Method and system for operation and maintenance task allocation assisted by agent intelligent agent

Abstract

The application provides an operation and maintenance task allocation method and system assisted by an agent intelligent body, first, an agent intelligent body cooperation network is established; then, a to-be-allocated operation and maintenance task demand set containing an execution target, resource demand and constraint condition is acquired; then, each agent intelligent body in the agent intelligent body cooperation network carries out interactive evaluation on the task demand set, and an adaptability evaluation result of the task and the agent intelligent body is determined; a preliminary operation and maintenance task allocation plan is generated by a coordination agent intelligent body according to the evaluation result; finally, a verification agent intelligent body carries out feasibility verification on the preliminary plan, and obtains a final operation and maintenance task allocation scheme after confirming that all constraint conditions are met, and outputs the final operation and maintenance task allocation scheme to an operation and maintenance management system. The application can realize efficient and reasonable allocation of smart home operation and maintenance tasks, and improve operation and maintenance efficiency and system stability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of operation and maintenance, in particular to an operation and maintenance task allocation method and system assisted by an Agent intelligent agent. BACKGROUND

[0002] At present, the number of smart devices in the family is increasing with the vigorous development of the Internet of Things smart home, covering smart security, smart lighting, smart home appliances, smart environmental monitoring and other categories. These devices together build a complex and interconnected smart home system. With the expansion of system size and the increasing complexity of functions, operation and maintenance tasks have become heavy and diversified, including device troubleshooting and repair, system software update, device configuration adjustment, security vulnerability repair, etc.

[0003] The traditional smart home operation and maintenance task allocation method has many drawbacks. On the one hand, it mainly relies on manual task allocation, and operation and maintenance personnel need to manually allocate tasks to appropriate execution personnel or devices based on their own experience and understanding of the system. However, due to the complexity and dynamics of the smart home system, manual allocation is difficult to fully consider various factors such as the real-time state of the device, the skill level and workload of the operation and maintenance personnel, and the dependency relationship between tasks, which can lead to unreasonable task allocation and cause some operation and maintenance personnel to be overburdened while others are idle, affecting operation and maintenance efficiency. On the other hand, existing automated task allocation methods often lack intelligence and flexibility, and are usually based on fixed rules or simple algorithms for allocation, and cannot dynamically adjust according to real-time changes and sudden situations in the smart home system. For example, when a critical device suddenly fails and needs to be handled urgently, the traditional method may not be able to reallocate tasks in time to prioritize the repair of the faulty device, thereby affecting the normal operation of the entire smart home system. In addition, the traditional method is also not up to the task in handling multi-task concurrency and resource conflict problems, and is difficult to achieve optimal allocation of operation and maintenance resources, which cannot meet the needs of modern smart home systems for efficient and stable operation and maintenance. SUMMARY

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present application, the embodiments of the present application provide an operation and maintenance task allocation method assisted by an Agent intelligent agent, which comprises:

[0005] Establishing an Agent intelligent agent collaboration network;

[0006] Obtaining a set of operation and maintenance task requirements to be allocated, which includes the execution target, resource requirements and constraint conditions of the operation and maintenance task;

[0007] The task adaptability evaluation results of each operation and maintenance task and each agent are determined by interaction evaluation of the operation and maintenance task demand set by each agent in the agent collaboration network.

[0008] According to the task adaptability evaluation results, a preliminary operation and maintenance task allocation plan is generated by a coordination agent in the agent collaboration network.

[0009] The feasibility of the preliminary operation and maintenance task allocation plan is verified, and the plan is confirmed to meet all constraint conditions in the operation and maintenance task demand set by a verification agent in the agent collaboration network, so that a final operation and maintenance task allocation scheme is obtained and output to the operation and maintenance management system.

[0010] In another aspect, the embodiment of the present application also provides an operation and maintenance task allocation system assisted by an agent, which comprises a processor and a machine readable storage medium, the machine readable storage medium is connected with the processor, the machine readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine readable storage medium to realize the above-mentioned method.

[0011] Based on the above aspects, by establishing an agent collaboration network, each agent in the agent collaboration network has autonomous perception, interaction and decision-making capabilities, can obtain state information and operation and maintenance task demands of the smart home system in real time, obtain an operation and maintenance task demand set containing execution targets, resource demands and constraint conditions, accurately define the specific requirements and restrictions of each operation and maintenance task, and make the task allocation more targeted and reasonable. Each agent in the agent collaboration network evaluates the operation and maintenance task demand set, fully utilizes the professional knowledge and analysis capabilities of each agent, comprehensively evaluates the adaptability of the task and the agent from multiple dimensions, ensures the accuracy and comprehensiveness of the evaluation results, and can effectively avoid the task allocation deviation caused by a single evaluation method. The coordination agent generates a preliminary operation and maintenance task allocation plan according to the task adaptability evaluation results, realizes the preliminary planning and coordination of task allocation, and improves the efficiency and rationality of task allocation. The verification agent verifies the feasibility of the preliminary plan, ensures that the plan meets all constraint conditions in the operation and maintenance task demand set, avoids problems such as resource conflict and task execution failure caused by unreasonable task allocation, and guarantees the smooth execution of operation and maintenance tasks. The final operation and maintenance task allocation scheme can fully consider the complexity and dynamics of the smart home system, realize the optimal allocation of operation and maintenance resources and efficient execution of tasks, effectively improve the operation and maintenance efficiency and quality of the smart home system, and provide more stable and reliable smart home services for users. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 is an execution flow schematic diagram of the operation and maintenance task allocation method assisted by the Agent intelligent agent provided by an embodiment of the present application.

[0013] Figure 2 is a schematic diagram of exemplary hardware and software components of the operation and maintenance task allocation system assisted by the Agent intelligent agent provided by an embodiment of the present application. DETAILED DESCRIPTION

[0014] The present application will be described in detail below with reference to the accompanying drawings of the specification, Figure 1 is a flow schematic diagram of the operation and maintenance task allocation method assisted by the Agent intelligent agent provided by an embodiment of the present application. The operation and maintenance task allocation method assisted by the Agent intelligent agent will be described in detail below.

[0015] Step S110: Establish an Agent intelligent agent collaboration network.

[0016] In this embodiment, the Agent intelligent agent collaboration network is established taking the operation and maintenance task allocation of the Internet of Things smart home system as an application scenario. The Agent intelligent agent collaboration network needs to integrate various types of Agent intelligent agents, clearly define the interaction rules, communication links and collaboration priorities among the intelligent agents, and ensure that the allocation of various operation and maintenance tasks in the smart home system, such as the allocation of tasks such as smart lamp fault repair, smart temperature control system upgrade, and smart door lock maintenance, can be efficiently and collaboratively completed.

[0017] Step S111: Determine the Agent intelligent agent types participating in the operation and maintenance task allocation, including task analysis Agent intelligent agents, resource scheduling Agent intelligent agents, capability evaluation Agent intelligent agents, coordination Agent intelligent agents, and verification Agent intelligent agents, to obtain an Agent intelligent agent type set.

[0018] In combination with the characteristics of the operation and maintenance tasks of the Internet of Things smart home system, the Agent intelligent agent types participating in the operation and maintenance task allocation are determined. The task analysis Agent intelligent agents are responsible for disassembling and classifying the received operation and maintenance task requirements; the resource scheduling Agent intelligent agents are responsible for overall planning of the tools, personnel and other resources required for operation and maintenance; the capability evaluation Agent intelligent agents are responsible for evaluating the capability adaptability of each Agent intelligent agent for executing tasks; the coordination Agent intelligent agents are responsible for coordinating the work of each Agent intelligent agent and generating a task allocation plan; and the verification Agent intelligent agents are responsible for verifying the feasibility of the allocation plan. The above five Agent intelligent agent types are organized to form an Agent intelligent agent type set.

[0019] Step S112: Based on the set of Agent types, define the interaction rules between Agent types, including information transmission format, request response time limit, data sharing range, to obtain a set of Agent interaction rules.

[0020] Based on the set of Agent types, define the interaction rules between Agent types. In terms of information transmission format, it is stipulated that the operation and maintenance task information transmitted between all Agent types needs to adopt a structured data format, including fixed fields such as task identification, task type, and requirement details. In terms of request response time limit, it is stipulated that after the task analysis Agent sends task information to the capability evaluation Agent, the capability evaluation Agent needs to feedback the evaluation results within a specified time limit; after the coordination Agent sends the allocation plan to the verification Agent, the verification Agent needs to complete the verification and reply within a specified time limit. In terms of data sharing range, it is stipulated that the task analysis Agent can share task disassembly results with all other Agent types, and the resource scheduling Agent only shares resource state data with the coordination Agent and the capability evaluation Agent, to avoid irrelevant data leakage. The above rules are integrated to form a set of Agent interaction rules.

[0021] Step S113: According to the set of Agent interaction rules, construct the communication links between Agent types, to ensure that different types of Agent can transmit operation and maintenance task related information in real time, to obtain a set of Agent communication links.

