An electric power intelligent agent screening method, device, equipment, medium and product

Abstract

The application discloses a power intelligent entity screening method and device, equipment, medium and product, and relates to the technical field of artificial intelligence and power system intelligent technology. The method comprises the following steps: obtaining a target message of a power business task; analyzing and extracting task features from the target message; matching the task features and portrait data of each intelligent entity in an intelligent entity portrait pool based on an intelligent entity selection model to obtain at least one candidate intelligent entity; confirming the priority of the at least one candidate intelligent entity based on a dynamic weight to obtain the processing priority of the at least one candidate intelligent entity; determining a target intelligent entity from the at least one candidate intelligent entity according to task concurrency computing power and data processing scale and in combination with the processing priority; and automatically processing the power business task based on the target intelligent entity. Through the above technical solution, the power business task and the intelligent entity are accurately matched, and the utilization rate of cluster computing power resources is improved.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and intelligent power system technology, and in particular to a method, apparatus, equipment, medium and product for screening intelligent power agents. Background Technology

[0002] Currently, task allocation and agent selection in power multi-agent collaborative analysis systems generally adopt two traditional modes: manual assignment or fixed grouping. Maintenance personnel rely on manual experience to assign analysis tasks, or static agent groups are pre-defined at the beginning of the project to undertake single tasks. The system simply records task success or failure results, without establishing a unified agent capability profile, lacking a quantitative evaluation mechanism, and the selection process lacks visual monitoring and a task feedback iteration link.

[0003] Existing technologies suffer from several drawbacks: First, task allocation relies entirely on manual operation. When faced with batch tasks such as massive concurrent load forecasting of the power grid, fault assessment of distribution areas, and user electricity consumption analysis, scheduling efficiency is low, labor costs are high, and task backlog and scheduling delays are easily encountered, failing to meet the high timeliness requirements of real-time power monitoring and immediate assessment. Second, there is a lack of a unified and standardized intelligent agent management system. The computing power, business expertise, and historical performance of intelligent agents are not quantitatively recorded. Task matching with intelligent agents relies solely on subjective human judgment, resulting in low matching accuracy. High-performance intelligent agents are often idle, while low-performance intelligent agents handle complex big data tasks, leading to a serious waste of cluster computing resources. Third, the static fixed grouping mode has extremely poor flexibility. When new power businesses are added, emergency repair tasks are issued, or some intelligent agents fail and go offline, it is impossible to quickly reorganize collaborative units, causing task interruptions, a significant increase in analysis latency, and insufficient adaptability. Fourth, the selection process lacks a standardized quantitative scoring model, and the selection of intelligent agents lacks objective data support, resulting in poor decision interpretability. When analysis biases or task timeouts occur, it is difficult to trace the root cause of problems in the selection process, making fault diagnosis difficult.

[0004] In summary, the traditional intelligent agent selection and allocation model suffers from multiple technical bottlenecks, including strong reliance on manual intervention, imbalanced resource matching, insufficient scheduling flexibility, lack of quantitative evaluation, lack of visual supervision, and lack of autonomous iteration capabilities. These limitations make it unable to support the large-scale, high-concurrency, and multi-scenario adaptive collaborative analysis business needs of the next generation of power big data. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, medium, and product for screening power intelligent agents, in order to solve the problems of traditional power intelligent agent task allocation relying on manual operation, low efficiency in scheduling massive concurrent tasks, high labor costs, and insufficient real-time performance.

[0006] According to one aspect of the present invention, a method for screening power intelligent agents is provided, the method comprising: Obtain the target message for the power business task; wherein, the target message includes at least one of the following: monitoring of electricity consumption by elderly people living alone, load forecasting of distribution areas, analysis of distribution network faults, and statistics on housing vacancy rates; The target message is parsed and task features are extracted; wherein, the task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level; Based on the agent selection model, the task features are matched with the profile data of each agent in the agent profile pool to obtain at least one candidate agent. Based on dynamic weights, the priority of the at least one candidate agent is determined to obtain the processing priority of the at least one candidate agent. Based on the concurrent computing power and data processing scale of the task, and in conjunction with the processing priority, the target intelligent agent is determined from the at least one candidate intelligent agent; The power business tasks are processed automatically based on the target intelligent agent.

