Task offloading method and system based on game theory in air-ground integrated network

Abstract

The present invention relates to the field of communication technology, and discloses a task offloading method and system based on game theory in an integrated air-ground-space network, including: S1, according to the time delay and computing resource requirements of each task generated by the current ground user equipment, respectively calculating the offloading cost of each task at the edge of the low-orbit satellite network, the edge of the drone network and the local area; S2, with the goal of minimizing cost overhead, finding an offloading strategy with the minimum offloading cost overhead; S3, according to the offloading strategy, assigning each task to the edge of the low-orbit satellite network and / or the edge of the drone network and / or the local area for offloading. The present invention proposes a Nash equilibrium iterative algorithm based on game theory to solve the problem of computational offloading, and obtains the optimal offloading combination strategy by finding the Nash equilibrium point, thereby minimizing the system cost of offloading.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a task unloading method and system based on game theory in an air-ground integrated network, which is used to solve the problem of task unloading in remote areas and unloading of intensive and delay-sensitive tasks in developed cities. Background Art

[0002] With the rapid development of Internet of Things (IoT) technology, latency-sensitive applications such as augmented reality, face recognition, and video analysis are emerging in an endless stream. However, due to the limited computing power of mobile IoT devices, the computing resources and latency requirements of these applications cannot be met. Therefore, the mobile edge computing (MEC) technology is proposed to transfer the computing tasks of IoT users to edge servers for execution, which can not only reduce the latency of task processing, but also meet the computing resource requirements of tasks. Considering that there is no ground network in some remote areas of deserts, oceans, and mountains, mobile IoT devices in these remote areas cannot process some tasks with high real-time requirements and large computing resource requirements. As a supplement to the ground network, the low earth orbit (LEO) satellite network has the characteristics of global seamless coverage and low transmission latency, and has been listed as a key supporting technology for the future 6G by the International Telecommunication Union (ITU). For computing tasks with high real-time requirements and large computing resource requirements for remote IoT users, edge servers can be deployed on LEO satellites to provide them with satellite edge computing services to effectively reduce the processing latency of tasks and improve the service quality of users. The edge computing network of aerial drone networks can further assist the satellite edge computing network, providing computing resources for intensive and delay-sensitive tasks in developed cities, and alleviating the pressure of LEO satellite edge computing. As for some satellite-ground fusion networks, they only consider a single satellite or multi-satellite collaboration. When faced with a large number of remote mountainous areas or delay-sensitive computing tasks, the computing resources of the satellite network may be insufficient or the delay may be long, causing serious network congestion, and therefore unable to meet the computing resource and delay requirements of the task. In the scenario of an integrated air-ground-space network, how to meet the user's task delay and computing resource requirements while minimizing system energy consumption and other cost overheads and perform task offloading is the current focus of research in academia and industry. Summary of the invention

[0003] The present invention provides a task offloading method and system based on game theory in an integrated air-ground-space network, which offloads tasks generated by ground user equipment to a satellite network, a UAV network cluster network or local processing to minimize the total system overhead, thereby solving the problems of poor communication quality and insufficient computing resources in remote areas, as well as network congestion and delay in intensive and delay-sensitive tasks in developed cities.

[0004] The present invention is achieved through the following technical solutions:

[0005] A task offloading method based on game theory in an air-ground integrated network comprises:

[0006] S1. Calculate the offloading cost of each task at the edge of the low-orbit satellite network, the edge of the drone network and locally, based on the latency and computing resource requirements of each task generated by the current ground user equipment;

[0007] S2, with the goal of minimizing cost overhead, find the offloading strategy with the lowest offloading cost overhead;

[0008] S3. According to the offloading strategy, each of the tasks is assigned to the edge of the low-orbit satellite network and / or the edge of the drone network and / or locally for offloading.

[0009] As an optimization, in S1, the specific steps of calculating the offloading cost of the task to the edge of the low-orbit satellite network are:

[0010] A1. Calculate the effective communication time T of each task generated by the ground user equipment s The first delay consumption of unloading to the edge of the low-orbit satellite network The first delay consumption It is composed of propagation delay, transmission delay and processing delay;

[0011] A2. Calculate the first transmission energy consumption of each task offloaded to the edge of the low-orbit satellite network And the first processing energy consumption E of each task at the edge of the low-orbit satellite network u，s ;

[0012] A3. According to the first delay consumption First transmission energy consumption And the first processing energy consumption E u，s Obtain the first unloading cost of each task to the edge of the low-orbit satellite network

[0013]

[0014] Wherein, μ1 is the weight factor of the first delay, and λ1 is the weight factor of the first transmission energy consumption and the first processing energy consumption.

