Resource scheduling method and device for edge computing

Abstract

The present invention provides a resource scheduling method and device for edge computing, which relates to the field of wireless communication technology. The method includes: constructing a target resource scheduling model, determining optimal scheduling parameters based on the target resource scheduling model; and scheduling multiple tasks to be scheduled based on the optimal scheduling parameters; wherein the target resource scheduling model is constructed based on the task completion delay, the central server load, and the loads of multiple edge servers; the tasks to be scheduled include disassembled tasks and non-disassembled tasks. The resource scheduling method and device for edge computing provided by the present invention construct a target resource scheduling model by considering the task completion delay, the central server load, and the loads of multiple edge servers, and determine the optimal scheduling parameters based on the target resource scheduling model to achieve a scheduling method that minimizes the user service completion delay, minimizes the central server load, and balances the edge server load.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a resource scheduling method and device for edge computing. Background Art

[0002] The existing edge computing technology is to place the edge server at the edge of the network and establish a direct link with the user. The user can offload the computing tasks directly to the edge server for calculation. After the calculation is completed, the server will return the calculation results.

[0003] Existing edge computing technologies, when running on multiple edge servers, primarily consider the total energy consumption and latency of completing user tasks, without considering server load balancing and fairness. This can result in excessive pressure on a single edge server, while the remaining servers experience minimal computing pressure. Furthermore, when central servers are present, these approaches focus solely on edge server scheduling, without considering how to alleviate pressure on central servers through edge servers.

[0004] Therefore, it is of great significance to propose a method to minimize the user service completion delay, ensure the fairness of edge server resource scheduling, and alleviate the service pressure of the central server as much as possible for edge computing scenarios with one central server, multiple edge servers, and multiple tasks. Summary of the Invention

[0005] The present invention provides a resource scheduling method and device for edge computing, which is used to solve the technical problem in the prior art that it is impossible to ensure the minimization of service completion delay and the fairness of edge server resource scheduling in edge computing while alleviating the service pressure of the central server.

[0006] In a first aspect, the present invention provides a resource scheduling method for edge computing, comprising:

[0007] Constructing a target resource scheduling model, and determining optimal scheduling parameters according to the target resource scheduling model;

[0008] According to the optimal scheduling parameters, the scheduling of multiple tasks to be scheduled is realized;

[0009] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0010] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0011] In one embodiment, the building of the target resource scheduling model includes:

[0012] The target resource scheduling model is constructed based on the amount of tasks to be scheduled, the computing power required to calculate the amount of tasks, the load of the central server, the computing power remaining after the edge server tasks are assigned, and the average delay in task completion.

[0013] In one embodiment, when the task to be scheduled is a decomposable task, the target resource scheduling model is:

[0014]

[0015]

[0016]

[0017]

[0018] Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, The average delay factor for task completion.

[0019] In one embodiment, when the task to be scheduled is an indivisible task, the target resource scheduling model is:

[0020]

[0021]

[0022]

[0023]

[0024] Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, The average delay factor for task completion.

[0025] In one embodiment, a slack non-negative variable β is introduced to transform the target resource scheduling model into:

[0026]

[0027]

[0028]

[0029]

[0030]

[0031] β≥0.

[0032] In one embodiment, a slack non-negative variable β is introduced to transform the target resource scheduling model into:

[0033]

[0034]

[0035]

[0036]

[0037]

[0038] β≥0.

[0039] In one embodiment, determining the optimal scheduling parameters according to the target resource scheduling model includes:

[0040] Generate α according to genetic algorithm k,n , in determining α k,n After the value is obtained, optimize β and f k,n The optimal scheduling parameters of the target resource scheduling model are determined by performing iterative calculation according to the preset number of iterations and convergence conditions.

[0041] In a second aspect, the present invention further provides a resource scheduling device for edge computing, comprising:

[0042] A model building module is used to build a target resource scheduling model and determine optimal scheduling parameters according to the target resource scheduling model;

[0043] A task scheduling module, configured to schedule multiple tasks to be scheduled according to the optimal scheduling parameters;

[0044] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0045] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0046] In a third aspect, the present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of any one of the above-mentioned edge computing resource scheduling methods are implemented.

[0047] In a fourth aspect, the present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of any of the above-mentioned edge computing resource scheduling methods are implemented.