[0007] According to another aspect of the present invention, a power intelligent agent screening device is provided, the device comprising: The target message acquisition module is used to acquire the target message of the power business task; wherein, the target message includes at least one of the following: power consumption monitoring of elderly people living alone, load prediction of transformer area, distribution network fault analysis, and housing vacancy rate statistics; The feature extraction module is used to parse the target message and extract task features; wherein, the task features include at least one of the following: task type, data volume to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level. The candidate agent determination module is used to match the task features with the profile data of each agent in the agent profile pool based on the agent selection model to obtain at least one candidate agent. The priority determination module is used to determine the priority of the at least one candidate agent based on dynamic weights, and obtain the processing priority of the at least one candidate agent. The target agent determination module is used to determine the target agent from the at least one candidate agent based on the task's concurrent computing power and data processing scale, and in conjunction with the processing priority. The task processing module is used to automate the processing of the power business tasks based on the target intelligent agent.

[0008] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the power agent screening method according to any embodiment of the present invention.

[0009] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the power intelligent agent screening method according to any embodiment of the present invention.

[0010] According to another aspect of the present invention, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements the power intelligent agent screening method according to any embodiment of the present invention.

[0011] The technical solution of this invention involves acquiring target messages for power business tasks; these target messages include at least one of the following: monitoring electricity consumption by elderly people living alone, load forecasting for transformer areas, distribution network fault analysis, and housing vacancy rate statistics; parsing the target messages and extracting task features; these task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level; matching the task features with the profile data of each intelligent agent in the intelligent agent profile pool based on an intelligent agent selection model to obtain at least one candidate intelligent agent; prioritizing the at least one candidate intelligent agent based on dynamic weights to obtain the processing priority of the at least one candidate intelligent agent; determining the target intelligent agent from the at least one candidate intelligent agent based on the task's concurrent computing power and data processing scale, combined with the processing priority; and automating the processing of power business tasks based on the target intelligent agent. This technical solution fully automates the selection and allocation of intelligent agents for power business tasks, completely replacing the manual assignment mode, significantly improving the scheduling response speed in batch concurrent task scenarios, significantly reducing maintenance manpower costs, and fully meeting the high-timeliness business requirements of real-time power grid monitoring and immediate fault assessment.

[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart of a power intelligent agent screening method provided by an embodiment of the present invention; Figure 2 This is a flowchart of a power intelligent agent screening method provided by an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a power intelligent agent screening device according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device that implements the power intelligent agent screening method of this invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0016] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] Furthermore, it should be noted that the collection, storage, use, processing, transmission, provision, and disclosure of data related to power business tasks and intelligent agents involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0018] Figure 1This is a flowchart of a power agent screening method according to an embodiment of the present invention. This embodiment is applicable to situations where power tasks are intelligently allocated to agents. The method can be executed by a power agent screening device, which can be implemented in hardware and / or software. This device can be configured in an electronic device carrying power agent screening functionality, such as a power task processing system in a server. Figure 1 As shown, the method includes: S110, Obtain the target message for the power business task.

[0019] The target message includes at least one of the following: electricity consumption monitoring of elderly people living alone, load forecasting of distribution areas, distribution network fault analysis, and housing vacancy rate statistics.

[0020] Specifically, when power-related business tasks such as monitoring electricity consumption by elderly people living alone, forecasting load in transformer areas, analyzing distribution network faults, and compiling statistics on housing vacancy rates are issued, the target message of the power-related business task is obtained.

[0021] S120. Parse the target message and extract task features.

[0022] The task characteristics include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level.

[0023] Specifically, the system automatically parses task messages and extracts multi-dimensional parameters such as task type, data volume to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level. These multi-dimensional parameters are then structurally transformed to generate a structured task feature vector, which serves as the task feature.

[0024] S130. Based on the agent selection model, match the task features with the profile data of each agent in the agent profile pool to obtain at least one candidate agent.

[0025] The agent selection model is a multi-dimensional feature similarity weighted matching model for screening agents in the power big data market. It is a customized multi-dimensional scoring and screening model, a structured quantitative scoring model for power business, and consists of three layers: an input layer that receives dual-path standardized feature vector input; a matching calculation layer that fuses similarity operators and historical matching coefficients for scoring; and a filtering output layer that generates a set of candidate agents based on threshold filtering. It should be noted that the agents include agents for transformer area load prediction, distribution network fault assessment, electricity consumption monitoring for elderly people living alone, and transformer area vacancy rate statistics, etc. These are deployed online by the power system according to the analysis business needs. After deployment, they automatically collect data on computing power, business adaptation, performance, etc., to generate profiles and store them in a dynamic profile pool for the scheduling model to screen and match.