[0015] As an optimization, the effective communication time T s The calculation steps include:

[0016] A1.1. Calculate the geocentric angle β corresponding to the coverage area of ​​the low-orbit satellite network according to the geometric relationship of the orbit model of the low-orbit satellite network and the given elevation angle α between the ground user equipment and the low-orbit satellite network;

[0017] A1.2. Calculate the communication arc length L between the ground user equipment and the low-orbit satellite network according to the geocentric angle β;

[0018] A1.3, according to the communication arc length L and the speed v of the low-orbit satellite network s Calculate the communication time between the ground user equipment and the low-orbit satellite network

[0019] As an optimization, in S1, the specific steps of calculating the offloading cost of the task to the edge of the drone network are:

[0020] B1. Calculate the second delay consumption of each task unloaded to the edge of the drone network The second delay consumes It is composed of propagation delay, uplink transmission delay and processing delay;

[0021] B2. Calculate the second transmission energy consumption of each task offloaded to the edge of the drone network And the second processing energy consumption E of each task at the edge of the drone network u，m ;

[0022] B3. According to the second delay consumption Second transmission energy consumption Second processing energy consumption E u，m Get the second unloading cost of each task to the edge of the drone network

[0023]

[0024] Wherein, μ2 is the weight factor of the second delay, and λ2 is the weight factor of the second transmission energy consumption and the second processing energy consumption.

[0025] As an optimization, in S1, the specific steps of calculating the offloading cost of the task to the local are:

[0026] C1. Calculate the third delay consumption of each task generated by the ground user equipment and offloaded to the local The third delay consumption It is composed of processing delays;

[0027] C2. Calculate the third processing consumption of each task offloaded to the local

[0028] C3, according to the third delay consumption and the third processing consumption Get the third unloading cost of unloading locally

[0029] Wherein, μ3 is the weight factor of the third delay, and λ3 is the weight factor of the third processing energy consumption.

[0030] As an optimization, the specific steps of S2 are:

[0031] S2.1. Convert the offloading cost problem into a cost optimization problem under the constraints of delay and energy consumption:

[0032]

[0033] Where K is the total number of ground user equipment, are the first unloading cost, the second unloading cost and the third unloading cost respectively, a u Represents the computing task W u Whether to unload to a low-orbit satellite, where a u =1 indicates the computing task W u to a low-orbit satellite, otherwise, a u =0 means computing task W u Not carried on low-orbit satellite; b u Represents the computing task W u Whether to unload to the edge of the drone, where b u =1 indicates computing task W u Load to the edge of the drone, otherwise, b u =0 means computing task W u Not loaded onto the edge of the drone;

[0034] S2.2. Based on the game theory method, the problem of minimizing the computational offloading cost composed of delay and energy consumption is expressed as a non-cooperative strategy game G. The strategy game G is expressed as G = {U, (X u ) u∈U ,(Υ u (x u ,x -u )) u∈U};

[0035] Wherein, U represents the set of game participants, and the game participants are ground user equipments, (X u ) u∈U represents the policy set of user device participant u, (Υ u (x u ,x -u )) u∈U represents the participants’ benefits, i.e., the system cost function, x -u represents the uninstallation policy for all devices except the user device participant u;

[0036] S2.3. Each ground user equipment independently adjusts the strategy game G under the constraints of resources and delay to maximize the benefits of the ground user equipment, that is, to minimize the cost:

[0037]

[0038] S2.4. Find the Nash equilibrium point of the non-cooperative game through an iterative unloading algorithm based on game theory, that is, find the unloading strategy with the lowest unloading cost. Make in, It represents the optimal offloading policy of other users except the user equipment u under the optimal offloading policy of the user equipment u.

