Mine dispatching question and answer method, device and system

Abstract

The embodiment of the present application provides a mine dispatching question and answer method, device and system, the method comprises the following steps: obtaining the mine dispatching question input by a target user in a front-end page; determining a target prompt text based on the mine dispatching question, configuration information of each MCP server in a plurality of MCP servers and a preset prompt template; inputting the target prompt text into a target language model for processing to obtain a target MCP server used for processing the mine dispatching question and target tool description information required in the target MCP server; calling a target tool to the target MCP server based on the target tool description information, and returning the target tool calling result to the target language model for information integration to obtain target reply content corresponding to the mine dispatching question; and returning the target reply content to the front-end page and displaying the target reply content in the front-end page. The technical scheme improves the docking efficiency of natural language and the dispatching system and improves the accuracy of front-end information display.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a mine scheduling question-and-answer method, apparatus and system. Background Technology

[0002] A mine dispatching system is an essential tool for scheduling mine production, determining the efficiency and safety of mine operations. Therefore, improving the efficiency and accuracy of the mine dispatching system is of paramount importance.

[0003] Existing mine dispatching systems are mostly information display platforms that only statistically analyze data from a single stage, and dispatching decisions rely on the dispatcher's experience and subjective judgment. Furthermore, the systems depend on telephone, broadcast, or push notifications to issue dispatching instructions.

[0004] Then, when faced with massive amounts of real-time data such as equipment status, environmental monitoring, and personnel location, dispatchers struggle to quickly and comprehensively specify the optimal scheduling plan. Simultaneously, data silos between systems, a lack of deep cross-domain data integration and correlation analysis, and inefficient generation, issuance, and execution feedback of scheduling instructions make it difficult to ensure that instructions reach the target personnel and equipment. Summary of the Invention

[0005] This invention provides a mine scheduling question-and-answer method, device, and system to achieve intelligent question-and-answer in mine scheduling, thereby improving the efficiency and accuracy of mine scheduling.

[0006] In a first aspect, embodiments of the present invention provide a mine scheduling question-and-answer method, including:

[0007] Obtain the mine scheduling question entered by the target user on the front-end page;

[0008] Based on the mine scheduling problem, the configuration information and preset prompt templates of each MCP server in multiple MCP servers, the target prompt text is determined. Among them, the MCP server is a server that follows the model context protocol, and the MCP server corresponds one-to-one with the scheduling subsystem or scheduling submodule in the mine scheduling system.

[0009] The target prompt text is input into the target language model for processing to obtain the target MCP server for handling mine scheduling problems and the target tool description information required by the target MCP server.

[0010] Based on the target tool description information, the target tool is invoked to the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration in order to obtain the target response content corresponding to the mine scheduling problem.

[0011] Return the target response content to the front-end page so that the target response content can be displayed on the front-end page.

[0012] Secondly, embodiments of the present invention also provide a mine dispatching question-and-answer device, the device comprising:

[0013] The mine scheduling problem acquisition module is used to acquire mine scheduling problems entered by target users on the front-end page;

[0014] The target prompt text determination module is used to determine the target prompt text based on the mine scheduling problem, the configuration information of each MCP server in multiple MCP servers and the preset prompt template. Here, the MCP server is a server that follows the model context protocol, and the MCP server corresponds one-to-one with the scheduling subsystem or scheduling submodule in the mine scheduling system.

[0015] The target tool description information acquisition module is used to input the target prompt text into the target language model for processing, so as to obtain the target MCP server for handling the mine scheduling problem and the target tool description information required by the target MCP server.

[0016] The target response content acquisition module is used to call the target tool to the target MCP server based on the target tool description information, and return the target tool call result to the target language model for information integration in order to obtain the target response content corresponding to the mine scheduling problem.

[0017] The target response content display module is used to return the target response content to the front-end page for display.

[0018] Thirdly, embodiments of the present invention also provide a mine scheduling question-and-answer system, including:

[0019] Front-end page, MCP client and multiple MCP servers;

[0020] The MCP client is used to implement a mine scheduling question-and-answer method as provided in any embodiment of the present invention.