[0026] Specifically, the similarity between task features and the profile data of each agent in the agent profile pool is calculated based on the agent selection model, and at least one candidate agent is selected based on the similarity.

[0027] S140. Prioritize at least one candidate agent based on dynamic weights to obtain the processing priority of at least one candidate agent.

[0028] In this embodiment, dynamic weight refers to the priority weight used to adjust the processing tasks of the intelligent agent; optionally, dynamic weight includes business adaptability weight, historical performance weight, average response speed weight, remaining computing power weight, and historical success rate weight for matching similar tasks. Processing priority refers to the priority of the intelligent agent in processing tasks.

[0029] Specifically, for each candidate agent, its business adaptability, historical performance, average response speed, remaining computing power, and historical success rate of matching similar tasks are obtained. The business adaptability, historical performance, average response speed, remaining computing power, and historical success rate of matching similar tasks are weighted and summed with their corresponding weights to obtain an adaptation score. At least one candidate agent is ranked according to its adaptation score to obtain the corresponding processing priority.

[0030] S150. Based on the concurrent computing power and data processing scale of the task, and in combination with the processing priority, determine the target intelligent agent from at least one candidate intelligent agent.

[0031] The target intelligent agent refers to the intelligent agent that handles power business tasks, and can be one or more intelligent agents.

[0032] Based on the concurrent computing power and data processing scale of the task, one or more intelligent agents are selected in order of priority to form the optimal collaborative unit, thus obtaining the target intelligent agent.

[0033] S160. Automated processing of power business tasks based on target intelligent agents.

[0034] Specifically, power business tasks are automatically distributed to the target intelligent agent for automated processing.

[0035] The technical solution of this invention involves acquiring target messages for power business tasks; these target messages include at least one of the following: monitoring electricity consumption by elderly people living alone, load forecasting for transformer areas, distribution network fault analysis, and housing vacancy rate statistics; parsing the target messages and extracting task features; these task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level; matching the task features with the profile data of each intelligent agent in the intelligent agent profile pool based on an intelligent agent selection model to obtain at least one candidate intelligent agent; prioritizing the at least one candidate intelligent agent based on dynamic weights to obtain the processing priority of the at least one candidate intelligent agent; determining the target intelligent agent from the at least one candidate intelligent agent based on the task's concurrent computing power and data processing scale, combined with the processing priority; and automating the processing of power business tasks based on the target intelligent agent. This technical solution fully automates the selection and allocation of intelligent agents for power business tasks, completely replacing the manual assignment mode, significantly improving the scheduling response speed in batch concurrent task scenarios, significantly reducing maintenance manpower costs, and fully meeting the high-timeliness business requirements of real-time power grid monitoring and immediate fault assessment.

[0036] Based on the above embodiments, as an optional approach of the present invention, it further includes: dividing the profile tag system according to capability tags, performance tags and status tags, and using sharded storage nodes in a distributed cluster to store different types of tag data; automatically collecting the basic configuration, business expertise, computing power specifications and data parsing adaptation capabilities of the intelligent agent through the full-domain power business intelligent agent, generating a preliminary standardized profile of the intelligent agent, and storing it in the profile pool.

[0037] Specifically, a standardized profile tagging system is established, divided into three functional modules: capability tags, performance tags, and status tags. Tag fields are expandable and compatible with all types of power business intelligent agents, including load forecasting, equipment fault analysis, and user electricity consumption statistics. A distributed, sharded database cluster is used to deploy the profile pool, splitting it into multiple storage nodes to handle different types of tag data. A separate central scheduling and management node centrally manages profile read / write, incremental updates, global indexes, and historical snapshot storage, decoupling management and storage operations. This supports horizontal scaling and adapts to high-speed read / write of massive amounts of intelligent agent profile data. Secondly, all power business intelligent agents are fully integrated, automatically collecting basic configurations, business strengths, computing power specifications, and data parsing and adaptation capabilities to generate initial standardized profiles, which are then stored in the profile pool. Profile initialization is automatically completed when a new intelligent agent is launched, while offline intelligent agents retain historical profile snapshots, achieving unified management and control of data throughout the entire intelligent agent lifecycle.

