Work order management method and device, computer equipment, storage medium and program product
Through the analysis of content of network operation and maintenance work orders and the quality inspection of large language models, the repeated order dispatch and matching operation and maintenance parties are identified, the problem of human error in network operation and maintenance work order quality inspection is solved, and the accuracy of work order processing is improved Sex and stability.
Patent Information
- Application Number
- CN202410971670.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-12
AI Technical Summary
During the full life cycle quality inspection of network operation and maintenance work orders, there is a lack of an effective quality control mechanism, which leads to errors introduced by human factors and affects the effectiveness of work order processing.
By analyzing the content of the online work order to be distributed, using a large language model to identify and match the operation and maintenance parties, quality inspection information is generated, ensuring that the work orders are effectively inspected before distribution, and processing efficiency and effectiveness are improved.
It realizes standardized quality inspection of network operation and maintenance work orders, reduces human errors, improves the accuracy and stability of work order processing, and improves operation and maintenance efficiency.
Smart Images

Figure CN118941232A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of network operation and maintenance technology, and in particular to a work order management method, device, computer equipment, storage medium and program product. Background Art
[0002] With the rapid development of modern communication technology, especially the widespread coverage and commissioning of 5G (5th Generation Mobile Communication Technology) networks, the scale of network infrastructure is becoming increasingly large and the complexity is constantly increasing. The corresponding network operation and maintenance tasks are showing an exponential growth trend.
[0003] At present, in the quality inspection process of the entire life cycle of network operation and maintenance work orders, the quality control in the early stage of network operation and maintenance work order distribution relies on manual review, which is prone to errors introduced by human factors and lacks an effective quality control mechanism, resulting in poor processing results of network operation and maintenance work orders. Summary of the invention
[0004] Based on this, it is necessary to provide a work order management method, device, computer equipment, storage medium and program product that can perform standardized quality inspection on network operation and maintenance work orders, thereby improving the processing effect of network operation and maintenance work orders, in response to the above technical problems.
[0005] In a first aspect, the present application provides a work order management method, comprising:
[0006] Analyze the work order content of the target network work order to be dispatched, and obtain the alarm information of the target network work order; wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result;
[0007] When it is determined that the target network work order meets the dispatching conditions based on the duplicate dispatching identification results, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching results.
[0008] In one embodiment, the work order content of the target network work order to be dispatched is analyzed to obtain the alarm information of the target network work order, including:
[0009] Generate target question information for the target network work order according to the question information template and the work order content of the target network work order to be dispatched;
[0010] Input the target question information and the work order content of the target network work order into the pre-trained large language model to obtain the response information of the target question information;
[0011] Extract the alarm information of the target network work order from the response information.
[0012] In one embodiment, the work order content of the target network work order to be dispatched is analyzed to obtain the alarm information of the target network work order, including:
[0013] Determine the similarity between the work order content of the target network work order to be dispatched and the work order content of each historical network work order;
[0014] If any content similarity among the content similarities is higher than the similarity threshold, a duplicate dispatch identification result is generated, indicating that there is a duplicate dispatch of the target network work order.
[0015] In one embodiment, the work order content of the target network work order to be dispatched is analyzed to obtain the alarm information of the target network work order, including:
[0016] According to the work order content of the target network work order to be dispatched, determine the target operation and maintenance scenario to which the target network work order belongs from various operation and maintenance scenarios;
[0017] Select the target operation and maintenance party from various operation and maintenance parties based on the target operation and maintenance scenario and the work order content of the target network work order;
[0018] According to the operation and maintenance strategy associated with the target operation and maintenance party and the target operation and maintenance scenario, the corresponding operation and maintenance party matching result is generated.
[0019] In one embodiment, the method further comprises:
[0020] In response to the operation and maintenance record of the target network work order provided by the target operation and maintenance party, the operation and maintenance record of the target network work order is analyzed to obtain quality inspection information of the target network work order.
[0021] In one of the embodiments, the operation and maintenance record includes key time, and the key time includes at least two of the work order response time, the cause location time of the network failure corresponding to the target network work order, and the solution provision time;
[0022] Accordingly, the operation and maintenance records of the target network work order are analyzed to obtain the quality inspection information of the target network work order, including:
[0023] Input the operation and maintenance records into the quality inspection model to obtain the timeliness value corresponding to each key time;
[0024] According to the timeliness coefficient corresponding to each key time, the timeliness value corresponding to each key time is weighted and summed to obtain the quality inspection score of the target network work order;
[0025] Generate corresponding quality inspection information based on the quality inspection score.