[0048] The resource scheduling method, device, electronic device, and storage medium for edge computing provided by the present invention construct a target resource scheduling model by considering task completion delay, central server load, and multiple edge server loads. Based on the target resource scheduling model, optimal scheduling parameters are determined to achieve a scheduling method that minimizes user service completion delay, minimizes central server load, and balances edge server load. While ensuring minimal service completion delay, the load of edge services is balanced as much as possible and the load of central servers is reduced, thereby alleviating central server service pressure and ensuring fairness in edge server resource scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 A flow chart of the resource scheduling method for edge computing provided by the present invention;

[0051] Figure 2 A schematic diagram of the framework of the edge computing system provided by the present invention;

[0052] Figure 3 A schematic diagram of the two-stage optimization solution process provided by the present invention;

[0053] Figure 4 This is a schematic diagram of the task offloading amount of each server under different task completion delay weights in the task decomposable scenario provided by the present invention;

[0054] Figure 5 A schematic diagram of the task offloading amount of each server under different task completion delay weights in the scenario where tasks cannot be disassembled provided by the present invention;

[0055] Figure 6 A schematic diagram showing the relationship between the average service completion delay factor and the task completion delay weight provided by the present invention;

[0056] Figure 7 A schematic diagram of the structure of the resource scheduling device for edge computing provided by the present invention;

[0057] Figure 8 This is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0058] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0059] Figure 1 This is a flow chart of the resource scheduling method for edge computing provided by the present invention. Figure 1 , the resource scheduling method for edge computing provided by the present invention may include:

[0060] S110, constructing a target resource scheduling model, and determining optimal scheduling parameters according to the target resource scheduling model;

[0061] S120, scheduling multiple tasks to be scheduled according to the optimal scheduling parameters;

[0062] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0063] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0064] The execution subject of the resource scheduling method for edge computing provided by the present invention can be an electronic device, a component in an electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. For example, the mobile electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle electronic device, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc., and the non-mobile electronic device can be a server, a network attached storage (NAS), a personal computer (PC), a television (TV), an ATM or a self-service machine, etc., which is not specifically limited by the present invention.

[0065] The following takes the resource scheduling method for edge computing provided by the present invention executed by a computer as an example to describe the technical solution of the present invention in detail.

[0066] It should be noted that the central server can be a central cloud server. An edge server is a server located at the edge of the network, establishing a direct link with users. Users can offload their computing tasks directly to the edge server for computation, which then returns the results. Compared to cloud server computing, which requires users to upload and download data across multiple layers of the internet, edge computing requires significantly less latency.

[0067] In step S110 , a target resource scheduling model is constructed, and optimal scheduling parameters are determined according to the target resource scheduling model.

[0068] Specifically, if Figure 2The edge computing system provided by the present invention is shown in the framework diagram, which consists of a central server and multiple edge servers. When multiple tasks arrive, it is necessary to determine how to allocate these multiple tasks to one central server and multiple edge servers, and how to allocate and process them in a more appropriate manner. After determining the requirements, a target resource scheduling model is constructed, and the optimal scheduling parameters are determined to implement task scheduling and allocation.

[0069] Optionally, a target resource scheduling model is constructed according to requirements, where the requirements may be one or more of minimum task completion delay, minimum central server load, and load balancing of multiple edge servers.

[0070] In step S120 , optimal scheduling parameters are determined according to the target resource scheduling model in step S110 , and scheduling of multiple tasks to be scheduled is achieved according to the optimal scheduling parameters.

[0071] Specifically, after the optimal scheduling parameters are determined, multiple tasks to be scheduled can be allocated according to the optimal scheduling parameters and allocated to the central server and multiple edge servers for execution.

[0072] It is understandable that, according to the determined optimal scheduling parameters, multiple tasks to be scheduled can be processed by the central server, can be processed by multiple edge servers, or can be assigned to the central server and multiple edge servers for joint processing.

[0073] The target resource scheduling model is constructed based on the task completion delay, the central server load and the loads of multiple edge servers; the tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0074] Optionally, when constructing a target resource scheduling model based on task completion delay, central server load, and multiple edge server loads, a target resource scheduling model with minimum task completion delay, minimum central server load, and balanced load on multiple edge servers can be determined.