[0038] Understandably, a modular, sharded, distributed intelligent agent dynamic profile pool is constructed, which stores multi-dimensional attributes of intelligent agents in a modular manner according to three major tags: capability, performance, and status. It is equipped with a profile incremental real-time update engine to uniformly manage the entire life cycle data of power intelligent agents across the entire domain. The distributed sharded architecture supports the horizontal expansion of massive intelligent agents, taking into account the read and write performance of massive profile data and storage scalability, and solving the problems of unquantified management of intelligent agent capabilities and data silos in traditional solutions.

[0039] Based on the above embodiments, as an optional aspect of the present invention, it further includes: real-time retrieval of portrait pool data, model adaptation scores, task scheduling logs, and agent runtime load data, and displaying them in chart form; automatically collecting multi-dimensional feedback indicator data after each task is completed by the agent; wherein, the feedback indicator data includes at least one of the following: analysis result accuracy, total task processing time, agent computing resource consumption, whether timeout or error occurred, data deviation degree, and matching effect of similar tasks; updating the agent's performance label, status label, or capability label in the portrait pool in real time according to the feedback indicator data; and automatically adjusting the dynamic weights and the similarity matching operator of the agent's model selection using a gradient optimization algorithm based on the batch feedback indicator data.

[0040] Specifically, firstly, real-time data from the profile pool, model adaptation scores, task scheduling logs, and agent runtime load data are retrieved. Agent-related data is displayed in multiple formats, including topology diagrams, bar charts, time-series curves, and detailed lists. Examples include all agent profile tags, candidate agent adaptation score rankings, final selected agent combinations, real-time task progress, agent CPU / memory load, and filtered anomaly alarm records. Historical filtering records are fully retained, and matching scores, weight parameters, and allocation results are traceable and interpretable throughout the process, facilitating quick location of scheduling anomalies by operations and maintenance personnel.

[0041] Understandably, this enables full-link visual monitoring of screening and scheduling, fully displaying profile data, task allocation, agent load, and historical records. Screening decisions are intuitive, verifiable, and fully explainable, significantly reducing the difficulty of troubleshooting and supporting manual intervention by operations and maintenance personnel.

[0042] Secondly, after identifying the completion of all single tasks, multi-dimensional feedback indicator data is automatically collected. This data includes the accuracy of the analysis results, total task processing time, agent computing resource consumption, timeout / error occurrence, data deviation degree, and matching effect of similar tasks. The feedback data is structured and stored as the input source for closed-loop optimization. Finally, a two-layer closed-loop iterative optimization mechanism is established: In the first layer, feedback performance indicators are synchronously written back to the dynamic profile pool, updating the corresponding agent performance tags and historical matching effect coefficients; in the second layer, the closed-loop optimization engine, based on batch task feedback data, automatically adjusts the state weights and the similarity matching operator of the agent selection model using a gradient optimization algorithm.

[0043] Understandably, by building a two-layer closed-loop self-optimization system, the first layer dynamically updates the performance data of the intelligent agent profile based on task feedback; the second layer iteratively optimizes the dynamic weight rules and matching calculation model based on the batch execution effect, and the screening and matching accuracy improves autonomously and steadily over a long period of operation. For business scenarios with consistently low matching accuracy, the system autonomously optimizes the weight adjustment logic, continuously improving screening and matching accuracy, thus achieving self-learning and self-optimization.

[0044] Figure 2 This is a flowchart of a power intelligent agent screening method according to an embodiment of the present invention. Based on the above embodiment, this embodiment further optimizes the following steps: "matching task features with the profile data of each intelligent agent in the intelligent agent profile pool based on an intelligent agent selection model to obtain at least one candidate intelligent agent" and "prioritizing at least one candidate intelligent agent based on dynamic weights to obtain the processing priority of at least one candidate intelligent agent," providing an optional implementation method. Figure 2 As shown, the method includes: S210, Obtain the target message for the power business task.