[0026] In one embodiment, the quality inspection model is trained in the following manner:
[0027] The large language model is used as the teacher model, and the knowledge distillation technology is used to build the student model corresponding to the large language model to obtain the initial quality inspection model;
[0028] According to the operation and maintenance records of the sample network work orders, the initial quality inspection model is trained for multiple rounds of iterations to obtain a trained quality inspection model.
[0029] In a second aspect, the present application also provides a work order management device, including:
[0030] The content analysis module is used to analyze the work order content of the target network work order to be dispatched, and obtain the alarm information of the target network work order; wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result;
[0031] The work order dispatching module is used to dispatch the target network work order to the corresponding target operation and maintenance party according to the matching result of the operation and maintenance party when it is determined that the target network work order meets the dispatching conditions based on the duplicate order identification result.
[0032] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method of the first aspect when executing the computer program.
[0033] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method of the first aspect when the computer program is executed by a processor.
[0034] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the steps of the method of the first aspect when executed by a processor.
[0035] The above-mentioned work order management method, device, computer equipment, storage medium and program product analyze the work order content of the target network work order to be dispatched, and obtain the alarm information of the target network work order; wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result; when it is determined that the target network work order meets the dispatch conditions according to the duplicate dispatch identification result, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching result. This application analyzes the work order content of the target network work order to achieve effective quality inspection of the target network work order before dispatching it, especially in terms of duplicate dispatch identification and dispatch operation and maintenance party matching, so as to reasonably dispatch the target network work order to the target operation and maintenance party for processing, thereby improving the processing efficiency and processing effect of the target network work order. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0037] Figure 1 An application environment diagram of a work order management method in an embodiment;
[0038] Figure 2 A flowchart of a work order management method in one embodiment;
[0039] Figure 3 A schematic diagram of a process for determining alarm information in an embodiment;
[0040] Figure 4 A schematic diagram of a process for determining a matching result of an operation and maintenance party in an embodiment;
[0041] Figure 5 A schematic diagram of an interface of an operation and maintenance record of a target network work order in an embodiment;
[0042] Figure 6 A flowchart of a work order management method in another embodiment;
[0043] Figure 7 is a structural block diagram of a work order management device in an embodiment;
[0044] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0046] The work order management method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The server analyzes the work order content of the target network work order to be dispatched, and obtains alarm information including the duplicate dispatch identification result and the operation and maintenance party matching result. Then, according to the duplicate dispatch identification result, when it is determined that the target network work order meets the dispatch condition, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching result. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.
[0047] In an exemplary embodiment, Figure 2 As shown, a work order management method is provided, which is applied to Figure 1 The following is an example of a server in the example, including:
[0048] S201, analyzing the work order content of the target network work order to be distributed, and obtaining the alarm information of the target network work order.
[0049] The target network work order to be dispatched refers to a network work order in a network operation and maintenance environment. The target network work order may be automatically generated by the system after a possible network failure is identified, or may be generated in response to a processing request for a possible network failure.
[0050] Before dispatching the current network work order to the corresponding operation and maintenance party for processing, the work order content of the target network work order is first analyzed to obtain the alarm information of the target network work order, wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result.
[0051] It should be noted that factors such as system delays caused by communication problems, human operation errors in emergency situations, and errors at the software code level may all cause duplicate dispatch of target network work orders. The operation and maintenance party is responsible for handling network work orders. Specifically, the operation and maintenance party is responsible for handling various situations in the network. Different operation and maintenance parties may be responsible for different areas, such as network equipment fault diagnosis, server performance optimization, and cloud platform resource allocation.
[0052] Exemplarily, a large language model is used to analyze the work order content of the target network work order to be dispatched. The large language model is trained and fine-tuned based on the work order content of different network work orders in a network operation and maintenance environment. It can more accurately understand the question-and-answer interaction in the network operation and maintenance environment, vector-encode the work order content of the network work order, and extract key semantics from it to determine the corresponding alarm information.
[0053] S202, when it is determined that the target network work order meets the dispatching conditions according to the duplicate dispatching identification result, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching result.
[0054] When the duplicate order identification result is that there is a duplicate order, it is determined that the target network work order does not meet the dispatching conditions, and the target network work order is refused to be dispatched to any operation and maintenance party for processing; when the duplicate order identification result is that there is no duplicate order, it is further confirmed whether the target network work order meets the dispatching conditions.
