Power distribution automation terminal task scheduling method and device based on power distribution network edge computing nodes, electronic equipment and storage medium

Abstract

The invention discloses a power distribution automation terminal task scheduling method and device based on a power distribution network edge computing node, electronic equipment and a storage medium, and belongs to the field of power distribution network automation, and the method comprises the steps: obtaining the total amount of special and public containers of an edge node and the decomposition configuration and operation parameters of a target power distribution automation task; the method comprises the following steps of: decomposing a task into micro-applications, constructing a calling graph, counting the repetition degree of calling edges, combining the edges with high repetition degree into an undetachable micro-application chain, and calculating and distributing exclusive containers according to the repetition degree and the total number of the exclusive containers; and calculating weights according to task calling times, priorities and concurrency numbers, and calculating and distributing the public containers based on the weights and the total number of the public containers. Therefore, by implementing the method and the device, the problems that high-frequency fixed service flow resources are severely scrambled and the utilization rate of discrete task resources is low in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of distribution network automation technology, and specifically to a method, device, electronic device, and storage medium for scheduling terminal tasks of distribution automation based on edge computing nodes of distribution networks. Background Technology

[0002] In the digital transformation of power distribution networks, task scheduling technology for distribution automation terminals based on edge computing nodes is a core component ensuring real-time response and stable operation of power grid services. As the complexity of edge-side services increases, traditional monolithic applications are gradually decoupled into multiple micro-applications within a microservice architecture. These micro-applications have complex calling dependencies and collaborative relationships. Efficient task scheduling not only reduces communication latency but also maximizes business processing capabilities under resource-constrained edge conditions, which is crucial for achieving critical services such as rapid fault isolation and real-time status monitoring.

[0003] However, existing task scheduling methods for distribution automation terminals often lack in-depth analysis of the topology and calling logic between micro-applications within a business process, typically managing all micro-applications as independent entities. This approach ignores the objectively existing differences between "high-frequency fixed call chains" and "low-frequency discrete tasks" during business execution, leading to a "one-size-fits-all" approach in resource allocation. This breaks up the fixed business flows that should be tightly coupled and executed continuously, forcing them to compete for computing resources indiscriminately with discrete tasks in the same resource pool. This disordered resource allocation not only causes response jitter in critical fixed business flows due to resource contention, compromising their execution continuity and stability, but also makes it difficult for discrete tasks to obtain matching resource quotas based on their actual load intensity. Ultimately, this results in low overall resource utilization of edge nodes, making it difficult to meet the scheduling requirements of high concurrency and high stability in distribution networks. Summary of the Invention

[0004] This invention provides a method, device, electronic device, and storage medium for scheduling tasks of distribution automation terminals based on edge computing nodes in a distribution network. It can solve the problems in the prior art where the lack of analysis of the internal topology and calling logic of the business leads to the use of an indiscriminate resource allocation method, resulting in severe resource contention for high-frequency fixed business flows and low resource utilization of discrete tasks.

[0005] An embodiment of the present invention provides a task scheduling method for distribution automation terminals based on edge computing nodes of a distribution network, comprising: Furthermore, the total number of dedicated container resources and the total number of public container resources of the distribution network edge computing nodes are obtained, and the task decomposition configuration information and operating status parameters of the target distribution automation terminal task are obtained; the operating status parameters include the number of task calls of the target distribution automation terminal task and the number of concurrent calls of the micro-application vertex; the dedicated container resources and public container resources are allocated by the total container resource pool. Based on the task decomposition configuration information, the multiple micro-applications contained in the target power distribution automation terminal task and the calling order between the micro-applications are parsed out; the parsed multiple micro-applications are mapped as micro-application vertices, and the calling order is mapped as calling edges connecting the micro-application vertices to construct a micro-application graph; Traverse the call edges in the micro-application graph, count the frequency of each call edge to generate the edge redundancy between micro-applications; mark the call edges with edge redundancy greater than a preset redundancy threshold as key redundancy edges, and integrate the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains. Based on the edge repetition degree of the key repeating edge and the total number of dedicated container resources, calculate the first number of containers required for each indivisible micro-application chain, and allocate dedicated containers to each indivisible micro-application chain from the total container resource pool according to the first number of containers. Micro-applications that are not integrated into the indivisible micro-application chain in the micro-application graph are identified as discrete micro-applications; the weight of each discrete micro-application is calculated based on the number of task calls, the preset task priority, and the number of concurrent calls; the number of second containers to be allocated to each discrete micro-application is calculated based on the weight of each discrete micro-application and the total number of public container resources, and public containers are allocated to each discrete micro-application from the total container resource pool according to the number of second containers.

[0006] Furthermore, the step of marking call edges with a redundancy greater than a preset redundancy threshold as key repeating edges, and integrating the micro-applications connected by the key repeating edges into corresponding indivisible micro-application chains, includes: Traverse each call edge in the micro-application graph and compare the edge repetition of the current call edge with a preset repetition threshold; If the edge repetition degree of the current call edge is greater than the preset repetition threshold, the current call edge will be marked as a critical repetition edge. Based on the connectivity of all marked key repeating edges in the micro-application graph, search and identify the connected subgraphs formed by the key repeating edges; All micro-application vertices within each connected subgraph are identified as members of the same indivisible micro-application chain, and the connected subgraph is defined as a corresponding indivisible micro-application chain.

[0007] Furthermore, the step of calculating the number of first containers required for each indivisible micro-application chain based on the edge repetition degree of the key repeating edges and the total number of dedicated container resources includes: Determine the weight of the edge redundancy of each indivisible micro-application chain as a percentage of the total redundancy of all key repeating edges; The total number of dedicated container resources is mapped to the baseline container quota of each indivisible micro-application chain according to the aforementioned proportion weight. The baseline container quota is numerically normalized and corrected based on a preset minimum resource threshold to obtain the first number of containers required for each indivisible micro-application chain.

