Operation ticket auditing rule self-learning method, operation ticket auditing method, system and related device

Abstract

Embodiments of the present application provide an operation ticket auditing method and system and related equipment. In the method, a user interacts with a self-learning intelligent agent, the self-learning intelligent agent uses a generative AI model to identify the intent of the current conversation information input by the user in the interaction, and if the intent category of the current conversation information is identified as a new operation ticket auditing rule, the self-learning intelligent agent uses the generative AI model in combination with a knowledge graph and a dispatch professional knowledge base to analyze the new operation ticket auditing rule in the current conversation information and obtain rule analysis content of the new operation ticket auditing rule. Thus, the self-learning of the new operation ticket auditing rule is completed, and an efficient, intelligent and self-learning operation ticket auditing rule self-learning solution is provided. In this way, in the subsequent operation ticket auditing stage, the self-learned operation ticket auditing rule can be used to efficiently and accurately automatically audit the operation ticket.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a self-learning method for operation ticket review rules, an operation ticket review method, a system, and related equipment. Background Technology

[0002] Operation tickets, as crucial written documentation for ensuring safe and standardized operations, are widely used in various industries, such as power systems, petrochemicals, and rail transportation. Operation tickets meticulously record every step of a work task, guiding operators to standardize work procedures, ensuring operational safety, reducing operational errors, improving work efficiency, and protecting the safety of personnel and equipment.

[0003] Currently, the review of operation tickets still relies primarily on manual, item-by-item verification. However, manual review is inefficient, and due to the complexity and diversity of operation tickets, reviewers are prone to overlooking crucial details due to fatigue, lack of experience, or distraction, thus affecting the safety and standardization of operations. Therefore, it is essential to research how to develop automated operation ticket review rules for efficient and accurate review. Summary of the Invention

[0004] This application provides a self-learning method for operation ticket review rules, an operation ticket review method, a system, and related equipment. It provides an efficient, intelligent, and self-learning solution for operation ticket review rules, so that during the operation ticket review stage, the self-learned operation ticket review rules can be used to perform efficient and accurate automated review of operation tickets.

[0005] This application provides a self-learning method for operation ticket review rules, applied to a self-learning intelligent agent. The method includes: acquiring the current dialogue information input by the user in the current dialogue interaction; using a generative AI model to perform intent recognition on the current dialogue information to obtain the intent category of the current dialogue information; if the intent category of the current dialogue information is a new operation ticket review rule, then using a generative AI model combined with a knowledge graph and a scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule.

[0006] This application provides a method for reviewing operation tickets, including: obtaining current dialogue information input by the user in the current dialogue interaction; using a generative AI model to perform intent recognition on the current dialogue information; if the intent category of the current dialogue information is identified as a new operation ticket review rule, then using the generative AI model in combination with a knowledge graph and a scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule; using the generative AI model to review the operation instructions in the operation ticket to be reviewed according to the rule parsing content, and outputting the review result of the operation ticket to be reviewed.

[0007] This application provides an operation ticket processing system, including: a terminal device and a server. The server runs a self-learning intelligent agent and an auditing intelligent agent. The terminal device is used to respond to user input operations on a dialogue interaction interface, obtain the current dialogue information input by the user in the current dialogue interaction and send it to the server, and display the audit result of the operation ticket to be audited returned by the server on the dialogue interaction interface. The server is used to use the self-learning intelligent agent to perform intent recognition on the current dialogue information using a generative AI model. If the intent category of the current dialogue information is identified as a new operation ticket audit rule, the server uses the generative AI model combined with a knowledge graph and a scheduling professional knowledge base to parse the new operation ticket audit rule in the current dialogue information to obtain the rule parsing content of the new operation ticket audit rule. The auditing intelligent agent uses the generative AI model to audit the operation instructions in the operation ticket to be audited according to the rule parsing content, and sends the audit result of the operation ticket to be audited to the terminal device.

[0008] This application also provides an electronic device, including: a memory and a processor; the memory for storing a computer program; the processor coupled to the memory for executing the computer program to perform steps in the operation ticket review rule self-learning method or the operation ticket review method.

[0009] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in the operation ticket review rule self-learning method or the operation ticket review method.

[0010] This application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, enables the processor to implement the steps in the operation ticket review rule self-learning method or the operation ticket review method.

[0011] The technical solution provided in this application involves a user interacting with a self-learning intelligent agent. The agent utilizes a generative AI model to identify the intent of the user's input dialogue information. If the intent category of the current dialogue information is identified as a rule for adding an operation ticket review, the agent uses the generative AI model combined with a knowledge graph and a scheduling professional knowledge base to parse the rule for adding an operation ticket review in the current dialogue information, obtaining the rule parsing content. This completes the self-learning of the rule for adding an operation ticket review, providing an efficient, intelligent, and self-learning solution for operation ticket review rules. Subsequently, during the operation ticket review stage, the self-learned operation ticket review rules can be used to efficiently and accurately automate the review of operation tickets, providing an efficient, intelligent, and self-learning operation ticket review solution that effectively ensures the security, standardization, and correctness of operation tickets. Attached Figure Description

[0012] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0013] Figure 1a An exemplary application scenario diagram provided for an embodiment of this application;

[0014] Figure 1b A flowchart illustrating a self-learning method for operation ticket review rules provided in this application embodiment;

[0015] Figure 2 A flowchart of an operation ticket review method provided in this application embodiment;

[0016] Figure 3 A flowchart illustrating another operation ticket review method provided in this application embodiment;

[0017] Figure 4 This is another exemplary application scenario diagram provided for embodiments of this application;

[0018] Figure 5 A system architecture diagram of an operation ticket review system provided in this application embodiment;

[0019] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the access relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone, where A and B can be singular or plural. In the textual description of this application, the character " / " generally indicates that the preceding and following associated objects have an "or" relationship. Furthermore, in the embodiments of this application, "first," "second," "third," etc., are only used to distinguish the content of different objects and have no other special meaning.

[0022] It should be noted that, in the cases involving user information in the embodiments of this application, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse. In addition, the various models involved in this application (including but not limited to language models or large models) comply with relevant laws and standards.

[0023] First, let me introduce some terms used in the embodiments of this application:

[0024] AI Agent: A software or hardware entity that has the ability to autonomously perceive its environment, make decisions, and execute actions, with the aim of completing specific tasks or achieving goals in a simulated or real world.

[0025] A language model (LM) is a model that learns general language structures and knowledge through unsupervised or self-supervised pre-training on large-scale text data. Language model architectures include, but are not limited to: bidirectional encoder representation from transformers (BERT), autoregressive language models, and generative pre-trained transformers (GPT). The Transformer module is a neural network structure based on a self-attention mechanism, which significantly improves model performance through parallel processing and self-attention.

[0026] Large Language Models (LLMs), also known as large-scale language models, refer to a class of natural language processing models with an extremely large number of parameters. LLMs are typically based on deep learning architectures, especially the Transformer architecture, which learns the complex structure of language and rich contextual information through pre-training on massive amounts of text data. The Transformer architecture addresses the bottleneck problem of traditional neural network models when processing long sequences by introducing a self-attention mechanism, and its highly parallelizable nature greatly improves training efficiency. The Transformer architecture includes either an encoder or a decoder.

[0027] Reasoning models are language models with reasoning capabilities. They can perform complex thinking processes such as logical inference, causal analysis, and relationship identification based on existing information, and can also derive reasonable conclusions or solutions when faced with complex problems through reasoning.

[0028] Generative AI models are algorithmic models that use artificial intelligence technology to automatically generate content such as text, images, audio, video, and code.

[0029] Chain of Thought (CoT) is a prompt engineering methodology that guides a language model to progressively break down complex problems and demonstrate the reasoning process, enabling it to better complete multi-step tasks. CoT emphasizes the interpretability of the language model, facilitating user understanding and verification of its reasoning results.

[0030] A knowledge graph (KG) is a graph-based knowledge representation used to describe relationships between entities. It uses nodes (entities) and edges (relationships) to represent entities and the relationships between them.

[0031] Graph databases are databases specifically designed for storing, managing, and querying graph data models (composed of nodes and edges). They provide the ability to efficiently handle complex network structures and are suitable for applications requiring rapid retrieval of related information. Knowledge graphs typically use graph databases as their underlying storage mechanism.

[0032] Retrieval Augmented Generation (RAG) is a deep learning architecture that integrates information retrieval and text generation, aiming to improve the understanding and generation capabilities of language models in specific tasks.

