Task scheduling and order optimization method and system in uav-assisted mobile edge computing network

Abstract

The present application belongs to the field of unmanned aerial vehicle assisted edge computing network resource allocation, and particularly relates to a task scheduling and order optimization method and system under an unmanned aerial vehicle assisted mobile edge computing network. On the basis of unmanned aerial vehicle assisted mobile edge computing, the present application acquires time costs of relaying and computing of different tasks according to the difference between the computing capacity of unmanned aerial vehicle nodes and ground nodes, and selects the optimal scheduling strategy. The present application solves the heterogeneity of different computing nodes, completes the scheduling of tasks among different nodes, realizes the reasonable utilization of resources, and improves the resource utilization rate of the system.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of unmanned aerial vehicle assisted edge computing network resource allocation, and particularly relates to a task scheduling and order optimization method and system in an unmanned aerial vehicle assisted mobile edge computing network. BACKGROUND

[0002] With the rapid development of network communication technology and the exponential growth of terminal access equipment, a large amount of data is generated, which brings great pressure to network and computing. With the popularization of digitization and intelligentization, the number of devices in various industries will maintain a high growth rate, and the global data volume will experience an even more explosive growth. How to effectively process these data will be a challenge to be solved in the future. At the same time, intelligent services such as autonomous driving, virtual reality, metaverse and smart city are rapidly iterating, which puts higher requirements on the performance and storage resources of terminal devices. In view of various terminal demands, the current mainstream computing mode is mobile cloud computing, which is extended from cloud computing services based on mobile terminal devices to mobile cloud computing services supporting various terminal devices. To solve the shortage of computing and storage resources of terminal devices, mobile cloud computing transmits high-resource-demand tasks to cloud computing centers to improve the ability of terminal devices to process tasks and reduce the energy consumption of terminal devices. This computing mode is widely used in current Internet of Things terminal devices, which alleviates the contradiction between massive data generation and limited processing capacity of intelligent terminal devices. However, with the further development of intelligent demand, the shortcomings of cloud computing services gradually appear. Since cloud computing needs to transmit data to the core network cloud server, it will bring a huge load to the wireless information transmission of the mobile network. When the network load is too large, it will cause high network delay. Limited by the spectrum resources of the wireless channel, the delay generated when data is transmitted is inevitable.

[0003] To cope with these challenges, mobile edge computing (MEC) is proposed as a new computing mode. As a way to expand processing power and storage to the edge of the network, the basic principle of MEC is to decentralize the core network cloud computing capability to the base station side close to the terminal, thereby reducing the communication distance between the computing center and the terminal demand side, improving resource utilization, reducing core network load and improving overall system performance. The resources of MEC deployed at the base station side are relatively fixed, while the resource demand of the terminal is dynamically changing. It is crucial to handle the contradiction between the two. Unmanned aerial vehicles (UAVs) have strong mobility, so they can use UAVs to carry MEC as an auxiliary of the MEC system to provide services for terminal devices. Compared with traditional static MEC systems, UAV-assisted MEC systems have high spatial mobility, fast response, and low use cost.

[0004] In the UAV-assisted MEC system, although the UAV-MEC enhances the ability of flexible allocation of resources in the spatial dimension, its performance still lags behind the existing ground MEC. However, current researches mostly focus on optimizing the energy consumption of UAV, and ignore the experience of users in system delay. There are two main ways about UAV-MEC task scheduling, which are computing node type and relay type. Computing node type: UAV-MEC uses its own resources to operate the terminal task, and finally returns the calculation result. This working mode is similar to the conventional ground MEC, but due to the relatively limited performance of UAV-MEC, there will be a large delay in processing large tasks, which cannot be applied to all use scenarios. Relay type: choose UAV as a relay node, UAV can use the transmission rate advantage of visual channel to relay the task to the MEC node with stronger performance, this working mode can effectively use the idle high-performance node, but the relay will have a certain delay cost. Both of the two task scheduling methods have their limitations and can only be applied to specific scenarios. At the same time, the existing task scheduling research only considers the scheduling of tasks between different computing nodes, and does not consider the reasonable scheduling of tasks within a single computing node. Different working methods of tasks will affect the processing efficiency of tasks, and then affect the resource utilization efficiency of the system. SUMMARY