[0008] Furthermore, the distribution automation terminal task scheduling method based on distribution network edge computing nodes also includes: Obtain task execution monitoring data of the edge computing nodes of the power distribution network concurrently executing several target power distribution automation terminal task instances; For each discrete micro-application, extract the actual calculation completion time and computational complexity of the current discrete micro-application when it performs calculation processing on the several target power distribution automation terminal task instances in response to the task execution monitoring data. Based on the actual calculation completion time, the preset latency requirement, and the computational complexity, calculate the latency performance deviation of the current discrete micro-application for each target power distribution automation terminal task instance, and record the target power distribution automation terminal task instance whose latency performance deviation is greater than the preset abnormal threshold as an abnormal task instance. The latency performance deviations are accumulated to generate the latency performance deviation sum of the current discrete micro-application; When the time delay performance deviation is greater than the preset total deviation threshold, the proportion of the computational complexity of the abnormal task instance to the computational complexity of the plurality of target power distribution automation terminal task instances is calculated based on the computational complexity of the abnormal task instance. Based on the computational complexity ratio and the total number of containers currently held by the discrete micro-application in the shared container, calculate the number of third containers to be allocated to the dedicated channel of the current discrete micro-application. Based on the latency requirements, the abnormal task instances are clustered to generate task groups. Calculate the sum of latency performance deviations for all abnormal task instances within each task group; Configure a dedicated channel for each task group, and allocate the containers corresponding to the third number of containers from the total container resource pool to each dedicated channel based on the sum of the latency performance deviations of each task group. Each task group will be migrated to its corresponding dedicated channel for execution.

[0009] Furthermore, the step of calculating the latency performance deviation of the current discrete micro-application for each target power distribution automation terminal task instance based on the actual calculation completion time, the preset latency requirement, and the computational complexity includes: Calculate the numerical difference between the actual calculation completion time and the preset delay requirement; The portion of the numerical difference that is greater than zero is determined as the effective delay duration; Calculate the ratio of the effective delay duration to the computational complexity, and determine the ratio as the delay performance deviation.

[0010] Furthermore, the step of calculating the proportion of the computational complexity of the abnormal task instance to the computational complexity of the plurality of target distribution automation terminal task instances based on the computational complexity of the abnormal task instance includes: Based on the computational complexity of each abnormal task instance, determine the total abnormal computational load carried by the current discrete micro-application; Based on the computational complexity of several target distribution automation terminal task instances, determine the total computational load carried by the current discrete micro-application; Calculate the proportion of the total abnormal computing load in the total total computing load, and determine the proportion as the computing complexity ratio.

[0011] Furthermore, the step of clustering the abnormal task instances according to the latency requirements to generate task groups includes: Calculate the numerical proximity between the preset latency requirements of each abnormal task instance; Based on the numerical proximity, abnormal task instances that meet the proximity condition for preset latency requirements are grouped into the same task group.

[0012] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0013] One embodiment of the present invention provides a task scheduling device for distribution automation terminals based on edge computing nodes of distribution networks, including: a parameter acquisition module, a micro-application graph construction module, a micro-application chain integration module, a dedicated resource allocation module, and a public resource allocation module; The parameter acquisition module is used to acquire the total number of dedicated container resources and the total number of public container resources of the distribution network edge computing node, and to acquire the task decomposition configuration information and operating status parameters of the target distribution automation terminal task; the operating status parameters include the number of task calls of the target distribution automation terminal task and the number of concurrent calls of the micro-application vertex; the dedicated container resources and public container resources are allocated by the total container resource pool; The micro-application graph construction module is used to parse out multiple micro-applications contained in the target power distribution automation terminal task and the calling order between the micro-applications according to the task decomposition configuration information; map the parsed multiple micro-applications as micro-application vertices, and map the calling order as calling edges connecting the micro-application vertices to construct a micro-application graph. The micro-application chain integration module is used to traverse the call edges in the micro-application graph, count the frequency of each call edge to generate the edge redundancy between micro-applications; mark the call edges with edge redundancy greater than a preset redundancy threshold as key redundancy edges, and integrate the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains. The dedicated resource allocation module is used to calculate the number of first containers required by each indivisible micro-application chain based on the edge repetition degree of the key repeated edges and the total number of dedicated container resources, and allocate dedicated containers to each indivisible micro-application chain from the total container resource pool according to the number of first containers. The public resource allocation module is used to identify micro-applications that are not integrated into the indivisible micro-application chain in the micro-application graph as discrete micro-applications; calculate the weight of each discrete micro-application based on the number of task calls, the preset task priority, and the number of concurrent calls; calculate the number of second containers that should be allocated to each discrete micro-application based on the weight of each discrete micro-application and the total number of public container resources, and allocate public containers to each discrete micro-application from the total container resource pool according to the number of second containers.

[0014] Based on the above method embodiments, the present invention provides corresponding electronic device embodiments.

[0015] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power distribution automation terminal task scheduling method based on the edge computing node of the power distribution network as described in any of the above-described method embodiments.

[0016] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0017] One embodiment of the present invention provides a storage medium storing a computer program thereon, wherein, when the computer program is running, it controls the device where the storage medium is located to execute any of the above-described method embodiments of the distribution automation terminal task scheduling method based on the edge computing node of the distribution network.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a method, apparatus, electronic device, and storage medium for scheduling terminal tasks of distribution automation based on edge computing nodes in a distribution network. The method obtains the total number of dedicated and public containers at the edge nodes, as well as the decomposition configuration and operating parameters (number of calls, concurrency) of the target distribution automation task. The task is decomposed into micro-applications, and a call graph is constructed. The redundancy of call edges is statistically analyzed; edges with high redundancy are merged into an indivisible chain of micro-applications, and dedicated containers are calculated and allocated based on redundancy and the total number of dedicated containers. The remaining micro-applications are discrete; weights are calculated based on the number of task calls, priority, and concurrency; and public containers are calculated and allocated based on these weights and the total number of public containers.