[0045] The target message includes at least one of the following: electricity consumption monitoring of elderly people living alone, load forecasting of transformer areas, distribution network fault analysis, and housing vacancy rate statistics; S220. Parse the target message and extract task features.

[0046] The task characteristics include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level. S230. Based on the agent selection model, match the task features with the profile data of each agent in the agent profile pool to obtain at least one candidate agent.

[0047] S240. Prioritize at least one candidate agent based on dynamic weights to obtain the processing priority of at least one candidate agent.

[0048] S250. Based on the concurrent computing power of the task and the scale of data processing, and in combination with the processing priority, determine the target intelligent agent from at least one candidate intelligent agent.

[0049] S260. Automated processing of power business tasks based on target intelligent agents.

[0050] An alternative approach involves matching task features with the profile data of each agent in the agent profile pool based on an agent selection model to obtain at least one candidate agent. This includes: calculating the similarity between task features and the profile data of each agent in the agent profile pool using a similarity matching operator in the agent selection model to obtain the profile similarity between the power business task and the agent; adding the profile similarity to the agent's historical matching coefficients to obtain the agent's basic adaptation score; wherein the historical matching coefficients include the average accuracy and average latency of the agent in undertaking historical tasks of the same type as the power business task; and filtering agents based on an adaptation threshold and the basic adaptation score to obtain at least one candidate agent.

[0051] Specifically, based on the similarity matching operator built into the agent selection model, similarity is calculated between the task features and the profile data of each agent in the agent profile pool to obtain the profile similarity between the power business task and the agent. Then, the profile similarity is added to the agent's historical matching coefficient to obtain the agent's basic fit score. Finally, agents with a basic fit score less than the fit threshold are removed, along with invalid agents that do not meet the task requirements in terms of computing power, have mismatched permissions, or are offline, resulting in at least one candidate agent. This reduces the scope of subsequent scheduling calculations and lowers the model's computing power consumption.

[0052] Understandably, by introducing historical matching effect coefficients to calculate the basic adaptation score and setting compliance thresholds to filter invalid agents, the standardization and quantitative evaluation of agent screening can be achieved, completely replacing traditional human experience judgment and significantly improving the accuracy of task-agent matching.

[0053] An optional approach involves prioritizing at least one candidate agent based on dynamic weights to obtain the processing priority of at least one candidate agent. This includes: adjusting the dynamic weights according to real-time business information; wherein the dynamic weights include business adaptability weights, historical performance weights, average response speed weights, remaining computing power weights, and historical success rate weights for matching similar tasks; weighting and correcting the basic adaptation scores of the candidate agents based on the dynamic weights to obtain a comprehensive adaptation score; and prioritizing the candidate agents according to their comprehensive adaptation scores from high to low to obtain the processing priority of at least one candidate agent.

[0054] First, it should be noted that each candidate intelligent agent has a basic adaptation score based on its business adaptability, historical performance, average response speed, remaining computing power, and historical success rate in matching similar tasks.

[0055] Specifically, the dynamic weights are adjusted based on real-time business information. Then, based on the adjusted dynamic weights, the basic adaptation scores of the candidate agents are weighted and corrected to obtain a comprehensive adaptation score. The candidate agents are prioritized according to the comprehensive adaptation scores from high to low to obtain the processing priority of at least one candidate agent.

[0056] For example, adjusting dynamic weights based on real-time business information includes: determining the business scenario, judging the task urgency, and detecting the cluster load to obtain the global cluster load based on the real-time business information; and adjusting the dynamic weights based on the business scenario, the task urgency, and the global cluster load.

[0057] The weights of business adaptability, historical performance, average response speed, remaining computing power, and historical success rate of matching similar tasks are added together to form a total of 1.

[0058] Specifically, real-time business information is identified to determine the business scenario and task urgency. Simultaneously, the cluster's computing power utilization is detected to obtain the overall cluster load. If the business scenario is a high real-time load prediction scenario, the weight of average response speed is increased; if the business scenario is a complex fault assessment scenario, the weight of business adaptability is increased; if the task urgency is an emergency early warning task, the weight of remaining computing power is increased; if the task urgency is a routine statistical task, the weight of historical performance is increased; and if the overall cluster load is high, the weight of remaining computing power is increased.