[0005] The purpose of the present application is to overcome the above-mentioned deficiencies, provide a task scheduling and order optimization method and system under a UAV-assisted mobile edge computing network, which adjusts the scheduling of tasks between nodes, preferentially selects the scheduling strategy with the minimum processing cost, optimizes the execution order of each task within the node, and optimizes the two-stage task scheduling.

[0006] In order to achieve the above-mentioned purpose, the task scheduling and order optimization method under the UAV-assisted mobile edge computing network comprises the following steps:

[0007] Under the multi-task request, a mobile edge computing network model and a task processing cost evaluation model are established according to the relationship between UAV assistance and MEC system;

[0008] The task processing cost evaluation model respectively evaluates the scheduling cost of UAV-MEC as task relay and ground MEC as task relay, and determines the optimal scheduling scheme according to the scheduling cost;

[0009] The mobile edge computing network model analyzes the task processing process of any UAV-MEC server, establishes a task processing model based on parallel workflow, and executes data transmission and calculation in parallel as a new task scheduling decision according to the task processing model based on parallel workflow;

[0010] The task execution strategy is obtained based on the task scheduling decision, the time block of each stage is calculated, the processing order of tasks is optimized, and the task scheduling scheme is implemented according to the two-stage algorithm of task scheduling and order adjustment.

[0011] The hovering position of the UAV is (x u ,y u ,h u The UAV-MEC collects information on all pending tasks within its coverage area. The task descriptions are as follows:

[0012] Task i ={Coordinates i ,d i}

[0013] Among them, Coordinates i The coordinates of the device to which task i belongs are: d i This indicates the amount of data that task i needs to compute.

[0014] The cost of task i in the calculation is expressed as follows:

[0015]

[0016] Among them, F uav For the CPU cycle of UAV-MEC, F mec For the CPU cycle of MEC, d i Let I be the amount of data that task i needs to compute, and let I be the set of tasks to be processed.

[0017] Channel gain for communication transmission between terminal equipment and UAV-MEC for:

[0018]

[0019] Where β0 is the received power and d is the distance between the terminal device and the UAV-MEC;

[0020] Transmission rate between terminal equipment and UAV-MEC for:

[0021]

[0022] Where B is the channel bandwidth, p is the antenna transmit power of the terminal, and σ 2 This represents the noise power at the receiving antenna.

[0023] The cost of transmitting data from the terminal to the UAV-MEC for:

[0024]

[0025] Channel gain g of UAV-MEC and ground MEC for communication transmission m is:

[0026]

[0027] wherein β0 is the received power, and L is the distance between UAV-MEC and ground MEC;

[0028] Transmission rate r of UAV-MEC and ground MEC m is:

[0029]

[0030] wherein B is the channel bandwidth, p is the antenna transmission power of the terminal, and σ 2 is the noise power at the receiving antenna;

[0031] Cost of data transmission from UAV-MEC to other idle MECs is:

[0032]

[0033] wherein di i represents the data amount required for task i to be calculated.

[0034] The time cost required when UAV-MEC acts as a task relay is:

[0035]

[0036] wherein di is the ratio of the data amount required for task i to be calculated to the CPU cycle of UAV-MEC, is the ratio of the data amount required for task i to be calculated to the wireless transmission rate,

[0037] The time cost required when ground MEC acts as a task relay is:

[0038]

[0039] wherein di is the cost of transmission from UAV-MEC to other idle MECs, is the ratio of the data amount required for task i to be calculated to the CPU cycle of MEC, and I is the set of tasks to be processed;

[0040] The scheduling decision of task i is represented by a matrix:

[0041]

[0042] Where A i = [1, 0] when the ground MEC acts as task relay, A i = [0, 1] when the UAV-MEC acts as task relay, there is only one case for single task scheduling, with the constraint that:

[0043] A i A i T = 1

[0044] Therefore, the optimal scheduling scheme is represented as follows:

[0045]

[0046] According to the optimal task scheduling decision matrix A * , the time length required to execute each task is obtained:

[0047]

[0048] Since the transmission and computation of tasks are performed sequentially, we obtain:

[0049] s i + i = i ,∈I

[0050] where represents the start time vector of the two different working stages of the ith task in execution, represents the time length vector of the two different working stages of the ith task in execution, represents the end time vector of the two different working stages of the ith task in execution.