[0021] Fourthly, embodiments of the present invention also provide an electronic device, comprising:

[0022] At least one processor; and

[0023] A memory that is communicatively connected to at least one processor; wherein,

[0024] The memory stores a computer program that can be executed by at least one processor, such that the at least one processor can execute a mine scheduling question-and-answer method as provided in any embodiment of the present invention.

[0025] Fifthly, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions for causing a processor to execute a mine scheduling question-and-answer method as provided in any embodiment of the present invention.

[0026] This invention, through its embodiments, obtains a mine scheduling question input by a target user on a front-end page. Based on the mine scheduling question, configuration information of multiple MCP servers, and preset prompt templates, it determines the target prompt text. Further, the target prompt text is input into a target language model for processing, obtaining the target MCP server used to handle the mine scheduling question and the description information of the target tools required by the target MCP server. Based on the target tool description information, the target tool is invoked from the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration to obtain the target answer content corresponding to the mine scheduling question. Finally, the target answer content is returned to the front-end page for display. By encapsulating the scheduling subsystem and scheduling submodule of the mine scheduling system into an MCP server that conforms to the Model Context Protocol, intelligent docking between the target language model and the mine scheduling system is achieved based on multiple MCP servers. This enables intelligent question-and-answer functionality for mine scheduling using the target language model, improving the efficiency and accuracy of mine scheduling.

[0027] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 A flowchart of a mine scheduling question-and-answer method provided in an embodiment of the present invention;

[0030] Figure 2 This is a schematic diagram of a mine scheduling method provided in an embodiment of the present invention;

[0031] Figure 3 A flowchart of a mine scheduling question-and-answer method provided in an embodiment of the present invention;

[0032] Figure 4 A flowchart of a mine scheduling question-and-answer method provided in an embodiment of the present invention;

[0033] Figure 5 This is an overall framework diagram of a mine scheduling question-and-answer method provided in an embodiment of the present invention;

[0034] Figure 6 This is a schematic diagram of the structure of a mine dispatching question-and-answer device provided in an embodiment of the present invention;

[0035] Figure 7 This is a schematic diagram of the structure of a mine dispatching question-and-answer system provided in an embodiment of the present invention;

[0036] Figure 8 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. Detailed Implementation

[0037] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0038] It should be noted that the terms "target," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0039] Figure 1 This is a flowchart illustrating a mine scheduling question-and-answer method provided in an embodiment of the present invention. This embodiment is applicable to situations involving mine scheduling responses. The method can be executed by a mine scheduling question-and-answer device, which can be implemented in hardware and / or software. This mine scheduling question-and-answer device can be configured in a computing device. Figure 1 As shown, the method includes:

[0040] S110. Obtain the mine scheduling question entered by the target user on the front-end page.

[0041] In this embodiment, the target users are personnel who use the mine scheduling system during the mine production process. These personnel can include dispatchers and production staff. The front-end page is the interface through which the target users interact with the system. The mine scheduling system can display question-and-answer results to the target users through the front-end page. Mine scheduling questions are natural language queries related to mine production scheduling operations. These queries can be keywords or a piece of text, used to query the mine scheduling system for information or request the execution of a certain operation.

[0042] Among them, mine production scheduling involves multiple links such as "mining, tunneling, transportation, ventilation and drainage". That is, mine scheduling problems can be related to the specific scenarios, real-time status, personnel location and resource allocation of each link such as mining, tunneling, transportation, ventilation and drainage.

[0043] Specifically, the target user enters a piece of text or at least one keyword related to mine scheduling business on the front-end page, and the mine scheduling system obtains the mine scheduling question entered by the target user.

[0044] S120. Based on the mine scheduling problem, the configuration information of each MCP server among multiple MCP servers, and the preset prompt template, determine the target prompt text.

[0045] Among them, the MCP server is a server that follows the Model Context Protocol (MCP). The MCP server corresponds one-to-one with the scheduling subsystem or scheduling submodule in the mine scheduling system.