[0075] It's understandable that building a target resource scheduling model based on the principles of minimizing task completion time and central server load can alleviate the workload on central servers while minimizing task completion time. Furthermore, considering load balancing across multiple edge servers when building the target resource scheduling model ensures fairness in edge server resource scheduling.

[0076] Scheduled tasks can be divided into two types: splittable tasks and non-splittable tasks. A splittable task is a single task that can be split into multiple parts, which can then be scheduled and assigned to different servers for execution. Non-splittable tasks are single tasks that cannot be split into multiple parts and can only be assigned to a single server for execution.

[0077] The edge computing resource scheduling method provided by the present invention constructs a target resource scheduling model by considering task completion latency, central server load, and the loads of multiple edge servers. Based on this target resource scheduling model, optimal scheduling parameters are determined to achieve a scheduling method that minimizes user service completion latency, minimizes central server load, and balances edge server load. While ensuring minimal service completion latency, the load of edge services is balanced as much as possible and the load of central servers is reduced, thereby alleviating service pressure on central servers and ensuring fairness in resource scheduling for edge servers.

[0078] In one embodiment, a target resource scheduling model is constructed, including: constructing the target resource scheduling model based on the task volume of the task to be scheduled, the computing power required to calculate the task volume, the load of the central server, the computing power remaining after the edge server task is allocated, and the average delay in task completion.

[0079] Optionally, an objective function can be constructed based on the amount of tasks to be scheduled, the computing power required to calculate the amount of tasks, the load of the central server, the computing power remaining after the edge server tasks are allocated, and the average delay in task completion. Constraints can be set based on minimizing the user service completion delay, minimizing the central server load, and balancing the edge server load to determine the optimal scheduling parameters as the target resource scheduling model.

[0080] The resource scheduling method for edge computing provided by the present invention constructs a target resource scheduling model by utilizing parameters such as the task volume of the task to be scheduled and the computing power required to calculate the task volume, and determines the optimal scheduling parameters based on the target resource scheduling model to achieve a scheduling method that minimizes the user service completion delay, minimizes the central server load, and balances the edge server load.

[0081] In one embodiment, when the task to be scheduled is a decomposable task, the target resource scheduling model is:

[0082]

[0083]

[0084]

[0085]

[0086] Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, is the average delay measurement factor for task completion.

[0087] Specifically, by constructing a target resource scheduling model with an objective function, the optimal scheduling parameters can be determined, thereby achieving a scheduling method that minimizes service completion delay, minimizes central server load, and balances edge server load. The objective function means that by optimizing α k,n With f k,n Maximize the minimum remaining computing power of each edge server, that is, ensure the load balance of the server. By adjusting the load pressure weight of the central server ω c and the task completion delay weight ω t The objective function can adjust the degree of emphasis on the central server load pressure and task delay.

[0088] Among them, the constraints in the objective function For computing power constraints, The delay calculated for the task must be less than the required delay of the task. Assign factor constraints to tasks, that is, tasks must be distributed completely.

[0089] It is understandable that the task is a splittable task, that is, a single task can be split into multiple parts and assigned to the server for execution. k,n ∈[0,1] represents the proportion of servers sharing task k, and its value can be a decimal between 0 and 1. For example, α k,n =0.5 means that server n can share half of task k.

[0090] The resource scheduling method for edge computing provided by the present invention constructs a target resource scheduling model of the objective function according to the decomposable characteristics of the task when the task to be scheduled is decomposable, and determines the optimal scheduling parameters according to the target resource scheduling model to achieve a scheduling method that minimizes the completion delay of the user's decomposable service, minimizes the central server load, and balances the edge server load.

[0091] In one embodiment, when the task to be scheduled is an indivisible task, the target resource scheduling model is:

[0092]

[0093]

[0094]

[0095]

[0096] Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, is the average delay measurement factor for task completion.

[0097] Specifically, by constructing a target resource scheduling model with an objective function, the optimal scheduling parameters can be determined, thereby achieving a scheduling method that minimizes service completion delay, minimizes central server load, and balances edge server load. The objective function means that by optimizing α k,n With f k,n Maximize the minimum remaining computing power of each edge server, that is, ensure the load balance of the server. By adjusting the load pressure weight of the central server ω c and the task completion delay weight ω tThe objective function can adjust the degree of emphasis on the central server load pressure and task delay.