[0059] Understandably, by combining business scenarios, task urgency, and cluster global load in real time to adjust multi-dimensional weight coefficients and dynamically correct the intelligent agent's adaptation score, and by implementing differentiated scheduling for different power businesses, the shortcomings of fixed grouping mode—insufficient flexibility and unbalanced allocation of computing resources—are completely resolved.

[0060] The technical solution of this invention involves acquiring target messages for power business tasks; these target messages include at least one of the following: monitoring electricity consumption by elderly people living alone, load forecasting for transformer areas, distribution network fault analysis, and housing vacancy rate statistics; parsing the target messages and extracting task features; these task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level; matching the task features with the profile data of each intelligent agent in the intelligent agent profile pool based on an intelligent agent selection model to obtain at least one candidate intelligent agent; determining the priority of at least one candidate intelligent agent based on dynamic weights to obtain the processing priority of at least one candidate intelligent agent; determining the target intelligent agent from at least one candidate intelligent agent based on the task's concurrent computing power and data processing scale, combined with the processing priority; and automating the processing of power business tasks based on the target intelligent agent. This technical solution fully automates the selection and allocation of intelligent agents for power business tasks, completely replacing the manual assignment mode, significantly improving the scheduling response speed in batch concurrent task scenarios, significantly reducing maintenance manpower costs, and fully meeting the high-timeliness business requirements of real-time power grid monitoring and immediate fault assessment.

[0061] This invention automates the selection and allocation of intelligent agents for power business tasks, completely replacing manual assignment. It significantly improves scheduling response speed in batch concurrent task scenarios, substantially reduces maintenance manpower costs, and fully meets the high-efficiency business requirements of real-time power grid monitoring and immediate fault assessment. A distributed dynamic profile pool enables standardized and quantitative management of the capabilities, performance, and status of all intelligent agents. Relying on a quantitative matching model, it accurately matches tasks with intelligent agents, significantly improving the utilization rate of cluster computing resources and avoiding resource waste caused by idle high-performance intelligent agents and low-performance intelligent agents handling complex tasks. Furthermore, the dynamic weighted adaptive scheduling mechanism abandons static fixed grouping and can quickly reorganize the optimal intelligent agent collaboration unit based on new business, urgent tasks, and offline intelligent agent failures, adapting to diverse power business scenarios such as load forecasting, user analysis, and equipment maintenance, greatly enhancing scheduling flexibility. Moreover, the visual monitoring module fully displays the entire intelligent agent selection process data, with matching scores, allocation criteria, and intelligent agent loads clearly visible. The selection decision has complete interpretability, and faults such as task timeouts and analysis deviations can be quickly traced back to their root causes in the selection process, significantly improving maintenance and troubleshooting efficiency. Furthermore, the overall architecture adopts a distributed design, and both the portrait pool and the intelligent agent cluster can be horizontally expanded. It can support the concurrent scheduling of massive intelligent agents across the entire provincial power grid and millions of daily analysis tasks, and is suitable for large-scale collaborative analysis scenarios of power big data.

[0062] Figure 3This is a schematic diagram of a power intelligent agent screening device according to an embodiment of the present invention. This embodiment is applicable to situations where power tasks are intelligently allocated by intelligent agents. The power intelligent agent screening device can be implemented in hardware and / or software, and can be configured in an electronic device that carries the power intelligent agent screening function, such as a server. Figure 3 As shown, the device includes: The target message acquisition module 310 is used to acquire the target message of the power business task; wherein, the target message includes at least one of the following: power consumption monitoring of elderly people living alone, load forecasting of transformer area, distribution network fault analysis, and housing vacancy rate statistics; The feature extraction module 320 is used to parse the target message and extract task features; among which, the task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level. The candidate agent determination module 330 is used to match task features with the profile data of each agent in the agent profile pool based on the agent selection model to obtain at least one candidate agent. The priority determination module 340 is used to determine the priority of at least one candidate agent based on dynamic weights, so as to obtain the processing priority of at least one candidate agent. The target agent determination module 350 is used to determine the target agent from at least one candidate agent based on the task's concurrent computing power and data processing scale, and in combination with processing priority. The task processing module 360 ​​is used to automate the processing of power business tasks based on the target intelligent agent.