[0022] According to the set of Agent interaction rules, construct the communication links between Agent types. A distributed communication architecture is adopted to establish a dedicated communication channel for each pair of Agent types that have interaction needs, for example, a bidirectional communication link is established between the task analysis Agent and the capability evaluation Agent, and a multi-directional communication link is established between the coordination Agent and all other types of Agent. Each communication link is configured with a data encryption transmission mechanism to ensure the security of operation and maintenance task information during transmission. At the same time, a state monitoring module is set for the communication link to monitor the connectivity and data transmission rate of the link in real time, and when the link is interrupted or congested, it automatically switches to a backup communication link. All constructed communication links and their configuration information are integrated to form a set of Agent communication links.

[0023] Step S114: based on the set of agent types and the set of agent communication links, setting the cooperation priority of each agent in the cooperation process, determining the decision weight of each agent in different scenarios, and obtaining an agent cooperation priority list.

[0024] In combination with the set of agent types and the set of agent communication links, the cooperation priority and decision weight of each agent are set according to different scenarios of the Internet of Things smart home system operation and maintenance. For example, in the emergency fault repair scenario of the smart door lock, the decision weight of the coordination agent and the capability evaluation agent is higher; in the regular maintenance scenario of the smart temperature control system, the decision weight of the task analysis agent and the resource scheduling agent is relatively higher. The agent priority and decision weight in different scenarios are sorted to form an agent cooperation priority list.

[0025] Step S1141: the coordination agent collects common scenario types of operation and maintenance task allocation, the scenario types include emergency fault repair scenario, regular maintenance scenario, system upgrade scenario and resource shortage scenario, and obtains a set of operation and maintenance task scenario types.

[0026] The coordination agent collects common scenario types of operation and maintenance task allocation by calling the historical operation and maintenance records of the Internet of Things smart home system. The emergency fault repair scenario refers to a scenario in which a smart device suddenly fails and affects normal use of the user, such as a smart door lock that cannot be unlocked, a smart lamp that suddenly goes out, etc.; the regular maintenance scenario refers to a scenario in which a smart device is regularly checked and maintained according to a plan, such as regular debugging of a smart camera, safety detection of a smart socket, etc.; the system upgrade scenario refers to a scenario in which the software or firmware of the smart home system is updated, such as system upgrade of a smart gateway, firmware update of a smart speaker, etc.; the resource shortage scenario refers to a scenario in which the operation and maintenance tools, personnel, etc. are insufficient, such as a scenario in which smart devices in multiple regions need to be maintained at the same time and tools are limited. The above scenario types are sorted to form a set of operation and maintenance task scenario types.

[0027] Step S1142: for each scenario type in the set of operation and maintenance task scenario types, the coordination agent analyzes the core capability required to complete the operation and maintenance task in the corresponding operation and maintenance task scenario, determines the agent type that has a guiding role in the corresponding operation and maintenance task, and obtains a scenario-agent key association result.

[0028] For each scene type in the set of operation and maintenance task scene types, the coordination agent analyzes the core capabilities required to complete the operation and maintenance task in the scene. In the emergency fault repair scene, the core capabilities are rapid assessment of agent capabilities, efficient coordination of resources and task allocation, so the agent types with a guiding role are coordination agent and capability evaluation agent; in the regular maintenance scene, the core capabilities are accurate task decomposition and reasonable resource scheduling, so the agent types with a guiding role are task analysis agent and resource scheduling agent; in the system upgrade scene, the core capabilities are accurate analysis of upgrade requirements and verification of the feasibility of the upgrade scheme, so the agent types with a guiding role are task analysis agent and verification agent; in the resource shortage scene, the core capabilities are optimization of resource allocation and coordination of task priority, so the agent types with a guiding role are coordination agent and resource scheduling agent. The association between the scene type and the corresponding key agent type is sorted to form a scene-agent key association result.

[0029] Step S1143: Based on the scene-agent key association result, set a decision weight for each agent type in each scene type, the decision weight of the key agent type is higher than that of the non-key agent type, and the decision weight is used to determine the opinion influence of each agent in the collaborative decision-making process.

[0030] Based on the scene-agent key association result, set a decision weight for each agent type in each scene type. In the emergency fault repair scene, the coordination agent and the capability evaluation agent are key types, and their decision weights are set to the same high value; the task analysis agent, the resource scheduling agent and the verification agent are non-key types, and their decision weights are set to the same low value; in the regular maintenance scene, the decision weights of the task analysis agent and the resource scheduling agent are set to a high value, and the decision weights of the other three types are set to a low value; in the system upgrade scene, the decision weights of the task analysis agent and the verification agent are set to a high value, and the decision weights of the other three types are set to a low value; in the resource shortage scene, the decision weights of the coordination agent and the resource scheduling agent are set to a high value, and the decision weights of the other three types are set to a low value. The high and low of the decision weight directly determines the influence degree of the opinion proposed by each agent in the collaborative decision-making on the final result.

[0031] Step S1144: The coordination Agent entity associates the Agent entity type and the corresponding decision weight under each scene type to form a scene-agent decision weight entry in combination with the Agent entity type set.

[0032] The coordination Agent entity associates each scene type with all Agent entity types and the corresponding decision weight under the scene according to the Agent entity type set. For example, the scene-agent decision weight entry of the emergency fault repair scene contains the decision weight of each of the task analysis Agent entity, the resource scheduling Agent entity, the capability evaluation Agent entity, the coordination Agent entity, and the verification Agent entity under the scene; the entry of the routine maintenance scene also contains the five Agent entity types and the corresponding decision weight, and so on. Each entry clearly shows the correspondence among the scene type, the Agent entity type, and the decision weight.

[0033] Step S1145: The coordination Agent entity checks whether all operation and maintenance task scene types have generated corresponding scene-agent priority entries, and if there is a missing scene type, the core capability and the key Agent entity type under the missing scene type are analyzed to generate a corresponding scene-agent priority entry.

[0034] The coordination Agent entity checks all scene types in the operation and maintenance task scene type set to confirm whether each scene type has generated a corresponding scene-agent decision weight entry. If a missing scene type is found, such as a newly added intelligent device access debugging scene, the core capability required to complete the operation and maintenance task under the scene is analyzed, the key Agent entity type with a guiding role is determined, and then the decision weight of each type of Agent entity under the scene is set to generate a corresponding scene-agent decision weight entry, ensuring that there is no missing scene type.

[0035] Step S1146: The coordination Agent entity sorts all scene-agent decision weight entries in descending order of the frequency of occurrence of the scene type to form an Agent entity scene decision weight list.

[0036] The coordination Agent entity sorts all scene-agent decision weight entries in descending order of the frequency of occurrence of the scene type in the historical operation and maintenance record. For example, if the routine maintenance scene has the highest frequency of occurrence, followed by the system upgrade scene, the emergency fault repair scene, and the resource shortage scene has the lowest frequency of occurrence, the entries are arranged in this order to form an Agent entity scene decision weight list, which facilitates quick access to the decision weight configuration of the corresponding scene in actual task allocation.

[0037] Step S115: integrating the Agent intelligent body type set, the Agent intelligent body interaction rule set, the Agent intelligent body communication link set and the Agent intelligent body cooperation priority list to generate an Agent intelligent body cooperation network structure document, and completing the establishment of the Agent intelligent body cooperation network.

[0038] The Agent intelligent body type set, the Agent intelligent body interaction rule set, the Agent intelligent body communication link set and the Agent intelligent body cooperation priority list are integrated to compile an Agent intelligent body cooperation network structure document. The Agent intelligent body cooperation network structure document details the Agent intelligent body types participating in cooperation and their respective functions, the interaction rules between different intelligent bodies, the construction method and backup mechanism of the communication link, the intelligent body cooperation priority and decision weight in different scenarios, and the like. The overall architecture and operation rules of the Agent intelligent body cooperation network are clarified through the document, and the establishment of the Agent intelligent body cooperation network is completed.

[0039] Step S120: obtaining a set of to-be-assigned operation and maintenance task requirements, the set of operation and maintenance task requirements including an execution target, resource requirements and constraint conditions of the operation and maintenance task.

[0040] In this embodiment, the set of to-be-assigned operation and maintenance task requirements is obtained from the operation and maintenance management system of the Internet of Things smart home system. The above operation and maintenance tasks cover various types such as fault repair, routine maintenance and system upgrade of intelligent devices, and each task needs to clearly define the execution target, required resources and constraint conditions in the execution process.

[0041] Step S121: receiving an operation and maintenance task requirement request sent from the operation and maintenance management system, analyzing the requirement source identifier in the operation and maintenance task requirement request, determining the business system or department initiating the operation and maintenance task requirement, and obtaining operation and maintenance task requirement source information.