[0033] The technical solutions of this application and how they solve the aforementioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The technical solutions provided by each embodiment of this application are described in detail below with reference to the accompanying drawings.

[0034] Figure 1a This is an exemplary application scenario diagram provided for an embodiment of this application. See also... Figure 1a The intelligent agent can include a self-learning intelligent agent or an operation ticket review intelligent agent. A self-learning intelligent agent is one that uses artificial intelligence technology to learn operation ticket review rules; an operation ticket review intelligent agent is one that uses artificial intelligence technology to automatically review operation tickets. Supported by a knowledge graph built on a graph database and a scheduling professional knowledge base, the self-learning intelligent agent integrates a generative AI model (which can be a reasoning model) to construct an efficient, intelligent, and self-learning intelligent operation ticket processing system. Through one or more rounds of interactive feedback with users (e.g., dispatchers), the self-learning intelligent agent forms a closed-loop mechanism of "knowledge learning - reasoning verification - feedback iteration," which can accumulate newly added operation ticket review rules and other content into long-term memory, continuously optimize the rule base, and constantly improve the intelligence level of the operation ticket processing system.

[0035] In practical applications, graph databases can provide various operation instructions, anti-misoperation rules, and topologies (e.g., power topologies). Topologies reflect the relationships between operational objects, typically represented as graphs showing the connections between them (e.g., switches, disconnectors, busbars, transformers). Operation instructions are specific commands issued to achieve particular operational goals and are the basic units in operation tickets. Anti-misoperation rules are a set of logical rules used to prevent erroneous operations during the operation process. Taking power dispatching as an example, these rules are usually based on power grid safety regulations, five-prevention rules, and historical accident cases. Multiple knowledge graphs can be constructed based on graph databases, including but not limited to: topology graphs reflecting the relationships between operational objects, operation instruction graphs, and anti-misoperation rule graphs.

[0036] Taking the power dispatching industry as an example, the power topology diagram reflects the connection relationships and hierarchical structure between various devices in the power grid, including but not limited to the electrical connections between devices such as substations, lines, transformers, circuit breakers, and disconnectors. Entities in the power topology diagram include, but are not limited to, substations, busbars, lines, switches, disconnectors, and transformers; relationships between entities include, but are not limited to, "connected to," "belongs to," "outgoing line," "feedback," "supplying to," and "parallel operation." Entity attribute information includes, but is not limited to, voltage level, equipment status (operating / out of service), and capacity.

[0037] The "Connected to" relationship type indicates an electrical connection between devices, for example, "5011 switch → connected to → 110kV bus". The "Belong to" relationship type indicates a subordinate relationship between devices, for example, "5011 switch → belongs to → substation A". The "Power to" relationship type indicates the direction of energy transfer between devices, for example, "110kV line L1 → power to → substation B". The "Parallel operation" relationship type indicates that multiple devices are operating in parallel, for example, "Main transformer 1 → parallel operation → main transformer 2". The "Outgoing line" relationship type indicates an output line from a substation, for example, "Substation A → outgoing line → L1". The "Input / Output" relationship type indicates the voltage conversion relationship of a transformer; for example, "Main transformer high-voltage side → input → 110kV; main transformer low-voltage side → output → 10kV". The "Feed-in" relationship type indicates that the distribution network supplies power to the load, for example, "10kV feeder F1 → feed-in → user distribution box".

[0038] In this embodiment, the operation instruction graph is a knowledge graph used to express and manage various operation instructions and their logical relationships. It connects entities such as operation verbs, operation objects, and state descriptions through relationships (such as operation order, dependency relationship, and security constraints) to form a complex knowledge network that supports functions such as intelligent analysis and decision assistance.

[0039] Taking the power dispatching industry as an example, the entities in an operation instruction diagram include: operation verbs, operation objects, and status descriptions. Operation verbs include, but are not limited to: "close," "open," "engage," "exit," "switch," "test for voltage," and "connect grounding wire." Operation objects refer to the specific equipment to be operated, including, but not limited to: switches, disconnectors, lines, busbars, transformers, and protection pressure plates. Status descriptions describe the current or expected state of the equipment, including, but not limited to: "running," "power outage," "disconnected," "opened," and "equipment restored to normal." Relationships between entities include, but are not limited to: operation objects, operation sequence, dependencies, safety constraints, prerequisites, and consequences. Operation objects indicate which equipment the operation applies to; operation sequence indicates the order in which different operations occur, such as "disconnect switch 5011 first, then open disconnector 5011-1." Dependencies indicate whether an operation depends on the completion of other operations, such as "the main transformer switching operation can only be performed after all load transfers are completed." Safety constraints are operations permitted or prohibited under specific conditions, such as "it is strictly forbidden to open disconnectors while the circuit is energized." Preconditions represent the conditions that must be met before performing the operation; consequences represent the impact or result of performing the operation. Entity attribute information includes, but is not limited to, execution steps, risk assessment information, etc. Preconditions represent the conditions that must be met before performing the operation, such as "confirm that the adjacent area has been de-energized." Execution steps reflect detailed operating guidelines, including necessary checkpoints and precautions. Risk assessment information is used to analyze the potential risks of the operation and corresponding control measures.

[0040] For example, the "Main Transformer Power Supply" operation process is represented in the operation instruction diagram as follows: {[Close]—(Operation Object)—>[5011 Switch]; [Close]—(Prerequisite)—>[5011-1 Disconnect Switch Opened]; [Close]—(Consequence)—>[High Voltage Side Electrified]; [Close]—(Operation Object)—>[5012 Switch]; [Close]—(Prerequisite)—>[5011 Switch Closed]; [Close]—(Consequence)—>[Low Voltage Side Electrified]; [Activate]—(Operation Object)—>[No. 1 Main Transformer Protection Panel]; [Activate]—(Prerequisite)—>[Main Transformer Powered and No Abnormalities]; [Activate]—(Consequence)—>[Protection Function Activated]; [Check]—(Object)—>[10kV Bus Voltage]; [Check]—(Consequence)—>[Confirm Power Supply Normal]}.

[0041] In this embodiment, the anti-misoperation rule graph is a knowledge graph that semantically models the rules for preventing misoperation. It can express which operations cannot be performed, under what conditions they are prohibited from being performed, what consequences will occur if an operation is violated, and what control measures should be taken to avoid misoperation.

[0042] Taking the power dispatching industry as an example, the entities in the anti-misoperation rule graph include, but are not limited to: the name of the anti-misoperation rule, the triggering conditions, the risk consequences, the control measures, the operation objects, etc.

[0043] The name of the anti-misoperation rule is used to identify the specific anti-misoperation rule. Anti-misoperation rules include, but are not limited to: preventing the operation of a live switch, preventing accidental entry into a live compartment, preventing accidental closing of a grounding switch, and preventing accidental activation of a protection switch, etc.

[0044] A trigger condition is a condition that causes a rule preventing misoperation to take effect. Identifying trigger conditions is crucial for determining whether a specific operation is permitted. Trigger conditions include, but are not limited to: a live line, a switch not being disconnected, incomplete maintenance, failure to verify voltage, and equipment being in operation. For example, "[Preventing live disconnection of switch]—Trigger condition → [Live line]".

[0045] This section describes the potential consequences or impacts of violating the relevant anti-misoperation rules. This helps assess the potential risk level and take appropriate preventative measures. Risk consequences include, but are not limited to: short-circuit tripping, equipment damage, personal injury, protection malfunction, or area power outage.

[0046] Control measures are specific procedures or steps designed to prevent misoperation. These measures typically require manual verification or adherence to specific operating procedures. Examples of control measures include, but are not limited to: requiring switches to be disconnected first, requiring voltage testing and recording, requiring safety barriers, setting up warning signs, or requiring verification by a responsible person.

[0047] The object of operation refers to the specific equipment or component involved in the execution of the operation. Objects of operation include, but are not limited to: disconnectors, grounding disconnectors, switches, busbars, transformers, or protection pressure plates. Relationships between entities in the anti-misoperation rule diagram include, but are not limited to: triggering conditions, causes, control measures, or belonging to. The "cause" relationship reflects the possible consequences of a violation; the "belong to" relationship reflects which type of anti-misoperation system a particular rule belongs to.