[0055] Optionally, check whether the elements in the work order content of the target network work order (such as fault location, network element name, etc.) are complete, and whether the content description (such as fault phenomenon, etc.) is clear, and then if the work order content check of the target network work order passes, determine that the target network work order meets the dispatch conditions.
[0056] Optionally, when the duplicate dispatch identification result is that there is no duplicate dispatch, it is directly determined that the target network work order meets the dispatching conditions.
[0057] Further, according to the matching result of the operation and maintenance party, the target operation and maintenance party that is most suitable for processing the target network work order is determined, and the target network work order is dispatched to the target operation and maintenance party for processing.
[0058] In the above work order management method, the work order content of the target network work order to be dispatched is analyzed to obtain the alarm information of the target network work order; wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result; when it is determined that the target network work order meets the dispatch conditions according to the duplicate dispatch identification result, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching result. This embodiment analyzes the work order content of the target network work order to achieve effective quality inspection before dispatching the target network work order, especially in terms of duplicate dispatch identification and dispatch operation and maintenance party matching, so as to reasonably dispatch the target network work order to the target operation and maintenance party for processing, thereby improving the processing efficiency and processing effect of the target network work order.
[0059] In an exemplary embodiment, Figure 3 As shown, the above S201 includes:
[0060] S301, generating target question information for the target network work order according to the question information template and the work order content of the target network work order to be dispatched.
[0061] Among them, the question information template is a preset prompt project in the network operation and maintenance environment. Its form is such as "Please call the query function get_recent_order_content to retrieve the work order content in the database in the past 2 hours, and use the text similarity function get_similarity_score to calculate the similarity information one by one. The similarity threshold is set to A. If the threshold is exceeded, it can be considered that there is a duplicate order. The detailed alarm information of the work order is as follows: B." It can be seen that this prompt project is used to indicate how to obtain the duplicate order identification result of the target network work order.
[0062] Furthermore, the conventional question information template is combined with the work order content of the target network work order to be dispatched to generate target question information for the target network work order.
[0063] Optionally, according to the work order content of the target network work order to be dispatched, "A" and "B" in the above example are set to obtain the target question information of the target network work order.
[0064] S302, input the target question information and the work order content of the target network work order into a pre-trained large language model to obtain response information of the target question information.
[0065] The target question information and the response information are represented as one or more matching question-answer pairs.
[0066] The initial large language model is trained and fine-tuned using the work order contents of different network work orders in the network operation and maintenance environment to obtain a trained large language model. When managing the target network work order, the target question information for the target network work order and the work order content of the target network work order are input into the pre-trained large language model together, and the output of the large language model is the response information for the target question information of the target network work order.
[0067] Exemplarily, the above-mentioned prompt project for indicating how to obtain the duplicate dispatch identification result of the target network work order, the response information format that matches the target question information is such as "After comparison, according to the work order alarm information provided, there is a duplicate dispatch, work order id: C, this dispatch is rejected." or "After comparison, according to the work order alarm information provided, there is no duplicate dispatch, and the dispatch process can be entered."
[0068] Exemplarily, a prompt project is used to indicate how to obtain the operation and maintenance party matching result of the target network work order, and the target question information is in the form of "Please analyze the content of this work order, and select the most suitable operation and maintenance party from {operation and maintenance party 1, ..., operation and maintenance party n}, and summarize the main contents of the relevant work orders that the operation and maintenance party has processed for reference. The detailed alarm information of the work order is as follows: D." Correspondingly, the response information is in the form of "Based on the work order alarm information provided, it is planned to dispatch the order to operation and maintenance party m for processing."
[0069] Optionally, a prompt project is used to indicate how to obtain the scenario identification result of the target network work order. The target question information format is such as "Please analyze the problem described in this work order. Which of {alarm scenario name 1,..., alarm scenario name n} is most applicable? Please answer with the scenario name. The following are the glossaries of the relevant scenarios: Alarm scenario name 1: E,..., alarm scenario name n: F, and the detailed alarm information of the work order is as follows: G." Correspondingly, the response format is such as "Based on the work order alarm information provided, the scenario name may be H, the cause of the fault may be I, and the reference solution is J."
[0070] Optionally, the question information template and the work order content of the target network work order are directly input into a pre-trained large language model to obtain response information of the target question information.
[0071] S303: extract the alarm information of the target network work order from the response information.
[0072] Exemplarily, for the duplicate dispatch identification result, the existence or non-existence of duplicate dispatch is extracted from the response information, and for the operation and maintenance party matching result, the target operation and maintenance party to which the order is to be dispatched is extracted from the response information, thereby obtaining the alarm information of the target network work order.
