Multi-dimensional resource scheduling method, device and platform for power calculation network and storage medium

Abstract

The invention provides an electric power computing network multi-dimensional resource scheduling method, device and platform and a storage medium, and belongs to the technical field of electric power computing power networks, and the method comprises the steps: obtaining the request information of a current service terminal and the computing power resource information of a cloud side computing node; determining an end-to-end time delay constraint, a cloud edge computing power resource constraint and a cloud edge end network bandwidth constraint of the intelligent server; according to the cloud edge computing power unit price and the network bandwidth unit price, determining an objective function of computing network multi-dimensional resource scheduling optimization by taking minimization of the network bandwidth and the cloud edge computing power overhead as an optimization objective; solving the plurality of scheduling constraint models to obtain an optimal solution; and outputting a computing network multi-dimensional resource scheduling strategy according to the optimal solution. According to the method, under the condition that different computing network resource service quality requirements of multiple services are considered, by dynamically deploying multi-dimensional resources such as network bandwidth, cloud computing power and edge computing power, the low response delay requirements of key services such as environmental risks and equipment faults in power grid operation are effectively met, and meanwhile it is ensured that the computing network resource consumption economical efficiency is optimal.

Description

Technical Field

[0001] This invention relates to a method, apparatus, platform, and storage medium for multi-dimensional resource scheduling in power computing networks, belonging to the field of power computing network technology. Background Technology

[0002] In intelligent power grid operations such as power transmission and transformation equipment inspection and online monitoring, drone inspection and video surveillance have developed towards autonomy and intelligence. Drones, equipped with infrared thermal imagers, LiDAR, and high-definition cameras, collect clear images and videos of power transmission and transformation equipment. Combined with deep learning-based algorithms to build intelligent visual analysis capabilities, they can achieve millimeter-level defect detection of transmission lines, such as insulator cracks and broken conductor strands.

[0003] Unmanned aerial vehicle (UAV) hubs and substation server rooms, serving as edge computing facilities, are interconnected with unified video and artificial intelligence platforms in the cloud via optical transmission networks or communication networks such as 5G, forming a computing network infrastructure that supports intelligent operations. The intelligent development of power grid operations increasingly relies on the support of a computing network infrastructure composed of heterogeneous networks and computing facilities, mainly to complete the perception of power grid equipment and environmental information, the integration and access of heterogeneous networks, and autonomous cognitive computing.

[0004] The traditional approach of relying on cloud computing centers to centrally complete cognitive computing results in high redundant investment in order to meet peak computing power demand. It also has problems such as high bandwidth consumption by the edge and cloud computing centers, and high cloud response latency due to multi-task concurrency.

[0005] Existing research has proposed cloud-edge collaborative computing methods. By deploying edge computing nodes such as MEC devices and nested nodes on the substation side, some computing tasks in autonomous cognitive computing are offloaded to the edge, such as feature extraction and semantic understanding of devices and environment. This enables localized cognitive computing, reduces the peak-valley demand difference of cloud computing power, and effectively reduces computing power investment.

[0006] However, cloud-edge collaborative computing rarely considers the extremely high real-time requirements imposed by millisecond-level response needs for abnormal events such as tower crane intrusions and wildfires in power grid operations. It also focuses more on the dynamic allocation of computing resources on the cloud and edge sides under conditions where communication resources are not limited, without considering the joint dynamic allocation of computing and network resources under conditions of multiple latency-sensitive tasks running concurrently and communication resources being limited. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, device, platform and storage medium for multi-dimensional resource scheduling of power grid computing. It takes the end-to-end latency and communication bandwidth of intelligent services such as UAV inspection and online equipment monitoring as constraints, minimizes the network transmission link and cloud-edge computing power overhead, and solves the problem that cloud-edge collaborative computing does not consider the conditions of concurrent multi-latency sensitive tasks and limited communication resources.