[0046] The MCP server is a dedicated service program that follows specific standards, such as the Model Context Protocol. Each MCP server encapsulates a specific business scope's functions or data and provides relevant data and functions to the model according to standard protocols. Business scopes include equipment status, environmental monitoring, and personnel location. Configuration information is a set of parameters defining the capabilities of the MCP server, containing basic information about each MCP server. Basic MCP server information includes, but is not limited to, server name, function description, server type, and interface information, which is loaded from the configuration file. The preset prompt template is a pre-defined structured text framework used to handle mine scheduling issues. The preset prompt template contains pre-defined text used to transform the questions entered by the target user on the front-end page into prompts that can be understood by the model and the MCP server. The target prompt text is an executable instruction text that can be directly input into the language model to guide the language model in analyzing and outputting preset results. The scheduling subsystems or scheduling submodules of the mine scheduling system are components responsible for handling a specific business scope. The mine scheduling system contains multiple scheduling subsystems or scheduling submodules. The aforementioned scheduling subsystems or modules work together to manage the entire mining production process. Each scheduling subsystem or module is encapsulated as an MCP server.

[0047] In this embodiment, the configuration file is a text file written in YAML format, used to define the basic attributes and capabilities of the MCP server. The configuration file may contain a toolset name, tool description, and link address, where the toolset name is the MCP server name. For example, taking the MCP server encapsulating the "Device Status" service function as an example, its corresponding configuration file contains the following core parameters: toolset name, tool description, and link address. The toolset name is "Device Data Acquisition," the tool description provides interfaces for acquiring data from device services, mainly including queries for devices such as personnel base stations, broadcast systems, fixed-line phones, and cameras. The link address is xxxxxx.

[0048] For example, the default prompt template is: "You are a server routing assistant. Your task is to select the most suitable processing server based on the user's problem. Available server list (each data entry is composed of configuration file content), the combination rule is that the name and description fields in the MCP server are connected by a colon character)." Example: "Device Data Acquisition: Interfaces for acquiring data from device services, mainly including queries for personnel, base stations, broadcasting, landlines, and cameras. Front-end Page Interaction: Responsible for controlling the front-end page, including all necessary front-end page operations. Important Rules: 1. Must and can only return one server name from the above list; 2. Do not return any explanations, punctuation marks, or extra text; 3. If no suitable dedicated server is available, return 'default'."

[0049] Specifically, the process involves obtaining the mine scheduling problem input by the target user. Simultaneously, it integrates the configuration information of all MCP servers and designs a preset prompt template. Based on the mine scheduling problem, the configuration information of each server in the MCP server, and the pre-designed prompt template, the target prompt template is obtained.

[0050] S130. Input the target prompt text into the target language model for processing to obtain the target MCP server for handling mine scheduling problems and the target tool description information required by the target MCP server.

[0051] The target language model is a pre-trained language model used to receive input information, analyze it, and generate structured output. This language model can perform semantic understanding and intent analysis on mine scheduling problems, and identify multi-dimensional information such as business objects, operation types, and urgency levels within the problem. The target MCP server is a specific MCP server determined by the target language model to be suitable for handling mine scheduling problems. There can be one or more target MCP servers. The target tool description information is the tool description in the target MCP server configuration information. Specifically, the target MCP server and the target tool description information required by the target server are the MCP server and tool description information in the multiple MCP server configuration information in the target prompt text.

[0052] Specifically, the determined target prompt text is input into a pre-trained target language model. The target language model analyzes the target prompt text and outputs the target MCP server and its corresponding target tool description information that can be used to handle mine scheduling problems.

[0053] S140. Based on the target tool description information, call the target tool to the target MCP server, and return the target tool call result to the target language model for information integration to obtain the target response content corresponding to the mine scheduling problem.

