether the task is processed locally at the HAP or offloaded to the LEO to minimize energy consumption and optimize resource utilization; formulate the optimization problem as a problem of optimizing computing resources and power allocation, and adopt the whale optimization algorithm to process the constraints by introducing a penalty function, conduct local and global searches, explore the global optimum, and gradually approach the optimal solution through fitness evaluation. The penalty function Λ(p) is expressed as:

[0119]

[0120] Among them, δ1(t) and η1(t) represent the adaptive penalty factors of key constraints, which change dynamically with time slots; δ2(t) and η2(t) are used for the adaptive penalty factors of secondary constraints; Represents the indicator function, which is used to judge the constraints Whether it is violated, when When the indicator function G h =1, otherwise G h =0; Represents the indicator function, which is used to judge the constraints Whether it is violated, when When H s =1, otherwise H s = 0; r and z are parameters for adjusting the violation degree index, usually taking the value of 2 to increase the penalty for solutions with a larger violation degree; Indicates the degree of violation A logarithmic penalty of , which applies to a smaller range of violations; Indicates the degree of violation An exponential penalty of 1% applies to rapidly increasing levels of violations; represents the objective function, Represents the power weight factor.

[0121] S3: The results of HAP selection association, task offloading decision and resource allocation are jointly iteratively optimized, and the weighted sum of the total energy consumption of the system obtained in each iteration is used as the fitness evaluation criterion until it converges to the optimal weighted sum of the total energy consumption of the system.

[0122] S3.1: The weighted sum of the total energy consumption of the system obtained by iteration is used as the fitness evaluation criterion. In each iteration, the optimal solution is gradually approached by optimizing the HAP association selection, task offloading path and resource allocation configuration. The iterative process updates the fitness evaluation criterion after each round of iteration until the weighted sum of the total energy consumption of the system reaches the convergence condition, thereby obtaining the optimal weighted sum of the total energy consumption of the system.

[0123] The present invention aims at the problem of task offloading of intensive tasks. Due to the sharp increase in the number of terminal devices and calculations, the complexity of communication infrastructure in some areas, etc., the dynamic association relationship of HAP is considered, and its large-scale access acts as a relay and an air base station at the same time, providing edge computing services for intensive tasks, and the edge server on the satellite can also provide edge computing services. A high-efficiency hybrid offloading strategy is proposed, which realizes the efficiency and flexibility of resource allocation by jointly optimizing the selection association of HAP, the task offloading path and reasonably allocating computing resources. The present invention effectively improves the energy efficiency of the system.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A resource allocation method for intensive tasks, characterized in that: The following steps are involved: S1. The intensive tasks of the ground terminal equipment are collected to the High Altitude Platform (HAP) or forwarded by the HAP to the Low Earth Orbit (LEO) satellite for processing, and a weighted and mathematical model of the total energy consumption of the system is constructed; S2. Determine the optimal path selection by constructing the bidirectional preference relationship between the ground terminal equipment and HAP and the association between HAP and LEO; decouple the total energy consumption weighted and mathematical model of the system into three sub-problems: HAP selection association, task offloading decision and resource allocation, and solve them separately; S3. The solution results of the three sub-problems of HAP selection association, task offloading decision and resource allocation are jointly iteratively optimized, and the weighted sum of the total energy consumption of the system obtained in each iteration is used as the fitness evaluation criterion until it converges to the optimal weighted sum of the total energy consumption of the system.

2. The resource allocation method for intensive tasks according to claim 1, characterized in that: In S1: The HAP acts as a base station and relay node. Under the maximum delay constraint, the weighted sum mathematical model of the total energy consumption of the construction system is expressed as: Among them, E U,H and E H,L They represent the transmission energy consumption from the terminal device to HAP and from HAP to LEO, respectively. α1 and α2 are weight coefficients, which respectively measure the different contributions of the transmission energy consumption from the terminal device to HAP and from HAP to LEO to the optimization objective. E H represents the energy consumption of HAP calculation; E L represents LEO computing energy consumption; β and γ are constant positive weights of HAP and LEO transmission energy consumption, respectively; H represents H HAPs; h represents HAP index, from 1 to H; t represents time slot index, from 1 to T; o represents HAP offloading decision constraint: P represents the maximum power constraint of terminal equipment and HAP; Indicates whether the task is processed locally in HAP; Indicates whether the task is offloaded to LEO for processing; P represents the maximum power constraint set of the terminal device and HAP; Indicates that the power of path X should not exceed the maximum power constraint; U indicates that there are U terminal devices; u indicates the terminal device index, from 1 to U; S indicates that there are S LEOs; s indicates the LEO index, from 1 to S; Represented as the bit size of the current computing task; represents the computing resources of the current path X; F represents the maximum computing resource constraint of HAP and LEO; d u Indicates the actual completion delay of the task; Indicates the maximum tolerable delay of the task; C1 indicates that the terminal device can only choose one task offloading mode in a time slot; C2 indicates that the power of the terminal device and HAP does not exceed the maximum power limit; C3 indicates that the resources used by HAP and LEO do not exceed the maximum resource constraint; C4 indicates the maximum tolerable delay constraint for task processing; The association constraint is expressed as: in, Indicates whether the terminal device u is associated with the HAP index at time slot t; Indicates whether the HAP index is associated with the LEO index at time slot t; The maximum power constraint P of the terminal device and HAP is expressed as: in, Indicates the maximum power of the terminal equipment; Indicates the maximum power of the HAP; o is represented by: in, when The task will be processed directly on the HAP. Then the task will be relayed to LEO via HAP for indirect processing; F is represented by: in, Indicates the maximum available computing resources of HAP; represents the maximum available computing resources of LEO; The path set X from the terminal device to the HAP and from the HAP to the LEO is expressed as: X={(i,j)|i∈U∪H,j∈H∪S}.