[0073] Optionally, for the scene recognition result, the scene name of the corresponding scene is extracted from the response information.
[0074] Optionally, the above three types of order verification processes are executed serially in the order of duplicate order identification results, scenario identification results and operation and maintenance party matching results to obtain alarm information of the target network work order.
[0075] In this embodiment, before dispatching the target network work order, the generative question-answering capability of the large language model is used to implement question-answering interaction in the network work order dispatch verification. The alarm information of the target network work order is accurately extracted through the large language model in the network operation and maintenance environment. Compared with rule-based keyword matching and text vectorization-based similarity matching, the alarm information is determined based on the cross-domain knowledge and context understanding ability of the large language model, which can be applied to the network work order dispatch verification in a complex network operation and maintenance environment, and significantly improve the accuracy and stability of work order management. In addition, this embodiment conducts evaluation around the judgment of the applicability of alarm scenarios, the identification of duplicate dispatches, and the professional matching of dispatches, breaking the limitations of the traditional method that relies on manual review, and can identify and correct deviations in work order management, providing an automation and quality control method for artificial intelligence technology in network operation and maintenance scenarios.
[0076] In an exemplary embodiment, the above S201 includes:
[0077] Determine the content similarity between the work order content of the target network work order to be dispatched and the work order content of each historical network work order; if any content similarity among the content similarities is higher than the similarity threshold, generate a duplicate dispatch identification result indicating that the target network work order has duplicate dispatch.
[0078] Among them, the historical network work order is a network work order that has been dispatched to the corresponding operation and maintenance party for processing. The historical network work order can be stored in the associated database for calling when the large language model determines the repeated dispatch identification result. The historical time corresponding to the historical network work order can be defined by the prompt project, and this embodiment does not limit this.
[0079] Specifically, a large language model is run to determine the content similarity between the work order content of the target network work order to be dispatched and the work order content of each historical network work order under the instruction of the prompt project, and then each content similarity is compared with the preset similarity threshold; if there is any content similarity higher than the similarity threshold, it means that there is a high similarity between the work order content of the historical network work order and the work order content of the target network work order, and it is identified that there is a duplicate dispatch of the target network work order, and a duplicate dispatch recognition result is generated to characterize that there is a duplicate dispatch of the target network work order; if there is no content similarity higher than the similarity threshold, it means that there is a high similarity between the work order content of the historical network work order and the work order content of the target network work order, and it is identified that there is no duplicate dispatch of the target network work order, and a duplicate dispatch recognition result is generated to characterize that there is no duplicate dispatch of the target network work order.
[0080] In this embodiment, historical network work orders are introduced, and by comparing the similarity between the work order contents, it is determined whether there is a duplicate dispatch of the target network work order. Compared with directly comparing the work order ID (identity) and other identification information of the network work order, the determination based on the similarity between the work order contents has higher accuracy, can effectively verify the target network work order before dispatching, and improve the work order management effect.
[0081] In an exemplary embodiment, Figure 4 As shown, the above S201 includes:
[0082] S401, determining a target operation and maintenance scenario to which the target network work order belongs from various operation and maintenance scenarios according to the work order content of the target network work order to be dispatched.
[0083] Among them, operation and maintenance scenarios refer to various specific scenarios in the network operation and maintenance environment, such as optical cable disconnection, IP (Internet Protocol) circuit failure, and board operating temperature exceeding the limit.
[0084] According to the work order content of the target network work order to be dispatched, context information in the work order content is extracted, and then the target operation and maintenance scenario is identified based on the context information. Optionally, the confidence value corresponding to each operation and maintenance scenario is determined based on the context information, and the operation and maintenance scenario with the highest confidence value is used as the target operation and maintenance scenario.
[0085] Exemplarily, based on a prompt project for indicating how to obtain a scenario recognition result of a target network work order, the corresponding target question information and the work order content of the target network work order are input into a pre-trained large language model together, so that the large language model determines the target operation and maintenance scenario to which the target network work order belongs from various operation and maintenance scenarios, and outputs response information containing the target operation and maintenance scenario.
[0086] S402: Select a target operation and maintenance party from various operation and maintenance parties according to the target operation and maintenance scenario and the work order content of the target network work order.
[0087] Among them, the operation and maintenance party generally refers to the operation and maintenance team in the network operation and maintenance scenario, such as the network equipment fault diagnosis team, the server performance optimization team, and the cloud platform resource allocation team.