[0054] The target tool refers to the tool within the target MCP server that can be used to handle mine scheduling problems. The target tool description information includes the tools that can be invoked within that MCP server. The target tool invocation result is the data returned by the target MCP server after executing the target tool. This data result can be structured data, such as JSON format. Information integration can be understood as the target language model analyzing and understanding the target tool invocation result after receiving it. The target response content is the natural language text generated by the target language model. This natural language text represents the final answer to the corresponding mine scheduling problem and can be directly presented to the user.

[0055] Specifically, based on the target tool description information output by the target language model, a call command is sent to the target MCP server to invoke the target tool. Upon receiving the target tool invocation result, it is input into the target language model. The target language model analyzes and processes the result to generate the target response content required by the target user. For example, Figure 2 This is a schematic diagram of a mine scheduling method provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the AI ​​front-end assistant (front-end page) is designed as an MCP service host using the MCP protocol. The core AI service is designed as an MCP client embedded within the MCP host, acting as an intermediary connecting the MCP host and the MCP server. The MCP client is the component used by the MCP host to connect to the MCP server and handle protocol interactions. The capabilities provided by each subsystem and submodule in the mine scheduling system are encapsulated into independent MCP servers. Relying on the LLM large model (target language model), the configuration information in the MCP server is identified, and the target language model outputs tool calls and corresponding tool parameters, which are then invoked by the MCP client and returned to the MCP service host for subsequent actions.

[0056] S150. Return the target response content to the front-end page so that the target response content can be displayed on the front-end page.

[0057] Specifically, the system receives the target response content output by the target language model, and then displays the target response content on the front-end page for the user to use.

[0058] Optionally, the method also includes:

[0059] The target response is parsed to obtain the business scheduling instructions, and the corresponding business scheduling is executed based on the business scheduling instructions.

[0060] Among them, business scheduling instructions include page jump instructions and / or element location instructions.

[0061] In this embodiment, the business scheduling instruction is an instruction parsed by the system from the target response content, which can be directly executed by the business system. The page jump instruction controls the front-end application to jump to other page modules; when the business system front-end receives a page jump instruction, it executes the corresponding page jump. The element positioning instruction instructs the front-end page to locate and highlight a specified interface element. For example, when the business system receives a page jump instruction, it executes the corresponding page jump; when the business system receives an element positioning instruction, it scrolls the view to the specified area and highlights the specified element. Business scheduling refers to the business system running a specified operation after receiving a business scheduling instruction.

[0062] For example, if a user inputs a mine scheduling question as "Query the total number of people going down the mine today, display it on the large screen numbered 001, and locate the large screen's location in the Geographic Information System (GIS)." The system receives the target response content generated by the target language model, parses the target response content, obtains the business scheduling instructions related to the front-end page such as [{"action":"open","page":"GIS location"}{"action":"location","number":"001","type":"LED"}], and executes the corresponding page jump and element location based on the business scheduling instructions.

[0063] Specifically, the system receives the target response content output by the target language model, parses the response content, and generates a business scheduling instruction. The business system receives the business scheduling instruction and executes the scheduling operation specified in the instruction. By parsing the target response content and generating business scheduling instructions, business scheduling is executed automatically, improving interaction efficiency and the accuracy of system operations.

[0064] The technical solution provided by this invention obtains a mine scheduling question input by a target user on a front-end page. Based on the mine scheduling question, configuration information of multiple MCP servers, and preset prompt templates, a target prompt text is determined. Further, the target prompt text is input into a target language model for processing, obtaining the target MCP server used to handle the mine scheduling question and the description information of the target tools required by the target MCP server. Based on the target tool description information, the target tool is invoked from the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration to obtain the target answer content corresponding to the mine scheduling question. Finally, the target answer content is returned to the front-end page for display. By encapsulating the scheduling subsystem and scheduling submodule in the mine scheduling system into an MCP server that conforms to the Model Context Protocol, intelligent docking between the target language model and the mine scheduling system is achieved based on multiple MCP servers. This enables intelligent question-and-answer functionality for mine scheduling using the target language model, improving the efficiency and accuracy of mine scheduling.