[0048] For example, the knowledge information in the anti-misoperation rule diagram is as follows: {[Preventing live disconnection of switch] —(Triggering condition) —>[Live line]; [Preventing live disconnection of switch] —(Cause) —>[Short circuit trip]; [Preventing live disconnection of switch] —(Control measures) —>[Switch must be disconnected first]; [Preventing accidental closing of grounding switch] —(Triggering condition) —>[Live line]; [Preventing accidental closing of grounding switch] —(Cause) —>[Short circuit fault]; [Preventing accidental closing of grounding switch] —(Control measures) —>[Electrical test and record must be performed first]}.

[0049] In practical applications, a dispatching expertise base refers to a systematic collection of information established within a specific field (such as power systems, petrochemicals, and rail transportation) to support the efficient execution of dispatching work. The dispatching expertise base includes, but is not limited to, the following information: operational procedures, expert experience, and rule bases.

[0050] Work procedures and specifications typically refer to a series of operational guidelines, standards, and regulations established in a specific industry or work environment to ensure the safety, quality, and efficiency of work processes. These procedures and specifications may be developed internally by a company or published by industry associations, national standards bodies, or international organizations to guide employees on how to correctly and safely complete their work tasks.

[0051] In practical applications, various industries have their own operating procedures and standards. Taking the power dispatching industry as an example, operating procedures and standards include, but are not limited to: power system dispatching procedures, power dispatching operation procedures, power grid accident handling procedures, power system stable operation procedures, relay protection and automatic device operation procedures, power grid dispatch automation system operation procedures, or new energy grid access dispatching procedures.

[0052] The Power System Dispatching Regulations define the principles of power system dispatching and management, the responsibilities of dispatching agencies at all levels, and their coordination mechanisms. It includes provisions on power system operation mode arrangements, load forecasting and management, generation planning, frequency and voltage adjustments, and other related aspects.

[0053] The power dispatching operation procedures detail the code of conduct for power dispatchers in their daily operations, including specific steps for operating switching equipment and switching lines. They emphasize the safety checks and simulations that must be conducted before any operation, as well as the principles to be followed during operation, such as the "three comparisons" and "three prohibitions."

[0054] The power grid accident handling procedures guide dispatchers on how to quickly and effectively respond to sudden or emergency events in the power grid, including fault identification, isolation, and power restoration processes. The aim is to minimize the scope of power outages and restore normal power supply as soon as possible.

[0055] The regulations for stable operation of power systems are formulated to meet the specific requirements for maintaining the stability of power systems, including but not limited to monitoring and regulation measures for voltage stability, frequency stability, etc.

[0056] The operating procedures for relay protection and automatic devices cover the configuration principles, setting calculations, commissioning and maintenance of relay protection devices and automatic control systems (such as automatic reclosing devices), aiming to protect the power system from the impact of faults and quickly restore the system to normal operation.

[0057] The Operating Procedures for Power Grid Dispatch Automation Systems describe the relevant technical standards and workflows for achieving power grid dispatch automation using modern information technology, including the application and maintenance of Energy Management Systems (EMS), Distribution Management Systems (DMS), and so on.

[0058] With the development of renewable energy, specific technical requirements and dispatching strategies have been formulated for the grid connection of new energy sources such as wind power and photovoltaics to ensure the coordinated development of new energy and traditional energy.

[0059] In practical applications, the expert experience knowledge in the dispatching professional knowledge base varies depending on the industry. Expert experience knowledge refers to the experience and knowledge accumulated and summarized by professionals with a strong theoretical foundation and rich practical experience in a specific field. In the power dispatching industry, expert experience knowledge is a valuable resource accumulated through long-term practice. It not only covers technical knowledge but also includes the ability to judge complex situations, experience in handling emergencies, and methods for optimizing power grid operation.

[0060] In this embodiment, the rule base primarily provides multiple standardized operation ticket review rules and related information for each rule. Operation ticket review rules are guidelines used to ensure that operation tickets meet all necessary safety and operational requirements.

[0061] It should be noted that, Figure 1a The application scenario shown is merely an example, and the embodiments of this application do not limit the application scenario.

[0062] Figure 1b A flowchart illustrating a self-learning method for operation ticket review rules is provided in this application embodiment. This method can be applied to a self-learning intelligent agent. See [link / reference]. Figure 1b The method may include the following steps:

[0063] 101. Obtain the current dialogue information entered by the user in this dialogue interaction;

[0064] 102. Use a generative AI model to identify the intent of the current dialogue information and obtain the intent category of the current dialogue information;

[0065] 103. If the intent category of the current dialogue information is to add an operation ticket review rule, then the generative AI model combined with the knowledge graph and scheduling professional knowledge base is used to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule.

[0066] Optionally, the rule parsing content includes: operation instruction review type, operation instruction filtering mind chain, and operation instruction review mind chain; operation instruction review type is used to indicate whether a single operation instruction is reviewed independently or multiple operation instructions are reviewed jointly; operation instruction filtering mind chain is used to indicate the filtering method for selecting at least one target operation instruction to be reviewed from at least one operation instruction included in the operation ticket to be reviewed; operation instruction review mind chain is used to indicate the review process for adding a new operation ticket review rule.

[0067] Optionally, the above also includes: in response to the user confirming an error in the review result of the operation ticket to be reviewed, obtaining the next dialogue information entered by the user in the next dialogue interaction; wherein, the review result of the operation ticket to be reviewed is obtained by reviewing the operation instructions in the operation ticket to be reviewed according to the rule parsing content; using a generative AI model to perform intent recognition on the next dialogue information; if the intent category of the next dialogue information is identified as updating the rule parsing content of the new operation ticket review rule, then the generative AI model is used to update the rule parsing content of the new operation ticket review rule according to the next dialogue information.

[0068] Optionally, the rule parsing content of the newly added operation ticket review rule is updated based on the next dialogue information, including: if the intent category of the next dialogue information is to supplement example information, then example information is added to the rule parsing content of the newly added operation ticket review rule. The example information includes positive example information and / or negative example information. Positive example information includes an example of the operation instruction and its review result after passing the operation ticket review rule example review; negative example information includes an example of the operation instruction and its review result after failing the operation ticket review rule example review. If the intent category of the next dialogue information is to supplement background knowledge, then background knowledge is added to the rule parsing content of the newly added operation ticket review rule. Background knowledge refers to the knowledge information required to assist in reviewing the operation instructions of the operation ticket to be reviewed using the newly added operation ticket review rule. If the intent category of the next dialogue information is to modify the operation instruction filtering mind chain, the operation instruction filtering mind chain in the rule parsing content of the newly added operation ticket review rule is modified according to the next dialogue information. If the intent category of the next dialogue information is to modify the operation instruction review mind chain, the operation instruction review mind chain in the rule parsing content of the newly added operation ticket review rule is modified according to the next dialogue information.

[0069] Optionally, the above method further includes: if the intent category of the current dialogue information is not identified, then output guiding information, which is used to guide the user to input dialogue information related to the operation ticket review; if the intent category of the current dialogue information is identified as an intent category unrelated to the operation ticket review, then output a fallback statement.

[0070] For details on how the self-learning intelligent agent automatically learns the operation ticket review rules, please refer to the following content.

[0071] The technical solution provided in this application involves a user interacting with a self-learning intelligent agent. The agent utilizes a generative AI model to identify the intent of the user's input dialogue information. If the intent category of the current dialogue information is identified as a rule for adding an operation ticket review, the agent uses the generative AI model combined with a knowledge graph and a scheduling professional knowledge base to parse the rule for adding an operation ticket review in the current dialogue information, obtaining the rule parsing content. This completes the self-learning of the rule for adding an operation ticket review, providing an efficient, intelligent, and self-learning solution for operation ticket review rules. Subsequently, during the operation ticket review stage, the self-learned operation ticket review rules can be used to efficiently and accurately automate the review of operation tickets, providing an efficient, intelligent, and self-learning operation ticket review solution that effectively ensures the security, standardization, and correctness of operation tickets.

[0072] Figure 2 A flowchart illustrating an operation ticket review method provided in this application embodiment. See also... Figure 2 The method may include the following steps:

[0073] 201. Obtain the current dialogue information entered by the user in this dialogue interaction.

[0074] 202. Use a generative AI model to identify the intent of the current dialogue information. If the intent category of the current dialogue information is identified as a new operation ticket review rule, then use the generative AI model in combination with the knowledge graph and scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information and obtain the rule parsing content of the new operation ticket review rule.

[0075] 203. Use a generative AI model to parse the content of the operation instructions in the operation ticket to be reviewed according to the rules, review the operation instructions, and output the review result of the operation ticket to be reviewed.

[0076] In practical applications, steps 201 to 203 can be executed by a self-learning agent, and step 203 can be executed by an operation ticket review agent.