[0051] For single task, the data transmission is completed first, followed by the subsequent task processing procedure, with the constraint that:

[0052]

[0053] For multiple tasks, each stage of the task is executed after the corresponding stage of the previous task is completed, with the constraint that:

[0054]

[0055] The minimum task processing time of the UAV-MEC is the difference between the data transmission start time of the task scheduling and the end time of the task computation.

[0056] The specific method for optimizing the processing order of the task is as follows:

[0057] S41, list the required transmission time and the time block length required for completion of computation of all current tasks, data transmission is The required time is calculated as

[0058] S42, the time block min(t) with the shortest time is selected i , if the time block is a data transmission stage, the task i is arranged with the highest priority; if the time block belongs to a calculation stage, the task i is arranged with the lowest priority;

[0059] S43, the task i with the shortest time block is excluded from the set of all tasks;

[0060] S44, S42 and S43 are repeated until all tasks are arranged.

[0061] The task scheduling and order optimization system under the UAV-assisted mobile edge computing network comprises:

[0062] A model construction module is configured to, under multiple task requests, establish a mobile edge computing network model and a task processing cost evaluation model according to the relationship between UAV assistance and the MEC system;

[0063] A scheduling scheme construction module is configured to evaluate the scheduling cost of the UAV-MEC as a task relay and the ground MEC as a task relay respectively through the task processing cost evaluation model, and determine an optimal scheduling scheme according to the scheduling cost;

[0064] A task scheduling decision generation module is configured to analyze the task processing process of any UAV-MEC server through the mobile edge computing network model, establish a task processing model based on parallel workflow, and execute data transmission and calculation in parallel as a new task scheduling decision according to the task processing model based on parallel workflow;

[0065] An optimization module is configured to obtain a task execution strategy according to the task scheduling decision, calculate the time blocks of each stage, optimize the processing order of the tasks, and implement the task scheduling scheme according to the two-stage algorithm of task scheduling and order adjustment.

[0066] The optimization module comprises:

[0067] A time block calculation module is configured to list the required transmission time of all current tasks and the length of the time block required for completion of calculation, the data transmission being The required time is calculated as

[0068] A task priority arrangement module is configured to select the time block min(t) with the shortest time i , if the time block is a data transmission stage, the task i is arranged with the highest priority; if the time block belongs to a calculation stage, the task i is arranged with the lowest priority;

[0069] Shortest task elimination module, used for eliminating the task i with the shortest time block from all task sets.

[0070] Compared with the prior art, the unmanned aerial vehicle assisted mobile edge computing is used as the basis, the time cost of relaying and computing of different tasks is obtained according to the difference between the operation capacity of the unmanned aerial vehicle node and the ground node, and the optimal scheduling strategy is selected. The heterogeneity of different computing nodes is solved, the scheduling of tasks between different nodes is completed, the reasonable use of resources is realized, and the resource utilization rate of the system is improved.

[0071] Further, the node-in scheduling optimization is performed on the tasks that have been determined to be scheduled, the execution process of the tasks is considered, the data transmission and data calculation are considered separately, the two task execution stages are executed in parallel, and the processing rate of the tasks can be effectively improved. Through analysis of the parallel work flow, it is found that in this case, the computing node carried by the unmanned aerial vehicle will randomly select tasks to execute, but because the task order optimization is not performed, there is a certain time gap between the task data transmission and the task execution calculation. The execution order of the tasks can be adjusted, the time gap between the tasks is optimized, and the execution time of the tasks is reduced.