[0065] Figure 3 This is a flowchart illustrating a mine scheduling question-and-answer method provided in an embodiment of the present invention. This embodiment further refines the training and fine-tuning process of the target language model based on the above-described optional implementation methods. For example... Figure 3 As shown, the method includes:

[0066] S210. Based on the mine scheduling information database, construct a labeled question-and-answer dataset for mine scheduling.

[0067] The mine scheduling information database is a domain database containing a large amount of information related to mine scheduling, used to store all historical information related to mine scheduling operations. The mine scheduling labeled question-and-answer dataset is a dataset used to train the target language model. This dataset consists of a large number of question-and-answer pairs related to mine scheduling.

[0068] In this embodiment, the construction process of the mine scheduling information database includes capability vocabulary extraction and capability classification construction. Capability vocabulary extraction involves identifying and extracting keywords representing mine business functions and operations from the natural language descriptions of mine scheduling business scenarios. For example, keywords such as "personnel," "on duty," and "query" are extracted from "common personnel information queries and duty status queries." Capability classification construction involves classifying the extracted capability vocabulary into predefined categories according to their respective business domains. Predefined categories include, but are not limited to, personnel management, equipment monitoring, sensor data, page interaction, and data visualization.

[0069] Specifically, natural language descriptions of mine scheduling business scenarios are obtained, and capability terms are extracted based on these descriptions. These capability terms are then categorized according to pre-defined categories to generate a mine scheduling information database. This database contains multiple capability terms and their corresponding category information. Based on this database, multiple question-and-answer pairs related to mine scheduling are formed, creating a labeled question-and-answer dataset for mine scheduling.

[0070] Optionally, each sample question in the mine scheduling labeled question-and-answer dataset includes: the question time and chained thinking prompts for the sample question; the label content corresponding to each sample question and the output of the pre-trained language model are both structured data.

[0071] The sample questions are the question samples from each sample pair in the labeled question-and-answer dataset for mine scheduling. Each sample contains a natural language question about mine scheduling. The question time is the specific timestamp recording when the question was asked. This timestamp is used to analyze the correlation between the question and a specific date. Integrating the question time into the labeled question-and-answer dataset for mine scheduling introduces the concept of time information into the training process of the pre-trained language model. Chain-thinking prompts are text prompts that guide the language model's thinking and analysis. These prompts break down the solution process into multiple intermediate steps, guiding the language model to organize the logical steps and further output the final answer. The label content corresponding to the sample question is the standard answer to that sample question. This standard answer is a set of structured data, stored in a machine-readable format such as a JSON object, for direct use by the program later. The output of the pre-trained language model is the answer generated by the pre-trained language model based on the sample questions in the labeled question-and-answer dataset for mine scheduling during the training and fine-tuning process. This answer has the same data format as the labels, both being structured data.

[0072] Specifically, the labeled question-and-answer dataset for mine scheduling consists of multiple question-and-answer sample pairs. Each sample pair includes not only the query text related to mine scheduling, but also the specific question's question time and chain-thinking prompts to guide the language model's analysis. By adding question time information to the labeled question-and-answer dataset for mine scheduling, the target language model's understanding of the timeliness of scheduling tasks is enhanced after training and fine-tuning. Simultaneously, the introduction of chain-thinking prompts guides the target language model to perform step-by-step reasoning to capture user intent. The addition of question time information and chain-thinking prompts improves the accuracy of the target language model's final generated answer.

[0073] Optionally, structured data includes multiple fields and field information for each field.

[0074] The fields include: the type of operation requested by the user, the time when the operation occurred, the target object of the operation, the priority field, and the location field.