[0088] There is a certain correlation between the operation and maintenance scenario and the operation and maintenance party. The large language model can identify the target operation and maintenance party by combining the context information in the target operation and maintenance scenario and the work order content of the target network work order. Optionally, the confidence value corresponding to each operation and maintenance party is determined based on the context information, and the operation and maintenance party with the highest confidence value is used as the target operation and maintenance party.
[0089] Exemplarily, based on a prompting project for indicating how to obtain the operation and maintenance party matching result of the target network work order, the corresponding target question information and the work order content of the target network work order are input into a pre-trained large language model together, so that the large language model determines the target operation and maintenance party from various operation and maintenance parties, and outputs response information containing the target operation and maintenance party.
[0090] Optionally, firstly select multiple highly associated operation and maintenance parties from various operation and maintenance parties according to the target operation and maintenance scenario, and then select the target operation and maintenance party from multiple highly associated operation and maintenance parties according to the work order content of the target network work order.
[0091] Optionally, the target operation and maintenance party is directly selected from various operation and maintenance parties based on the work order content of the target network work order.
[0092] S403: Generate a corresponding operation and maintenance party matching result according to the operation and maintenance strategy associated with the target operation and maintenance party and the target operation and maintenance scenario.
[0093] There is a certain correlation between operation and maintenance scenarios and operation and maintenance strategies. Operation and maintenance strategies are used to represent reference solutions to network failures. Operation and maintenance strategies can be preset or solutions to network failures adopted by other historical network work orders. For the same or similar operation and maintenance scenarios, the same operation and maintenance strategy can be used to solve the problems in the operation and maintenance scenario.
[0094] According to the operation and maintenance strategy associated with the target operation and maintenance party and the target operation and maintenance scenario, an operation and maintenance party matching result is generated. It can be understood that when the steps of this embodiment are executed by a large language model, the operation and maintenance party matching result is the response information of the target question information.
[0095] In this embodiment, before dispatching the target network work order, the applicability of the target network work order to the alarm scenario and the professional matching degree of the dispatch order are evaluated, which can determine a more accurate operation and maintenance party matching result, thereby dispatching the target network work order to the target operation and maintenance party in the operation and maintenance party matching result for processing, and providing the target operation and maintenance party with a reference operation and maintenance strategy, which is conducive to improving the efficiency of subsequent processing of the target network work order and comprehensively improving the effect of work order management.
[0096] In an exemplary embodiment, the work order management method further includes:
[0097] In response to the operation and maintenance record of the target network work order provided by the target operation and maintenance party, the operation and maintenance record of the target network work order is analyzed to obtain quality inspection information of the target network work order.
[0098] Among them, the operation and maintenance record of the target network work order is used to intuitively display the processing flow of the target network work order. The operation and maintenance record of the target network work order can be uploaded by the target operation and maintenance party, or it can be generated in real time in response to the input operation of the target operation and maintenance party on the operation and maintenance terminal.
[0099] By analyzing the operation and maintenance records of the target network work order, a reasonable evaluation of the processing effect of the target network work order can be made, thereby obtaining the quality inspection information of the target network work order.
[0100] For example, Figure 5 As shown, both Role A and Role B are target operation and maintenance parties. For example, Role A can be a dispatching center or a monitoring center, and Role B can be subsidiaries at all levels where the fault alarm is located. In response to the feedback from Role A and Role B, the operation and maintenance record of the target network work order is automatically generated.
[0101] Optionally, the operation and maintenance record includes critical time, and the critical time includes at least two of the work order response time, the cause location time of the network failure corresponding to the target network work order, and the solution provision time.
[0102] The work order response time is used to indicate the time of the first effective response to the target network work order, such as "2023-03-02 23:50:18"; the cause location time of the network fault corresponding to the target network work order, such as "2023-03-02 23:50:18"; the solution provision time of the network fault corresponding to the target network work order, such as "2023-03-02 23:50:18". Optionally, the key time may also include the fault handling time of the network fault corresponding to the target network work order, such as "2023-03-02 23:57:33".
[0103] Optionally, the operation and maintenance records are input into the quality inspection model to obtain the timeliness value corresponding to each key time; according to the timeliness coefficient corresponding to each key time, the timeliness value corresponding to each key time is weighted and summed to obtain the quality inspection score of the target network work order; according to the quality inspection score, the corresponding quality inspection information is generated.
[0104] The quality inspection model is built on the basis of the large language model using knowledge distillation technology. The quality inspection model extracts key times from the operation and maintenance records of the target network work orders, such as the work order response time t f , Cause location time t c and the time t of providing the solution s , and then determine the timeliness value corresponding to each key time.