[0063] The technical solution of this invention involves acquiring target messages for power business tasks; these target messages include at least one of the following: monitoring electricity consumption by elderly people living alone, load forecasting for transformer areas, distribution network fault analysis, and housing vacancy rate statistics; parsing the target messages and extracting task features; these task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level; matching the task features with the profile data of each intelligent agent in the intelligent agent profile pool based on an intelligent agent selection model to obtain at least one candidate intelligent agent; determining the priority of at least one candidate intelligent agent based on dynamic weights to obtain the processing priority of at least one candidate intelligent agent; determining the target intelligent agent from at least one candidate intelligent agent based on the task's concurrent computing power and data processing scale, combined with the processing priority; and automating the processing of power business tasks based on the target intelligent agent. This technical solution fully automates the selection and allocation of intelligent agents for power business tasks, completely replacing the manual assignment mode, significantly improving the scheduling response speed in batch concurrent task scenarios, significantly reducing maintenance manpower costs, and fully meeting the high-timeliness business requirements of real-time power grid monitoring and immediate fault assessment.

[0064] Optionally, the candidate agent determination module 330 is used for: Based on the similarity matching operator in the agent selection model, the similarity between the task features and the portrait data of each agent in the agent portrait pool is calculated to obtain the portrait similarity between the power business task and the agent. The basic adaptation score of the agent is obtained by adding the portrait similarity score and the historical matching coefficient of the agent; among which, the historical matching coefficient includes the average accuracy and average latency of the agent in undertaking historical tasks of the same type as the power business task. Based on the adaptation threshold and the basic adaptation score, the agents are screened to obtain at least one candidate agent.

[0065] Optionally, the priority determination module 340 is used for: The dynamic weights are adjusted based on real-time business information. These dynamic weights include business adaptability weights, historical performance weights, average response speed weights, remaining computing power weights, and historical success rates of matching similar tasks. The basic adaptation scores of candidate agents are weighted and corrected based on dynamic weights to obtain a comprehensive adaptation score. Candidate agents are prioritized according to their comprehensive adaptation scores from high to low to obtain the processing priority of at least one candidate agent.

[0066] Optionally, the priority determination module 340 is specifically used for: Based on real-time business information, determine the business scenario, assess the urgency of tasks, and detect cluster load to obtain the global cluster load. The dynamic weights are adjusted based on the business scenario, task urgency, and global cluster load.

[0067] Optionally, the device also includes an image pool construction module for: The profile tag system is divided into capability tags, performance tags, and status tags, and different types of tag data are stored in sharded storage nodes in a distributed cluster. The intelligent agent for the entire power business automatically collects the agent's basic configuration, business expertise, computing power specifications, and data parsing and adaptation capabilities to generate a preliminary standardized profile of the agent, which is then stored in the profile pool.

[0068] Optionally, the device also includes a visualization and update module for: Real-time retrieval of profile pool data, model adaptation scores, task scheduling logs, and agent runtime load data, and display in chart form; After each task is completed by the agent, multi-dimensional feedback indicator data is automatically collected. The feedback indicator data includes at least one of the following: accuracy of analysis results, total processing time of the task, consumption of computing resources of the agent, whether timeout or error occurred, degree of data deviation, and matching effect of similar tasks. The performance tags, status tags, or capability tags of agents in the profile pool are updated in real time based on feedback indicator data. Based on batch feedback index data, a gradient optimization algorithm is used to automatically adjust the dynamic weights and the similarity matching operator of the agent selection model.

[0069] The power intelligent agent screening device provided in the embodiments of the present invention can execute the power intelligent agent screening method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0070] According to embodiments of the present invention, the present invention also provides an electronic device, a readable storage medium, and a computer program product.

[0071] Figure 4 This is a schematic diagram of the structure of an electronic device that implements the power intelligent agent screening method of this invention. Figure 4A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0072] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0073] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0074] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the power agent screening method.

[0075] In some embodiments, the power agent screening method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the power agent screening method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the power agent screening method by any other suitable means (e.g., by means of firmware).