[0075] In this embodiment, structured data refers to data formatted according to a predefined model, which can be a collection of "field-field information" pairs. This structured data can be directly used for program parsing and invocation. A field is the basic unit in structured data; each field describes a specific category, and each field has a field name and a corresponding field information. In a mine scheduling scenario, fields may include, but are not limited to, the user-requested operation type, operation occurrence time, target object of the operation, priority field, and location field. The user-requested operation type indicates the operation the user intends to perform, such as "start," "stop," and "query." The operation occurrence time is the time when the user requests the operation to be executed. The target object of the operation is the object to which the operation is applied, used to determine the scope of the operation, such as the mine equipment number, the area of ​​effect, etc. The priority field characterizes the importance of the operation in processing; the field information in the priority field is usually a predefined importance value. The location field describes location information related to the operation, such as the location of the operation object and / or the location information of the area where the operation occurs. The field information corresponding to each field is the specific content of the corresponding field. If the field represents the type of operation requested by the user, the field information can be "Start", "Stop", or "Query", etc.

[0076] Specifically, a structured data format is pre-defined. The structured data includes multiple fields and corresponding field information for each field. The fields in the structured data include, but are not limited to, the type of operation requested by the user, the time when the operation occurred, the target object of the operation, priority fields, and location fields.

[0077] S220, based on the labeled question-and-answer dataset for mine scheduling and the low-rank adaptation fine-tuning method, fine-tunes the pre-trained language model to obtain the target language model.

[0078] The low-rank adaptation fine-tuning method is a model fine-tuning approach. For a specific fully connected layer in the pre-trained language model, such as the query, key, and value matrices in the attention mechanism, the original weights of the pre-trained language model are frozen, and two small low-rank matrices are introduced. During the fine-tuning of the pre-trained language model, only these two low-rank matrices are trained, and their product is used for incremental updates to the original weights. The pre-trained language model is a language model pre-trained on a large-scale general dataset, possessing the ability to understand language and generate text in general scenarios. The target language model is the language model obtained by fine-tuning the pre-trained language model using the low-rank adaptation fine-tuning method on a labeled question-answering dataset for mine scheduling. This target language model retains the language understanding ability in general scenarios and also incorporates professional knowledge from the mine scheduling domain.

[0079] In this embodiment, during the training and fine-tuning of the pre-trained language model, firstly, sample questions from the labeled question-and-answer dataset for mine scheduling are input into the pre-trained language model. The model then generates answers to the corresponding sample questions, which are structured data containing multiple fields and their information. Subsequently, a JavaScript Object Notation (JSON) parser is used to parse both the model's output and the corresponding standard answers from the labeled question-and-answer dataset for mine scheduling, converting them into programmable data structures. To optimize model parameters, the system executes a two-stage comparison strategy. First, the output data undergoes a field accuracy comparison, checking whether all fields in the pre-trained language model's output answer are completely consistent with the field set in the sample answers in the dataset, ensuring the standardization of the output structure. Second, some fields in the sample answers are pre-defined as fields with strong semantic information, and the fields with strong semantic information in the model's output answer are further compared with those in the sample answers. The cosine distance method is used to calculate the semantic similarity of the fields with strong semantic information in the two answers. Based on the above two comparison results, a loss function is constructed, and the model parameters are iteratively optimized through backpropagation. Finally, a target language model that conforms to the mine scheduling scenario is trained.

[0080] For example, in the model fine-tuning phase, field consistency checks are first performed to ensure that the structured answers generated by the pre-trained language model for the sample questions completely include all fields defined in the corresponding sample answers of the mine scheduling labeled answer dataset. Based on this, semantic comparisons are further performed on specific fields. For instance, the field "user-requested operation type" is identified as a strong semantic information field, and the model output for this field is "alarm," while the sample answer shows "notification." The semantic distance between the two field information "alarm" and "notification" is quantified using cosine distance. This semantic similarity calculation result, along with the field consistency check result, participates in the construction of the loss function. Similarity calculations are performed, and the model parameters of the pre-trained language model are optimized based on the calculation structure.

[0081] Specifically, a labeled question-and-answer dataset for mine scheduling is obtained. Sample questions from this dataset are input into a pre-trained language model, which is then trained and fine-tuned using low-rank adaptation. During training, the answers output by the pre-trained language model are compared a second time with the sample answers in the dataset. Based on the comparison results, the model parameters are optimized to ultimately obtain the target language model.