[0042] The operation and maintenance task requirement request sent by the operation and maintenance management system is received, and the requirement source identifier field is included in the operation and maintenance task requirement request. By analyzing the field, the initiator of the operation and maintenance task requirement is determined. For example, if the requirement source identifier is "smart lock management module", it is determined that the requirement is initiated from the smart lock management business system; if the identifier is "living room area operation and maintenance group", it is determined that the requirement is initiated from the operation and maintenance department of the living room area. The above requirement initiator information is sorted to form the operation and maintenance task requirement source information.

[0043] Step S122: Based on the operation and maintenance task demand source information, the core content in the operation and maintenance task demand request is extracted, including the execution target that needs to be achieved by the operation and maintenance task, the resource demand required for completing the task, and the constraint condition that needs to be followed in the task execution process, to obtain an operation and maintenance task core information set.

[0044] Based on the operation and maintenance task demand source information, the core content is extracted from the operation and maintenance task demand request. For the demand request from the intelligent door lock management business system, the execution target can be "repairing the fingerprint identification fault of the intelligent door lock of a certain user", the resource demand can be "carrying the special debugging tool for the intelligent door lock and replacing the fingerprint identification module", and the constraint condition can be "completing within the time period agreed by the user and not leaking the door lock information of the user". For the demand request from the living room area operation and maintenance group, the execution target can be "periodically maintaining all intelligent lamps in the living room", the resource demand can be "carrying the lamp detection instrument and insulation tool", and the constraint condition can be "executing during the non-peak electricity consumption period and ensuring the safety of power-off during the maintenance process". The execution target, resource demand, and constraint condition of all operation and maintenance tasks are sorted to form an operation and maintenance task core information set.

[0045] Step S123: The execution target in the operation and maintenance task core information set is classified, the operation and maintenance tasks are divided into hardware maintenance class tasks, software upgrade class tasks, fault repair class tasks, and system optimization class tasks according to the business attribute of the execution target, and an operation and maintenance task type classification result is obtained.

[0046] According to the business attribute of the execution target, all operation and maintenance tasks in the operation and maintenance task core information set are classified. The hardware maintenance class task refers to the inspection, maintenance, and replacement of the hardware components of the intelligent device, such as cleaning the lens of the intelligent camera and repairing the internal circuit of the intelligent socket. The software upgrade class task refers to updating the operating system, application program, or firmware of the intelligent device, such as upgrading the system version of the intelligent sound box and updating the firmware of the intelligent temperature controller. The fault repair class task refers to troubleshooting and repairing the functional faults of the intelligent device, such as repairing the intelligent door lock that cannot be unlocked and repairing the control failure of the intelligent window curtain. The system optimization class task refers to optimizing the overall performance of the smart home system, such as optimizing the data transmission rate of the intelligent gateway and adjusting the logic of multi-device linkage. The classified task list is sorted to form an operation and maintenance task type classification result.

[0047] Step S124: For each type of operation and maintenance task in the operation and maintenance task type classification result, the corresponding constraint condition is labeled, including the time window of task execution, the required professional skill level, and the resource usage limit, to obtain an operation and maintenance task constraint condition labeling result.

[0048] For each type of operation and maintenance task in the operation and maintenance task type classification result, the constraint conditions are labeled one by one. The constraint conditions of the hardware maintenance task may include "time window is non-use period of the device, required professional skill level is intermediate and above, and resource usage limit is that the dedicated maintenance tool is only used for specified tasks"; the constraint conditions of the software upgrade task may include "time window is night system low load period, required professional skill level is senior and above, and resource usage limit is that the upgrade server bandwidth is preferentially guaranteed for the task"; the constraint conditions of the fault repair task may include "time window is user emergency demand response period, required professional skill level is intermediate and above, and resource usage limit is that the standby spare parts are preferentially allocated"; and the constraint conditions of the system optimization task may include "time window is non-user active period, required professional skill level is senior and above, and resource usage limit is that the optimization tool cannot affect the normal operation of the system". The constraint conditions of each type of task are sorted to form an operation and maintenance task constraint condition labeling result.

[0049] Step S125: The operation and maintenance task core information set, operation and maintenance task type classification result and operation and maintenance task constraint condition labeling result are integrated to form a to-be-assigned operation and maintenance task demand set.

[0050] The execution target and resource demand of each task in the operation and maintenance task core information set, the task classification in the operation and maintenance task type classification result, and the constraint condition information in the operation and maintenance task constraint condition labeling result are associated and integrated. Each to-be-assigned operation and maintenance task includes information in four dimensions of execution target, resource demand, task type and constraint condition. The above information is sorted in a unified format to form a to-be-assigned operation and maintenance task demand set.

[0051] Step S130: The operation and maintenance task demand set is interactively evaluated by each Agent in the Agent collaboration network to determine a task adaptability evaluation result of each operation and maintenance task and each Agent.

[0052] In this embodiment, the task analysis Agent, the capability evaluation Agent and the coordination Agent in the Agent collaboration network interact and cooperate through the communication link to evaluate each task in the operation and maintenance task demand set, analyze the adaptability of each task to different Agents, and finally form a task adaptability evaluation result.

[0053] Step S131: The task analysis Agent in the Agent collaboration network receives the operation and maintenance task demand set, and based on the operation and maintenance task type classification result in the operation and maintenance task demand set, the operation and maintenance task demand set is disassembled into a single independent operation and maintenance task to obtain an independent operation and maintenance task list.

[0054] The task analysis Agent receives a set of operation and maintenance task requirements to be allocated, and according to the operation and maintenance task type classification result, the complex task in the set is disassembled into a single independent operation and maintenance task. For example, if a task is “comprehensive maintenance of the living room intelligent system, including intelligent lamp maintenance, intelligent curtain debugging and intelligent temperature controller upgrade”, according to the task type classification result, it is disassembled into “intelligent lamp hardware maintenance”, “intelligent curtain hardware maintenance” and “intelligent temperature controller software upgrade” three independent operation and maintenance tasks. The independent operation and maintenance task list is arranged in order.

[0055] Step S132: The task analysis Agent sends the independent operation and maintenance task list to the capability evaluation Agent through the Agent communication link set, and simultaneously transmits the execution target, resource requirement and constraint condition corresponding to each independent operation and maintenance task.

[0056] The task analysis Agent sends the independent operation and maintenance task list to the capability evaluation Agent through the dedicated communication link in the pre-constructed Agent communication link set. While sending the list, the execution target, resource requirement and constraint condition corresponding to each independent operation and maintenance task are simultaneously transmitted. For example, when sending the “intelligent door lock fingerprint identification fault repair” task, the execution target of the task is to repair the fingerprint identification function, the resource requirement is a special debugging tool and a replacement module, and the constraint condition is the user's agreed time period and other information, so as to ensure that the capability evaluation Agent obtains complete task information.

[0057] Step S133: After receiving the independent operation and maintenance task list and the corresponding task information, the capability evaluation Agent extracts the Agent capability parameters stored by itself, and the Agent capability parameters include the skill coverage range, resource control amount and task processing history record of each Agent, to obtain an Agent capability parameter set.

[0058] After receiving the independent operation and maintenance task list and the corresponding task information, the capability evaluation Agent starts the parameter extraction function of itself, and extracts the stored Agent capability parameters.

[0059] Step S1331: The capability evaluation Agent starts the parameter extraction module of itself, sends a capability parameter acquisition request to all Agents in the Agent collaboration network through the Agent communication link set, and the capability parameter acquisition request contains the time range of parameter acquisition and parameter type requirement.

[0060] The capability evaluation agent intelligent entity starts a built-in parameter extraction module, sends a capability parameter acquisition request to the task analysis agent intelligent entity, the resource scheduling agent intelligent entity, the coordination agent intelligent entity and the verification agent intelligent entity in the collaborative network through the multi-directional communication link in the agent intelligent entity communication link set. The request specifies that the time range of parameter acquisition is the capability data in a period of time, and the parameter type requires three types of skill coverage range, resource control amount and task processing history record. For example, the request specifies to provide the skill update of each agent intelligent entity in a period of time, the number and type of currently schedulable resources, the details of the operation and maintenance tasks completed in a period of time and other data, to ensure that each agent intelligent entity clearly knows the parameter content to be fed back.

[0061] Step S1332: After each agent intelligent entity receives the capability parameter acquisition request, the skill coverage range of the agent intelligent entity in the time range is extracted according to the time range of parameter acquisition, the skill coverage range includes mastered operation and maintenance technology types, authentication qualifications and skill update records, and the skill coverage range information of each agent intelligent entity is obtained.

[0062] After each agent intelligent entity receives the capability parameter acquisition request, the skill coverage range information of the agent intelligent entity is extracted according to the time range specified in the request. Taking the resource scheduling agent intelligent entity as an example, the skill coverage range it extracts includes mastered operation and maintenance technology types such as intelligent device resource planning technology and operation and maintenance tool scheduling algorithm, authentication qualifications such as resource manager authentication, and skill update records such as participation in intelligent resource scheduling system training in a period of time. The task analysis agent intelligent entity extracts mastered operation and maintenance technology types such as task decomposition algorithm and operation and maintenance task classification technology, authentication qualifications such as task management authentication, and training records such as task analysis model optimization training in a period of time, to form the skill coverage range information of each agent intelligent entity.