[0077] In practical applications, a complete operation ticket typically includes the following information: operation ticket number, operation task, operation time, operator, or list of operation steps. The list of operation steps includes multiple ordered operation steps, and the step information for each step typically includes: step number (indicating the operation sequence), operation instruction, device name (used to specify the operation object), and operation status (e.g., pending execution / in execution / completed).

[0078] For example, the pending operation ticket in a power dispatching scenario is as follows:

[0079] {

[0080] Operation ticket number: OPE-20250604-014

[0081] Operational task: Switch L1 line from cold standby to operation and perform series power supply operation.

[0082] Operation time: 2025-06-04 09:00~11:00

[0083] Operator: Zhang San; Guardian: Li Si

[0084] Operation steps list:

[0085] Operation Step 1: {Operation Command: Close the disconnect switches on both sides of L1 line; Equipment Name: Disconnect Switches on both sides of L1 line; Operation Status: Pending Execution};

[0086] Operation Step 2: {Operation Command: Check L1 line for no voltage; Device Name: L1 line; Operation Status: Pending};

[0087] Operation Step 3: {Operation Instruction: Close L1 Line Switch 101; Device Name: L1 Line Switch 101; Operation Status: Pending};

[0088] Operation Step 4: {Operation Instruction: Switch L1 to run and supply power in series; Equipment Name: Switch L1 to run; Operation Status: Pending};

[0089] }

[0090] To ensure the security and compliance of operational procedures, it is necessary to review the operational instructions in the operational ticket using operational ticket review rules to guarantee the security and standardization of the entire operational process. Operational ticket review rules can review the operational instructions in the operational ticket from one or more dimensions, such as standardization checks, completeness checks, or correctness checks. Understandably, the more dimensions involved, the better the security and compliance of the operational ticket.

[0091] The standardization inspection mainly checks whether the language, terminology, and equipment naming in the operation instructions on the operation ticket comply with relevant operating procedures and standards. By standardizing language and naming conventions, human error in understanding is reduced, and operational safety and compliance are improved.

[0092] Optionally, the standardization check may include one or more of the following items: terminology consistency check, double naming check, ambiguous terminology check, and format uniformity check.

[0093] The terminology consistency check item is mainly used to check whether standard scheduling terms (such as "open", "close", "disconnect") are used. For example, "open the knife switch" in the operation instruction should be written as "pull open the knife switch".

[0094] The dual naming check is mainly used to check whether the equipment name uses the format of number and equipment type. For example, "switch" in the operation instruction should be written as "101 switch".

[0095] The ambiguous terminology check item is mainly used to check whether there are uncertain or ambiguous descriptions (such as "may", "probably", "should") in the operation instructions. For example, the operation instruction "should open the grounding switch" → "open the #1 main transformer grounding switch" should be written as "open the #1 main transformer grounding switch".

[0096] The format consistency check item is mainly used to check whether the writing format of each operation instruction is consistent. For example, the operation instruction "confirm status after disconnecting switch 101" should be written in two steps: "disconnect switch 101" and "confirm that switch 101 has been disconnected".

[0097] The integrity check primarily examines whether any items are missing from the operation ticket or instructions, determining if any mandatory operational steps required by the work procedures have been omitted. Omissions include, but are not limited to: not opening the grounding switch, not activating protection devices, or not checking the switch status. For example, if the "grounding switch not opened" omission exists, there is a risk of closing the circuit with the ground wire connected if the grounding switch was not opened before the line was switched from maintenance to operation. If the "protection device not activated" omission exists, operation without protection will occur if the short-lead protection was not activated before the power outage. If the "switch status not checked" omission exists, failure to check that switch 101 is disconnected before operating the isolating switch may result in opening the switch under load.

[0098] The correctness check is mainly used to verify whether the sequence of operation instructions or steps is reasonable. For example, opening the grounding switch first and then closing the switches on both sides of the switch is a reasonable operation sequence.

[0099] In this embodiment, users can add operation ticket review rules that have not yet been added to the rule base (referred to as new operation ticket review rules). If the review result of the new operation ticket review rule meets the user's expectations, the new operation ticket review rule can be added to the rule base for routine review in the future. For ease of distinction, operation ticket review rules that have already been added to the rule base are referred to as existing operation ticket review rules.

[0100] In practical applications, operation tickets requiring review are referred to as pending operation tickets. When a user interacts with the self-learning intelligent agent, the user can input current dialogue information in natural language on the dialogue interface. The intent category of the current dialogue information varies depending on the user's needs. For example, the intent category of the current dialogue information may be "adding operation ticket review rules," "supplementary content," or "irrelevant intent." Specifically, when the intent category is "supplementary content," it mainly reflects the user's need for supplementary explanations of the added operation ticket review rules. When the intent category is "intent unrelated to operation ticket review," it mainly reflects that the user's input of the current dialogue information is unrelated to operation ticket review; the current dialogue information may be, for example, casual conversation, general questions and answers, etc.

[0101] In some optional embodiments, the current dialogue information may include an intent category marker, which can be a keyword identifying the intent category to which the current dialogue information belongs. For example, the intent category marker could be "add rule" or "supplementary content". By carrying the intent category marker in the current dialogue information, the self-learning agent can efficiently and accurately identify the intent category to which the current dialogue information belongs, thereby improving the efficiency and quality of operation ticket review.

[0102] For example, the current dialogue message is "Add rule: When a line with a disconnect switch is de-energized, short-lead protection should be activated before the switch is switched to parallel operation." The "Add rule" in the current dialogue message indicates that the intent category of the current dialogue message is to add an operation ticket review rule.

[0103] For example, a user might input the following natural language information into the interactive interface: "Supplementary content: This rule only applies to power outages or restorations of lines with disconnect switches, and the operation content is 'the line is switched from cold standby to operation and series supply.' Operation instructions that do not meet the above conditions do not require review." The "supplementary content" in this current dialogue message indicates that the intent of the current dialogue message is to supplement the rules for reviewing newly added operation tickets.

[0104] For example, a user enters natural language information on the dialogue interface as: "Irrelevant content: How's the weather today?" The intent category of the "irrelevant content" in the current dialogue information is "intent unrelated to operation ticket review".

[0105] In this embodiment, the self-learning agent acquires the current dialogue information input by the user in this dialogue interaction and performs intent recognition on the current dialogue information.

[0106] Specifically, generative AI models leverage their powerful semantic understanding and reasoning capabilities to deeply understand the current conversational information input by the user. They can also perform semantic reasoning based on the dialogue context to accurately identify the intent category of the current conversational information. The dialogue context can be understood as the recorded dialogue content or other information from previous conversations; it can be considered short-term memory information that assists the generative AI model in its decision-making. Furthermore, the generative AI model identifies the intent of the current conversational information based on the intent category labels within the current dialogue information.

[0107] Optionally, the knowledge graph includes at least one of the following: a topological structure graph reflecting the relationships between operational objects, an operational instruction graph, and a rule graph for preventing misoperation; the scheduling professional knowledge base includes at least one of the following: work procedure specifications, expert experience knowledge, and a rule base.

[0108] In practical applications, if the generative AI model fails to identify the intent category of the current dialogue, it outputs guidance information. This guidance guides the user to input dialogue information related to the operation ticket review process. This natural and user-friendly approach guides the user to input content relevant to the operation ticket review task, ensuring the effectiveness of the interaction and the continuity of the process, preventing the user from straying from the core scenario of operation ticket review for an extended period. For example, if the user's current dialogue input is "Add a rule," the generative AI model might output guidance information such as "Please describe the rule you want to add, for example, differential protection should be deactivated before the main transformer is de-energized." Or, if the user's current dialogue input is "Check if this is correct," the generative AI model might output guidance information such as "Which rule or operation ticket are you referring to? Can you provide more details?"

[0109] In practical applications, if the generative AI model identifies the intent category of the current dialogue as unrelated to the operation ticket review, it outputs a fallback message to guide the user back to the operation ticket review scenario. For example, if the user's current conversation input is "How's the weather today?", the fallback message would be "I am a self-learning assistant. My current task is to add new operation ticket review rules and learn from them. I'm sorry I cannot answer your question."

[0110] In practical applications, if the generative AI model identifies the intent category of the current dialogue information as a new operation ticket review rule, the generative AI model combines the knowledge graph and the scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information and obtain the rule parsing content of the new operation ticket review rule.