3. The resource allocation method for intensive tasks according to claim 2, characterized in that: Said It is expressed as: Said It is expressed as: Among them, u t represents the terminal device u that needs to process tasks in time slot t; h t represents the HAP index participating in the task processing in time slot t; s t represents the LEO index participating in the processing task in time slot t; Each terminal device can only be associated with one HAP, expressed as: Each HAP can only be associated with one satellite, expressed as: Each HAP has a limited load, which satisfies: Among them, N H is the maximum number of associations of HAP.

4. The resource allocation method for intensive tasks according to claim 1, characterized in that: The S2 is specifically: Preference list P of terminal devices on HAP u,h , expressed as: Among them, φ u Expressed as task priority weight; It is expressed as the transmission rate from the terminal device to the HAP; It represents the available computing resources of HAP; t represents the time slot index, from 1 to T; U represents the number of terminal devices; u represents the terminal device index, from 1 to U; u t U represents the terminal device u that needs to process tasks in time slot t; t represents the set of all terminal devices that need to process tasks in time slot t; HAP preference list P of the terminal device h,u , expressed as: in, Indicates the bit size of the current computing task; h t H represents the HAP index participating in the processing task in time slot t; t represents the set of all HAPs participating in processing tasks in time slot t; The comprehensive preference objective is defined as: Where M is the matching set, λ and μ are weight coefficients; using the blocking pair theory, by determining whether the matching relationship between the terminal device and the HAP meets the blocking pair condition, M is gradually adjusted; in each iteration, if a blocking pair is found, resources are reallocated until all blocking pairs are eliminated, so that M reaches a stable state and meets the constraint conditions; The matching problem between HAP and LEO is modeled as a many-to-many matching problem, which is solved by using the GUROBI optimizer to minimize the transmission energy consumption between HAP and LEO; Based on the transmission and computing characteristics of tasks, a task offloading decision model is established, and the optimal offloading path is solved by particle swarm optimization algorithm; Based on HAP selection association and task offloading decision, the problem of optimizing resource allocation is expressed as the problem of optimizing computing resources and power allocation. Under the condition of ensuring that the allocated computing resources do not exceed the maximum computing capacity, the whale optimization algorithm is adopted to deal with the constraints by introducing the penalty function, conduct local and global searches, explore the global optimum, and gradually approach the optimal solution through fitness evaluation. The penalty function Λ(p) is expressed as: Among them, δ1(t) and η1(t) represent the adaptive penalty factors of key constraints, which change dynamically with time slots; δ2(t) and η2(t) are used for the adaptive penalty factors of secondary constraints; Represents the indicator function, which is used to judge the constraints whether it has been violated; Represents the indicator function, which is used to judge the constraints Whether it is violated; r and z are parameters for adjusting the violation degree index, which are used to increase the penalty for solutions with greater violations; Indicates the degree of violation The logarithmic penalty is used for a smaller range of violations; Indicates the degree of violation The exponential penalty is used for rapidly growing violation levels; Q(p) represents the objective function, Represents the power weight factor.

5. The method for allocating resources for intensive tasks according to claim 4, characterized in that: The task offloading decision model is specifically:

6. The method for allocating resources for intensive tasks according to claim 4, characterized in that: like Then the indicator function G h =1; otherwise G h =0.

7. The method for allocating resources for intensive tasks according to claim 4, characterized in that: like Then H s =1; otherwise H s =0.

8. The method for allocating resources for intensive tasks according to claim 4, characterized in that: The objective function Q(p) is specifically:

9. The resource allocation method for intensive tasks according to claim 1, characterized in that: S3 specifically comprises: taking the weighted sum of the total system energy consumption obtained by the iteration as a fitness evaluation criterion, and gradually approaching the optimal solution by optimizing HAP association selection, task offloading path and resource allocation configuration in each iteration; The iterative process updates the fitness evaluation criterion after each round of iteration until the weighted sum of the total energy consumption of the system reaches a convergence condition to obtain the optimal weighted sum of the total energy consumption of the system.

Images