[0019] This invention constructs a micro-application graph and calculates edge redundancy to accurately identify high-frequency call logic with strong coupling relationships within a business process, effectively solving the problem of blind resource allocation caused by the lack of business topology analysis in existing technologies. Based on this, by integrating micro-applications connected by key repeating edges into an indivisible micro-application chain and allocating dedicated containers, it directly eliminates resource contention and response jitter in high-frequency fixed business flows under a hybrid competition mode, ensuring the continuity of core business. Simultaneously, it allocates shared containers to discrete micro-applications based on call counts and concurrency levels, breaking the rigid average allocation model and achieving on-demand quota allocation for discrete tasks while ensuring the stability of critical links, significantly improving the overall resource utilization efficiency of edge computing nodes. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a task scheduling method for distribution automation terminals based on edge computing nodes of a distribution network, provided by an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the structure of a power distribution automation terminal task scheduling device based on a power distribution network edge computing node according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] like Figure 1As shown, to address the problems in existing technologies where the lack of analysis of the internal topology and calling logic of services leads to indiscriminate resource allocation, resulting in severe resource contention for high-frequency fixed service flows and low resource utilization for discrete tasks, an embodiment of the present invention provides a task scheduling method for distribution automation terminals based on distribution network edge computing nodes, comprising at least the following steps: Step S1: Obtain the total number of dedicated container resources and the total number of public container resources of the distribution network edge computing nodes, and obtain the task decomposition configuration information and operating status parameters of the target distribution automation terminal task; the operating status parameters include the number of task calls of the target distribution automation terminal task and the number of concurrent calls of the micro-application vertex; the dedicated container resources and public container resources are allocated by the total container resource pool. Specifically, the steps involve obtaining the total number of dedicated container resources and the total number of public container resources of the distribution network edge computing nodes, and obtaining the task decomposition configuration information and operating status parameters of the target distribution automation terminal task to be scheduled.

[0024] First, regarding the acquisition and allocation of hardware resources, the distribution network edge computing node acts as a resource scheduling center, maintaining a total container resource pool. The total number of containers in the total container resource pool is defined as follows: To achieve physical resource isolation between high-frequency fixed service flows and discrete service flows during the static configuration phase, the system introduces a preset first-part resource reservation ratio. The total container resource pool is divided into two parts: one part consists of dedicated container resources for ensuring the stability of fixed chains, and the other part consists of public container resources for carrying discrete services. The total number of dedicated container resources... With the total number of public container resources Based on the total number of containers in the total container resource pool The allocation is performed, and the calculation logic is shown in the following formula: in, This represents the total number of containers in the edge computing nodes of the distribution network, which is the upper limit of physical resources; This indicates the preset first part of the resource reservation ratio. This parameter is usually configured during system design based on historical data statistics or business importance requirements. This represents the total number of dedicated container resources calculated. This represents the total number of public container resources calculated. The floor function is used to ensure that the number of containers is an integer. By explicitly stating that both the dedicated container resources and the public container resources are allocated from the total container resource pool, boundary constraints for resource scheduling are established.

[0025] Secondly, regarding the acquisition of software business characteristics, the task decomposition configuration information is static data describing the business logic structure. If the set of tasks for the power distribution automation terminal is denoted as... The target power distribution automation terminal task to be scheduled For set One of the elements. The task decomposition configuration information defines in detail the target power distribution automation terminal task. The microservice architecture consists of which micro-application units (corresponding to vertices in the subsequent micro-application graph) constitute the task, and what kind of calling logic exists between these micro-application units (corresponding to edges in the subsequent micro-application graph).

[0026] Secondly, regarding the acquisition of dynamic load characteristics, the operational status parameters reflect the heat and pressure of the business in the actual operating environment. The system collects raw data within a set statistical time window through a built-in business call link tracer and resource monitoring agent. To avoid instantaneous traffic spikes misleading subsequent weight calculations, the system uses percentile statistics (such as the P95 value) to smooth the raw data. Specifically, this includes: the number of task calls. This indicates the target distribution automation terminal task within the statistical period. The statistical value of the total frequency of initiation is used to measure the overall activity of the task; the number of concurrent calls to the micro-application vertex. Used to measure the target power distribution automation terminal task The overall activity level of the system within the statistical time window. The target power distribution automation terminal task is recorded within. The sequence of the number of times the task was initiated, and a set of task call counts is constructed. The system calculates the 95th percentile (P95) of this set and determines it as the number of times the task is invoked. Its definition is as follows: in, This means taking the values ​​in the set that are arranged in ascending order and located at the 95th percentile, thereby eliminating the top 5% of extremely high-frequency call noise.

[0027] Concurrent call count of the micro-application vertex , used to measure the first component of the task Micro-application vertex The system calculates the load pressure. The system uses a resource monitoring agent to sample at high frequency within the statistical time window. Internal record of the micro-application vertex The number of instances currently in execution is used to construct a set of concurrent instance counts. The system also calculates the 95th percentile of this set and determines it as the number of concurrent calls to the microapplication vertex. Its definition is as follows: By defining the parameters based on statistics as described above, the system can obtain the true load characteristics after removing accidental factors, thereby ensuring that the resource weights calculated based on these parameters can meet the needs of most peak scenarios without causing excessive redundancy in resource allocation due to extreme value effects.

[0028] Step S2: Based on the task decomposition configuration information, parse out the multiple micro-applications contained in the target power distribution automation terminal task and the calling order between the micro-applications; map the parsed multiple micro-applications as micro-application vertices, and map the calling order as calling edges connecting the micro-application vertices to construct a micro-application graph; Specifically, the steps involve parsing out multiple micro-applications included in the target power distribution automation terminal task and the calling order between the micro-applications based on the task decomposition configuration information; mapping the parsed multiple micro-applications as micro-application vertices; mapping the calling order as calling edges connecting the micro-application vertices; and constructing a micro-application graph.

[0029] In the processing logic of edge computing nodes, distribution automation terminal tasks are considered as a decomposable complex system. Let the target distribution automation terminal task to be scheduled be denoted as... ,in As a task index, the distribution network edge computing node first reads the task decomposition configuration information and identifies the tasks that make up the target distribution automation terminal through a parsing algorithm. The system comprises all independent functional units, namely the multiple micro-applications. Subsequently, the system analyzes the data flow direction and logical dependencies between these micro-applications to determine the calling order among them. Based on the above analysis results, the system uses a graph theory model to construct the target power distribution automation terminal task. The corresponding micro-application diagram The micro-application diagram The mathematical expression for it is defined as: in, This represents the set of vertices in the microapplication graph. This represents the set of call edges in the microapplication graph.

[0030] In the process of mapping the multiple micro-applications into micro-application vertices, the system instantiates each parsed micro-application into the vertex set. One of the elements in the vertex set. The definition is as follows: in, Indicates the target power distribution automation terminal task The first in A micro-application vertex Indicates the target power distribution automation terminal task The total number of micro-application vertices included. To clarify the correspondence between vertices and specific functions, if the micro-application vertices... The corresponding micro-application collection The first in Micro-applications Then establish a mapping relationship. .