[0039] As an optimization, the specific steps of S2.4 are:

[0040] S2.4.1. Initialize the strategy set, set the maximum number of iterations to perform algorithm iterations, traverse all ground user devices in each iteration, and find the current offloading strategy x′ corresponding to the minimum cost of computing task offloading for each ground user device. u ;

[0041] S2.4.2, iterate and set the current unloading strategy x′ of all the ground user equipments u The cost and the historical tentative optimal unloading strategy x u The cost is compared with that of the current unloading strategy x′. u The cost is greater than the historical tentative optimal unloading strategy x u The cost spent, then continue to temporarily set the optimal unloading strategy x for this history u Set as the historical tentative optimal unloading strategy x u ; If the current uninstallation policy x′ u The cost is less than the historical tentative optimal unloading strategy x u The cost spent, then the current uninstallation strategy x′ u Set as the new historical tentative optimal unloading strategy x u ;

[0042] S2.4.3. Determine the current uninstallation strategy x′ u and historical tentative optimal unloading strategy x u Is the error between them less than the set threshold? If so, stop the iteration and change the current uninstallation strategy x′ u As the optimal unloading strategy X * , Make This means that this strategy is the optimal unloading strategy and can minimize the system cost. No other strategy can be found with a system cost smaller than this one. Otherwise, return to S2.4.2 and continue to the next iteration.

[0043] A task offloading system based on game theory in an air-ground-integrated network is used to implement the above-mentioned task offloading method based on game theory in an air-ground-integrated network, characterized by comprising:

[0044] A calculation module, used to calculate the offloading cost of each task at the edge of the low-orbit satellite network, the edge of the drone network and locally according to the latency and computing resource requirements of each task generated by the current ground user equipment;

[0045] A strategy formulation module is used to find the offloading strategy with the lowest offloading cost with the goal of minimizing the cost;

[0046] An execution module is used to allocate each of the tasks to the edge of the low-orbit satellite network and / or the edge of the drone network and / or locally for unloading according to the unloading strategy.

[0047] As an optimization, the low-orbit satellite network includes several low-orbit satellites, and each of the low-orbit satellites is provided with a lightweight MEC server for carrying a computing module, and one of the low-orbit satellites is also provided with a global controller for carrying a policy formulation module.

[0048] As an optimization, the drone network includes several drone clusters, each of which is equipped with multiple drones, and each drone is equipped with a lightweight MEC server for carrying a computing module; an inter-domain controller is provided in each drone cluster for resource allocation and task scheduling of the entire drone cluster.

[0049] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0050] 1. The present invention proposes a three-layer edge computing network architecture of low-orbit satellite-UAV cluster-ground. The computing tasks in remote areas can be offloaded to satellites for processing, which can well solve the task computing problems in harsh environmental areas. The intensive and delay-sensitive tasks in developed cities can be offloaded to UAV clusters for processing, which can not only meet the resource requirements and delay requirements of computing tasks, but also further alleviate the pressure on satellite computing resources and network congestion.

[0051] 2. A global controller is set up in the satellite network of this three-layer network, which can globally manage and centrally control the inter-domain controllers of the aerial drone cluster; and the inter-domain controllers are deployed in the drone cluster, which can allocate resources and schedule tasks for the drone cluster. The global controller and the inter-domain controller can ensure the stable and high-speed operation of the entire air-ground integrated network.

[0052] 3. The present invention combines lightweight MEC servers with satellite networks and drone cluster networks, and utilizes the resources of MEC servers to provide heterogeneous computing services for ground user devices. In addition, the limited computing power and coverage time of low-orbit satellites are also considered, and then a task offloading minimization cost model and optimization objectives are proposed.

[0053] 4. The present invention proposes a Nash equilibrium iterative algorithm based on game theory to solve the problem of computational offloading. By finding the Nash equilibrium point, the optimal offloading combination strategy is obtained, thereby minimizing the system cost of offloading;

[0054] 5. The integrated air-space-ground network of the present invention can well solve the problem of computing resources and delay requirements required for the task, and also alleviate the computing pressure of the satellite network. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0056] Figure 1 An example diagram of the satellite-UAV cluster-ground air-ground integrated network architecture provided in Example 1;

[0057] Figure 2 The low-orbit satellite orbit coverage map provided in Example 2;

[0058] Figure 3 This is the main flow chart of unloading based on game theory provided in Example 3. DETAILED DESCRIPTION

[0059] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0060] Example 1

[0061] like Figure 1As shown in the figure, the construction of the integrated air-space-ground network is mainly to provide a reliable and efficient communication method for the computing tasks of user equipment in remote areas due to poor ground network conditions, lack of computing resources and high latency. At the same time, it can also meet the computing resource needs and latency requirements of intensive and delay-sensitive tasks in cities, and on the other hand, it can also reduce network congestion and alleviate the pressure on satellite computing resources.