[0063] Step S1333: Each agent intelligent entity simultaneously extracts the resource control amount of the agent intelligent entity in the time range, the resource control amount includes the number of schedulable hardware devices, software tool permissions and human resource allocation, and the resource control amount information of each agent intelligent entity is obtained.

[0064] While extracting the skill coverage information, each Agent entity synchronously extracts the resource control information. The resource control of the coordination Agent entity includes the number of schedulable multi-type operation and maintenance coordination terminal devices, the advanced operation permission of intelligent task allocation software, the number and scheduling of full-time coordination personnel, and the like; the verification Agent entity includes the number of schedulable plan verification servers, the use permission of operation and maintenance scheme detection software, the configuration of verification personnel, and the like. Each Agent entity extracts and organizes its own resource control information according to the standard.

[0065] Step S1334: Each Agent entity further extracts its own task processing history record in the time range, and the task processing history record includes the completed operation and maintenance task type, the task completion quality evaluation, and the task processing time consumption statistics, to obtain the task processing history record information of each Agent entity.

[0066] Each Agent entity further extracts the task processing history record information. The task processing history record of the task analysis Agent entity includes the operation and maintenance task type such as the completed intelligent lamp maintenance task and intelligent temperature control system upgrade task in a period of time, the five-star evaluation of the user on the task disassembly accuracy, and the time statistics of each task disassembly; the task processing history record of the capability evaluation Agent entity includes the Agent entity adaptability evaluation task type completed in a period of time, the recognition evaluation of the collaborative network on the evaluation result, and the time consumption statistics of the evaluation process, to form the respective task processing history record information.

[0067] Step S1335: Each Agent entity feeds back its own skill coverage information, resource control information, and task processing history record information to the capability evaluation Agent entity through the Agent entity communication link set.

[0068] Each Agent entity extracts and organizes the completed skill coverage information, resource control information, and task processing history record information, and feeds back to the capability evaluation Agent entity through the respective corresponding communication link in the Agent entity communication link set. For example, the task analysis Agent entity sends information through the bidirectional communication link between the task analysis Agent entity and the capability evaluation Agent entity, and the verification Agent entity also completes information feedback through the exclusive communication link, to ensure the accuracy and timeliness of information transmission.

[0069] Step S1336: After receiving the feedback information from each Agent entity, the capability evaluation Agent entity structures and organizes the information, stores the skill coverage, resource control amount and task processing history record of each Agent entity according to the Agent entity identifier, forms the capability parameter item corresponding to each Agent entity, integrates all the capability parameter items, and obtains the Agent entity capability parameter set.

[0070] After receiving the feedback information from all Agent entities, the capability evaluation Agent entity starts the structured organization process. According to the unique identifier of each Agent entity, an independent capability parameter item is established for each Agent entity, and each item contains three parts of skill coverage, resource control amount and task processing history record of the Agent entity. For example, the item established for the resource scheduling Agent entity contains skill information such as intelligent device resource planning technology, resource information such as the number of schedulable operation and maintenance terminals, and historical information such as completed resource allocation task records. All the capability parameter items of the Agent entities are integrated together to form the Agent entity capability parameter set.

[0071] Step S134: The capability evaluation Agent entity compares each independent operation and maintenance task in the independent operation and maintenance task list with the Agent entity capability parameter set, analyzes the matching degree of the execution target of each independent operation and maintenance task with the skill coverage of the Agent entity, the fitting degree of the resource requirement with the resource control amount of the Agent entity, and the adaptation degree of the constraint condition with the task processing history record of the Agent entity, and obtains the original analysis result of single task multi-agent in each evaluation dimension.

[0072] The capability evaluation Agent entity compares and analyzes each independent operation and maintenance task in the independent operation and maintenance task list with the Agent entity capability parameter set one by one. Taking the independent operation and maintenance task of "smart door lock fingerprint identification fault repair" as an example, the matching degree of the execution target "repair fingerprint identification fault" with the skill coverage of each Agent entity is analyzed. If an Agent entity masters the smart door lock fingerprint module repair technology, the matching degree is higher. The fitting degree of the resource requirement "special debugging tool, replacement module" with the resource control amount of each Agent entity is analyzed. If an Agent entity can schedule the above tools and modules, the fitting degree is higher. The adaptation degree of the constraint condition "user agreed time period" with the task processing history record of each Agent entity is analyzed. If an Agent entity has a record of completing a task in a similar time period, the adaptation degree is higher. Each independent operation and maintenance task is analyzed from these three dimensions to obtain the original analysis result of single task multi-agent in each evaluation dimension.

[0073] Step S135: The capability evaluation Agent entity standardizes the original analysis results of each dimension corresponding to each independent operation and maintenance task, converts them into a unified evaluation scale, and performs weighted synthesis based on the preset weights of each dimension to obtain the comprehensive adaptation score of each independent operation and maintenance task and each Agent entity. The capability evaluation Agent entity sends the analysis results containing the comprehensive adaptation score to the coordination Agent entity through the Agent entity communication link set. The coordination Agent entity combines the Agent entity cooperation priority list, sorts the Agent entities according to the comprehensive adaptation score from high to low, determines the adaptation priority of each independent operation and maintenance task and each Agent entity, and generates the task adaptation evaluation result of each operation and maintenance task and each Agent entity.

[0074] The capability evaluation Agent entity standardizes the original analysis results of each independent operation and maintenance task, converts the analysis results of different dimensions into a unified evaluation scale, and eliminates the influence of different dimension data differences. Then, weighted synthesis is performed according to the preset weights of each dimension, for example, the matching degree weight is set to be the highest, the fitting degree is the second, and the adaptation degree is the lowest. The standardized dimension results are multiplied by the corresponding weights and summed to obtain the comprehensive adaptation score of each independent operation and maintenance task and each Agent entity.

[0075] The capability evaluation Agent entity sends the analysis results containing the comprehensive adaptation score to the coordination Agent entity through the communication link. The coordination Agent entity combines the Agent entity cooperation priority list, sorts the Agent entities according to the comprehensive adaptation score from high to low, determines the adaptation priority of each independent operation and maintenance task and each Agent entity. For example, in the "smart door lock fingerprint identification fault repair" task, the Agent entity with the highest comprehensive adaptation score is the first adaptation priority, and the same applies to the following. After sorting the adaptation priorities of all independent operation and maintenance tasks, the task adaptation evaluation result of each operation and maintenance task and each Agent entity is generated.

[0076] Step S140: According to the task adaptation evaluation result, the coordination Agent entity in the Agent entity cooperation network generates a preliminary operation and maintenance task allocation plan.

[0077] In this embodiment, the coordination Agent entity is based on the task adaptation evaluation result, and combines the real-time resource load of each Agent entity to allocate the independent operation and maintenance task, form a preliminary operation and maintenance task allocation plan, and ensure the rationality and feasibility of task allocation.

[0078] Step S141: The coordination agent receives the task adaptability evaluation result sent by the capability evaluation agent, and extracts the highest adaptation priority agent corresponding to each independent operation and maintenance task based on the adaptation priority of each independent operation and maintenance task in the task adaptability evaluation result, to obtain a single-task adaptation agent preliminary selection result.

[0079] After the coordination agent receives the task adaptability evaluation result, the result is analyzed, and for each independent operation and maintenance task, the first-ranked agent, i.e., the highest adaptation priority agent, is extracted from the adaptation priority of each agent corresponding to the task. For example, the highest adaptation priority agent of the "smart lamp hardware maintenance" task is a certain task analysis agent, and the highest adaptation priority agent of the "smart temperature controller software upgrade" task is a certain resource scheduling agent. The correspondence between each independent operation and maintenance task and the corresponding highest adaptation priority agent is recorded to obtain a single-task adaptation agent preliminary selection result.

[0080] Step S1411: The coordination agent analyzes the task adaptability evaluation result and extracts all independent operation and maintenance task identifiers contained therein. The corresponding adaptation priority information is filtered one by one according to the independent operation and maintenance task identifier to obtain a multi-agent adaptation priority list of each independent operation and maintenance task.

[0081] The coordination agent analyzes the task adaptability evaluation result and extracts all unique identifiers of independent operation and maintenance tasks in the result, such as "task 001-smart door lock fingerprint repair" and "task 002-smart camera debugging". All agent adaptation priority information corresponding to each independent operation and maintenance task is filtered one by one according to these identifiers, for example, the adaptation priority information corresponding to "task 001" is "Agent A (priority 1), Agent B (priority 2), and Agent C (priority 3)", forming a multi-agent adaptation priority list of each independent operation and maintenance task.