[0008] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0009] In a first aspect, the present invention provides a multi-dimensional resource scheduling method for power computing networks, applied to a cloud-edge network composed of cloud-edge computing nodes and service terminals; the multi-dimensional resource scheduling method for power computing networks includes: S1, obtaining request information of the current service terminal and computing power resource information of each cloud-edge computing node; wherein, the request information includes: intelligent service type, requested data transmission volume, and intelligent service end-to-end latency information;

[0010] S2. Based on the request information and computing resource information, determine multiple scheduling constraint models, specifically including: determining the intelligent server end-to-end latency constraint; determining cloud-edge-device computing resource constraints; and determining cloud-edge-device network bandwidth constraints.

[0011] S3. Based on the unit price of computing power and the unit price of network bandwidth of each cloud-edge computing node, determine the objective function for multi-dimensional resource scheduling optimization of the power computing network; whereby the optimization objective is defined as: under the constraints of end-to-end latency of intelligent service, cloud-edge computing power resources, and cloud-edge bandwidth, minimize network bandwidth and cloud-edge computing power overhead through cloud-edge computing resource scheduling of the power computing network.

[0012] S4. Based on the optimization objective, solve multiple scheduling constraint models to obtain the optimal solution for each scheduling constraint model;

[0013] S5. Based on the optimal solutions of the multiple scheduling constraint models, output the multi-dimensional resource scheduling strategy of the power grid.

[0014] Furthermore, the end-to-end latency constraint of the intelligent server in S2 is defined as follows:

[0015] ;

[0016] in, For business terminal index variables, The number of service terminals currently requesting allocation of computing network resources; For cloud edge computing node index variables, To calculate the number of nodes for the cloud edge, Indicates a cloud computing node; For the occurrence of variables, Indicates the business terminal Through cloud edge computing nodes Data transmission and intelligent computing, otherwise ; The amount of data transmitted to the business terminal; The computing resource requirements for the amount of data per intelligent service unit are set according to the type of intelligent service. For business terminals Access cloud edge computing nodes Network bandwidth; cloud-edge computing nodes For business terminals Allocated computing resources; For business terminals The end-to-end latency of intelligent services.

[0017] Furthermore, the cloud-edge-device computing power resource constraints in S2 are defined as follows:

[0018] ;

[0019] Cloud-edge-device network bandwidth constraints are defined as follows:

[0020] ;

[0021] in, For business terminal index variables, The number of service terminals currently requesting allocation of computing network resources; For cloud edge computing node index variables, To calculate the number of nodes for the cloud edge, Indicates a cloud computing node; For the occurrence of variables, Indicates the business terminal Through cloud edge computing nodes Data transmission and intelligent computing, otherwise ; cloud-edge computing nodes For business terminals Allocated computing resources; cloud-edge computing nodes Computing resources; For business terminals Access cloud edge computing nodes Network bandwidth; cloud-edge computing nodes Total bandwidth.

[0022] Furthermore, the objective function for the multi-dimensional resource scheduling optimization of the power grid is:

[0023] ;

[0024] in, For business terminal index variables, The number of service terminals currently requesting allocation of computing network resources; For cloud edge computing node index variables, To calculate the number of nodes for the cloud edge, Indicates a cloud computing node; For the occurrence of variables, Indicates the business terminal Through cloud edge computing nodes Data transmission and intelligent computing, otherwise ; cloud-edge computing nodes For business terminals Allocated computing resources; cloud-edge computing nodes The unit price of computing power; For business terminals Access cloud edge computing nodes Network bandwidth; This refers to the unit price of network bandwidth.

[0025] Furthermore, the step of solving multiple scheduling constraint models based on the optimization objective includes:

[0026] S31, Let the loop variable... Randomly generate a set of initial solutions that satisfy all scheduling constraints, denoted as . Then the current optimal objective function ;

[0027] S32, Regarding the current situation , in response to any occurring variable Change from 0 to 1, obtain and to Obtain by performing neighborhood changes ;

[0028] In response to Satisfy all scheduling constraints, Updated to Otherwise, re-execute S32;

[0029] S33. Calculate the corresponding objective function. If the objective function Then update Find the optimal solution and update the optimal objective function. ;

[0030] S34. Repeat S32 to S33 until the iteration stopping condition is met.