[0111] Specifically, the rule parsing content of the newly added operation ticket review rules can be understood as the structured rule description obtained after parsing the natural language description of the newly added operation ticket review rules input by the user.

[0112] In some optional embodiments, in order to improve the efficiency and accuracy of operation ticket review, the rule parsing content may include at least one of the following information: operation instruction review type, operation instruction filtering thought chain, and operation instruction review thought chain.

[0113] In some optional embodiments, regarding the instruction filtering thought chain and the operation instruction review thought chain, if the generative AI model has deep thinking capabilities, it can be considered a reasoning model, and thus the generative AI model with deep thinking capabilities can be used to parse the instruction filtering thought chain and the operation instruction review thought chain. If the generative AI model does not have deep thinking capabilities, it can also directly parse the instruction filtering thought chain and the operation instruction review thought chain. Of course, the generative AI model with deep thinking capabilities can parse the instruction filtering thought chain and the operation instruction review thought chain more accurately.

[0114] In practical applications, appropriate prompts can be designed through prompt engineering, and these prompts can be used to guide generative AI models in parsing the rules for reviewing new operation tickets. The prompts can guide the generative AI model to analyze various rule parsing aspects of the new operation ticket review rules, such as the operation instruction review type, the operation instruction filtering thought chain, and the operation instruction review thought chain.

[0115] To ensure the effectiveness of the prompt information, it is necessary to supplement it with knowledge information from the knowledge graph and the scheduling professional knowledge base. In practical applications, the generative AI model analyzes the rules for reviewing new operation tickets to determine preliminary rule analysis results, such as the operation instructions involved in the rules and the relationships between them. Based on the preliminary rule analysis results, relevant knowledge information, including equipment topology information, is queried through the knowledge graph interface. Based on the preliminary rule analysis results, retrieval augmented generation (RAG) technology is used to recall relevant knowledge information from the scheduling professional knowledge base, and the recalled knowledge information is summarized to obtain the recall summary knowledge information. The prompt information is supplemented with background knowledge, including the knowledge information retrieved from the knowledge graph and the recall summary knowledge information from the scheduling professional knowledge base. In this way, guided by this background knowledge, the generative AI model can more accurately perform structured parsing of the rules for reviewing new operation tickets.

[0116] For example, the new operation ticket review rule is "Electrical verification is required before closing the circuit breaker." The background knowledge obtained from the knowledge graph is as follows: 1. Anti-misoperation rule: "Prevent pulling the knife switch while the line is energized," the trigger condition is that the line is energized, and the control measure is that electrical verification must be performed first; 2. Operation verbs: "Verify electrical verification," "Close the circuit breaker"; 3. Equipment information: "Circuit breaker status," "Bus voltage level." The summary information of Article X of the "Power Dispatch Operation Regulations" obtained from the dispatching professional knowledge base is: "It is strictly forbidden to close the circuit breaker without verification of electrical voltage." The generative AI model combines this background knowledge to analyze the rule.

[0117] In this embodiment, the operation instruction review type is used to indicate whether a single operation instruction is reviewed independently or multiple operation instructions are reviewed jointly.

[0118] Specifically, if the operation instruction review type indicates that each operation instruction should be reviewed independently, it means that the review object of the new operation ticket review rule is singular, focusing only on the individual operation instruction itself. Therefore, each operation instruction is reviewed separately using the new operation ticket review rule. For example, if there are five operation instructions that need review, each instruction is reviewed separately using the new operation ticket review rule. Taking the new operation ticket review rule as "The drafted operation instruction ticket should have a clear task, a clear ticket design, and correctly use dual equipment naming and scheduling terminology" as an example, only the correct use of dual equipment naming and scheduling terminology in a single operation instruction needs to be checked.

[0119] If the operation instruction review type indicates that multiple operation instructions will be jointly reviewed, it means that the review objects of the newly added operation ticket review rules are diversified, focusing on multiple operation instructions and the operation sequence, dependencies, or logical relationships between them. Therefore, each time, the newly added operation ticket review rules are used to review the combined behavior of two or more operation instructions in the operation ticket. For example, if the newly added operation ticket review rule is "First open isolating switch A, then close isolating switch B, reverse operation is prohibited," then multiple operation instructions need to be reviewed simultaneously.

[0120] In this embodiment, the operation instruction filtering mind chain is used to indicate the filtering method for selecting at least one target operation instruction to be reviewed from at least one operation instruction included in the operation ticket to be reviewed. Specifically, the operation instruction filtering mind chain is a mind chain used to filter operation instructions. For example, if the review rule for a new operation ticket is "Electrical verification must be performed before closing the circuit breaker", the operation instruction filtering mind chain is "Identify and extract all operation instructions involving 'closing the circuit breaker' from the operation ticket for subsequent review and judgment on 'whether to verify electrical status'". As another example, if the review rule for a new operation ticket is "Open disconnector A first, then close disconnector B, and reverse operation is prohibited", the operation instruction filtering mind chain is "Filter all 'open' operations involving 'disconnector A' and 'close' operations involving 'disconnector B' from the operation ticket". This process ensures that only operation instructions related to the review rule for the new operation ticket are selected so that subsequent joint review of sequence and compliance can be performed.

[0121] In this embodiment, the operation instruction review mind chain is used to indicate the review process for new operation ticket review rules. Specifically, the operation instruction review mind chain is a mind chain used to indicate the review process for new operation ticket review rules. For example, if the new operation ticket review rule is "voltage must be checked before closing the circuit breaker," the operation instruction review mind chain is: "If the current operation instruction is a circuit breaker closing operation, Step 1: Check if there is a corresponding voltage check operation before the circuit breaker closing operation; Step 2: Determine whether the voltage check operation occurred before the closing operation; Step 3: Verify whether the interval between the voltage check time and the closing time is within the allowable range (e.g., within 5 minutes); Step 4: If any of the above conditions are not met, the operation instruction is determined to be non-compliant (i.e., the review result is not passed). If all the above conditions are met, the operation instruction is determined to be compliant (i.e., the review result is passed)." For example, the new operation ticket review rule is "Open isolating switch A first, then close isolating switch B; reverse operation is prohibited." The operation instruction review thought process is as follows: "Step 1: Confirm whether the operation ticket contains an operation instruction to open isolating switch A; Step 2: Confirm whether the operation ticket contains an operation instruction to close isolating switch B; Step 3: Verify the operation sequence to ensure that the opening operation of isolating switch A occurs before the closing operation of isolating switch B; Step 4: If any of the above conditions are not met, multiple operation instructions are deemed non-compliant (i.e., the review result is not passed). If all the above conditions are met, multiple operation instructions are deemed compliant (i.e., the review result is passed)."

[0122] In this embodiment, after the self-learning agent uses a generative AI model to parse the rules for reviewing new operation tickets, it provides the rule parsing content of the new operation ticket review rules to the inference model. The inference model, combined with the knowledge graph and scheduling professional knowledge base, reviews the operation instructions in the operation ticket to be reviewed according to the rule parsing content and outputs the review result of the operation ticket to be reviewed.

[0123] In some scenarios, the self-learning agent can provide the rule parsing content of the newly added operation ticket review rules to the operation ticket review agent. The operation ticket review agent uses a generative AI model to review the operation instructions in the operation ticket to be reviewed according to the rule parsing content, and outputs the review result of the operation ticket to be reviewed.

[0124] In practical applications, the review result of a pending operation ticket can include: approved or rejected. Approval means the pending operation ticket has passed the review of the new operation ticket review rules and is a qualified operation ticket; rejection means the pending operation ticket has failed the review of the new operation ticket review rules and is an unqualified operation ticket. The review result may also include an analysis of the reasons for approval or rejection, detailed information about the review process, etc.

[0125] In practical applications, the review results of a pending operation ticket can include the review results of each target operation instruction. Target operation instructions are those selected from the pending operation tickets that require review. If the review result of a target operation instruction is "approved," it means that the target operation instruction has passed the review of the newly added operation ticket review rules and is a qualified operation instruction. If the review result of a target operation instruction is "failed," it means that the target operation instruction has not passed the review of the newly added operation ticket review rules and is an unqualified operation instruction.

[0126] In practical applications, the review result of the pending operation ticket can be obtained by summarizing the review results of each target operation instruction. If all target operation instructions pass the review, the review result of the pending operation ticket is "passed"; if any target operation instruction fails the review, the review result of the pending operation ticket is "failed".

[0127] In some optional embodiments, in order to further improve the security and standardization of the operation ticket, the inference model reviews the operation instructions in the operation ticket to be reviewed from multiple dimensions according to the rule parsing content; the multiple dimensions include: completeness check, standardization check or correctness check.