[0072] Further, the execution time cost of the tasks is evaluated, the difference between the computing capacities of the nodes is combined, the appropriate task scheduling strategy is selected, the scheduling strategy is determined, the processing flow of the tasks is analyzed, the parallel working mode is introduced to process the tasks, the execution efficiency of the tasks is improved, finally the processing time of the tasks is reduced through optimization of the execution order of the tasks, and the resource utilization rate of the unmanned aerial vehicle assisted mobile edge computing system is improved. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 The flowchart of the present application;

[0074] Figure 2 The system model architecture diagram of the present application;

[0075] Fig. 3(a) is a schematic diagram of task serial processing in the present application;

[0076] Fig. 3(b) is a schematic diagram of task execution through a parallel working mode in the present application;

[0077] Fig. 3(c) is a schematic diagram of task execution of a parallel working flow after optimization of the task order in the present application;

[0078] Figure 4 The system diagram of the present application. DETAILED DESCRIPTION

[0079] The present application will be further described below in combination with the drawings.

[0080] AsFigure 1 As shown, the task scheduling and order optimization method under UAV-assisted mobile edge computing network specifically includes the following steps:

[0081] Step 1: Establish a UAV-assisted MEC system model for multi-task scenarios, including a network model and a task processing cost evaluation model.

[0082] Figure 2 The image shows a UAV-assisted MEC system model in a multi-tasking scenario. The UAV carries an MEC server to provide services to the terminal devices, and the hovering position is (x u ,y u ,h u The UAV-MEC collects information on all pending tasks within its coverage area. The task descriptions are as follows:

[0083] Task i ={Coordinates i ,d i}

[0084] Among them, Coordinates i The coordinates of the device to which task i belongs are: d i Let F represent the amount of data that task i needs to compute. There are n tasks waiting to be processed, and I is the set of these tasks. The processing time of a computation task in the MEC service is related to the CPU cycles required for the computation. At a given CPU frequency, the more cycles a computation task requires, the more complex the task and the longer it takes. Define F. uav F represents the CPU cycle of the UAV-MEC. mec This represents the CPU cycle of MEC. The computational cost of task i is expressed as follows:

[0085]

[0086] Due to the excellent mobility of UAVs, UAV-MECs can obtain line-of-sight communication channels when communicating. To highlight this feature of UAV-MECs, this paper only considers large-scale fading of the channel. Let B be the channel bandwidth, p be the terminal's antenna transmit power, and β0 be the receive power σ. 2 Let the noise power at the receiving antenna be the channel gain for communication transmission between the terminal device and the UAV-MEC, then:

[0087]

[0088] in, This refers to the distance between the terminal device and the UAV-MEC.

[0089] Thus, according to Shannon formula, the transmission rate of terminal device to UAV-MEC can be represented by the following formula:

[0090]

[0091] The cost of data transmission from terminal to UAV-MEC is:

[0092]

[0093] Similarly, assuming the coordinates of the remote idle ground MEC are (L, 0, 0), the channel gain of the transmission between UAV-MEC and ground MEC is represented as

[0094]

[0095] The transmission rate of UAV-MEC to ground MEC can be represented by the following formula:

[0096]

[0097] The cost of data transmission from UAV-MEC to other idle MEC is:

[0098]

[0099] Step 2, task scheduling decision scheme:

[0100] The computing capability of ground MEC is much stronger than that of UAV-MEC, but the time for ground MEC to obtain task data is much longer. Therefore, the main problem of task scheduling is to consider where to execute the task, i.e., to execute in UAV-MEC or in ground MEC, to minimize the task completion time.