[0031] Furthermore, the iteration stopping condition is: the loop variable reaches a preset upper limit or the improvement in the objective function value is continuous. The number of times less than the preset threshold, where, This is the default value.

[0032] Secondly, the present invention provides a power grid resource scheduling device for implementing the power grid multi-dimensional resource scheduling method described in the first aspect. The device includes: a data acquisition module for acquiring request information from the current business terminal and computing power resource information from each cloud-edge computing node; wherein the request information includes: intelligent service type, amount of data to be sent, and intelligent service end-to-end latency information.

[0033] The constraint model determination module is used to determine multiple scheduling constraint models based on the request information and computing resource information, specifically including: intelligent server end-to-end latency constraint, cloud-edge-device computing resource constraint, and cloud-edge-device network bandwidth constraint.

[0034] The objective function determination module is used to determine the objective function for optimizing the multi-dimensional resource scheduling of the power computing network based on the unit price of computing power and network bandwidth of each cloud-edge computing node. The optimization objective is defined as: under the constraints of end-to-end latency of the intelligent server, cloud-edge computing power resources, and cloud-edge bandwidth, to minimize network bandwidth and cloud-edge computing power overhead through the scheduling of cloud-edge computing resources of the power computing network.

[0035] The objective solving module is used to solve multiple scheduling constraint models based on the optimization objective, and obtain the optimal solution for multiple scheduling constraint models;

[0036] The strategy output module is used to output a multi-dimensional resource scheduling strategy for the power grid based on the optimal solution of the multiple scheduling constraint models.

[0037] Thirdly, the present invention provides a server including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the steps of the power grid multi-dimensional resource scheduling method described in the first aspect.

[0038] Fourthly, the present invention provides a computing network resource scheduling platform, comprising: a server as described in the third aspect, wherein the server is signal-connected to the business terminal and each cloud-edge computing node, and is used to obtain request information from each business terminal and computing power resource information from each cloud-edge computing node; and is also used to transmit the generated power computing network multi-dimensional resource scheduling strategy to each business terminal and each cloud-edge computing node.

[0039] Fifthly, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, performs the steps of the power grid multi-dimensional resource scheduling method described in the first aspect.

[0040] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: The multi-dimensional resource scheduling method for power grid provided by the present invention, under the conditions of intelligent operation scenarios such as UAV inspection, online equipment monitoring, and emergency repair, and considering the different service quality requirements of multi-service power grid resources, effectively meets the low response latency requirements of key services such as environmental risks and equipment failures in power grid operations by dynamically allocating multi-dimensional resources such as network bandwidth, cloud computing power, and edge computing power, while ensuring the optimal economic efficiency of power grid resource consumption. Attached Figure Description

[0041] Figure 1 This is a diagram of the computing network resource scheduling platform architecture in Embodiment 1 of the present invention;

[0042] Figure 2 This is a flowchart of the multi-dimensional resource scheduling method for power grid in Embodiment 1 of the present invention. Detailed Implementation

[0043] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or apparatuses. The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0044] Example 1

[0045] This embodiment provides a multi-dimensional resource scheduling method for power grid computing networks, applied to a cloud-edge-device network composed of cloud-edge computing nodes and service terminals; such as Figure 1 As shown, the multi-dimensional resource scheduling method for power computing networks provided in this embodiment involves a computing network resource scheduling platform, a cloud-edge computing node consisting of edge computing nodes and cloud computing power centers, and business terminals such as drones and video.

[0046] The computing network resource scheduling platform is responsible for receiving real-time parameters of the intelligent service type of business terminals, the computing resource usage of edge computing nodes and cloud computing centers, and allocating edge or cloud computing resources to business terminals according to their intelligent service type. It also ensures that the processing latency of each business terminal's intelligent service and the total computing resource usage of edge computing nodes and cloud computing centers meet the constraints. Furthermore, it transmits the generated multi-dimensional computing network resource scheduling strategy to business terminals, edge computing nodes, and cloud computing centers. In summary, it has the functions of data collection, resource monitoring, resource allocation, and resource notification.