[0082] Step S1412: For the multi-agent adaptation priority list of each independent operation and maintenance task, the coordination agent sorts the adaptation priority in descending order, and identifies the agent at the top of the list. This agent is the highest adaptation priority agent corresponding to the independent operation and maintenance task.

[0083] For each independent operation and maintenance task, the multi-agent adaptation priority list is generated. The coordination agent confirms that the list is sorted by adaptation priority from high to low, and directly identifies the agent with the highest priority in the list. For example, the adaptation priority list of the task "003 - intelligent window curtain control failure repair" is "AgentD (priority 1), AgentE (priority 2)", and the AgentD with the highest priority in the list is the highest adaptation priority Agent for the task.

[0084] Step S1413: The coordination agent records the mapping relationship between each independent operation and maintenance task identifier and the corresponding highest adaptation priority Agent identifier, and associates the execution target, resource requirement and constraint condition of the independent operation and maintenance task to form a single task-agent mapping entry.

[0085] The coordination agent records the mapping relationship between each independent operation and maintenance task identifier and the corresponding highest adaptation priority Agent identifier, such as "task 001 - intelligent door lock fingerprint repair Agent A" and "task 002 - intelligent camera debugging Agent F". At the same time, the execution target, resource requirement and constraint condition of each independent operation and maintenance task are associated with the corresponding mapping relationship, for example, the mapping entry of "task 001 Agent A" also contains "execution target: repair fingerprint recognition failure; resource requirement: special debugging tool, replacement module; constraint condition: user agreed time period", forming a single task-agent mapping entry.

[0086] Step S1414: The coordination agent checks whether all independent operation and maintenance tasks have generated corresponding single task-agent mapping entries. If there are independent operation and maintenance tasks that have not generated mapping entries, the multi-agent adaptation priority list of the independent operation and maintenance task is reanalyzed, and the corresponding single task-agent mapping entry is supplemented.

[0087] The coordination agent compares the independent operation and maintenance task identifier in the single task-agent mapping entry with the identifier in the independent operation and maintenance task list to check whether there is any omission. If it is found that "task 005 - intelligent gateway system upgrade" in the independent operation and maintenance task list has not generated a corresponding mapping entry, the multi-agent adaptation priority list of the task is reanalyzed, the highest adaptation priority Agent is extracted, and the mapping entry containing the execution target, resource requirement and constraint condition is supplemented to ensure that all independent operation and maintenance tasks have corresponding records.

[0088] Step S1415: The coordination agent arranges all single task-agent mapping entries in the order of independent operation and maintenance task identifier to form a preliminary selection result of single task adaptation Agent.

[0089] The coordination agent arranges all generated single-task-agent mapping entries in the order of the independent operation and maintenance task identifiers, such as arranging the corresponding mapping entries in the order of "task 001, task 002, task 003, …", to form a single-task adaptation agent preliminary selection result with clear structure, facilitating subsequent resource load analysis.

[0090] Step S142: Based on the single-task adaptation agent preliminary selection result, the coordination agent aggregates all independent operation and maintenance tasks to be allocated to each agent, extracts the total resource demand of the independent operation and maintenance tasks, acquires real-time resource control amount margin information of each agent, analyzes the resource load of each agent, determines whether there is a case where the total task resource demand allocated to an agent exceeds its current resource control amount margin, and obtains an agent resource load analysis result.

[0091] According to the single-task adaptation agent preliminary selection result, the coordination agent aggregates all independent operation and maintenance tasks to be allocated to each agent according to the agent identifier. For example, the tasks of "task 001-intelligent door lock fingerprint repair" and "task 004-intelligent socket repair" to be allocated to Agent A are aggregated, and the total resource demand of these tasks is extracted, such as "2 sets of special debugging tools, 1 fingerprint module, and 3 socket accessories".

[0092] At the same time, the coordination agent sends a real-time resource query request to each agent through a communication link to acquire real-time resource control amount margin information of each agent, such as "1 set of special debugging tools, 0 fingerprint modules, and 2 socket accessories" remaining in Agent A. The total resource demand of each agent is compared with its real-time resource control amount margin to determine whether there is a case where the total resource demand exceeds the margin, for example, the total demand for special debugging tools and fingerprint modules of Agent A exceeds the margin, which is determined as resource load exceeding the standard, and finally an agent resource load analysis result is formed.

[0093] Step S143: If the Agent intelligent agent resource load analysis result shows that there is an Agent intelligent agent with resource load exceeding the standard, the coordination Agent intelligent agent reassigns part of the independent operation and maintenance tasks on the Agent intelligent agent to the Agent intelligent agent with the second highest adaptation priority based on the task adaptation evaluation result, and updates the single-task adaptation intelligent agent preliminary selection result to obtain the single-task adaptation intelligent agent adjustment result; if the Agent intelligent agent resource load analysis result shows that the resource loads of all Agent intelligent agents are within the reasonable range, the single-task adaptation intelligent agent preliminary selection result is directly taken as the single-task adaptation intelligent agent determination result.

[0094] If the Agent intelligent agent resource load analysis result shows that there is an Agent intelligent agent with resource load exceeding the standard, such as Agent A, the coordination Agent intelligent agent checks the Agent intelligent agent with the second highest adaptation priority for the task assigned to Agent A in the task adaptation evaluation result. For example, the Agent intelligent agent with the second highest adaptation priority for the task "task 001 - fingerprint repair of intelligent door lock" is Agent B, and the task is reassigned to Agent B. At the same time, the corresponding Agent intelligent agent for "task 001" in the single-task adaptation intelligent agent preliminary selection result is updated to Agent B to obtain the single-task adaptation intelligent agent adjustment result.

[0095] If the resource load analysis result shows that the total resource demand of all Agent intelligent agents does not exceed the real-time resource control amount, i.e., the resource load is within the reasonable range, no adjustment is needed, and the single-task adaptation intelligent agent preliminary selection result is directly taken as the single-task adaptation intelligent agent determination result.

[0096] Step S144: The coordination Agent intelligent agent integrates the single-task adaptation intelligent agent adjustment result or single-task adaptation intelligent agent determination result with the execution target, resource demand, and constraint condition corresponding to each independent operation and maintenance task, arranges the allocation order of each independent operation and maintenance task according to the time window of task execution, and generates a preliminary operation and maintenance task allocation plan.

[0097] The coordination Agent intelligent agent integrates the single-task adaptation intelligent agent adjustment result or determination result with the execution target, resource demand, and constraint condition of each independent operation and maintenance task. For example, after integration, "task 001 - fingerprint repair of intelligent door lock" corresponds to Agent B, the execution target is to repair the fingerprint recognition fault, the resource demand is a special debugging tool and a fingerprint module, and the constraint condition is the user-specified time period; "task 002 - intelligent camera debugging" corresponds to Agent F, the execution target is to debug the camera parameters, the resource demand is a debugging instrument, and the constraint condition is the night period.

[0098] Subsequently, the allocation sequence is arranged in the order of the time windows in the constraint conditions of each independent operation and maintenance task, such as performing the task 001 with the time window in the morning first, then performing the task 002 with the time window in the afternoon, and finally performing the task 003-intelligent temperature controller software upgrade with the time window at night, to form a preliminary operation and maintenance task allocation plan.

[0099] Step S150: The feasibility of the preliminary operation and maintenance task allocation plan is verified, the verification Agent intelligent agent in the Agent intelligent agent cooperation network confirms that the plan meets all the constraint conditions in the operation and maintenance task demand set, obtains a final operation and maintenance task allocation scheme, and outputs the scheme to the operation and maintenance management system.

[0100] In this embodiment, the verification Agent intelligent agent comprehensively verifies the preliminary operation and maintenance task allocation plan, mainly checks whether all the constraint conditions are met, and feeds back the part that does not meet the requirements to the coordination Agent intelligent agent for adjustment until the plan completely meets the requirements, forms a final scheme, and outputs the scheme.

[0101] Step S151: The coordination Agent intelligent agent sends the preliminary operation and maintenance task allocation plan to the verification Agent intelligent agent through the Agent intelligent agent communication link set, and simultaneously transmits all the constraint conditions in the operation and maintenance task demand set.

[0102] The coordination Agent intelligent agent sends the preliminary operation and maintenance task allocation plan to the verification Agent intelligent agent through the exclusive communication link between the coordination Agent intelligent agent and the verification Agent intelligent agent. Meanwhile, the constraint conditions of all the independent operation and maintenance tasks in the operation and maintenance task demand set are transmitted, including the time window, the professional skill level, the resource use limitation, and the like, to ensure that the verification Agent intelligent agent has complete verification basis.

[0103] Step S152: After receiving the preliminary operation and maintenance task allocation plan and the constraint conditions, the verification Agent intelligent agent disassembles the preliminary operation and maintenance task allocation plan into single operation and maintenance task allocation items, each allocation item contains the allocated Agent intelligent agent identifier, the task execution target, the resource demand, and the constraint conditions, and obtains an operation and maintenance task allocation item list.