[0101] There are two main scheduling methods. The time cost required for task i to execute in UAV-MEC can be represented as:

[0102]

[0103] The time cost required for task i to execute in remote MEC through UAV-MEC relay can be represented as:

[0104]

[0105] The scheduling decision of task i is represented by a matrix:

[0106]

[0107] wherein, when task i is calculated in UAV-MEC, A i = [1, 0], when task i is relayed in UAV-MEC, Ai = [0, 1]. There is only one case for single task scheduling, and the constraint is:

[0108] A i A i T = 1

[0109] By the above analysis, we transform the problem into the optimal task scheduling scheme. The minimum task completion time cost problem proposed in this patent can be expressed as follows:

[0110]

[0111] Step 3 Task Order Optimization Algorithm:

[0112] As shown in Figure 3(a), in the traditional UAV-assisted MEC system, when processing each task, it is usually executed in series. Each device can only use the resources of UAV-MEC after the previous device's task processing is completed, which will greatly increase the time for MEC system to process tasks. To further improve efficiency, data transmission and task calculation are divided, allowing data transmission and calculation to be executed in parallel, thereby improving the efficiency of MEC.

[0113] Analyzing the process of the mth UAV-MEC server task processing, a task processing model based on parallel workflow is given, as shown in Figure 3(b). Executing data transmission and calculation in parallel can effectively improve the efficiency of the system.

[0114] Figure 3 illustrates a task scheduling model without execution order optimization. In this case, MEC will randomly select tasks for execution, but due to the lack of task order optimization, there is a certain time gap between task data transmission and task execution calculation. As shown in Figure 3(c), after order optimization, the execution time of the task can be greatly reduced.

[0115] According to the task scheduling principle of the previous stage, the optimal task scheduling decision matrix A * is obtained. The length of time required to execute each task

[0116]

[0117] N tasks are in the UAV-MEC server waiting queue, waiting for UAV-MEC to process or relay to other nodes. S represents the execution order of the task. Task I i is calculated on the UAV-MEC server

[0118] Let denote the starting time vector of the two different working stages of the ith task in execution. Let denotes the vector of the length of time duration of the two different working stages of the ith task in execution. Let denotes the vector of the time of the end of the two different working stages of the ith task in execution.

[0119] Because the transmission and the calculation of the task are carried out in sequence, so:

[0120] s i + i = i , ∈ I

[0121] In a single task, the data transmission is carried out first, and the subsequent task processing process can be carried out only after the data transmission is completed. These constraints are:

[0122]

[0123] For multiple tasks, each task is also executed in a serial order, and each stage of the task must be executed after the completion of the corresponding stage of the previous task. The constraints are:

[0124]

[0125] According to the figure, the minimum task processing time of the UAV-MEC is the difference between the data transmission start time of the task scheduling and the end time of the task calculation. Because the data transmission start time is Therefore

[0126]

[0127] Therefore, the objective function is:

[0128]

[0129] The objective function as shown in the above formula is to solve the shortest time for the UAV-MEC to process the task under the energy depletion condition under the task flow sequence constraint condition. According to the above parallel working flow task processing model, the optimization order problem is a combination optimization problem. The number of times to be tried for solving the task execution order by using the exhaustion method is at least N!, and the time complexity for solving the optimal task order is O(N!). When the number of tasks to be processed is small, the exhaustion method can be used, but when the number of tasks is large, the time spent is too much, and the efficiency is very low. Analysis of the problem shows that the parallel task order scheduling can be transformed into a pipeline scheduling problem, so Johnson's algorithm is introduced to solve the execution order of the task.

[0130] The main optimization process is as follows:

[0131] (31) List the required transmission time and the length of the time block required for completing the calculation of all the tasks at present, the data transmission is The required time is calculated as

[0132] (32) The time block with the shortest time min(t) is selected. i If the time block is a data transmission stage, the task i is arranged with the highest priority. If the time block belongs to a calculation stage, the task i is arranged with the lowest priority.

[0133] (33) The task i with the shortest time block min(t) is excluded from the set of all tasks.

[0134] (34) Steps (32) and (33) are repeated until all tasks are arranged.

[0135] The parallel workflow execution is performed according to the task priority from high to low.