[0031] In the process of mapping the call order as call edges, the system establishes the call edge set based on the predecessor-successor relationship between micro-applications. The set of calling edges. The definition is as follows: in, Indicates from the vertex of the micro-application Pointing to the vertex of the micro-application The system establishes a one-way call relationship. To quantify this topology, the system introduces an indicator function for the call edge. The existence is defined as follows: Through the above construction process, the power distribution network edge computing node transforms abstract business logic into the micro-application graph with a clear topology. This enables subsequent quantitative analysis of the degree of coupling within the business based on graph theory algorithms, thereby accurately identifying critical paths and resource demand hotspots in the business process.

[0032] Step S3: Traverse the call edges in the micro-application graph, count the frequency of each call edge to generate the edge redundancy between micro-applications; mark the call edges with edge redundancy greater than a preset redundancy threshold as key redundancy edges, and integrate the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains. In a preferred embodiment, marking call edges with a redundancy greater than a preset redundancy threshold as key repeating edges, and integrating the micro-applications connected by the key repeating edges into corresponding indivisible micro-application chains, includes: Traverse each call edge in the micro-application graph and compare the edge repetition of the current call edge with a preset repetition threshold; If the edge repetition degree of the current call edge is greater than the preset repetition threshold, the current call edge will be marked as a critical repetition edge. Based on the connectivity of all marked key repeating edges in the micro-application graph, search and identify the connected subgraphs formed by the key repeating edges; All micro-application vertices within each connected subgraph are identified as members of the same indivisible micro-application chain, and the connected subgraph is defined as a corresponding indivisible micro-application chain.

[0033] Specifically, after the distribution network edge computing node constructs the micro-application graph based on the task decomposition configuration information, the system needs to perform quantitative analysis on the call edges in the micro-application graph to accurately identify the fixed coupling relationships existing in the business process. Specifically, the system traverses each call edge in the micro-application graph and counts the frequency of each call edge in historical operation data to generate an edge repetition degree that measures the tightness of calls between micro-applications. This edge repetition degree can quantitatively reflect the fixed probability that two micro-applications are continuously called during business execution.

[0034] To accurately calculate the edge repetition degree, the system adopts the following calculation model: in, Indicates the first in the set of micro-applications Micro-applications With the Micro-applications Edge repetition between them; This indicates the total number of distribution automation terminal services; For business indexing; Indicates the first Each business contains a set of micro-application vertices; and These are the vertex indices in the vertex set; Indicates the first From the top of each business To the top Does the calling edge exist? (1 if it exists, 0 if it does not exist). Indicates the first The first business The micro-applications corresponding to each vertex; For the characteristic function, when all the conditions within the parentheses are met (i.e., there exists a calling edge, and the starting point is a microapplication),... The endpoint is micro-applications. The function value is 1 if the condition is met, and 0 otherwise. Using the above formula, the system can accumulate the number of calls between specific micro-application pairs appearing in all business processes, thus obtaining the global edge redundancy.

[0035] After obtaining the edge repetition rate, the system executes the critical path determination logic. The system compares the calculated edge repetition rate with a preset repetition threshold. The repetition threshold is compared. These are empirical values ​​derived from statistical analysis of business call relationships in historical system operation data, used to distinguish between occasional and recurring calls. The system defines identifiers for critical repeating edges based on the comparison results. The calculation logic is as follows: in, Indicates micro-applications With micro-applications Whether the connection between them is a critical repeating edge; when the edge repeatability Greater than or equal to the repeatability threshold hour, A value of 1 indicates that the called edge is marked as a critical repeating edge; otherwise, a value of 0 is taken.

[0036] In a preferred embodiment, to transform scattered key repeating edges into independently manageable resource units, the distribution network edge computing node performs a connectivity-based chain integration operation. Specific steps include: the system traverses each calling edge in the micro-application graph, comparing the edge repeatability of the current calling edge with a preset repeatability threshold. When the edge repeatability of the current calling edge is greater than the preset repeatability threshold, the system marks the current calling edge as a key repeating edge. Subsequently, based on the topological connections of all marked key repeating edges in the micro-application graph, the system executes a graph theory search algorithm (such as depth-first search or breadth-first search) to search for and identify connected subgraphs formed by interconnected key repeating edges. Any two micro-application vertices within the connected subgraph are directly or indirectly connected through the key repeating edges, indicating that these micro-applications have extremely strong coupling in business logic. The system determines all micro-application vertices within each identified connected subgraph as members of the same indivisible micro-application chain, and defines the connected subgraph as a corresponding indivisible micro-application chain.

[0037] By integrating high-frequency fixed micro-applications into an indivisible chain of micro-applications and allocating dedicated containers as described above, the embodiments of the present invention can effectively avoid the system resource consumption caused by frequent splitting and combining of micro-applications. At the same time, physical isolation reduces resource competition between critical services and other discrete micro-applications, thereby significantly improving the continuity and stability of power distribution automation terminal tasks during execution.

[0038] Step S4: Calculate the number of first containers required for each indivisible micro-application chain based on the edge repetition degree of the key repeating edge and the total number of dedicated container resources, and allocate dedicated containers to each indivisible micro-application chain from the total container resource pool according to the number of first containers. In a preferred embodiment, calculating the number of first containers required for each indivisible micro-application chain based on the edge repetition degree of the key repeating edges and the total number of dedicated container resources includes: Determine the weight of the edge redundancy of each indivisible micro-application chain as a percentage of the total redundancy of all key repeating edges; The total number of dedicated container resources is mapped to the baseline container quota of each indivisible micro-application chain according to the aforementioned proportion weight. The baseline container quota is numerically normalized and corrected based on a preset minimum resource threshold to obtain the first number of containers required for each indivisible micro-application chain.

[0039] Specifically, the system calculates the number of first containers required for each indivisible microapplication chain based on the edge repetition degree of the key repeating edges and the total number of pre-acquired dedicated container resources, and allocates dedicated containers to each indivisible microapplication chain from the total container resource pool according to the calculated number of first containers. This process aims to establish a mapping relationship between the tightness of business coupling and the supply of physical resources, ensuring that high-frequency business links can exclusively occupy computing resources that match their load characteristics.