[0128] In some optional embodiments, to further improve the security and standardization of operation tickets, the implementation method of using a generative AI model to parse the content according to rules and review the operation instructions in the operation ticket to be reviewed is as follows: The generative AI model uses an operation instruction filtering thought chain to filter at least one target operation instruction to be reviewed from at least one operation instruction included in the operation ticket to be reviewed; if the operation instruction review type indicates that a single operation instruction is reviewed independently, then the single target operation instruction is reviewed independently according to the operation instruction review thought chain to obtain the review result of the single target operation instruction; and the review result of the operation ticket to be reviewed is generated based on the review result of at least one target operation instruction; if the operation instruction review type indicates that multiple operation instructions are reviewed jointly, then the multiple target operation instructions are reviewed jointly according to the operation instruction review thought chain to obtain the review result of multiple target operation instructions; and the review result of the operation ticket to be reviewed is generated based on the review result of multiple target operation instructions.

[0129] The technical solution provided in this application, through dialogue interaction with the user, utilizes a generative AI model to identify the intent of the current dialogue information input by the user during the dialogue interaction. If the intent category of the current dialogue information is identified as a new operation ticket review rule, then, combining a knowledge graph and a scheduling professional knowledge base, the new operation ticket review rule in the current dialogue information is parsed to obtain the rule parsing content of the new operation ticket review rule. The generative AI model then reviews the operation instructions in the operation ticket to be reviewed according to the rule parsing content and outputs the review result of the operation ticket to be reviewed. Therefore, an efficient, intelligent, and self-learning operation ticket review solution is provided, which can effectively ensure the security, standardization, and correctness of operation tickets.

[0130] Figure 3 A flowchart illustrating another operation ticket review method provided in this application embodiment. See also... Figure 3 The method may include the following steps:

[0131] 301. Obtain the current dialogue information entered by the user in this dialogue interaction.

[0132] 302. Use a generative AI model to identify the intent of the current dialogue information. If the intent category of the current dialogue information is identified as a new operation ticket review rule, then use the generative AI model in combination with the knowledge graph and scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information and obtain the rule parsing content of the new operation ticket review rule.

[0133] 303. Use a generative AI model to parse the content of the operation instructions in the operation ticket to be reviewed according to the rules, and output the review result of the operation ticket to be reviewed. Then execute step 304 or step 307.

[0134] Steps 301-302 can be executed by a self-learning agent, and step 303 can be executed by an operation ticket review agent.

[0135] In practical applications, after the intelligent agent for reviewing operation tickets outputs the review results, users can confirm whether the results meet their expectations. If the user confirms that the review results meet expectations, it can be considered that the review results are correct and reliable. Conversely, if the user confirms that the review results do not meet expectations, it can be considered that the review results are incorrect and unreliable. Based on user feedback, the rule parsing content for newly added operation ticket review rules can be continuously optimized, thereby improving the accuracy and adaptability of subsequent reviews.

[0136] 304. In response to a user's confirmation of an error in the review result of a pending operation ticket, obtain the next dialogue information entered by the user in the next dialogue interaction.

[0137] 305. Using a generative AI model, perform intent recognition on the next dialogue information. If the intent category of the next dialogue information is identified as updating the rule parsing content of the new operation ticket review rule, then update the rule parsing content of the new operation ticket review rule according to the next dialogue information.

[0138] Optionally, a generative AI model can be used in conjunction with a knowledge graph and a scheduling professional knowledge base to update the rule parsing content of the newly added operation ticket review rules based on the information from the next dialogue.

[0139] 306. Use a generative AI model to parse the content of the operation instructions in the operation ticket to be reviewed according to the updated rules, review the operation instructions in the operation ticket to be reviewed, and output the review result of the operation ticket to be reviewed.

[0140] Steps 304-305 can be executed by the self-learning agent, and step 306 can be executed by the operation ticket review agent.

[0141] In this embodiment, if the user confirms that the review result of the operation ticket to be reviewed is incorrect, the user can continue to interact with the self-learning agent. The self-learning agent obtains the next dialogue information entered by the user in the next dialogue interaction. The next dialogue interaction is also the dialogue interaction following the current dialogue interaction, and the next dialogue information is the dialogue information entered by the user in the next dialogue interaction.

[0142] In practical applications, users input information for the next dialogue as needed. Different information for the next dialogue will update the rule parsing content of the newly added operation ticket review rules differently, thereby achieving continuous optimization and personalized adaptation of the rule parsing content.

[0143] In some optional embodiments, the method for updating the rule parsing content of the newly added operation ticket review rule based on the next dialogue information is as follows: if the intent category of the next dialogue information is supplementary example information, then example information is added to the rule parsing content of the newly added operation ticket review rule. The example information includes positive example information and / or negative example information. Positive example information includes operation instruction examples and their review results that have passed the operation ticket review rule example review; negative example information includes operation instruction examples and their review results that have not passed the operation ticket review rule example review.

[0144] Specifically, the operation instruction example can be understood as a model operation instruction, the operation ticket review rule example can be understood as a model operation ticket review rule, and the operation instruction example is some operation instructions in an operation ticket that has already been reviewed using the operation ticket review rule example. By introducing positive and negative example information, generative AI models and inference models can better understand the newly added operation ticket review rules, thereby improving the accuracy and efficiency of operation ticket review.

[0145] In some optional embodiments, the method for updating the rule parsing content of the new operation ticket review rule based on the next dialogue information is as follows: if the intent category of the next dialogue information is to supplement background knowledge, then background knowledge is added to the rule parsing content of the new operation ticket review rule. Background knowledge refers to the knowledge information required to assist in reviewing the operation instructions of the operation ticket to be reviewed using the new operation ticket review rule.

[0146] Specifically, supplementing background knowledge can help generative AI models and inference models better understand the new operation ticket review rules, thereby improving the accuracy and efficiency of operation ticket review.

[0147] In some optional embodiments, the method for updating the rule parsing content of the new operation ticket review rule based on the next dialogue information is as follows: if the intent category of the next dialogue information is to modify the operation instruction filtering thought chain, modify the operation instruction filtering thought chain in the rule parsing content of the new operation ticket review rule based on the next dialogue information.

[0148] Specifically, it supports users to dynamically update the thought chain for filtering operation instructions, thereby more accurately identifying operation instructions that should be reviewed by the new operation ticket review rules, and realizing personalized adaptation and continuous optimization of the rule parsing content.

[0149] In some optional embodiments, the method for updating the rule parsing content of the new operation ticket review rule based on the next dialogue information is as follows: if the intent category of the next dialogue information is to modify the operation instruction review thought chain, modify the operation instruction review thought chain in the rule parsing content of the new operation ticket review rule based on the next dialogue information.

[0150] Specifically, it supports users to dynamically update the operational instruction review mindset, thereby more accurately reviewing operational tickets and achieving personalized adaptation and continuous optimization of rule parsing content.

[0151] In practical applications, after each round of dialogue interaction, the operation ticket review agent outputs the review result of the operation ticket to be reviewed. The user can then confirm whether the review result of the operation ticket to be reviewed meets expectations. If it does not meet expectations, the process returns to steps 304 to 306 until it is confirmed that the review result of the operation ticket to be reviewed meets expectations.

[0152] 307. In response to the user's confirmation that the review result of the pending operation ticket is correct, the newly added operation ticket review rules and the correct rule parsing content are associated and stored in the rule base.

[0153] Step 307 can be performed by a self-learning agent.

[0154] In practical applications, after each round of dialogue interaction, the operation ticket review agent outputs the review result for the operation ticket to be reviewed. The user can then confirm whether the review result meets expectations. If it does, the newly added operation ticket review rules and their corresponding parsed content can be stored in the rule base. This process of embedding the newly added operation ticket review rules and their corresponding parsed content into long-term memory allows for continuous optimization of the rule base. During daily management, the operation ticket review rules in the rule base can be used for automated review of operation tickets. Parsed content that meets expectations can be considered correct rule parsing content confirmed by the user; parsing content that does not meet expectations can be considered incorrect rule parsing content confirmed by the user.

[0155] In practical applications, the rule base can include at least one existing operation ticket review rule and the rule parsing content for each existing operation ticket review rule. Optionally, the rule base can also include reviewed operation tickets for each existing operation ticket review rule. Reviewed operation tickets refer to operation tickets that have completed review using existing operation ticket review rules. Thus, when a new operation ticket review rule is added and stored in long-term memory, the new operation ticket review rule, its expected rule parsing content, and the operation tickets to be reviewed can be stored in the rule base.