[0062] This embodiment 1 provides the construction of a satellite-UAV cluster-ground air-ground integrated network architecture, such as Figure 1 As shown, the following steps are included:

[0063] S(1) Build a low-orbit satellite network, each satellite is equipped with a lightweight MEC server with certain computing capabilities;

[0064] A global controller is provided in the low-orbit satellite network, and an unloading algorithm based on game theory is placed in the global controller, that is, a task unloading method based on game theory in an air-ground integrated network of the present invention is used for centralized resource management of inter-domain controllers and making optimal unloading decisions for tasks generated by ground users;

[0065] S(2) Build an aerial drone cluster network. Each drone cluster is equipped with multiple drones, and each drone is equipped with a lightweight MEC server with certain computing capabilities. In each drone cluster, an inter-domain controller is deployed to allocate resources and schedule tasks for the entire cluster.

[0066] S(3) Computing tasks generated on the ground are processed locally according to their own computing resource requirements and latency requirements, or offloaded to low-orbit satellites or aerial drone clusters for processing.

[0067] Example 2

[0068] Due to the high dynamics of low-orbit satellites, the communication time between them and the user equipment within their coverage area is limited. The ground user equipment must offload the task to the satellite within the period of satellite coverage, otherwise the data may be lost when the task is uploaded.

[0069] This embodiment 2 provides a low-orbit satellite orbit coverage map, such as Figure 2 As shown, the following steps are included:

[0070] S(1) Calculate the geocentric angle β corresponding to the low-orbit satellite coverage area based on the geometric relationship of the low-orbit satellite orbit model and the given elevation angle α between the ground user equipment and the satellite:

[0071]

[0072] Where R is the radius of the earth and H is the distance between the ground user equipment and the low-orbit satellite orbit.

[0073] S(2) According to the geocentric angle, the arc length L between the ground user equipment and the low-orbit satellite can be calculated:

[0074] S(3), and finally according to the communication arc length L and the speed v of the low-orbit satellite s The effective communication time between ground user equipment and low-orbit satellites can be calculated

[0075] Example 3

[0076] After building an integrated air-ground network and satellite orbit coverage map, it is necessary to offload tasks to satellites, drones, or local processing according to the network communication conditions in different regions. Since the cost of offloading tasks to satellites, drone clusters, and local processing is different, firstly, according to the computing resource requirements and latency requirements required for the computing tasks, the offloading costs of the tasks at these locations are calculated, and then with the goal of minimizing the total cost of offloading computing tasks for all users on the ground, an iterative offloading algorithm based on game theory is proposed, that is, a task offloading method based on game theory in an integrated air-ground network of the present invention.

[0077] The main flow chart of uninstallation based on game theory provided in this embodiment 3 is as follows: Figure 3 As shown, the following steps are included:

[0078] S(1) Build an integrated air-space-ground network;

[0079] S(2) computing tasks are generated by ground user equipment;

[0080] S(3) calculates the cost of offloading the task to a satellite, a drone cluster, or local processing;

[0081] S(4) First, according to the Shannon formula, the uplink transmission rate R from the ground user equipment to the low-orbit satellite is calculated. u,s and the uplink transmission rate R from ground user equipment to the UAV cluster u,m ;

[0082]

[0083] Among them, p s is the uplink transmission power from ground user equipment u to the satellite, h u,s represents the channel power gain from the ground user equipment u to the LEO satellite s, B1 represents the channel bandwidth from the ground user equipment to the LEO satellite, N0 represents the power spectral density of Gaussian white noise, p iFor the uplink transmission power of other user equipment unloaded to the low-orbit satellite, p is set here i This takes into account the problem of channel interference caused by transmissions from different user devices.