[0047] The business terminal is responsible for collecting images of the equipment and environment, and sending intelligent service type parameters to the computing network resource scheduling platform before obtaining computing network resources, and receiving the resource allocation results from the computing network resource scheduling platform; after obtaining computing network resources, it sends the intelligent service type parameters and the images to be processed to the allocated edge computing nodes and cloud computing power centers.

[0048] In this embodiment, the number of service terminals is set to... , For the business terminal index variable, the end-to-end latency of the intelligent server is... The amount of data requested by each service terminal is [amount]. The unit is Mb.

[0049] Edge computing nodes or cloud computing centers are responsible for receiving images to be processed from business terminals, selecting appropriate artificial intelligence models based on intelligent service type parameters, and providing relevant inference and analysis services.

[0050] In this embodiment, the cloud computing center and multiple edge computing nodes are interconnected through a high-speed communication link to form a... Each cloud-edge computing node allows any business terminal to obtain computing resources from the cloud computing center or a specific edge computing node. For the index variable of the cloud edge computing node, This refers to the cloud platform, while edge computing nodes are from... start.

[0051] The multi-dimensional resource scheduling method for power grid provided in this embodiment is implemented by servers included in the power grid resource scheduling platform, such as... Figure 2 As shown, it includes the following steps:

[0052] Step 1: Obtain the request information of the business terminal currently requesting allocation of computing network resources, as well as the computing power resource information of each cloud edge computing node.

[0053] Specifically, the request information includes: the type of intelligent service, the amount of data to be transmitted, and the end-to-end latency information of the intelligent service.

[0054] Generally, intelligent service types mainly include equipment defect identification, environmental risk identification, and operational safety risk identification; the artificial intelligence algorithms required for the first two types of intelligent services mainly include:

[0055] (1) Data preprocessing, such as image segmentation and noise reduction;

[0056] (2) Target detection, such as feature mining and semantic recognition;

[0057] (3) Risk assessment; requires data preprocessing, multi-target detection, intelligent spatial computing and security risk assessment.

[0058] It is important to note that the computing resource requirements of business terminals are dynamically changing. Therefore, the computing network resource scheduling platform receives request information from business terminals in real time and allocates cloud-edge computing resources accordingly. Furthermore, the computing resource requirements per unit data volume differ for different types of intelligent services, and the computing resource requirements for intelligent services are directly proportional to the amount of data to be processed. Therefore, this embodiment defines the computing resource requirements per unit data volume for intelligent services as follows: The unit is MFLO / Mb.

[0059] Step 2: Determine multiple scheduling constraint models based on the obtained request information and computing resource information, including: intelligent server end-to-end latency constraint, cloud-edge-device computing resource constraint, and cloud-edge-device network bandwidth constraint.

[0060] Specifically, currently Each business terminal will specify the type of intelligent service and the amount of data requested for transmission. The end-to-end latency of the intelligent service is ; After the parameters are sent to the cloud computing resource scheduling platform, the platform determines the following computing resource constraints for the cloud-edge computing nodes:

[0061]

[0062] in, For the occurrence of variables, Indicates the business terminal Through cloud edge computing nodes Data transmission and intelligent computing, otherwise ; cloud-edge computing nodes Computing resources.

[0063] The end-to-end latency constraints for all intelligent servers are as follows: ;

[0064] in, For business terminals Access cloud edge computing nodes Network bandwidth;

[0065] The bandwidth constraints for cloud-edge computing nodes are as follows:

[0066] in, cloud-edge computing nodes Total bandwidth.

[0067] Step 3: Determine the objective function for multi-dimensional resource scheduling optimization of the power grid based on the unit price of computing power and network bandwidth of each cloud edge computing node.