[0040] In a preferred embodiment, the calculation logic for the first container quantity follows the steps of weight determination, quota mapping, and numerical normalization and correction. First, the system determines the proportion weight of the edge repetition degree of each of the indivisible micro-application chains within the total repetition degree of all marked key repeating edges. The system traverses the micro-application graph, accumulates the edge repetition degrees of all marked key repeating edges to obtain the total repetition degree, and calculates the ratio of the edge repetition degree of the key repeating edge corresponding to the currently processed indivisible micro-application chain to the total repetition degree. Second, the system maps the total number of dedicated container resources to a baseline container quota for each of the indivisible micro-application chains according to the proportion weight; that is, by multiplying the ratio by the total number of dedicated container resources, a theoretical container allocation value is obtained. Finally, the system performs numerical normalization on the baseline container quota and corrects it based on a preset minimum resource threshold to obtain the final first container quantity. The numerical rounding process specifically involves rounding down to ensure that the number of allocated containers is an integer and the total allocation of all chains does not exceed the total number of dedicated resources. The correction process specifically involves comparing the rounded value with the minimum resource threshold and taking the larger value to prevent the allocation quantity from being zero due to the weight being too small, thus making it impossible to maintain the chain operation.

[0041] The above method for calculating the number of the first containers can be expressed by the following mathematical formula: in, Indicated by microapplications With micro-applications The number of the first containers required to form the indivisible microapplication chain; This indicates the total number of the dedicated container resources; Indicates micro-applications With micro-applications The degree of edge repetition between them; The key repeating edge is identified by a value of 1 when the edge repeatability is greater than a preset repeatability threshold, and 0 otherwise; the denominator is... This represents the sum of the edge repetition degrees of all key repeating edges in the microapplication graph; This represents the floor function operator; This represents the preset minimum resource threshold, which is typically set to 1. This represents the function that takes the maximum value.

[0042] By employing the resource allocation strategy based on edge repetition ratio, the embodiments of the present invention can ensure that high-frequency fixed business processes obtain dedicated computing resources that match their calling frequency. This not only avoids the system overhead caused by frequent splitting and combining of micro-applications, but also effectively reduces resource competition between critical business and other discrete business through physical isolation, thereby significantly improving the continuity and stability of power distribution automation terminal tasks executed at the edge.

[0043] Step S5: Identify the micro-applications in the micro-application graph that are not integrated into the indivisible micro-application chain as discrete micro-applications; calculate the weight of each discrete micro-application based on the number of task calls, the preset task priority, and the number of concurrent calls; calculate the number of second containers that should be allocated to each discrete micro-application based on the weight of each discrete micro-application and the total number of public container resources, and allocate public containers to each discrete micro-application from the total container resource pool according to the number of second containers.

[0044] Specifically, the steps involve identifying micro-applications in the micro-application graph that are not integrated into the indivisible micro-application chain as discrete micro-applications; calculating the weight of each discrete micro-application based on the number of task calls, the preset task priority, and the number of concurrent calls; calculating the number of second containers that should be allocated to each discrete micro-application based on the weight of each discrete micro-application and the total number of public container resources, and allocating public containers to each discrete micro-application from the total container resource pool according to the number of second containers.

[0045] First, the system traverses all vertices in the micro-application graph, identifies micro-applications that are not marked as key repeating edge-connected nodes, and defines these micro-applications as discrete micro-applications. For each discrete micro-application, the system first needs to aggregate its concurrent load pressure in different business call paths and calculate the discrete micro-application... Total number of concurrent calls The total number of concurrent calls. The calculation formula is as follows: in, Indicates the first The total number of concurrent calls to the discrete micro-applications described above; This indicates the total number of tasks in the power distribution automation terminal. For task indexing; Indicates the first Each task contains a set of micro-application vertices; For vertex index; Indicates the first The first task The number of concurrent calls for each micro-application vertex; For the characteristic function, when the condition inside the parentheses is true (i.e., the micro-application corresponding to that vertex is indeed a characteristic function), the characteristic function is defined as follows: The value is 1 when the condition is met, and 0 otherwise.

[0046] Secondly, in order to allocate limited public resources fairly and efficiently, the system constructs a multi-dimensional evaluation index system to determine the resource scheduling priority of each discrete micro-application, i.e., weight. This weight comprehensively considers the activity level of the business (number of task calls), the importance of the business (task priority), and the load of the microapplication itself (total number of concurrent calls). The weight... The calculation formula is as follows: in, Indicates the first The weights of the discrete micro-applications; Indicates the first The number of times the task is invoked for each task; Indicates the first The preset task priority for each task; This represents the total number of concurrent calls to the discrete micro-application obtained from the aforementioned calculations; These correspond to weighting coefficients for the number of task calls, task priority, and concurrent calls to the micro-application, respectively. To ensure the scientific validity and objectivity of these weighting coefficients, the weighting coefficient for task priority is... The weighting coefficients for the number of task calls are determined using the Analytic Hierarchy Process (AHP), calculated by constructing a judgment matrix and passing a consistency check. Weighting coefficients for the number of concurrent calls to the micro-application The entropy weight method is used to determine the information entropy and discrimination ability of the indicator by utilizing the degree of variation in the data.

[0047] Finally, after determining the weights of each discrete micro-application, the system divides the total number of public container resources according to the weight ratio and calculates the number of second containers that should be allocated to each discrete micro-application. This step enables weighted, on-demand allocation, ensuring that critical and busy services receive more computing resources. The second number of containers The calculation formula is as follows: in, Indicates assignment to the first The number of the second containers for each of the discrete micro-applications; This represents the total number of public container resources (i.e., the remaining portion after deducting dedicated resources from the total container resource pool). Indicates the first The weights of the discrete micro-applications; This represents the sum of the total weights of all discrete micro-applications; This indicates the floor function; This represents the preset minimum number of containers threshold (usually set to 1), used to ensure that each discrete micro-application is allocated at least one container to maintain basic operation; This represents the function that takes the maximum value. The system uses the calculated number of the second container as the basis for this function. The system retrieves a corresponding number of containers from the total container resource pool and marks them as public containers for use by the discrete micro-applications. Through the above steps, this embodiment of the invention can break the rigid mode of traditional static configuration and dynamically adjust the resource quotas of discrete services according to the actual popularity and importance of the services, maximizing the resource utilization efficiency of edge nodes while ensuring critical services.