[0084]

[0085] In the formula, p m is the uplink transmission power from ground user equipment u to the UAV, h u,m represents the channel power gain from the ground user equipment u to the UAV cluster, B2 represents the channel bandwidth from the ground user equipment to the UAV cluster, and p j In order to offload the uplink transmission power of other user devices in the drone cluster together, the problem of channel interference caused by transmissions from different user devices is taken into consideration.

[0086] S(5) calculates the cost of offloading the task to a low-orbit satellite, which mainly consists of delay consumption and energy consumption.

[0087] S(5.1) first calculates the first delay consumption of the computing task of the ground user equipment unloaded at the edge of the low-orbit satellite It is mainly composed of transmission delay, propagation delay and processing delay;

[0088] S(5.2) calculates the energy consumption of the first transmission of the computing task of the ground user equipment to the low-orbit satellite and the first processing energy consumption E of the task u，s ;

[0089] S(5.3) can be used to obtain the cost function of offloading tasks to the edge nodes of low-orbit satellites:

[0090]

[0091] Where μ1 is the weight factor of the first delay, λ1 is the weight factor of the first transmission energy consumption and the first processing energy consumption

[0092] S(6) calculates the cost of offloading to the drone cluster, which is mainly composed of delay consumption and energy consumption;

[0093] S(6.1) first calculates the second delay consumption of the computing task of the ground user equipment unloaded at the drone edge node It is mainly composed of transmission delay, propagation delay and processing delay;

[0094] S(6.2) calculates the second transmission energy consumption of offloading the computing task of the ground user equipment to the UAV The second processing energy consumption E of the task u，m ;

[0095] S(6.3) can be used to obtain the cost function of offloading tasks to the edge nodes of drones:

[0096]

[0097] Wherein, μ2 is the weight factor of the second delay, and λ2 is the weight factor of the second transmission energy consumption and the second processing energy consumption.

[0098] S(7) calculates the cost of local processing, which mainly consists of processing delay and task processing energy consumption;

[0099] S(7.1) calculates the local third latency consumption of the task

[0100] S(7.2) calculates the local third processing consumption of the task

[0101] S(7.3) can be used to obtain the cost function of the task calculated locally:

[0102]

[0103] Where μ3 is the weight factor of the third delay, and λ3 is the weight factor of the third processing energy consumption.

[0104] S(8) transforms the offloading problem into a system cost function optimization problem under the constraints of delay and energy consumption, which is defined as:

[0105] Where K is the total number of ground user equipment, are the first unloading cost, the second unloading cost and the third unloading cost respectively, a u Represents the computing task W u Whether to unload to a low-orbit satellite, where a u =1 indicates computing task W u to a low-orbit satellite, otherwise, a u =0 means computing task W u Not carried on low-orbit satellite; b u Represents the computing task W u Whether to unload to the edge of the drone, where b u =1 indicates computing task W u Load to the edge of the drone, otherwise, b u =0 means computing task W u Not loaded onto the edge of the drone;

[0106] S(10) Based on the game theory method, the computational offloading problem with the minimum system cost composed of delay and energy consumption is expressed as a non-cooperative strategy game G, which can be expressed as G = {U, (X u )u∈U ,(Υ u (x u ,x -u )) u∈U}, U represents the set of game participants, (X u ) u∈U represents the policy set of user device participant u, (Υ u (x u ,x -u )) u∈U represents the participants’ benefits, i.e., the system cost function, x -u represents the offloading strategy of all other devices except user device u. In a competitive game, each user device independently adjusts its strategy under the constraints of resources and delay to maximize its own benefits, that is, minimize the system cost:

[0107]

[0108] S(11) uses an iterative unloading algorithm based on game theory to find the Nash equilibrium of the non-cooperative game (if any participant changes their strategy, it will affect the optimal value of the strategy combination, so that all participants will not get better benefits). The specific steps are as follows:

[0109] Step 1: First, initialize the strategy set and set the maximum number of iterations to iterate the algorithm. In each iteration, traverse all user devices on the ground and find the offloading strategy x′ that minimizes the offloading cost of each user's computing task. u ;

[0110] Step 2: Set the current uninstall policy x′ of all user devices u The system cost and the historical tentative optimal unloading strategy x u The system cost is compared with that if the current unloading strategy x′ u The system cost is greater than the historical tentative optimal unloading strategy x u The system cost spent, then the current unloading strategy x′ is considered u It is not the best uninstallation strategy. The best uninstallation strategy is tentatively determined by history. u It is tentatively determined as the optimal unloading strategy; if the current unloading strategy x′ u The system cost is less than the historical tentative optimal unloading strategy x u The system cost spent, then the current unloading strategy x′ is temporarily determined u To obtain the optimal unloading strategy, continue to the next iteration.