[0105] Optionally, the following formulas are used to determine the work order response time t f The corresponding timeliness value E f , Cause location time t c The corresponding timeliness value E c and the time t of providing the solution s The corresponding timeliness value E s :
[0106] ;
[0107] ;
[0108] ;
[0109] Furthermore, the timeliness coefficient corresponding to each key time is introduced, such as the work order response time t f Corresponding timeliness coefficient α, cause location time t c The corresponding timeliness coefficient β and solution provision time t s The corresponding timeliness coefficient γ satisfies α+β+γ=1 and α, β and γ are all positive numbers. According to the timeliness coefficient corresponding to each key time, the timeliness value corresponding to each key time is weighted and summed.
[0110] Optionally, the following formula is used to perform weighted sum processing on the timeliness values corresponding to each key time to obtain the quality inspection score Q of the target network work order:
[0111] Q = αEf + βEc + γEs;
[0112] Among them, the quality inspection score Q is used to evaluate the efficiency of the target operation and maintenance party in handling the target network work order.
[0113] According to the quality inspection score, the work order management effect on the target network work order is analyzed to obtain quality inspection information including the quality inspection score, such as a quality inspection report.
[0114] In this embodiment, after the target network work order is dispatched, the operation and maintenance record of the target network work order is identified through the quality inspection model, so as to perform work order quality inspection on the target network work order, and an automated quality inspection information generation function is proposed to provide an instant feedback mechanism for work order management in network operation and maintenance scenarios, and to convert quality inspection information into a data-driven decision support tool, paving an intelligent path for the operation and maintenance party's training needs analysis and strategy planning.
[0115] In an exemplary embodiment, the above quality inspection model is trained in the following manner:
[0116] The large language model is used as the teacher model, and the knowledge distillation technology is used to construct a student model corresponding to the large language model to obtain the initial quality inspection model. According to the operation and maintenance records of the sample network work orders, the initial quality inspection model is iteratively trained for multiple rounds to obtain the trained quality inspection model.
[0117] The information extraction task can be converted into a sequence labeling problem, and question-answer pairs can be constructed on a small amount of labeled data. Then, the large language model can be iteratively trained in multiple rounds so that the large language model can gradually adapt to specific network operation and maintenance scenarios.
[0118] The trained large language model is used to infer the unlabeled data and generate soft labels with probability distribution. These soft labels and a small amount of original labeled data are integrated to help the student model learn information extraction capabilities and obtain an initial quality inspection model. Under the guidance of the large language model, the quality inspection model can not only imitate the output of the large language model, but also learn the decision logic of the large language model and perform semantic analysis on the operation and maintenance records of the target network work order.
[0119] Furthermore, for existing labeled data, the cross entropy loss function L is used ent Measures the difference between the probability distribution predicted by the model and the actual labeled information. For unlabeled data, the consistency loss L is introduced con , ensuring that the large language model can still help the quality inspection model fit the probability distribution pattern in the absence of direct label guidance. Optionally, based on the operation and maintenance records of the sample network work order, the cross entropy loss function L ent The initial quality inspection model is iteratively trained until the difference between the prediction result and the annotation information is lower than the preset threshold, the iteration end condition is met, and a trained quality inspection model is obtained.
[0120] In this embodiment, the large language model is combined with the quality inspection model to form a collaborative semi-supervised training mode, which is used to guide the quality inspection model to learn key information. At the same time, the feedback information of the quality inspection model flows back to the large language model, prompting the large language model to focus more on the learning and strengthening of entity-related knowledge. After successfully extracting key information (such as key time) from the operation and maintenance records of the network work order, this information is combined with the established quality inspection rules to achieve quantitative scoring of the work order quality inspection process. It can be seen that the knowledge distillation technology of the large language model is applied to the field of work order quality inspection. Through fine-tuning and sequence labeling task transformation, a collaborative semi-supervised training environment is constructed to complete the requirements of the models for key information extraction, which not only reduces the dependence on a large amount of labeled data, but also promotes information backflow between models, continuously optimizes model performance, and reflects the dual improvement from data utilization efficiency to model intelligence.
[0121] In an exemplary embodiment, Figure 6 As shown, a work order management method is provided, which is applied to Figure 1 The following is an example of a server in the example, including:
[0122] S601, generating target question information for the target network work order according to the question information template and the work order content of the target network work order to be dispatched.
[0123] S602, input the target question information and the work order content of the target network work order into a pre-trained large language model to obtain response information of the target question information.
[0124] S603: Extract the alarm information of the target network work order from the response information.