[0068] Specifically, the optimization objective is defined as follows: under the constraints of end-to-end latency of intelligent services, cloud-edge-device computing resources, and cloud-edge-device bandwidth, minimize network bandwidth and cloud-edge computing overhead through power grid cloud-edge-device computing resource scheduling.

[0069] In some embodiments, in order to improve the overall power grid resource utilization efficiency, the optimization objective function of the optimization problem is listed as follows:

[0070] ;

[0071] in, cloud-edge computing nodes The unit price of computing power; This refers to the unit price of network bandwidth.

[0072] Step 4: Based on the optimization objective, solve multiple scheduling constraint models to obtain the optimal solution of multiple scheduling constraint models.

[0073] This embodiment does not specifically limit the method for solving multiple scheduling constraint models. Specifically, existing technologies such as deep reinforcement learning and heuristic search can be used to solve multiple scheduling constraint models, thereby obtaining the optimal solutions for multiple scheduling constraint models.

[0074] Step 5: Based on the optimal solutions of multiple scheduling constraint models, output the multi-dimensional resource scheduling strategy of the power grid.

[0075] This embodiment uses the latency and communication bandwidth of intelligent operation business terminals such as UAV inspection and online equipment monitoring as constraints to perform artificial intelligence services. It adjusts the multi-dimensional computing resource allocation strategy of the power grid computing network in real time to adapt to the dynamic environment, minimize network transmission links and cloud-edge computing power overhead, and ultimately achieve the optimal allocation of cloud-edge collaborative computing network resources for power grid intelligent services.

[0076] Example 2

[0077] Based on Example 1, this example provides a process for solving multiple scheduling constraint models in real time using an iterative update search heuristic method.

[0078] In this embodiment, multiple scheduling constraint models are solved according to the optimization objective, including:

[0079] Step 30: Set the loop variable This is used to record the number of solutions to the objective function found during the iteration process.

[0080] Step 31: Randomly generate a set of initial solutions that satisfy all scheduling constraints, denoted as . ,calculate As the current optimal objective function ;

[0081] Step 32, for any , in response to any occurring variable If it changes from 0 to 1, then you get and to Obtain by performing neighborhood changes .

[0082] The neighborhood variation mentioned here refers to each variable randomly increasing or decreasing by a small number. .

[0083] In response to If all scheduling constraints are satisfied, then proceed with the loop variable. The update will Updated to Otherwise, repeat step 32.

[0084] It should be noted that the loop variable The update only represents the number of solutions that satisfy all scheduling constraints found during the iteration process. However, not every solution found can be updated to the current optimal solution, so the optimal solution needs to be verified next.

[0085] Step 33: Calculate the corresponding objective function. If the objective function Then update Find the optimal solution and update the optimal objective function. .

[0086] This step is to verify the optimal solution. If the objective function value corresponding to the currently found solution is the historical minimum, then the current solution is updated to the optimal solution, and its corresponding objective function value is updated to the optimal objective function.

[0087] Step 34: Repeat steps 32 to 33 until the iteration stopping condition is met.

[0088] In some embodiments, the iteration stopping condition is: the loop variable reaches 200, that is, 200 solutions that satisfy all scheduling constraints are found, and the solution with the smallest corresponding objective function value is taken as the optimal solution.

[0089] In some embodiments, the iteration stopping condition is: the improvement in the objective function value is continuous. The number of times is less than the preset threshold.

[0090] Example 3

[0091] This embodiment provides a power grid resource scheduling device for implementing the power grid multi-dimensional resource scheduling method described in Embodiment 1 or 2. It includes a data acquisition module for acquiring request information from the current business terminal and computing resource information from each cloud-edge computing node. The request information includes: intelligent service type, amount of data to be sent, and intelligent service end-to-end latency information.

[0092] The constraint model determination module is used to determine multiple scheduling constraint models based on the request information and computing resource information, specifically including: intelligent server end-to-end latency constraint, cloud-edge-device computing resource constraint, and cloud-edge-device network bandwidth constraint.