[0111] Step 3: Finally, the current uninstallation strategy x′ u The system cost and the historical tentative optimal unloading strategy x uIf the system cost is within the termination error ε, it means the termination of the iteration. At this time, the unloading strategy x′ u Is the optimal uninstall strategy X * , that is, the Nash equilibrium state is reached. That is, the optimal decision combination for unloading Make in, It represents the optimal offloading policy of other users except the user equipment u under the optimal offloading policy of the user equipment u.

[0112] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A task offloading method based on game theory in an air-ground integrated network, characterized in that: include: S1. Calculate the offloading cost of each task at the edge of the low-orbit satellite network, the edge of the drone network and locally, based on the latency and computing resource requirements of each task generated by the current ground user equipment; In S1, the specific steps of calculating the offloading cost of the task to the edge of the low-orbit satellite network are: A1. Calculate the effective communication time T of each task generated by the ground user equipment s The first delay consumption of unloading to the edge of the low-orbit satellite network The first delay consumption It is composed of propagation delay, transmission delay and processing delay; A2. Calculate the first transmission energy consumption of each task offloaded to the edge of the low-orbit satellite network And the first processing energy consumption E of each task at the edge of the low-orbit satellite network u,s ; A3. According to the first delay consumption First transmission energy consumption And the first processing energy consumption E u,s Obtain the first unloading cost of each task to the edge of the low-orbit satellite network Wherein, μ1 is the weight factor of the first delay, λ1 is the weight factor of the first transmission energy consumption and the first processing energy consumption; S2, with the goal of minimizing cost overhead, find the offloading strategy with the lowest offloading cost overhead; S3. According to the offloading strategy, each of the tasks is assigned to the edge of the low-orbit satellite network and / or the edge of the drone network and / or locally for offloading.

2. The task offloading method based on game theory in an air-ground integrated network according to claim 1 is characterized in that: The effective communication time T s The calculation steps include: A1.

1. Calculate the geocentric angle β corresponding to the coverage area of ​​the low-orbit satellite network according to the geometric relationship of the orbit model of the low-orbit satellite network and the given elevation angle α between the ground user equipment and the low-orbit satellite network; A1.

2. Calculate the communication arc length L between the ground user equipment and the low-orbit satellite network according to the geocentric angle β; A1.3, according to the communication arc length L and the speed v of the low-orbit satellite network s Calculate the communication time between the ground user equipment and the low-orbit satellite network 3. The task offloading method based on game theory in an air-ground integrated network according to claim 1 is characterized in that: In S1, the specific steps of calculating the offloading cost of the task to the edge of the drone network are: B1. Calculate the second delay consumption of each task unloaded to the edge of the drone network The second delay consumes It is composed of propagation delay, uplink transmission delay and processing delay; B2. Calculate the second transmission energy consumption of each task offloaded to the edge of the drone network And the second processing energy consumption E of each task at the edge of the drone network u,m ; B3. According to the second delay consumption Second transmission energy consumption Second processing energy consumption E u,m Get the second unloading cost of each task to the edge of the drone network Wherein, μ2 is the weight factor of the second delay, and λ2 is the weight factor of the second transmission energy consumption and the second processing energy consumption.

4. The task offloading method based on game theory in an air-ground integrated network according to claim 3 is characterized in that: In S1, the specific steps of calculating the offloading cost of the task to the local are: C1. Calculate the third delay consumption of each task generated by the ground user equipment and offloaded to the local The third delay consumption It is composed of processing delays; C2. Calculate the third processing consumption of each task offloaded to the local C3, according to the third delay consumption and the third processing consumption Get the third unloading cost of unloading locally Wherein, μ3 is the weight factor of the third delay, and λ3 is the weight factor of the third processing energy consumption.

5. The task offloading method based on game theory in an air-ground integrated network according to claim 4 is characterized in that: The specific steps of S2 are: S2.