[0136] Referring to Figure 4 , the unmanned aerial vehicle (UAV) assisted mobile edge computing (MEC) network task scheduling and order optimization system comprises:

[0137] A model construction module is configured to, under a plurality of task requests, establish a mobile edge computing network model and a task processing cost evaluation model according to the relationship between the UAV assistance and the MEC system.

[0138] A scheduling scheme construction module is configured to evaluate the scheduling cost of the UAV-MEC as a task relay and the ground MEC as a task relay respectively through the task processing cost evaluation model, and determine an optimal scheduling scheme according to the scheduling cost.

[0139] A task scheduling decision generation module is configured to analyze the task processing process of any UAV-MEC server through the mobile edge computing network model, establish a task processing model based on parallel workflow, and execute data transmission and calculation in parallel as a new task scheduling decision according to the task processing model based on parallel workflow.

[0140] An optimization module is configured to obtain a task execution strategy according to the task scheduling decision, calculate the time blocks of each stage, optimize the processing order of the tasks, and implement the task scheduling scheme according to the two-stage algorithm of task scheduling and order adjustment.

[0141] The optimization module comprises:

[0142] A time block calculation module is configured to list the required transmission time of all current tasks and the length of the time block required for completion of calculation, with the data transmission being The required time is calculated as

[0143] A task priority arrangement module is configured to select the time block with the shortest time min(t). i) if the time block is a data transmission phase, then the task i is scheduled with the highest priority; if the time block belongs to a computation phase, then the task i is scheduled with the lowest priority;

[0144] a shortest task elimination module for eliminating the task i with the shortest time block from the set of all tasks.

[0145] The present application is not limited to the above-mentioned embodiments, and based on the technical solutions disclosed in the present application, those skilled in the art can make some substitutions and modifications to some technical features without creative labor, and these substitutions and modifications are all within the protection scope of the present application.

Claims

1. A method for task scheduling and order optimization in a UAV-assisted mobile edge computing network, characterized in that, Includes the following steps: Under multi-task requests, a network model and task processing cost evaluation model for mobile edge computing are established based on the relationship between UAV assistance and MEC system. The task processing cost assessment model evaluates the scheduling costs of UAV-MEC as a task relay and ground MEC as a task relay, respectively, and determines the optimal scheduling scheme based on the scheduling costs. The time cost required when UAV-MEC is used as a task relay is: in, For the task The ratio of the amount of data to be computed to the CPU cycles of the UAV-MEC. For the task The ratio of the amount of data to be calculated to the wireless transmission rate. When a ground-based MEC acts as a mission relay, the required time cost is: in, The cost of transferring data from a UAV-MEC to other idle MECs, For the task The ratio of the amount of data to be computed to the CPU cycles of the MEC. A collection of tasks to be processed; Task The scheduling decision is represented by a matrix: When the ground-based MEC acts as a mission relay... When UAV-MEC is used as a mission relay, A single task can only be scheduled in one scenario, subject to the following constraints: Therefore, the optimal scheduling scheme is expressed as follows: Based on the optimal task scheduling decision matrix This gives the time required to execute each task: in, For the task The amount of data to be calculated For the transmission rate between the terminal device and the UAV-MEC, CPU cycles for UAV-MEC; Since the transmission and computation of tasks are performed sequentially, we obtain: in, Indicates the first The start time vector of two different working phases in the execution of a task. Indicates the first The vector of durations of two different work phases in the execution of a task. Indicates the first The time vector of the end of two different working phases in the execution of a task; For a single task, data transmission is completed first, followed by subsequent task processing. The constraint is as follows: For multi-tasking, each stage of a task is executed after the corresponding stage of the previous task is completed, with the following constraints: The minimum task processing time of UAV-MEC is the difference between the data transmission start time of task scheduling and the task computation end time. The network model analysis of mobile edge computing analyzes the task processing process of any UAV-MEC server, establishes a task processing model based on parallel workflow, and executes data transmission and computation in parallel according to the task processing model based on parallel workflow, as a new task scheduling decision. The task execution strategy is obtained based on the task scheduling decision, the time block of each stage is calculated, the processing order of tasks is optimized, and the task scheduling scheme is implemented according to the two-stage algorithm of task scheduling and order adjustment.