[0076] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include: implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0077] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0078] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0079] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0080] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0081] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0082] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0083] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for screening power intelligent agents, characterized in that, include: Obtain the target message for the power business task; wherein, the target message includes at least one of the following: monitoring of electricity consumption by elderly people living alone, load forecasting of distribution areas, analysis of distribution network faults, and statistics on housing vacancy rates; The target message is parsed and task features are extracted; wherein, the task features include at least one of the following: task type, volume of data to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level. Based on the agent selection model, the task features are matched with the profile data of each agent in the agent profile pool to obtain at least one candidate agent. Based on dynamic weights, the priority of the at least one candidate agent is determined to obtain the processing priority of the at least one candidate agent. Based on the concurrent computing power and data processing scale of the task, and in conjunction with the processing priority, the target intelligent agent is determined from the at least one candidate intelligent agent; The power business tasks are processed automatically based on the target intelligent agent.

2. The method according to claim 1, characterized in that, Based on the agent selection model, the task features are matched with the profile data of each agent in the agent profile pool to obtain at least one candidate agent, including: Based on the similarity matching operator in the agent selection model, the similarity between the task features and the portrait data of each agent in the agent portrait pool is calculated to obtain the portrait similarity between the power business task and the agent. The basic adaptation score of the agent is obtained by adding the portrait similarity score and the historical matching coefficient of the agent; wherein, the historical matching coefficient includes the average accuracy and average latency of the agent in undertaking historical tasks of the same type as the power business task. Based on the adaptation threshold and the basic adaptation score, the agents are screened to obtain at least one candidate agent.

3. The method according to claim 2, characterized in that, Prioritizing the at least one candidate agent based on dynamic weights to obtain the processing priority of the at least one candidate agent includes: The dynamic weights are adjusted based on real-time business information; the dynamic weights include business adaptability weights, historical performance weights, average response speed weights, remaining computing power weights, and historical success rate weights for matching similar tasks. Based on the dynamic weights, the basic adaptation scores of the candidate agents are weighted and corrected to obtain the comprehensive adaptation score. Candidate agents are prioritized according to their comprehensive adaptation scores from high to low to obtain the processing priority of at least one candidate agent.

4. The method according to claim 3, characterized in that, Adjust dynamic weights based on real-time business information, including: Based on real-time business information, determine the business scenario, assess the urgency of tasks, and detect cluster load to obtain the global cluster load. The dynamic weights are adjusted based on the business scenario, the task urgency, and the cluster global load.

5. The method according to claim 1, characterized in that, Also includes: The profile tag system is divided into capability tags, performance tags, and status tags, and different types of tag data are stored in sharded storage nodes in a distributed cluster. The intelligent agent for the entire power business automatically collects the agent's basic configuration, business expertise, computing power specifications, and data parsing and adaptation capabilities to generate a preliminary standardized profile of the agent, which is then stored in the profile pool.

6. The method according to claim 1, characterized in that, Also includes: Real-time retrieval of portrait pool data, model adaptation scores, task scheduling logs, and agent runtime load data, and display in chart form; After each task is completed by the agent, multi-dimensional feedback indicator data is automatically collected. The feedback indicator data includes at least one of the following: accuracy of analysis results, total processing time of the task, consumption of computing resources of the agent, whether timeout or error occurred, degree of data deviation, and matching effect of similar tasks. The performance tags, status tags, or capability tags of the agents in the profile pool are updated in real time based on the feedback indicator data. Based on batch feedback index data, a gradient optimization algorithm is used to automatically adjust the dynamic weights and the similarity matching operator of the agent selection model.

7. A power intelligent agent screening device, characterized in that, include: The target message acquisition module is used to acquire the target message of the power business task; wherein, the target message includes at least one of the following: power consumption monitoring of elderly people living alone, load prediction of transformer area, distribution network fault analysis, and housing vacancy rate statistics; The feature extraction module is used to parse the target message and extract task features; wherein, the task features include at least one of the following: task type, data volume to be processed, difficulty level, real-time timeliness threshold, minimum computing power requirement, business urgency, and access security permission level. The candidate agent determination module is used to match the task features with the profile data of each agent in the agent profile pool based on the agent selection model to obtain at least one candidate agent. The priority determination module is used to determine the priority of the at least one candidate agent based on dynamic weights, and to obtain the processing priority of the at least one candidate agent. The target agent determination module is used to determine the target agent from the at least one candidate agent based on the task's concurrent computing power and data processing scale, and in conjunction with the processing priority. The task processing module is used to automate the processing of the power business tasks based on the target intelligent agent.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the power agent screening method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the power agent screening method according to any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the power intelligent agent screening method according to any one of claims 1-6.

Images