[0048] In a preferred embodiment, the distribution automation terminal task scheduling method based on distribution network edge computing nodes further includes: Obtain task execution monitoring data of the edge computing nodes of the power distribution network concurrently executing several target power distribution automation terminal task instances; For each discrete micro-application, extract the actual calculation completion time and computational complexity of the current discrete micro-application when it performs calculation processing on the several target power distribution automation terminal task instances in response to the task execution monitoring data. Based on the actual calculation completion time, the preset latency requirement, and the computational complexity, calculate the latency performance deviation of the current discrete micro-application for each target power distribution automation terminal task instance, and record the target power distribution automation terminal task instance whose latency performance deviation is greater than the preset abnormal threshold as an abnormal task instance. The latency performance deviations are accumulated to generate the latency performance deviation sum of the current discrete micro-application; When the time delay performance deviation is greater than the preset total deviation threshold, the proportion of the computational complexity of the abnormal task instance to the computational complexity of the plurality of target power distribution automation terminal task instances is calculated based on the computational complexity of the abnormal task instance. Based on the computational complexity ratio and the total number of containers currently held by the discrete micro-application in the shared container, calculate the number of third containers to be allocated to the dedicated channel of the current discrete micro-application. Based on the latency requirements, the abnormal task instances are clustered to generate task groups. Calculate the sum of latency performance deviations for all abnormal task instances within each task group; Configure a dedicated channel for each task group, and allocate the containers corresponding to the third number of containers from the total container resource pool to each dedicated channel based on the sum of the latency performance deviations of each task group. Each task group will be migrated to its corresponding dedicated channel for execution.

[0049] In a preferred embodiment, calculating the latency performance deviation of the current discrete micro-application for each target power distribution automation terminal task instance based on the actual calculation completion time, the preset latency requirement, and the computational complexity includes: Calculate the numerical difference between the actual calculation completion time and the preset delay requirement; The portion of the numerical difference that is greater than zero is determined as the effective delay duration; Calculate the ratio of the effective delay duration to the computational complexity, and determine the ratio as the delay performance deviation.

[0050] In a preferred embodiment, calculating the proportion of the computational complexity of the abnormal task instance to the computational complexity of the plurality of target distribution automation terminal task instances based on the computational complexity of the abnormal task instance includes: Based on the computational complexity of each abnormal task instance, determine the total abnormal computational load carried by the current discrete micro-application; Based on the computational complexity of several target distribution automation terminal task instances, determine the total computational load carried by the current discrete micro-application; Calculate the proportion of the total abnormal computing load in the total total computing load, and determine the proportion as the computing complexity ratio.

[0051] In a preferred embodiment, the step of clustering the abnormal task instances according to the latency requirements to generate task groups includes: Calculate the numerical proximity between the preset latency requirements of each abnormal task instance; Based on the numerical proximity, abnormal task instances that meet the proximity condition for preset latency requirements are grouped into the same task group.

[0052] Specifically, after the edge computing node of the power distribution network completes static resource configuration and is put into operation, in order to solve the performance interference problem caused by the mixed operation of multiple types of tasks in the shared container resource pool of discrete micro-applications, the system needs to implement real-time monitoring and dynamic splitting strategies.

[0053] First, the system acquires and extracts key indicators from task execution monitoring data. The distribution network edge computing node monitors the discrete micro-applications running in the shared container in real time. Since the discrete micro-applications may concurrently respond to multiple call requests from the target distribution automation terminal task instances within the same time period, the system uses a built-in resource monitoring agent and link tracer to extract the actual computation completion time corresponding to each discrete micro-application's computation processing in response to each target distribution automation terminal task instance. and computational complexity .in, This is the index of the task instance. The computational complexity... Specifically, this can be measured by the number of CPU cycles or instruction execution time consumed when the discrete micro-application processes the task instance, reflecting the actual consumption of computing resources by the task instance.

[0054] Secondly, the system calculates latency performance deviations and filters abnormal task instances. To quantitatively assess whether each task instance is affected by resource contention, the system combines the actual computation completion time... Preset latency requirements and the computational complexity Calculate the latency performance deviation of the discrete micro-application for each of the target power distribution automation terminal task instances. The specific calculation logic is shown in the following formula: During this calculation process, the system first calculates the actual calculation completion time. With preset latency requirements The numerical difference between them, i.e. Subsequently, the system determines the portion of the numerical difference that is greater than zero as the effective delay duration, i.e. This filters out normal task instances that meet the latency requirements, focusing only on timeout tasks. Finally, the system calculates the effective delay duration and the computational complexity. The ratio, and the ratio is determined as the time delay performance deviation. By normalizing by dividing by the computational complexity, the impact of the task's computational load on delay assessment can be eliminated, thus accurately identifying those disrupted tasks that have "low computational load but high delay." After calculation, the system will... The target power distribution automation terminal task instance that exceeds the preset abnormal threshold is marked as an abnormal task instance.

[0055] Next, the system performs total deviation accumulation and splitting trigger determination. The system calculates the latency performance deviation for all concurrently executed task instances within the discrete micro-application. Accumulate these values ​​to generate a latency performance deviation that measures the overall congestion level of the current discrete micro-applications. The calculation formula is as follows: When the time delay performance deviation and Greater than the preset total deviation threshold When this occurs, it indicates that the shared container containing the discrete micro-application has experienced severe performance interference and cannot meet the current concurrent processing requirements, and the system then triggers a dynamic splitting process.

[0056] Subsequently, the system calculates the computational complexity ratio. To determine how much load needs to be stripped from shared resources, the system sums the computational complexities of all instances marked as abnormal tasks, denoted as the total abnormal computational load; simultaneously, it sums the computational complexities of all currently concurrently processed target power distribution automation terminal task instances within the discrete micro-application, denoted as the total full computational load. The system calculates the proportion of the total abnormal computational load within the total full computational load and determines this proportion as the computational complexity ratio. The calculation formula is as follows: Next, the system calculates the number of third containers. The system bases this calculation on the computational complexity ratio μ and the total number of containers currently held by the discrete micro-application in the shared container (i.e., the second number of containers allocated in the preceding steps, denoted as μ). ), calculate the number of third containers to be allocated to the dedicated channels of the discrete micro-application. This step aims to allocate a corresponding proportion of computing resources from the total resource pool based on the percentage of abnormal load. The calculation formula is as follows: in This indicates rounding down. Ensure that at least one container is allocated.