[0125] Among them, the alarm information includes the duplicate order identification results and the operation and maintenance party matching results.
[0126] Specifically, determine the content similarity between the work order content of the target network work order to be dispatched and the work order content of each historical network work order; if any content similarity among the content similarities is higher than the similarity threshold, generate a duplicate dispatch identification result indicating that the target network work order has duplicate dispatch.
[0127] Specifically, according to the work order content of the target network work order to be dispatched, the target operation and maintenance scenario to which the target network work order belongs is determined from various operation and maintenance scenarios; according to the target operation and maintenance scenario and the work order content of the target network work order, the target operation and maintenance party is selected from various operation and maintenance parties; according to the operation and maintenance strategy associated with the target operation and maintenance party and the target operation and maintenance scenario, the corresponding operation and maintenance party matching result is generated.
[0128] S604, when it is determined that the target network work order meets the dispatching conditions according to the duplicate dispatching identification result, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching result.
[0129] S605, in response to the operation and maintenance record of the target network work order provided by the target operation and maintenance party, the operation and maintenance record is input into the quality inspection model to obtain the timeliness value corresponding to each key time.
[0130] The operation and maintenance record includes key time, which includes at least two of the work order response time, the cause location time of the network failure corresponding to the target network work order, and the solution provision time.
[0131] S606: Perform weighted sum processing on the timeliness values corresponding to each key time according to the timeliness coefficient corresponding to each key time, and obtain the quality inspection score of the target network work order.
[0132] S607, generating corresponding quality inspection information according to the quality inspection score.
[0133] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0134] Based on the same inventive concept, the embodiment of the present application also provides a work order management device for implementing the work order management method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more work order management device embodiments provided below can refer to the limitations on the work order management method above, and will not be repeated here.
[0135] In an exemplary embodiment, Figure 7 As shown, a work order management device is provided, including: a content analysis module 10 and a work order dispatching module 20, wherein:
[0136] The content analysis module 10 is used to analyze the work order content of the target network work order to be distributed and obtain the alarm information of the target network work order.
[0137] Among them, the alarm information includes the duplicate order identification results and the operation and maintenance party matching results.
[0138] The work order dispatching module 20 is used to dispatch the target network work order to the corresponding target operation and maintenance party according to the matching result of the operation and maintenance party when it is determined that the target network work order meets the dispatching conditions according to the duplicate dispatching identification result.
[0139] In an exemplary embodiment, the content analysis module 10 includes:
[0140] The question generating unit is used to generate target question information for the target network work order according to the question information template and the work order content of the target network work order to be dispatched.
[0141] The response determination unit is used to input the target question information and the work order content of the target network work order into a pre-trained large language model to obtain response information of the target question information.
[0142] The information extraction unit is used to extract the alarm information of the target network work order from the response information.
[0143] In an exemplary embodiment, the content analysis module 10 includes:
[0144] The similarity determination unit is used to determine the content similarity between the work order content of the target network work order to be dispatched and the work order content of each historical network work order.
[0145] The first analysis unit is configured to generate a duplicate dispatch identification result indicating that duplicate dispatch exists in the target network work order when any content similarity among the content similarities is higher than a similarity threshold.
[0146] In an exemplary embodiment, the content analysis module 10 includes:
[0147] The scenario determination unit is used to determine the target operation and maintenance scenario to which the target network work order belongs from various operation and maintenance scenarios according to the work order content of the target network work order to be dispatched.
[0148] The operation and maintenance party selection unit is used to select a target operation and maintenance party from various operation and maintenance parties according to a target operation and maintenance scenario and the work order content of a target network work order.
[0149] The second analysis unit is used to generate a corresponding operation and maintenance party matching result according to the operation and maintenance strategy associated with the target operation and maintenance party and the target operation and maintenance scenario.
[0150] In an exemplary embodiment, the work order management device further includes:
[0151] The work order quality inspection module is used to respond to the operation and maintenance records of the target network work order provided by the target operation and maintenance party, analyze the operation and maintenance records of the target network work order, and obtain quality inspection information of the target network work order.
[0152] In an exemplary embodiment, the operation and maintenance record includes key time, which includes at least two of the work order response time, the cause location time of the network failure corresponding to the target network work order, and the solution provision time; accordingly, the work order quality inspection module includes:
[0153] The timeliness determination unit is used to input the operation and maintenance records into the quality inspection model to obtain the timeliness value corresponding to each key time.