[0104] After receiving the preliminary operation and maintenance task allocation plan and constraint conditions, the verification Agent checks the plan, disassembles it, and generates a separate allocation item for each independent operation and maintenance task. Each allocation item contains detailed information such as the allocated Agent identifier, such as "Agent B"; the task execution target, such as "repairing the fingerprint recognition failure of the smart door lock"; resource requirements, such as "one set of special debugging tools and one fingerprint module"; and constraint conditions, such as "time window: Saturday morning this week, professional skill level: intermediate and above, resource usage limit: special tools only for this task". All allocation items are organized into an operation and maintenance task allocation item list.

[0105] Step S153: The verification Agent checks whether the constraint conditions in each allocation item in the operation and maintenance task allocation item list are met one by one, including checking whether the task execution time window matches the allocated Agent's idle time, whether the required professional skill level is consistent with the allocated Agent's skill coverage, and whether the resource usage limit is within the allocated Agent's resource control range, to obtain a single item verification result.

[0106] The verification Agent verifies each allocation item in the operation and maintenance task allocation item list one by one. Taking a certain allocation item as an example, first check whether the task execution time window "Saturday morning this week" matches the allocated Agent B's idle time by comparing it with Agent B's schedule data. If Agent B has no other allocated tasks on Saturday morning, the time window is determined to be matched; if there are already allocated tasks, the time window is determined to be not matched. Next, check whether the required professional skill level "intermediate and above" in this allocation item matches Agent B's skill coverage by checking Agent B's skill certification records and skill update records. If Agent B holds an intermediate or above level of smart door lock maintenance certification, the skill level is determined to be matched; otherwise, it is not matched. Finally, check whether the resource usage limit "special debugging tools only for this task" is within Agent B's resource control range by querying Agent B's schedulable tool list. If Agent B can currently schedule the special debugging tools and no other tasks are occupied, the resource usage limit is determined to be met; if the tools cannot be scheduled or are already occupied, it is determined to be not met. Based on the results of these three checks, a single item verification result for this allocation item is generated. If all three items are met, the result is "pass"; if any one item is not met, the result is "not pass".

[0107] Step S154: The verification Agent agent aggregates all single entry verification results, and if all single entry verification results show that the constraint conditions are met, it is determined that the preliminary operation and maintenance task allocation plan is feasible, and the preliminary operation and maintenance task allocation plan is taken as the final operation and maintenance task allocation scheme; if there is a single entry verification result showing that the constraint condition is not met, the allocation entry that does not meet the constraint condition is fed back to the coordination Agent agent, and the Agent agent allocation of the allocation entry is adjusted by the coordination Agent agent based on the task adaptability evaluation result to generate an adjusted preliminary operation and maintenance task allocation plan, which is sent to the verification Agent agent again for verification until all allocation entries meet the constraint condition, and the final operation and maintenance task allocation scheme is obtained;

[0108] The verification Agent agent aggregates all single entry verification results of all allocation entries, and counts the number of entries that pass and do not pass. If all single entry verification results are passed, it is indicated that the preliminary operation and maintenance task allocation plan meets all constraint conditions, and it is directly determined as the final operation and maintenance task allocation scheme. If there is an entry that does not pass, the verification Agent agent filters out the allocation entry that does not meet the constraint condition, and labels the constraint type that does not meet in detail, such as "time window mismatch", "skill level mismatch", etc., to form an abnormal allocation entry list, and feeds back the list to the coordination Agent agent. The coordination Agent agent selects appropriate Agent agents for the abnormal allocation entry according to the task adaptability evaluation result, generates an adjusted preliminary operation and maintenance task allocation plan, and sends it to the verification Agent agent again. The verification Agent agent re-verifies according to the same single entry verification process, and repeats the process until the verification results of all allocation entries are passed, and finally determines the scheme that meets all constraint conditions as the final operation and maintenance task allocation scheme.

[0109] For example, step S1541: the verification Agent agent marks the allocation entry that does not meet the constraint condition as an abnormal allocation entry, extracts the independent operation and maintenance task identifier, the currently allocated Agent agent identifier, and the constraint condition type that does not meet, and generates an abnormal allocation entry report.

[0110] The verification Agent filters the aggregated single-item verification results, and assigns a tag of "abnormal allocation item" to all allocation items with a verification result of "not passed". For each abnormal allocation item, the independent operation and maintenance task identifier, such as "task 006-intelligent curtain motor replacement", the currently assigned Agent identifier, such as "Agent C", and the constraint condition type that is not met, such as "resource usage limit not met (lack of special disassembly tool)", are extracted. The above information is arranged in a unified format to form an abnormal allocation item report, and each record in the report clearly corresponds to an abnormal item, a currently assigned Agent, and a problem type.

[0111] Step S1542: The verification Agent sends the abnormal allocation item report to the coordination Agent through the set of Agent communication links.

[0112] The verification Agent sends the generated abnormal allocation item report to the coordination Agent through the dedicated communication link between the verification Agent and the coordination Agent. During transmission, the data encryption mechanism configured by the communication link is used to encrypt the report content, ensuring that the abnormal information is not leaked or tampered with during transmission, and the link state monitoring module is used to confirm that the information has been successfully sent to the coordination Agent.

[0113] Step S1543: After receiving the abnormal allocation item report, the coordination Agent retrieves the multi-agent adaptation priority list corresponding to the independent operation and maintenance task based on the independent operation and maintenance task identifier in the abnormal allocation item.

[0114] After receiving the abnormal allocation item report, the coordination Agent retrieves the multi-agent adaptation priority list corresponding to the independent operation and maintenance task based on the independent operation and maintenance task identifier in the abnormal allocation item. For example, for "task 006-intelligent curtain motor replacement", the adaptation priority list is "Agent C (priority 1), Agent D (priority 2), Agent E (priority 3)".

[0115] Step S1544: The coordination Agent excludes the Agent currently assigned to the independent operation and maintenance task, and selects the Agent with the second highest adaptation priority from the multi-agent adaptation priority list as a new candidate Agent.

[0116] The coordination agent excludes the currently assigned agent that causes the verification failure from the multi-agent adaptation priority list. Taking task 006 as an example, after excluding the currently assigned agent C, the agent D with the second highest priority is selected from the list as a new candidate agent, ensuring that the new candidate agent still has a high priority in terms of adaptability.

[0117] Step S1545: The coordination agent extracts the capability parameters of the new candidate agent, including the skill coverage, resource control amount, and task processing history, and checks whether the candidate agent meets the constraint condition that is not met in the abnormal allocation item.

[0118] The coordination agent sends a capability parameter query request to the new candidate agent D through the agent communication link set, extracts the skill coverage, resource control amount, and task processing history of the agent D. For the constraint condition “lack of special disassembly tool” that is not met in the task 006 abnormal allocation item, the resource control amount of the agent D is checked to see whether it contains the special disassembly tool, and it is confirmed that the tool is not currently occupied by other tasks, so as to determine whether the agent D meets the constraint condition.

[0119] Step S1546: If the new candidate agent meets the constraint condition, the coordination agent updates the allocation item of the independent operation and maintenance task, replaces the currently assigned agent with the new candidate agent, and generates an adjusted allocation item. If the new candidate agent still does not meet the constraint condition, the next agent with the highest priority in the multi-agent adaptation priority list is selected, and the process of checking the constraint condition is repeated until a candidate agent that meets the constraint condition is found, and an adjusted allocation item is generated.

[0120] If it is found that the resource control amount of the candidate agent D contains the special disassembly tool required by task 006 and is not occupied, i.e., the constraint condition is met, the coordination agent updates the allocation item of task 006, replaces the currently assigned agent C with agent D, and generates an adjusted allocation item. If agent D also lacks the special disassembly tool, i.e., does not meet the constraint condition, the coordination agent continues to select the next agent E with the highest priority from the adaptation priority list as a new candidate agent, and repeats the process of extracting capability parameters and checking constraint conditions until a candidate agent that meets the constraint condition of “having a special disassembly tool” is found, and an adjusted allocation item is generated.

[0121] Step S1547: The coordination Agent replaces the abnormal allocation item in the preliminary operation and maintenance task allocation plan with the adjusted allocation item, and the remaining allocation items remain unchanged, to generate an adjusted preliminary operation and maintenance task allocation plan.

[0122] The coordination Agent replaces the abnormal allocation item in the preliminary operation and maintenance task allocation plan with the adjusted allocation item, and the remaining allocation items remain unchanged, to generate an adjusted preliminary operation and maintenance task allocation plan.

[0123] Step S1548: The coordination Agent sends the adjusted preliminary operation and maintenance task allocation plan to the verification Agent through the Agent communication link set, and the verification Agent re-verifies according to the single-item verification process.

[0124] The coordination Agent sends the adjusted preliminary operation and maintenance task allocation plan to the verification Agent through the Agent communication link set, and the verification Agent re-verifies according to the single-item verification process.