[0156] It is worth noting that the reviewed operation tickets are actual cases that have been reviewed based on the corresponding existing operation ticket review rules. They can be used to assist in the understanding of rules and the continuous optimization of rules, thereby improving the intelligence level of the self-learning agent.

[0157] The technical solution provided in this application involves an intelligent agent engaging in dialogue with the user. Utilizing generative AI models, inference models, knowledge graphs, and a scheduling-specific knowledge base, it offers an efficient, intelligent, and self-learning-capable operation ticket review solution, effectively ensuring the security, standardization, and correctness of operation tickets. Through one or more rounds of interactive feedback with the user, a closed-loop mechanism of "knowledge learning - inference verification - feedback iteration" is formed, achieving enhanced human-machine collaboration and dynamic rule updates, continuously improving the intelligence level of operation ticket review.

[0158] To facilitate understanding, the following will be combined with... Figure 4 This section introduces a specific scenario example. See [link / reference] Figure 4 First, the self-learning agent acquires the current dialogue information input by the user in the dialogue interaction interface and performs intent recognition on the current dialogue information.

[0159] If the current dialogue information does not explicitly express the user's intent, the intent recognition result is a follow-up intent, which can also be understood as the intent category of the current dialogue information not being recognized. At this time, the self-learning agent outputs guidance information on the dialogue interaction interface. The guidance information is used to guide the user to input dialogue information related to the operation ticket review, that is, the guidance information is used to guide the user back to the main task (i.e., the operation ticket review task).

[0160] If the intent recognition result is irrelevant, the self-learning agent outputs a fallback statement on the dialogue interface. If the intent recognition result is the main task, the self-learning agent can determine whether the intent of the main task is to add a new rule or supplement content.

[0161] If the main task's intent is to add new rules, the self-learning agent parses the rules for reviewing new operation tickets to obtain the rule parsing content. If the main task's intent is to supplement content, the self-learning agent updates the rule parsing content. The operation ticket review agent uses the final rule parsing content to review the operation ticket and outputs the review result on the dialogue interface.

[0162] If a user confirms that the review result of the operation ticket does not meet expectations, the user can continue to input supplementary content for the review rules of the newly added operation ticket. The self-learning agent uses the supplementary content to update the rule parsing content, and continues to review the operation ticket using the updated rule parsing content, and outputs the review result of the operation ticket on the dialogue interaction interface for the user to confirm whether it meets expectations.

[0163] If a user confirms that the review result of the operation ticket meets expectations, they can perform the rule entry operation, which means saving the newly added operation ticket review rule and its parsing content to the rule library so that the operation ticket can be reviewed using the operation ticket review rule in the subsequent daily management stage.

[0164] In this embodiment, efficient, intelligent, and self-learning operation ticket review can be achieved, which has the following advantages:

[0165] 1. Utilize the context learning and thought chain reasoning capabilities of generative AI models to achieve semantic understanding and rule parsing of operation ticket review rules in natural language form, complete the self-learning of operation ticket review rules, and support dynamic updates of the rule parsing content of operation ticket review rules.

[0166] 2. The self-learning intelligent agent forms a closed-loop mechanism of "knowledge learning - reasoning verification - feedback iteration" through one or more rounds of interactive feedback with users. This enables dynamic rule updates that enhance human-machine collaboration and continuously improve the intelligence level of operation ticket review. The relevant content of newly added operation ticket review rules can be stored in long-term memory for continuous optimization of the rule base.

[0167] 3. The knowledge graph and scheduling professional knowledge base dynamically verify the rationality of operation instructions to ensure the real-time and accurate nature of the audit results.

[0168] Figure 5 This is a system architecture diagram of an operation ticket processing system provided in an embodiment of this application. See also... Figure 5 The operation ticket processing system may include: terminal device 10 and server 20, the server 20 running a self-learning intelligent agent and an operation ticket review intelligent agent;

[0169] Terminal device 10 is used to respond to the user's input operation on the dialogue interaction interface, obtain the current dialogue information entered by the user in this dialogue interaction and send it to the server, and display the review result of the pending operation ticket returned by the server on the dialogue interaction interface.

[0170] Server 20 is used to identify the intent of the current dialogue information through a generative AI model using a self-learning intelligent agent. If the intent category of the current dialogue information is identified as a new operation ticket review rule, the generative AI model is used in combination with a knowledge graph and a scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule. The operation ticket review intelligent agent uses the generative AI model to review the operation instructions in the operation ticket to be reviewed according to the rule parsing content, and sends the review result of the operation ticket to be reviewed to the terminal device.

[0171] The detailed implementation methods and beneficial effects of each step in this embodiment have been described in detail in the foregoing embodiments, and will not be elaborated here.

[0172] It should be noted that the execution subject of each step of the method provided in the above embodiments can be the same device, or the method can be executed by different devices. For example, the execution subject of steps 201 to 203 can be device A; or the execution subject of steps 201 and 202 can be device A, and the execution subject of step 203 can be device B; and so on.

[0173] Furthermore, in some of the processes described in the above embodiments and accompanying drawings, multiple operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or they may be executed in parallel. The operation numbers, such as 201, 202, etc., are merely used to distinguish different operations and do not represent any execution order. Additionally, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first" and "second" in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0174] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 6 As shown, the electronic device includes: a memory 61 and a processor 62;

[0175] Memory 61 is used to store computer programs and can be configured to store various other data to support operation on the computing platform. Examples of this data include instructions for any application or method operating on the computing platform, data structures, contact data, phone book data, messages, pictures, videos, etc.

[0176] Processor 62, coupled to memory 61, is used to execute computer programs in memory 61 for: performing steps in the operation ticket review rule self-learning method or operation ticket review method.

[0177] Optional, such as Figure 6 As shown, the electronic device also includes other components such as a communication component 63, a display 64, a power supply component 65, and an audio component 66. Figure 6 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 6 The components shown. Additionally... Figure 6The components within the dashed box are optional, not mandatory, and their specific requirements depend on the product form of the electronic device. The electronic device in this embodiment can be a desktop computer, laptop computer, smartphone, or IoT (Internet of Things) device, or a server-side device such as a conventional server, cloud server, or server array. If the electronic device in this embodiment is a desktop computer, laptop computer, or smartphone, it may include... Figure 6 The components within the dashed box; if the electronic device in this embodiment is implemented as a conventional server, cloud server, or server array, etc., it may be omitted. Figure 6 The component within the dashed box.

[0178] The aforementioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0179] The aforementioned communication component is configured to facilitate wired or wireless communication between the device containing the communication component and other devices. The device containing the communication component can access wireless networks based on communication standards, such as 2G (2nd Generation), 3G (3rd Generation), 4G (4th Generation) / LTE (long Term Evolution), 5G (5th Generation), or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel.

[0180] The aforementioned display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a Touch Panel, the screen can be implemented as a touchscreen to receive input signals from the user. The Touch Panel includes one or more touch sensors to sense touches, swipes, and gestures on the Touch Panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation.

[0181] The aforementioned power supply components provide power to various components within the device in which they reside. These power supply components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device in which they reside.

[0182] The aforementioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.

[0183] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in the above-described method embodiments. The computer-readable storage medium may be volatile, non-volatile, or a combination thereof, and may be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), flash memory or other memory technologies, CD-ROM, digital video disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium.

[0184] Accordingly, this application also provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor is able to implement the steps in the above method embodiments. It should be understood that each step or combination of steps in the above method flow can be implemented by the computer program or instructions. In addition, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, so that the processor of the general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device can be implemented as a means to implement the corresponding functions in the above method embodiments.

[0185] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0186] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A self-learning method for operation ticket review rules, characterized in that, Applied to self-learning intelligent agents, the method includes: Obtain the current dialogue information entered by the user in this conversation interaction; Generative AI models are used to identify the intent of the current dialogue information to obtain the intent category of the current dialogue information; If the intent category of the current dialogue information is a new operation ticket review rule, then the generative AI model combined with the knowledge graph and scheduling professional knowledge base is used to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule; The rule parsing content includes: operation instruction review type, operation instruction screening mind chain, and operation instruction review mind chain; the operation instruction review type is used to indicate whether a single operation instruction is reviewed independently or multiple operation instructions are reviewed jointly; the operation instruction screening mind chain is used to indicate the screening method for selecting at least one target operation instruction to be reviewed from at least one operation instruction included in the operation ticket to be reviewed; the operation instruction review mind chain is used to indicate the review process of the newly added operation ticket review rule.