[0093] The objective function determination module is used to determine the objective function for optimizing the multi-dimensional resource scheduling of the power computing network based on the unit price of computing power and network bandwidth of each cloud-edge computing node. The optimization objective is defined as: under the constraints of end-to-end latency of the intelligent server, cloud-edge computing power resources, and cloud-edge bandwidth, to minimize network bandwidth and cloud-edge computing power overhead through the scheduling of cloud-edge computing resources of the power computing network.

[0094] The objective solving module is used to solve multiple scheduling constraint models based on the optimization objective, and obtain the optimal solution for multiple scheduling constraint models;

[0095] The strategy output module is used to output a multi-dimensional resource scheduling strategy for the power grid based on the optimal solution of the multiple scheduling constraint models.

[0096] The specific functions of each module are described in the relevant content of the method in Example 1, and will not be repeated here.

[0097] Example 4

[0098] This embodiment provides a server, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the steps of the power grid multi-dimensional resource scheduling method described in embodiment 1 or 2.

[0099] Example 5

[0100] This embodiment provides a computing network resource scheduling platform, including: a server as described in embodiment 4, wherein the server is signal-connected to each business terminal and each cloud-edge computing node, and is used to obtain request information from each business terminal and computing power resource information from each cloud-edge computing node; the server is also used to transmit the generated computing network multi-dimensional resource scheduling strategy to each business terminal and each cloud-edge computing node.

[0101] Example 6

[0102] This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the power grid multi-dimensional resource scheduling method described in Embodiment 1 or 2.

[0103] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0104] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for multi-dimensional resource scheduling of an electric power grid, characterized in that, The application is applied to a cloud-edge-terminal network composed of cloud-edge computing nodes and service terminals. The power algorithm network multi-dimensional resource scheduling method comprises the following steps: S1, obtaining request information of a current service terminal and algorithm resource information of each cloud-edge computing node; wherein the request information comprises intelligent service type, request transmission data volume, and intelligent service end-to-end delay information; S2, determining a plurality of scheduling constraint models according to the request information and the algorithm resource information, specifically comprising: determining intelligent service end-to-end delay constraints; determining cloud-edge-terminal algorithm resource constraints; determining cloud-edge-terminal network bandwidth constraints; S3, determining a target function of power algorithm network multi-dimensional resource scheduling optimization according to the algorithm unit price and the network bandwidth unit price of each cloud-edge computing node; wherein the optimization target is defined as: under the constraints of intelligent service end-to-end delay, cloud-edge-terminal algorithm resource, and cloud-edge-terminal bandwidth, minimizing network bandwidth and cloud-edge algorithm resource consumption through power algorithm network cloud-edge-terminal computing resource scheduling; S4, solving the plurality of scheduling constraint models according to the optimization target to obtain optimal solutions of the plurality of scheduling constraint models; S5, outputting a power algorithm network multi-dimensional resource scheduling strategy according to the optimal solutions of the plurality of scheduling constraint models.

2. The method of claim 1, wherein, The intelligent service end-to-end delay constraints in S2 are defined as: ； wherein, is an index variable for the service terminal, is the number of service terminals allocated with network resources for the current request; is an index variable for the cloud edge computing node, is the number of cloud edge computing nodes, represents a cloud edge computing node; is a variable that occurs, represents a service terminal passes through a cloud edge computing node transmits data and intelligent computing, otherwise ; is the amount of data transmitted by the service terminal; is the computing resource requirement per unit of data for intelligent services, the value of which is set according to the type of intelligent service; is the service terminal accesses the cloud edge computing node network bandwidth; is the cloud edge computing node is the computing resource allocated to the service terminal ; is the end-to-end latency of the intelligent service terminal .