[0125] Step S1549: Repeat the above process of abnormal allocation item feedback, Agent re-selection, plan adjustment and re-verification until all single-item verification results of the verification Agent show that the constraint condition is met, stop adjustment, and determine the preliminary operation and maintenance task allocation plan at this time as the final operation and maintenance task allocation scheme.

[0126] If the verification Agent re-verifies the adjusted plan and still has abnormal allocation items, then according to the process of steps S1541 to S1548, the abnormal item feedback, Agent re-selection, plan adjustment and verification are performed again. Repeat this process until the verification Agent's summary of all single-item verification results is "pass", i.e. all allocation items meet the constraint condition, at which point the adjustment process is stopped and the current preliminary operation and maintenance task allocation plan is officially determined as the final operation and maintenance task allocation scheme.

[0127] Step S155: The verification Agent agent sends the final operation and maintenance task allocation scheme to the operation and maintenance management system through the Agent agent communication link set, generates a scheme verification report, and outputs the final operation and maintenance task allocation scheme together.

[0128] The verification Agent agent establishes a connection with the operation and maintenance management system through a preset communication interface, sends the final operation and maintenance task allocation scheme to the operation and maintenance management system, and generates a scheme verification report. The report contains the summary of the verification process, the detailed verification results of each allocation item, the abnormal item adjustment record, the list of constraint conditions satisfied by the final scheme, and the like. The scheme verification report and the final operation and maintenance task allocation scheme are output to the operation and maintenance management system to ensure that the operation and maintenance personnel can accurately perform each operation and maintenance task according to the scheme.

[0129] Figure 2 A schematic diagram of exemplary hardware and software components of the operation and maintenance task allocation system 100 assisted by Agent agents, which can implement the idea of the present application, is shown. For example, the processor 120 can be used in the operation and maintenance task allocation system 100 assisted by Agent agents, and is used to perform the functions in the present application.

[0130] The operation and maintenance task allocation system 100 assisted by Agent agents can be a general server or a special-purpose server, both of which can be used to implement the operation and maintenance task allocation method assisted by Agent agents of the present application. Although only one server is shown in the present application, for the sake of convenience, the functions described in the present application can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0131] For example, the operation and maintenance task allocation system 100 assisted by Agent agents can include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and different forms of storage media 140, such as a disk, a ROM, or a RAM, or any combination thereof. Exemplarily, the operation and maintenance task allocation system 100 assisted by Agent agents can also include program instructions stored in a ROM, a RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present application can be implemented according to the program instructions. The operation and maintenance task allocation system 100 assisted by Agent agents also includes an I / O interface 150 between the computer and other input and output devices.

[0132] For the convenience of description, only one processor is described in the operation and maintenance task allocation system assisted by the agent intelligent agent. However, it should be noted that the operation and maintenance task allocation system assisted by the agent intelligent agent in the present application can also include multiple processors, so the steps performed by one processor described in the present application can also be jointly executed or separately executed by multiple processors. For example, if the processor of the operation and maintenance task allocation system assisted by the agent intelligent agent executes steps A and B, it should be understood that steps A and B can also be executed by two different processors together or separately in one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor execute steps A and B together.

[0133] In addition, the embodiment of the present application also provides a readable storage medium, wherein computer executable instructions are preset, and when a processor executes the computer executable instructions, the operation and maintenance task allocation method assisted by the agent intelligent agent is realized.

[0134] It should be noted that, in order to simplify the description of the present application and to help understand one or more embodiments of the present application, in the foregoing description of the embodiments of the present application, various features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. A method for operation and maintenance task allocation assisted by a combination of an agent and an intelligent agent, characterized in that, The method comprises: establishing an Agent intelligent agent cooperation network; obtaining a set of to-be-assigned operation and maintenance task requirements, the set of operation and maintenance task requirements comprising an execution target, resource requirements and constraint conditions of the operation and maintenance task; interacting and evaluating the set of operation and maintenance task requirements by each Agent intelligent agent in the Agent intelligent agent cooperation network to determine a task adaptability evaluation result of each operation and maintenance task and each Agent intelligent agent; generating a preliminary operation and maintenance task allocation plan by a coordination Agent intelligent agent in the Agent intelligent agent cooperation network according to the task adaptability evaluation result; verifying the feasibility of the preliminary operation and maintenance task allocation plan, confirming by a verification Agent intelligent agent in the Agent intelligent agent cooperation network that the plan meets all constraint conditions in the set of operation and maintenance task requirements, obtaining a final operation and maintenance task allocation scheme and outputting the final operation and maintenance task allocation scheme to an operation and maintenance management system; The establishment of the Agent intelligent agent cooperation network comprises: determining an Agent intelligent agent type participating in operation and maintenance task allocation, the Agent intelligent agent type comprising a task analysis Agent intelligent agent, a resource scheduling Agent intelligent agent, a capability evaluation Agent intelligent agent, a coordination Agent intelligent agent and a verification Agent intelligent agent, obtaining an Agent intelligent agent type set; defining an interaction rule between each type of Agent intelligent agent based on the Agent intelligent agent type set, the interaction rule comprising an information transmission format, a request-response time limit and a data sharing range, obtaining an Agent intelligent agent interaction rule set; constructing a communication link between each Agent intelligent agent according to the Agent intelligent agent interaction rule set to ensure that different types of Agent intelligent agents can transmit operation and maintenance task related information in real time, obtaining an Agent intelligent agent communication link set; based on the Agent intelligent agent type set and the Agent intelligent agent communication link set, setting a cooperation priority of each Agent intelligent agent in the cooperation process, determining a decision weight of each Agent intelligent agent in different scenarios, obtaining an Agent intelligent agent cooperation priority list; integrating the Agent intelligent agent type set, the Agent intelligent agent interaction rule set, the Agent intelligent agent communication link set and the Agent intelligent agent cooperation priority list to generate an Agent intelligent agent cooperation network structure document, and completing the establishment of the Agent intelligent agent cooperation network; based on the Agent intelligent agent type set and the Agent intelligent agent communication link set, setting a cooperation priority of each Agent intelligent agent in the cooperation process, determining a decision weight of each Agent intelligent agent in different scenarios, obtaining an Agent intelligent agent cooperation priority list, comprising: the coordination Agent intelligent agent collects common scene types of operation and maintenance task allocation, the scene types comprising an emergency fault repair scene, a regular maintenance scene, a system upgrade scene and a resource shortage scene, obtaining a set of operation and maintenance task scene types; For each scene type in the set of operation and maintenance task scene types, the coordination agent analyzes the core capabilities required to complete the operation and maintenance task under the corresponding operation and maintenance task scene, determines the agent type that has a guiding role under the corresponding operation and maintenance task, and obtains a scene-agent key association result; Based on the scene-agent key association result, set a decision weight for each agent type under each scene type, the decision weight of the key agent type is higher than that of the non-key agent type, and the decision weight is used to determine the opinion influence of each agent in the collaborative decision-making process; The coordination agent combines the set of agent types, associates the agent types under each scene type and the corresponding decision weight, and forms a scene-agent decision weight item; The coordination agent checks whether all operation and maintenance task scene types have generated corresponding scene-agent priority items, if there are missing scene types, supplement the analysis of the core capabilities and key agent types under the missing scene types, and generate corresponding scene-agent priority items; The coordination agent sorts all scene-agent decision weight items according to the frequency of scene types from high to low to form an agent scene decision weight list.

2. The method of claim 1, wherein, The operation and maintenance task demand set to be allocated includes: Receiving an operation and maintenance task demand request sent from an operation and maintenance management system, parsing the demand source identifier in the operation and maintenance task demand request, determining the business system or department that initiates the operation and maintenance task demand, and obtaining operation and maintenance task demand source information; Based on the operation and maintenance task demand source information, extract the core content in the operation and maintenance task demand request, including the execution target that needs to be achieved by the operation and maintenance task, the resource demand required to complete the task, and the constraint condition that needs to be followed in the task execution process, and obtain a set of operation and maintenance task core information; Classify the execution target in the set of operation and maintenance task core information, divide the operation and maintenance task into hardware maintenance task, software upgrade task, fault repair task and system optimization task according to the business attribute of the execution target, and obtain an operation and maintenance task type classification result; For each type of operation and maintenance task in the operation and maintenance task type classification result, label the corresponding constraint condition, including the time window of task execution, the required professional skill level, and the resource usage limit, and obtain an operation and maintenance task constraint condition labeling result; Integrate the set of operation and maintenance task core information, operation and maintenance task type classification result and operation and maintenance task constraint condition labeling result to form a set of operation and maintenance task demands to be allocated.