[0098] Among them, the constraints in the objective function For computing power constraints, The delay calculated for the task must be less than the required delay of the task. Assign factor constraints to tasks, that is, tasks must be distributed completely.

[0099] It is understandable that the task is an indivisible task, that is, a single task cannot be split into multiple parts and assigned to servers for execution, that is, a single task can only be completed by one server. k,n ∈{0,1} represents the proportion of servers sharing task k, and its value can only be 0 or 1. That is, server n performs task k or does not perform task k.

[0100] The resource scheduling method for edge computing provided by the present invention constructs a target resource scheduling model of the objective function based on the non-disassembly characteristics of the task when the task to be scheduled cannot be disassembled, and determines the optimal scheduling parameters based on the target resource scheduling model to achieve a scheduling method that minimizes the completion delay of the user's disassemblyable business, minimizes the central server load, and balances the edge server load.

[0101] In one embodiment, a slack non-negative variable β is introduced to transform the target resource scheduling model into:

[0102]

[0103]

[0104]

[0105]

[0106]

[0107] β≥0

[0108] Specifically, since the target resource scheduling model under the condition that the determined tasks can be decomposed is a non-convex problem, it can be transformed into a convex problem for solution. Introducing the slack non-negative variable β, replacing the Introducing constraints and β>0; becomes At the same time, ω t Change the minus sign before the .

[0109] The specific proof process is: Can be equivalent to Can be equivalent to and Minimum equivalent to Max. However is a non-concave function, so we need to use A concave function with the same monotonicity as the function for The perspective function of , so it is a concave function, and along with increases with the increase of have the same monotonicity. is α k,n increasing function, but due to the constraint Therefore, it will mainly be achieved by increasing To increase Based on this, the replaced problem is equivalent to the original problem and has the same addition and subtraction properties as the original problem.

[0110] The resource scheduling method for edge computing provided by the present invention, when the task to be scheduled is decomposable, after determining the target resource scheduling model, converts the target resource scheduling model of the non-convex problem into a convex problem equivalent to the original problem, thereby facilitating the solution of the optimal scheduling parameters.

[0111] In one embodiment, a slack non-negative variable β is introduced to transform the target resource scheduling model into:

[0112]

[0113]

[0114]

[0115]

[0116]

[0117] β≥0

[0118] Specifically, when the task to be scheduled is an indecomposable task, since the target resource scheduling model is a non-convex problem, a slack non-negative variable β is introduced to transform the target resource scheduling model into a convex problem for solution.

[0119] The resource scheduling method for edge computing provided by the present invention, when the task to be scheduled cannot be disassembled, after determining the target resource scheduling model, converts the target resource scheduling model of the non-convex problem into a convex problem equivalent to the original problem, thereby facilitating the solution of the optimal scheduling parameters.

[0120] In one embodiment, determining the optimal scheduling parameters according to the target resource scheduling model includes: generating α according to a genetic algorithmk,n , in determining α k,n After the value is obtained, optimize β and f k,n The optimal scheduling parameters of the target resource scheduling model are determined by performing iterative calculation according to the preset number of iterations and convergence conditions.

[0121] Specifically, a two-stage algorithm can be used, such as Figure 3 As shown in the two-stage optimization solution process diagram provided by the present invention, the maximum number of iterations and the algorithm convergence condition are preset first. Then, in the first stage, α is generated by the genetic algorithm. k,n The second stage is based on the α generated in the first stage k,n value, use interior point method or subgradient method to optimize β and f k,n , obtain and record the objective function value; repeat the above two stages until the output parameters converge or the number of iterations reaches the maximum value, and output the optimal scheduling parameters.

[0122] Optionally, in determining α k,n , β and f k,n After the value of , the correctness and effectiveness of the target resource scheduling model can be verified by setting simulation parameters. The simulation parameters used are as follows: N = 4, K = 5, r k =10000bit, C k =1000, τ k = 1 second. The simulation results are as follows Figure 4 The present invention provides a schematic diagram of the task offloading amount of each server under different task completion delay weights in the task decomposable scenario. Figure 5 The task unloading amount of each server under different task completion delay weights in the scenario where the task cannot be disassembled provided by the present invention is shown in the schematic diagram.