[0154] The weighted summation unit is used to perform weighted summation processing on the timeliness values corresponding to each key time according to the timeliness coefficient corresponding to each key time, so as to obtain the quality inspection score of the target network work order.
[0155] The third analysis unit is used to generate corresponding quality inspection information according to the quality inspection score.
[0156] In an exemplary embodiment, the work order management device further includes:
[0157] The knowledge distillation module is used to use the large language model as the teacher model and adopt the knowledge distillation technology to build a student model corresponding to the large language model to obtain the initial quality inspection model.
[0158] The iterative training module is used to perform multiple rounds of iterative training on the initial quality inspection model according to the operation and maintenance records of the sample network work orders to obtain a trained quality inspection model.
[0159] Each module in the above work order management device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.
[0160] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 8As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store historical network work order data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a work order management method is implemented.
[0161] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0162] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps of the above-mentioned work order management method when executing the computer program.
[0163] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned work order management method are implemented.
[0164] In one embodiment, a computer program product is provided, including a computer program, which implements the steps of the above-mentioned work order management method when executed by a processor.
[0165] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0166] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0167] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A work order management method, characterized in that: The method comprises: Analyze the work order content of the target network work order to be dispatched, and obtain the alarm information of the target network work order; wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result; When it is determined that the target network work order meets the dispatching conditions according to the duplicate dispatching identification result, the target network work order is dispatched to the corresponding target operation and maintenance party according to the operation and maintenance party matching result.
2. The method according to claim 1, characterized in that The analyzing the work order content of the target network work order to be dispatched to obtain the alarm information of the target network work order includes: Generate target question information for the target network work order according to the question information template and the work order content of the target network work order to be dispatched; Inputting the target question information and the work order content of the target network work order into a pre-trained large language model to obtain response information of the target question information; Extract the alarm information of the target network work order from the response information.
3. The method according to claim 1, characterized in that: The analyzing the work order content of the target network work order to be dispatched to obtain the alarm information of the target network work order includes: Determine the similarity between the work order content of the target network work order to be dispatched and the work order content of each historical network work order; If any content similarity among the content similarities is higher than the similarity threshold, a duplicate dispatch identification result is generated, indicating that duplicate dispatch exists in the target network work order.
4. The method according to claim 1, characterized in that: The analyzing the work order content of the target network work order to be dispatched to obtain the alarm information of the target network work order includes: According to the work order content of the target network work order to be dispatched, determining the target operation and maintenance scenario to which the target network work order belongs from various operation and maintenance scenarios; Selecting a target operation and maintenance party from various operation and maintenance parties according to the target operation and maintenance scenario and the work order content of the target network work order; According to the operation and maintenance strategy associated with the target operation and maintenance party and the target operation and maintenance scenario, a corresponding operation and maintenance party matching result is generated.
5. The method according to claim 2, characterized in that: The method further comprises: In response to the operation and maintenance record of the target network work order provided by the target operation and maintenance party, the operation and maintenance record of the target network work order is analyzed to obtain quality inspection information of the target network work order.
6. The method according to claim 5, characterized in that The operation and maintenance record includes key time, and the key time includes at least two of the work order response time, the cause location time of the network failure corresponding to the target network work order, and the solution provision time; Accordingly, the operation and maintenance record of the target network work order is analyzed to obtain quality inspection information of the target network work order, including: Input the operation and maintenance records into the quality inspection model to obtain the timeliness value corresponding to each key time; According to the timeliness coefficient corresponding to each key time, the timeliness value corresponding to each key time is weighted and summed to obtain the quality inspection score of the target network work order; According to the quality inspection score, corresponding quality inspection information is generated.
7. The method according to claim 6, characterized in that The quality inspection model is trained in the following way: The large language model is used as a teacher model, and the knowledge distillation technology is used to construct a student model corresponding to the large language model to obtain an initial quality inspection model; According to the operation and maintenance records of the sample network work orders, the initial quality inspection model is iteratively trained for multiple rounds to obtain a trained quality inspection model.
8. A work order management device, characterized in that: The device comprises: A content analysis module, used to analyze the work order content of the target network work order to be dispatched, and obtain the alarm information of the target network work order; wherein the alarm information includes the duplicate dispatch identification result and the operation and maintenance party matching result; The work order dispatching module is used to dispatch the target network work order to the corresponding target operation and maintenance party according to the matching result of the operation and maintenance party when it is determined that the target network work order meets the dispatching conditions according to the duplicate dispatching identification result.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
Citation Information
Cited By
Power supply service work order intelligent monitoring and early warning method and system based on large model
CN122656368A