2. The method according to claim 1, characterized in that, Also includes: In response to the user confirming that the review result of the operation ticket to be reviewed is incorrect, the system obtains the next dialogue information entered by the user in the next dialogue interaction; wherein, the review result of the operation ticket to be reviewed is obtained by reviewing the operation instructions in the operation ticket to be reviewed according to the content parsed by the rules; The generative AI model is used to identify the intent of the next dialogue information; If the intent category of the next dialogue information is identified as updating the rule parsing content of the newly added operation ticket review rule, then the generative AI model is used to update the rule parsing content of the newly added operation ticket review rule based on the next dialogue information.

3. The method according to claim 2, characterized in that, The rule parsing content of the newly added operation ticket review rule is updated based on the information from the next dialogue, including: If the intent category of the next dialogue information is supplementary example information, then the example information is added to the rule parsing content of the newly added operation ticket review rule. The example information includes positive example information and / or negative example information. The positive example information includes an example of an operation instruction and its review result after passing the operation ticket review rule example review. The negative example information includes an example of an operation instruction and its review result after failing the operation ticket review rule example review. If the intent category of the next dialogue information is to supplement background knowledge, then the background knowledge is added to the rule parsing content of the newly added operation ticket review rule. The background knowledge refers to the knowledge information required to assist in reviewing the operation instructions of the operation ticket to be reviewed using the newly added operation ticket review rule. If the intent category of the next dialogue information is to modify the operation instruction filtering thought chain, then modify the operation instruction filtering thought chain in the rule parsing content of the new operation ticket review rule according to the next dialogue information. If the intent category of the next dialogue information is to modify the operation instruction review thought chain, then modify the operation instruction review thought chain in the rule parsing content of the new operation ticket review rule according to the next dialogue information.

4. The method according to any one of claims 1 to 3, characterized in that, Also includes: If the intent category of the current dialogue information is not identified, guidance information is output to guide the user to input dialogue information related to the operation ticket review. If the intent category of the current dialogue information is identified as an intent category unrelated to the operation ticket review, then a fallback statement is output.

5. A method for verifying operation tickets, characterized in that, include: Obtain the current dialogue information entered by the user in this conversation interaction; Generative AI models are used to identify the intent of the current dialogue information; If the intent category of the current dialogue information is identified as a new operation ticket review rule, then the generative AI model combined with the knowledge graph and scheduling professional knowledge base is used to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule; The generative AI model is used to parse the content of the operation instructions in the operation ticket to be reviewed according to the rules, and the review result of the operation ticket to be reviewed is output. The rule parsing content includes: operation instruction review type, operation instruction filtering mind chain, and operation instruction review mind chain; the operation instruction review type is used to indicate whether a single operation instruction is reviewed independently or multiple operation instructions are reviewed jointly; the operation instruction filtering mind chain is used to indicate the filtering method for selecting at least one target operation instruction to be reviewed from at least one operation instruction included in the operation ticket to be reviewed; the operation instruction review mind chain is used to indicate the review process of the new operation ticket review rule.

6. The method according to claim 5, characterized in that, The generative AI model is used to parse the content according to the rules and review the operation instructions in the operation ticket to be reviewed, including: Using the generative AI model, at least one target operation instruction to be reviewed is selected from at least one operation instruction included in the operation ticket to be reviewed, according to the operation instruction filtering thought chain. If the operation instruction review type indicates that a single operation instruction should be reviewed independently, then the single target operation instruction should be reviewed independently according to the operation instruction review thought chain to obtain the review result of the single target operation instruction; and the review result of the operation ticket to be reviewed should be generated based on the review result of at least one target operation instruction. If the operation instruction review type indicates that multiple operation instructions should be jointly reviewed, then the multiple target operation instructions are jointly reviewed according to the operation instruction review thought chain to obtain the review results of the multiple target operation instructions; and the review result of the operation ticket to be reviewed is generated based on the review results of the multiple target operation instructions.

7. The method according to claim 5, characterized in that, After outputting the review result of the operation ticket to be reviewed, the following steps are also included: In response to the user's confirmation that the review result of the operation ticket to be reviewed is incorrect, the system obtains the next dialogue information entered by the user in the next dialogue interaction; The generative AI model is used to identify the intent of the next dialogue information. If the intent category of the next dialogue information is identified as updating the rule parsing content of the new operation ticket review rule, then the rule parsing content of the new operation ticket review rule is updated according to the next dialogue information. The generative AI model is used to parse the content according to the updated rules, review the operation instructions in the operation ticket to be reviewed, and output the review result of the operation ticket to be reviewed.

8. The method according to claim 7, characterized in that, The rule parsing content of the newly added operation ticket review rule is updated based on the information from the next dialogue, including: If the intent category of the next dialogue information is supplementary example information, then the example information is added to the rule parsing content of the newly added operation ticket review rule. The example information includes positive example information and / or negative example information. The positive example information includes an example of an operation instruction and its review result after passing the operation ticket review rule example review. The negative example information includes an example of an operation instruction and its review result after failing the operation ticket review rule example review. If the intent category of the next dialogue information is to supplement background knowledge, then the background knowledge is added to the rule parsing content of the newly added operation ticket review rule. The background knowledge refers to the knowledge information required to assist in reviewing the operation instructions of the operation ticket to be reviewed using the newly added operation ticket review rule. If the intent category of the next dialogue information is to modify the operation instruction filtering thought chain, modify the operation instruction filtering thought chain in the rule parsing content of the new operation ticket review rule according to the next dialogue information; If the intent category of the next dialogue information is to modify the operation instruction review thought chain, modify the operation instruction review thought chain in the rule parsing content of the new operation ticket review rule according to the next dialogue information.

9. The method according to claim 5, characterized in that, Also includes: In response to the user's confirmation that the review result of the operation ticket to be reviewed is correct, the review rules for the newly added operation ticket and the correct rule parsing content are associated with the operation ticket to be reviewed and stored in the rule base.

10. The method according to any one of claims 5 to 9, characterized in that, Also includes: If the intent category of the current dialogue information is not identified, guidance information is output to guide the user to input dialogue information related to the operation ticket review. If the intent category of the current dialogue information is identified as an intent category unrelated to the operation ticket review, then a fallback statement is output.

11. The method according to any one of claims 5 to 9, characterized in that, The review of the operation instructions in the operation ticket to be reviewed according to the parsed content of the aforementioned rules includes: The operation instructions in the operation ticket to be reviewed are examined from multiple dimensions according to the parsing rules described above; the multiple dimensions include: completeness check, standardization check, or correctness check.

12. A ticket processing system, characterized in that, include: Terminal equipment and server, wherein the server runs a self-learning intelligent agent and an operation ticket review intelligent agent; The terminal device is used to respond to the user's input operation on the dialogue interaction interface, obtain the current dialogue information entered by the user in this dialogue interaction and send it to the server, and display the review result of the pending operation ticket returned by the server on the dialogue interaction interface. The server is used to perform intent recognition on the current dialogue information through the self-learning intelligent agent using a generative AI model. If the intent category of the current dialogue information is identified as a new operation ticket review rule, the server uses the generative AI model in combination with a knowledge graph and a scheduling professional knowledge base to parse the new operation ticket review rule in the current dialogue information to obtain the rule parsing content of the new operation ticket review rule. The operation ticket review agent uses the generative AI model to parse the content according to the rules, reviews the operation instructions in the operation ticket to be reviewed, and sends the review result of the operation ticket to the terminal device. The rule parsing content includes: operation instruction review type, operation instruction screening mind chain, and operation instruction review mind chain; the operation instruction review type is used to indicate whether a single operation instruction is reviewed independently or multiple operation instructions are reviewed jointly; the operation instruction screening mind chain is used to indicate the screening method for selecting at least one target operation instruction to be reviewed from at least one operation instruction included in the operation ticket to be reviewed; the operation instruction review mind chain is used to indicate the review process of the newly added operation ticket review rule.

13. An electronic device, characterized in that, include: Memory and processor; The memory is used to store computer programs; The processor is coupled to the memory for executing the computer program to perform the steps of the method according to any one of claims 1-11.

14. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the steps of the method according to any one of claims 1-11.

15. A computer program product, characterized in that, Includes a computer program / instruction that, when executed by a processor, causes the processor to perform the steps of the method according to any one of claims 1-11.

Images