[0123] in, Figure 4 and Figure 5 The edge server task completion delay weight ω under the conditions of task decomposability and non-decomposability c Perform simulation tests. Since the delay factor and the load value have different magnitudes, the magnitudes of their weights are also different. For the convenience of display, Figure 4 The vertical axis is a logarithmic scale. Figure 5 is the natural scale. It can be seen that as ω c In both scenarios, the scheduling resources allocated to the central server gradually decrease, the amount of scheduling resources allocated to the edge server gradually increases, and the amount of data unloaded to each edge server task is basically the same, that is, the edge server can achieve load balancing.

[0124] like Figure 6 The relationship between the average service completion delay factor and the task completion delay weight provided by the present invention is shown in the schematic diagram. It can be seen that as the task completion delay weight ω tWith the increase of , the average delay factor in both the decomposable and non-decomposable task scenarios decreases, which verifies that the target resource scheduling model can reduce the task completion delay by adjusting the task completion delay weight, thereby minimizing the user completion delay.

[0125] The resource scheduling method for edge computing provided by the present invention generates α through genetic algorithm k,n , in determining α k,n After the value is obtained, optimize β and f k,n The optimal scheduling parameters of the target resource scheduling model are determined by iterative calculation based on the preset number of iterations and convergence conditions. Simulation parameters are set and the target resource scheduling model is verified to ensure the correctness and effectiveness of the model settings.

[0126] The present invention also provides a resource scheduling device for edge computing, which can correspond to the resource scheduling method for edge computing described above.

[0127] Figure 7 A schematic diagram of the structure of the resource scheduling device for edge computing provided by the present invention is shown as follows: Figure 7 As shown, the device includes:

[0128] A model building module 710 is used to build a target resource scheduling model and determine optimal scheduling parameters according to the target resource scheduling model;

[0129] A task scheduling module 720 is configured to schedule multiple tasks to be scheduled according to the optimal scheduling parameters;

[0130] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0131] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0132] The edge computing resource scheduling device provided by the present invention constructs a target resource scheduling model by considering task completion delay, central server load, and the loads of multiple edge servers. Based on this target resource scheduling model, it determines optimal scheduling parameters to achieve a scheduling method that minimizes user service completion delay, minimizes central server load, and balances edge server load. While ensuring minimal service completion delay, it balances the load of edge services and reduces the load of central servers as much as possible, thereby alleviating service pressure on central servers and ensuring fairness in resource scheduling for edge servers.

[0133] In one embodiment, the model building module 710 is specifically configured to:

[0134] The target resource scheduling model is constructed based on the amount of tasks to be scheduled, the computing power required to calculate the amount of tasks, the load of the central server, the computing power remaining after the edge server tasks are assigned, and the average delay in task completion.

[0135] In one embodiment, the model building module 710 is further configured to:

[0136] When the task to be scheduled is a decomposable task, the target resource scheduling model is:

[0137]

[0138]

[0139]

[0140]

[0141] Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, The average delay factor for task completion.

[0142] In one embodiment, the model building module 710 is further configured to:

[0143] When the task to be scheduled is an indivisible task, the target resource scheduling model is:

[0144]

[0145]

[0146]

[0147]

[0148] Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, The average delay factor for task completion.

[0149] In one embodiment, the model building module 710 is further configured to:

[0150] Introducing the slack non-negative variable β, the target resource scheduling model is transformed into:

[0151]

[0152]

[0153]

[0154]

[0155]

[0156] β≥0.

[0157] In one embodiment, the model building module 710 is further configured to:

[0158] Introducing the slack non-negative variable β, the target resource scheduling model is transformed into:

[0159]

[0160]

[0161]

[0162]

[0163]

[0164] β≥0.

[0165] In one embodiment, the model building module 710 is further configured to:

[0166] Generate α according to genetic algorithm k,n , in determining α k,n After the value is obtained, optimize β and f k,n The optimal scheduling parameters of the target resource scheduling model are determined by performing iterative calculation according to the preset number of iterations and convergence conditions.