2. The task scheduling and order optimization method under UAV-assisted mobile edge computing network according to claim 1, characterized in that, The hovering position of the UAV is The UAV-MEC collects information on all pending tasks within its coverage area. The task description is as follows: in, Indicates task The coordinates of the device are , Indicates task The amount of data that needs to be calculated.

3. The task scheduling and order optimization method under UAV-assisted mobile edge computing network according to claim 2, characterized in that, Task The cost of each task is expressed as follows: in, For UAV-MEC CPU cycles, For the CPU cycle of MEC, For the task The amount of data to be calculated This is a collection of tasks to be processed.

4. The task scheduling and order optimization method under UAV-assisted mobile edge computing network according to claim 3, characterized in that, Channel gain for communication transmission between terminal equipment and UAV-MEC for: in, For received power, The distance between the terminal device and the UAV-MEC; Transmission rate between terminal equipment and UAV-MEC for: in, For channel bandwidth, This refers to the antenna transmit power of the terminal. This represents the noise power at the receiving antenna. The cost of transmitting data from the terminal to the UAV-MEC for: 。 5. The task scheduling and order optimization method under UAV-assisted mobile edge computing network according to claim 3, characterized in that, Channel gain for communication transmission between UAV-MEC and terrestrial MEC for: in, For received power, This refers to the distance between the UAV-MEC and the ground-based MEC. Transmission rate between UAV-MEC and terrestrial MEC for: in, For channel bandwidth, This refers to the antenna transmit power of the terminal. This represents the noise power at the receiving antenna. The cost of transferring data from a UAV-MEC to other idle MECs for: in, Indicates task The amount of data that needs to be calculated.

6. The task scheduling and order optimization method under UAV-assisted mobile edge computing network according to claim 1, characterized in that, The specific methods for optimizing the task processing order are as follows: S41 lists the required transmission time for all current tasks and the time block length required to complete the computation. Data transmission is... Calculate the required time as ; S42, select the time block with the shortest time. If the time block is a data transmission phase, then schedule this task. This task is scheduled with the highest priority; if the time block belongs to the computation phase, then this task is scheduled. The priority is the lowest. S43, exclude the task with the shortest time block from the entire task set. ; S44, repeat S42 and S43 until all tasks are scheduled.

7. A task scheduling and order optimization system for UAV-assisted mobile edge computing networks, based on the task scheduling and order optimization method for UAV-assisted mobile edge computing networks as described in claim 1, characterized in that... include: The model building module is used to establish a network model and a task processing cost evaluation model for mobile edge computing based on the relationship between UAV assistance and MEC system under multi-task requests. The scheduling scheme construction module is used to evaluate the scheduling costs of UAV-MEC as a task relay and ground MEC as a task relay respectively through the task processing cost evaluation model, and determine the optimal scheduling scheme based on the scheduling costs. The task scheduling decision generation module is used to analyze the task processing process of any UAV-MEC server through the network model of mobile edge computing, establish a task processing model based on parallel workflow, and execute data transmission and computation in parallel according to the task processing model based on parallel workflow, as a new task scheduling decision. The optimization module is used to obtain task execution strategies based on task scheduling decisions, calculate time blocks for each stage, optimize the processing order of tasks, and implement the task scheduling scheme according to the two-stage algorithm of task scheduling and order adjustment.

8. The task scheduling and order optimization system under the UAV-assisted mobile edge computing network according to claim 7, characterized in that, The optimization modules include: The time block calculation module lists the required transmission time for all current tasks and the length of the time block needed to complete the calculation. Data transmission is... Calculate the required time as ; The task priority scheduling module is used to select the shortest time block. If the time block is a data transmission phase, then schedule this task. This task is scheduled with the highest priority; if the time block belongs to the computation phase, then this task is scheduled. The priority is the lowest. The shortest task exclusion module is used to exclude the task with the shortest time block from the entire task set. .

Images