[0057] Finally, the system performs task grouping and clustering, dedicated channel allocation, and task migration. To prevent tasks with different latency sensitivities from interfering with each other again in the new dedicated channel, the system clusters the abnormal task instances according to preset latency requirements. Specifically, the system calculates the numerical proximity between the preset latency requirements of each abnormal task instance, and groups abnormal task instances whose preset latency requirements meet proximity conditions (e.g., the same value or within the same set range) into the same task group (denoted as group). Subsequently, the system calculates the sum of latency performance deviations for all abnormal task instances contained within each task group. The system is based on the sum of the deviations of each of the task groups. Total deviation of all abnormal tasks The weight will be used to calculate the number of the third container. Assign a dedicated channel to each of the aforementioned task groups. Assign the first... Number of containers in each group The calculation is as follows: After the allocation is completed, the system calls the corresponding number of containers from the total container resource pool to build the dedicated channel, and modifies the traffic policy of each task group to point to the corresponding dedicated channel, thereby physically isolating the interfered services from the shared environment.

[0058] Through the above-mentioned dynamic monitoring and splitting mechanism, the embodiments of the present invention can perceive the performance bottleneck caused by the mixing of multiple types of services in real time during system operation, and dynamically open up an isolated dedicated operating environment for the affected key service flows based on the quantitative analysis of latency deviation and computational complexity. In this way, while ensuring the response speed of high-priority or low-latency services, the performance "noise" interference problem in the shared resource pool is effectively solved.

[0059] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0060] like Figure 2 As shown, an embodiment of the present invention provides a power distribution automation terminal task scheduling device based on a power distribution network edge computing node, including: a parameter acquisition module, a micro-application graph construction module, a micro-application chain integration module, a dedicated resource allocation module, and a public resource allocation module; The parameter acquisition module is used to acquire the total number of dedicated container resources and the total number of public container resources of the distribution network edge computing node, and to acquire the task decomposition configuration information and operating status parameters of the target distribution automation terminal task; the operating status parameters include the number of task calls of the target distribution automation terminal task and the number of concurrent calls of the micro-application vertex; the dedicated container resources and public container resources are allocated by the total container resource pool; The micro-application graph construction module is used to parse out multiple micro-applications contained in the target power distribution automation terminal task and the calling order between the micro-applications according to the task decomposition configuration information; map the parsed multiple micro-applications as micro-application vertices, and map the calling order as calling edges connecting the micro-application vertices to construct a micro-application graph. The micro-application chain integration module is used to traverse the call edges in the micro-application graph, count the frequency of each call edge to generate the edge redundancy between micro-applications; mark the call edges with edge redundancy greater than a preset redundancy threshold as key redundancy edges, and integrate the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains. The dedicated resource allocation module is used to calculate the number of first containers required by each indivisible micro-application chain based on the edge repetition degree of the key repeated edges and the total number of dedicated container resources, and allocate dedicated containers to each indivisible micro-application chain from the total container resource pool according to the number of first containers. The public resource allocation module is used to identify micro-applications that are not integrated into the indivisible micro-application chain in the micro-application graph as discrete micro-applications; calculate the weight of each discrete micro-application based on the number of task calls, the preset task priority, and the number of concurrent calls; calculate the number of second containers that should be allocated to each discrete micro-application based on the weight of each discrete micro-application and the total number of public container resources, and allocate public containers to each discrete micro-application from the total container resource pool according to the number of second containers.

[0061] It should be noted that the embodiments of the device described above correspond to the embodiments of the present invention described above, and can realize the distribution automation terminal task scheduling method based on the edge computing node of the distribution network as described above in any one of the present invention. Furthermore, the embodiments of the device described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort.

[0062] Based on the above-described method embodiments of the present invention, a corresponding embodiment of an electronic device is provided.

[0063] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power distribution automation terminal task scheduling method based on the edge computing node of the power distribution network as described in any one of the present invention, or the processor implements the functions of each module in the above-described device embodiments.

[0064] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device.

[0065] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0066] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0067] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0068] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments; Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is running, the device where the storage medium is located executes any of the above-described distribution automation terminal task scheduling methods based on distribution network edge computing nodes of the present invention.

[0069] The aforementioned storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0070] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0071] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for scheduling tasks in distribution automation terminals based on edge computing nodes in a distribution network, characterized in that, include: The system obtains the total number of dedicated container resources and the total number of public container resources of the edge computing nodes of the distribution network, and obtains the task decomposition configuration information and operating status parameters of the target distribution automation terminal task; the operating status parameters include the number of task calls of the target distribution automation terminal task and the number of concurrent calls of the micro-application vertex; the dedicated container resources and public container resources are allocated by the total container resource pool. Based on the task decomposition configuration information, the multiple micro-applications contained in the target power distribution automation terminal task and the calling order between the micro-applications are parsed out. The parsed micro-applications are mapped as micro-application vertices, and the call sequence is mapped as call edges connecting the micro-application vertices to construct a micro-application graph; Traverse the call edges in the micro-application graph, count the frequency of each call edge to generate the edge redundancy between micro-applications; mark the call edges with edge redundancy greater than a preset redundancy threshold as key redundancy edges, and integrate the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains. Based on the edge repetition degree of the key repeating edge and the total number of dedicated container resources, calculate the first number of containers required for each indivisible micro-application chain, and allocate dedicated containers to each indivisible micro-application chain from the total container resource pool according to the first number of containers. Micro-applications that are not integrated into the indivisible micro-application chain in the micro-application graph are identified as discrete micro-applications; the weight of each discrete micro-application is calculated based on the number of task calls, the preset task priority, and the number of concurrent calls. Based on the weight of each discrete micro-application and the total number of public container resources, calculate the number of second containers that should be allocated to each discrete micro-application, and allocate public containers to each discrete micro-application from the total container resource pool according to the number of second containers.

2. The task scheduling method for distribution automation terminals based on distribution network edge computing nodes as described in claim 1, characterized in that, The step of marking call edges with a redundancy greater than a preset redundancy threshold as key redundancy edges, and integrating the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains, includes: Traverse each call edge in the micro-application graph and compare the edge repetition of the current call edge with a preset repetition threshold; If the edge repetition degree of the current call edge is greater than the preset repetition threshold, the current call edge will be marked as a critical repetition edge. Based on the connectivity of all marked key repeating edges in the micro-application graph, search and identify the connected subgraphs formed by the key repeating edges; All micro-application vertices within each connected subgraph are identified as members of the same indivisible micro-application chain, and the connected subgraph is defined as a corresponding indivisible micro-application chain.