1. Convert the offloading cost problem into a cost optimization problem under the constraints of delay and energy consumption: Where K is the total number of ground user equipment, are the first unloading cost, the second unloading cost and the third unloading cost respectively, a u Represents the computing task W u Whether to unload to a low-orbit satellite, where a u =1 indicates the computing task W u to a low-orbit satellite, otherwise, a u =0 means computing task W u Not carried on low-orbit satellite; b u Represents the computing task W u Whether to unload to the edge of the drone, where b u =1 indicates the computing task W u Load to the edge of the drone, otherwise, b u =0 means computing task W u Not loaded onto the edge of the drone; S2.

2. Based on the game theory method, the problem of minimizing the computational offloading cost composed of delay and energy consumption is expressed as a non-cooperative strategy game G. The strategy game G is expressed as G = {U, (X u ) u∈U ,(γ u (x u ,x -u )) u∈U }; Wherein, U represents the set of game participants, and the game participants are ground user equipments, (X u ) u∈U represents the policy set of user device participant u, (γ u (x u ,x -u )) u∈U represents the participants’ benefits, i.e., the system cost function, x -u represents the uninstallation policy for all devices except the user device participant u; S2.

3. Each ground user equipment independently adjusts the strategy game G under the constraints of resources and delay to maximize the benefits of the ground user equipment, that is, to minimize the cost: S2.

4. Find the Nash equilibrium point of the non-cooperative game through an iterative unloading algorithm based on game theory, that is, find the unloading strategy with the lowest unloading cost. Make in, It represents the optimal uninstallation policy of users other than the user equipment u under the optimal uninstallation policy of the user equipment u.

6. The task offloading method based on game theory in an air-ground integrated network according to claim 5 is characterized in that: The specific steps of S2.4 are: S2.4.

1. Initialize the strategy set, set the maximum number of iterations to perform algorithm iterations, traverse all ground user devices in each iteration, and find the current offloading strategy x′ corresponding to the minimum cost of computing task offloading for each ground user device. u ; S2.4.2, iterate and set the current unloading strategy x′ of all the ground user equipments u The cost and the historical tentative optimal unloading strategy x u The cost is compared with that of the current unloading strategy x′. u The cost is greater than the historical tentative optimal unloading strategy x u The cost spent, then continue to temporarily set the optimal unloading strategy x for this history u Set as the historical tentative optimal unloading strategy x u ; If the current uninstallation policy x′ u The cost is less than the historical tentative optimal unloading strategy x u The cost spent, then the current uninstallation strategy x′ u Set as the new historical tentative optimal unloading strategy x u ; S2.4.

3. Determine the current uninstallation strategy x′ u and historical tentative optimal unloading strategy x u Is the error between them less than the set threshold? If so, stop the iteration and change the current uninstallation strategy x′ u As the optimal unloading strategy x * , Make Otherwise, return to S2.4.2 and continue to the next iteration.

7. A task unloading system based on game theory in an air-ground-integrated network, used to implement a task unloading method based on game theory in an air-ground-integrated network as described in any of claims 1-6, characterized in that: include: A calculation module, used to calculate the offloading cost of each task at the edge of the low-orbit satellite network, the edge of the drone network and locally according to the latency and computing resource requirements of each task generated by the current ground user equipment; A strategy formulation module is used to find the offloading strategy with the lowest offloading cost with the goal of minimizing the cost; An execution module is used to allocate each of the tasks to the edge of the low-orbit satellite network and / or the edge of the drone network and / or locally for unloading according to the unloading strategy.

8. The task offloading system based on game theory in an air-ground integrated network according to claim 7 is characterized in that: The low-orbit satellite network includes several low-orbit satellites, and each of the low-orbit satellites is provided with a lightweight MEC server for carrying a computing module, and one of the low-orbit satellites is also provided with a global controller for carrying a policy formulation module.

9. The task offloading system based on game theory in an air-ground integrated network according to claim 7 is characterized in that: The drone network includes several drone clusters, each of which is equipped with multiple drones, and each drone is equipped with a lightweight MEC server for carrying a computing module; an inter-domain controller is provided in each drone cluster for resource allocation and task scheduling of the entire drone cluster.

Images