3. The method of claim 1, wherein, The cloud-edge-terminal algorithm resource constraints in S2 are defined as: ； The cloud-edge-terminal network bandwidth constraints are defined as: ； wherein, is a service terminal index variable, is the number of service terminals assigned to allocate network resources for the current request; is a cloud edge computing node index variable, is the number of cloud edge computing nodes, represents a cloud edge computing node; is an occurrence variable, represents a service terminal through a cloud edge computing node transmits data and intelligent computing, otherwise ; is a cloud edge computing node is a service terminal assigned computing power resources; is the computing power resources of a cloud edge computing node ; is a service terminal accesses the network bandwidth of a cloud edge computing node ; is the total bandwidth of a cloud edge computing node .

4. The method of claim 1, wherein, The target function of power algorithm network multi-dimensional resource scheduling optimization is: ； wherein, is a service terminal index variable, is the number of service terminals allocated network resources for the current request; is a cloud edge computing node index variable, is the number of cloud edge computing nodes, represents a cloud edge computing node; is an occurrence variable, represents a service terminal through a cloud edge computing node transmits data and intelligent computing, otherwise ; is a cloud edge computing node is a service terminal allocated computing power resources; is the computing power unit price of a cloud edge computing node ; is a service terminal accesses a cloud edge computing node network bandwidth; is the network bandwidth unit price.

5. The method of claim 1, wherein, Solving the plurality of scheduling constraint models according to the optimization target comprises: S31, let the loop variable a set of initial solutions that satisfy all scheduling constraints, denoted as the current optimal objective function ; S32, to the current , in response to any occurrence variable from 0 to 1, get , and to neighborhood change to get ; in response to satisfying all scheduling constraints, updating to ; otherwise re-perform S32; S33, calculate the corresponding objective function , if the objective function , update to the optimal solution, and update the optimal objective function ; S34, repeating S32 to S33 until the iteration stopping condition is met.

6. The method of claim 5, wherein, The iteration stopping condition is that a loop variable reaches a preset upper limit or an improvement amount of a target function value is less than a preset threshold value for a preset number of times, wherein, the preset threshold value is less than a preset threshold value, wherein, the preset value is a preset value.

7. A power grid resource scheduling apparatus, characterized by comprising: The device for implementing the power algorithm network multi-dimensional resource scheduling method in any one of claims 1-6 comprises: a data acquisition module for obtaining request information of a current service terminal and algorithm resource information of each cloud-edge computing node; wherein the request information comprises intelligent service type, to-be-sent data volume, and intelligent service end-to-end delay information; a constraint model determination module for determining a plurality of scheduling constraint models according to the request information and the algorithm resource information, specifically comprising intelligent service end-to-end delay constraints, cloud-edge-terminal algorithm resource constraints, and cloud-edge-terminal network bandwidth constraints; an optimization target function determination module for determining a target function of power algorithm network multi-dimensional resource scheduling optimization according to the algorithm unit price and the network bandwidth unit price of each cloud-edge computing node; wherein the optimization target is defined as: under the constraints of intelligent service end-to-end delay, cloud-edge-terminal algorithm resource, and cloud-edge-terminal bandwidth, minimizing network bandwidth and cloud-edge algorithm resource consumption through power algorithm network cloud-edge-terminal computing resource scheduling; a target solving module for solving the plurality of scheduling constraint models according to the optimization target to obtain optimal solutions of the plurality of scheduling constraint models; a strategy output module for outputting a power algorithm network multi-dimensional resource scheduling strategy according to the optimal solutions of the plurality of scheduling constraint models.

8. A server comprising a memory and a processor, the memory having stored therein a computer program executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the power algorithm network multi-dimensional resource scheduling method in any one of claims 1-6.

9. A computing resource scheduling platform, comprising: Comprise: The server of claim 8, which is in signal connection with the service terminal and each cloud edge computing node respectively, is configured to acquire request information from each service terminal and acquire computing resource information from each cloud edge computing node, and is further configured to transmit the generated power-computing network multi-dimensional resource scheduling strategy to each service terminal and each cloud edge computing node.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, performs the steps of the power-computing network multi-dimensional resource scheduling method of any one of claims 1-6.

Images