3. The task scheduling method for distribution automation terminals based on distribution network edge computing nodes as described in claim 2, characterized in that, The step of calculating the number of first containers required for each indivisible micro-application chain based on the edge repetition degree of the key repeating edges and the total number of dedicated container resources includes: Determine the weight of the edge redundancy of each indivisible micro-application chain as a percentage of the total redundancy of all key repeating edges; The total number of dedicated container resources is mapped to the baseline container quota of each indivisible micro-application chain according to the aforementioned proportion weight. The baseline container quota is numerically normalized and corrected based on a preset minimum resource threshold to obtain the first number of containers required for each indivisible micro-application chain.

4. The task scheduling method for distribution automation terminals based on distribution network edge computing nodes as described in claim 3, characterized in that, Also includes: Obtain task execution monitoring data of the edge computing nodes of the power distribution network concurrently executing several target power distribution automation terminal task instances; For each discrete micro-application, extract the actual calculation completion time and computational complexity of the current discrete micro-application when it performs calculation processing on the several target power distribution automation terminal task instances in response to the task execution monitoring data. Based on the actual calculation completion time, the preset latency requirement, and the computational complexity, calculate the latency performance deviation of the current discrete micro-application for each target power distribution automation terminal task instance, and record the target power distribution automation terminal task instance whose latency performance deviation is greater than the preset abnormal threshold as an abnormal task instance. The latency performance deviations are accumulated to generate the latency performance deviation sum of the current discrete micro-application; When the time delay performance deviation is greater than the preset total deviation threshold, the proportion of the computational complexity of the abnormal task instance to the computational complexity of the plurality of target power distribution automation terminal task instances is calculated based on the computational complexity of the abnormal task instance. Based on the computational complexity ratio and the total number of containers currently held by the discrete micro-application in the shared container, calculate the number of third containers to be allocated to the dedicated channel of the current discrete micro-application. Based on the latency requirements, the abnormal task instances are clustered to generate task groups. Calculate the sum of latency performance deviations for all abnormal task instances within each task group; Configure a dedicated channel for each task group, and allocate the containers corresponding to the third number of containers from the total container resource pool to each dedicated channel based on the sum of the latency performance deviations of each task group. Each task group will be migrated to its corresponding dedicated channel for execution.

5. The task scheduling method for distribution automation terminals based on distribution network edge computing nodes as described in claim 4, characterized in that, The step of calculating the latency performance deviation of the current discrete micro-application for each target power distribution automation terminal task instance based on the actual calculation completion time, the preset latency requirement, and the computational complexity includes: Calculate the numerical difference between the actual calculation completion time and the preset delay requirement; The portion of the numerical difference that is greater than zero is determined as the effective delay duration; Calculate the ratio of the effective delay duration to the computational complexity, and determine the ratio as the delay performance deviation.

6. The task scheduling method for distribution automation terminals based on distribution network edge computing nodes as described in claim 5, characterized in that, The step of calculating the proportion of the computational complexity of the abnormal task instance to the computational complexity of the plurality of target distribution automation terminal task instances based on the computational complexity of the abnormal task instance includes: Based on the computational complexity of each abnormal task instance, determine the total abnormal computational load carried by the current discrete micro-application; Based on the computational complexity of several target distribution automation terminal task instances, determine the total computational load carried by the current discrete micro-application; Calculate the proportion of the total abnormal computing load in the total total computing load, and determine the proportion as the computing complexity ratio.

7. The task scheduling method for distribution automation terminals based on distribution network edge computing nodes as described in claim 6, characterized in that, The step of clustering the abnormal task instances according to the latency requirements to generate task groups includes: Calculate the numerical proximity between the preset latency requirements of each abnormal task instance; Based on the numerical proximity, abnormal task instances that meet the proximity condition for preset latency requirements are grouped into the same task group.

8. A task scheduling device for distribution automation terminals based on distribution network edge computing nodes, characterized in that, include: The module includes a parameter acquisition module, a micro-application graph construction module, a micro-application chain integration module, a dedicated resource allocation module, and a public resource allocation module. The parameter acquisition module is used to acquire the total number of dedicated container resources and the total number of public container resources of the distribution network edge computing node, and to acquire the task decomposition configuration information and operating status parameters of the target distribution automation terminal task; the operating status parameters include the number of task calls of the target distribution automation terminal task and the number of concurrent calls of the micro-application vertex; the dedicated container resources and public container resources are allocated by the total container resource pool; The micro-application graph construction module is used to parse out the multiple micro-applications contained in the target power distribution automation terminal task and the calling order between the micro-applications based on the task decomposition configuration information. The parsed micro-applications are mapped as micro-application vertices, and the call sequence is mapped as call edges connecting the micro-application vertices to construct a micro-application graph; The micro-application chain integration module is used to traverse the call edges in the micro-application graph, count the frequency of each call edge to generate the edge redundancy between micro-applications; mark the call edges with edge redundancy greater than a preset redundancy threshold as key redundancy edges, and integrate the micro-applications connected by the key redundancy edges into corresponding indivisible micro-application chains. The dedicated resource allocation module is used to calculate the number of first containers required by each indivisible micro-application chain based on the edge repetition degree of the key repeated edges and the total number of dedicated container resources, and allocate dedicated containers to each indivisible micro-application chain from the total container resource pool according to the number of first containers. The public resource allocation module is used to identify micro-applications that are not integrated into the indivisible micro-application chain in the micro-application graph as discrete micro-applications; and to calculate the weight of each discrete micro-application based on the number of task calls, the preset task priority, and the number of concurrent calls. Based on the weight of each discrete micro-application and the total number of public container resources, calculate the number of second containers that should be allocated to each discrete micro-application, and allocate public containers to each discrete micro-application from the total container resource pool according to the number of second containers.

9. An electronic device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the distribution automation terminal task scheduling method based on the distribution network edge computing node as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is running, it controls the device where the storage medium is located to execute the power distribution automation terminal task scheduling method based on the edge computing node of the power distribution network as described in any one of claims 1 to 7.

Images