3. The method of claim 1, wherein, The operation and maintenance task demand set is interactively evaluated by each agent in the agent collaboration network to determine the task adaptability evaluation result of each operation and maintenance task and each agent, including: The task analysis Agent in the Agent collaboration network receives a set of operation and maintenance task requirements, and based on a task type classification result in the set of operation and maintenance task requirements, the set of operation and maintenance task requirements is decomposed into single independent operation and maintenance tasks to obtain a list of independent operation and maintenance tasks; The task analysis Agent sends the list of independent operation and maintenance tasks to the capability evaluation Agent through a set of Agent communication links, and simultaneously transmits the execution target, resource requirement and constraint condition corresponding to each independent operation and maintenance task; After receiving the list of independent operation and maintenance tasks and the corresponding task information, the capability evaluation Agent extracts the capability parameters of each Agent stored by itself, the capability parameters of each Agent including the skill coverage range, resource control amount and task processing history record of each Agent to obtain a set of Agent capability parameters; The capability evaluation Agent compares each independent operation and maintenance task in the list of independent operation and maintenance tasks with the set of Agent capability parameters, analyzes the matching degree of the execution target of each independent operation and maintenance task with the skill coverage range of the Agent, the fitting degree of the resource requirement with the resource control amount of the Agent and the adaptation degree of the constraint condition with the task processing history record of the Agent, and obtains the original analysis result of single-task multi-agent in each evaluation dimension; The capability evaluation Agent standardizes each dimension original analysis result corresponding to each independent operation and maintenance task, converts it into a unified evaluation scale, and performs weighted integration based on the preset weight of each dimension to obtain the comprehensive adaptation score of each independent operation and maintenance task and each Agent. The capability evaluation Agent sends the analysis result containing the comprehensive adaptation score to the coordination Agent through a set of Agent communication links, the coordination Agent combines the Agent collaboration priority list, sorts according to the comprehensive adaptation score from high to low, determines the adaptation priority of each independent operation and maintenance task and each Agent, and generates the task adaptation evaluation result of each operation and maintenance task and each Agent.

4. The method of claim 3, wherein, The capability evaluation Agent extracts the capability parameters of each Agent stored by itself, the capability parameters of each Agent including the skill coverage range, resource control amount and task processing history record of each Agent to obtain a set of Agent capability parameters, including: The capability evaluation Agent starts the parameter extraction module of itself, sends a capability parameter acquisition request to all Agent in the Agent collaboration network through a set of Agent communication links, the capability parameter acquisition request containing the time range of parameter acquisition and parameter type requirement; The Agent intelligent agent receives the capability parameter acquisition request, extracts the skill coverage range of itself in the time range according to the time range of parameter acquisition, the skill coverage range includes mastered operation and maintenance technology type, authentication qualification and skill update record, and obtains skill coverage range information of each Agent intelligent agent; Each Agent intelligent agent extracts the resource control amount of itself in the time range, the resource control amount includes the number of schedulable hardware devices, software tool permissions and human resource allocation, and obtains resource control amount information of each Agent intelligent agent; Each Agent intelligent agent also extracts the task processing history record of itself in the time range, the task processing history record includes the type of completed operation and maintenance tasks, task completion quality evaluation and task processing time consumption statistics, and obtains task processing history record information of each Agent intelligent agent; Each Agent intelligent agent feeds back the skill coverage range information, resource control amount information and task processing history record information of itself to the capability evaluation Agent intelligent agent through the Agent intelligent agent communication link set; The capability evaluation Agent intelligent agent receives the information fed back by each Agent intelligent agent, structures and organizes the information, stores the skill coverage range, resource control amount and task processing history record according to the Agent intelligent agent identifier, forms the capability parameter item corresponding to each Agent intelligent agent, integrates all capability parameter items, and obtains the Agent intelligent agent capability parameter set.

5. The method of claim 1, wherein, The preliminary operation and maintenance task distribution plan is generated by the coordination Agent intelligent agent in the Agent intelligent agent cooperation network according to the task adaptability evaluation result, which includes: The coordination Agent intelligent agent receives the task adaptability evaluation result sent by the capability evaluation Agent intelligent agent, extracts the highest adaptability priority Agent intelligent agent corresponding to each independent operation and maintenance task based on the adaptability priority of each independent operation and maintenance task and each Agent intelligent agent, and obtains the single task adaptability intelligent agent preliminary selection result; The coordination Agent intelligent agent summarizes all independent operation and maintenance tasks to be distributed to each Agent intelligent agent based on the single task adaptability intelligent agent preliminary selection result, extracts the total resource demand of the independent operation and maintenance tasks, acquires the real-time resource control amount of each Agent intelligent agent, analyzes the resource load of each Agent intelligent agent, judges whether there is an Agent intelligent agent whose total task resource demand exceeds the current resource control amount, and obtains the Agent intelligent agent resource load analysis result; If the Agent intelligent agent resource load analysis result shows that there is an Agent intelligent agent with resource load exceeding the standard, the coordination Agent intelligent agent reassigns part of the independent operation and maintenance tasks of the Agent intelligent agent to the Agent intelligent agent with the second highest adaptability priority based on the task adaptability evaluation result, updates the single task adaptability intelligent agent preliminary selection result, and obtains the single task adaptability intelligent agent adjustment result; If the Agent intelligent body resource load analysis result shows that all Agent intelligent body resource loads are within a reasonable range, the single-task adaptive Agent preliminary selection result is directly taken as the single-task adaptive Agent determination result; The coordination Agent integrates the single-task adaptive Agent adjustment result or the single-task adaptive Agent determination result with the execution target, resource demand and constraint condition corresponding to each independent operation and maintenance task, arranges the allocation sequence of each independent operation and maintenance task in the order of the time window of task execution, and generates a preliminary operation and maintenance task allocation plan.

6. The method of claim 5, wherein, The coordination Agent extracts the highest adaptive priority Agent intelligent body corresponding to each independent operation and maintenance task based on the adaptive priority of each independent operation and maintenance task in the task adaptability evaluation result and each Agent intelligent body, and obtains a single-task adaptive Agent preliminary selection result, including: The coordination Agent analyzes the task adaptability evaluation result, extracts all independent operation and maintenance task identifiers contained therein, and filters the corresponding adaptive priority information one by one according to the independent operation and maintenance task identifier, to obtain a multi-agent adaptive priority list of each independent operation and maintenance task; For the multi-agent adaptive priority list of each independent operation and maintenance task, the coordination Agent sorts the adaptive priorities in descending order, identifies the Agent intelligent body at the top of the list, which is the highest adaptive priority Agent intelligent body corresponding to the independent operation and maintenance task; The coordination Agent records the mapping relationship between each independent operation and maintenance task identifier and the corresponding highest adaptive priority Agent intelligent body identifier, and associates the execution target, resource demand and constraint condition of the independent operation and maintenance task to form a single-task-Agent mapping entry; The coordination Agent checks whether all independent operation and maintenance tasks have generated corresponding single-task-Agent mapping entries, and if there are independent operation and maintenance tasks that have not generated mapping entries, it reanalyzes the multi-agent adaptive priority list of the independent operation and maintenance task and supplements the generation of the corresponding single-task-Agent mapping entry; The coordination Agent arranges all single-task-Agent mapping entries in the order of independent operation and maintenance task identifiers to form a single-task adaptive Agent preliminary selection result.

7. The method of claim 1, wherein, The feasibility verification of the preliminary operation and maintenance task allocation plan is performed through the verification Agent intelligent body in the Agent intelligent body collaboration network to confirm that the plan meets all constraint conditions in the operation and maintenance task demand set, to obtain a final operation and maintenance task allocation scheme and output it to the operation and maintenance management system, including: The coordination Agent sends the preliminary operation and maintenance task allocation plan to the verification Agent intelligent body through the Agent intelligent body communication link set, and simultaneously transmits all constraint conditions in the operation and maintenance task demand set; The verification Agent checks whether the constraint condition in each allocation item in the list of the maintenance task allocation items is satisfied, including checking whether the task execution time window matches the idle time of the allocated Agent, whether the required professional skill level is consistent with the skill coverage range of the allocated Agent, and whether the resource usage limit is within the resource control range of the allocated Agent, to obtain a single-item verification result. The verification Agent aggregates all single-item verification results. If all single-item verification results show that the constraint condition is satisfied, it is determined that the preliminary maintenance task allocation plan is feasible, and the preliminary maintenance task allocation plan is taken as the final maintenance task allocation scheme. If there is a single-item verification result showing that the constraint condition is not satisfied, the allocation item that does not satisfy the constraint condition is fed back to the coordination Agent, which adjusts the Agent allocation of the allocation item based on the task adaptability evaluation result, generates an adjusted preliminary maintenance task allocation plan, and sends it to the verification Agent for verification again until all allocation items satisfy the constraint condition, obtaining the final maintenance task allocation scheme. The verification Agent sends the final maintenance task allocation scheme to the maintenance management system through the Agent communication link set, generates a scheme verification report, and outputs it together with the final maintenance task allocation scheme. The device comprises a processor and a memory, the memory and the processor are connected, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to realize the method for assisting the maintenance task allocation combined with the Agent according to any one of claims 1-7.

8. An operation and maintenance task allocation system assisted by a combination of Agent intelligent agents, characterized in that, ​

Images