[0167] The present invention also provides an electronic device, such as Figure 8 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the steps of the resource scheduling method for edge computing, for example, including:

[0168] Constructing a target resource scheduling model, and determining optimal scheduling parameters according to the target resource scheduling model;

[0169] According to the optimal scheduling parameters, the scheduling of multiple tasks to be scheduled is realized;

[0170] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0171] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0172] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0173] On the other hand, the present invention further provides a computer program product, comprising a computer program stored on a non-transitory computer-readable storage medium, wherein the computer program comprises program instructions. When the program instructions are executed by a computer, the computer can perform the steps of the resource scheduling method for edge computing provided in the above-mentioned method embodiments, for example, including:

[0174] Constructing a target resource scheduling model, and determining optimal scheduling parameters according to the target resource scheduling model;

[0175] According to the optimal scheduling parameters, the scheduling of multiple tasks to be scheduled is realized;

[0176] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0177] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0178] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the resource scheduling method for edge computing provided in the above-mentioned method embodiments are implemented, for example, including:

[0179] Constructing a target resource scheduling model, and determining optimal scheduling parameters according to the target resource scheduling model;

[0180] According to the optimal scheduling parameters, the scheduling of multiple tasks to be scheduled is realized;

[0181] The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers;

[0182] The tasks to be scheduled include disassembled tasks and non-disassembled tasks.

[0183] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0184] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0185] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A resource scheduling method for edge computing, characterized in that: include: Constructing a target resource scheduling model, and determining optimal scheduling parameters according to the target resource scheduling model; According to the optimal scheduling parameters, the scheduling of multiple tasks to be scheduled is realized; The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers; The tasks to be scheduled include disassembled tasks and non-disassembled tasks; When the task to be scheduled is a decomposable task, the target resource scheduling model is: Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, is the average delay measurement factor for task completion; When the task to be scheduled is an indivisible task, the target resource scheduling model is: Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, is the average delay measurement factor for task completion.

2. The resource scheduling method for edge computing according to claim 1, characterized in that: The target resource scheduling model is constructed, including: The target resource scheduling model is constructed based on the amount of tasks to be scheduled, the computing power required to calculate the amount of tasks, the load of the central server, the computing power remaining after the edge server tasks are assigned, and the average delay in task completion.

3. The resource scheduling method for edge computing according to claim 2, characterized in that: When the task to be scheduled is a decomposable task, the target resource scheduling model is: Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, is the average delay measurement factor for task completion.

4. The resource scheduling method for edge computing according to claim 2, characterized in that: When the task to be scheduled is an indivisible task, the target resource scheduling model is: Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, The average delay factor for task completion.

5. The resource scheduling method for edge computing according to claim 3, characterized in that: Introducing the slack non-negative variable β, the target resource scheduling model is transformed into:

6. The resource scheduling method for edge computing according to claim 4, characterized in that: Introducing the slack non-negative variable β, the target resource scheduling model is transformed into:

7. The resource scheduling method for edge computing according to claim 5 or 6, characterized in that: The determining of the optimal scheduling parameters according to the target resource scheduling model includes: Generate α according to genetic algorithm k,n , in determining α k,n After the value is obtained, optimize β and f k,n The optimal scheduling parameters of the target resource scheduling model are determined by performing iterative calculation according to the preset number of iterations and convergence conditions.

8. A resource scheduling device for edge computing, characterized in that: include: A model building module is used to build a target resource scheduling model and determine optimal scheduling parameters according to the target resource scheduling model; A task scheduling module, configured to schedule multiple tasks to be scheduled according to the optimal scheduling parameters; The target resource scheduling model is constructed based on the task completion delay, the central server load and the load of multiple edge servers; The tasks to be scheduled include disassembled tasks and non-disassembled tasks; When the task to be scheduled is a decomposable task, the target resource scheduling model is: Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, is the average delay measurement factor for task completion; When the task to be scheduled is an indivisible task, the target resource scheduling model is: Where n is the server, is a set of servers, including a central server and N edge servers, n = 0 represents the central server; k is the task to be scheduled, is the set of tasks to be scheduled; r k is the workload of task k, c k is the computing power required to calculate the unit task volume, τ k is the maximum delay of task k; α k,n The proportion of tasks k assigned to each server; f k,n The computing power allocated to server n for task k; ω c is the central server load pressure weight, ω t is the task completion delay weight; f n The remaining computing power of the nth server before the assigned task arrives; The minimum remaining computing power after allocating tasks to N edge servers, The computing power consumed by the central server, The average delay factor for task completion.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the resource scheduling method for edge computing as described in any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the resource scheduling method for edge computing as described in any one of claims 1 to 7 are implemented.

Images