[0082] The technical solution provided in this invention constructs a mine scheduling information database through capability vocabulary extraction and capability classification. Based on the mine scheduling information database, a labeled question-and-answer dataset for mine scheduling is constructed. Using this dataset, a low-rank adaptation fine-tuning method is employed to train and fine-tune a pre-trained language model. During training, the output data of the pre-trained language model is compared a second time with the sample answers in the dataset to ultimately obtain the trained target language model. Addressing the weaknesses of general pre-trained language models, such as time analysis and understanding of mine-specific terminology, a data construction method combining chain-thinking prompts and structured output is used to enable the model to learn the instruction parsing approach in the mine domain. This improves the accuracy of text generation in mine-domain scenarios. The low-rank adaptation fine-tuning method allows the pre-trained language model to quickly adapt to mine scheduling scenarios. The structured output information is used for subsequent scheduling execution operations, realizing a complete process from user request to execution feedback, thus improving scheduling automation and accuracy.

[0083] Figure 4 This is a flowchart illustrating a mine scheduling question-and-answer method provided in an embodiment of the present invention. In implementing the mine scheduling question-and-answer process, considering the sensitivity of tool queries to external data, user identity information is used to control data access permissions in the mine scheduling system during the target language model's tool invocation process. Based on the above optional implementation methods, this embodiment describes the specific methods for controlling data access permissions during the mine scheduling question-and-answer process, such as... Figure 4As shown, the method includes:

[0084] S310. Obtain the mine scheduling question entered by the target user on the front-end page, the target token corresponding to the target user, and the session identifier information for this round.

[0085] The target token for each user serves as a unique identifier and verification credential. Upon login, the server issues an encrypted string as the target token, which contains the relevant system permissions. The current session identifier is a string that uniquely identifies each individual conversation. When a user initiates a new conversation (opening a new chat window), the system generates a unique identifier for that session.

[0086] Specifically, after a target user logs in to the system, the server issues a target token corresponding to that user. This target token contains the user's corresponding operational permissions within the system. When a target user creates a new chat window, the system generates session identifier information for that round.

[0087] S320. Parse the target token to obtain the target user's basic token, the target user's authorized scheduling subsystem identifier information, and the target user's authorized scheduling submodule identifier information.

[0088] The basic token is a string used to identify the target user, typically a specific field, which can be used for subsequent permission verification. The authorized scheduling subsystem identifier information represents the business system code that the target user has the right to access, and the authorized scheduling submodule identifier information represents the submodule code that the target user specifically has the operation permissions within a certain subsystem. In this embodiment, the mine scheduling system contains multiple relatively independent subsystems and submodules.

[0089] Specifically, after obtaining the target user's target token, a JSON parser can be used to parse it, which can obtain the relevant permission information corresponding to the target user, such as the target user's basic token, the scheduling subsystem identifier information authorized by the target user, and the scheduling submodule identifier information authorized by the target user.

[0090] For example, after a target user completes login, the target token corresponding to the target user is obtained as follows: {"Token":"1234-1234-1234-1234","AuthorizationModule":"001,002","AuthorizationSystem":"A,B,C"}. Parsing the target token using a JSON parser yields the target user's basic token, the authorized scheduling subsystem identifier information, and the authorized scheduling submodule identifier information.

[0091] S330. Cache the target user's basic token, the target user's authorized scheduling subsystem identifier, the target user's authorized scheduling submodule identifier, and the current session identifier.

[0092] Caching can be a method of temporarily storing data in an easily accessible intermediate storage layer.

[0093] Specifically, the system obtains the target user's base token (parsed from the target token), the target user's authorized scheduling subsystem identifier information, and the target user's authorized scheduling submodule identifier information. After the user creates a new chat window, this information is cached together with the system-generated current-round session identifier information, maintaining the above data as inherent data cached when the current-round session identifier information remains unchanged.

[0094] S340. Based on the mine scheduling problem, the configuration information of each MCP server among multiple MCP servers, and the preset prompt template, determine the target prompt text.

[0095] S350. Input the target prompt text into the target language model for processing to obtain the target MCP server for handling mine scheduling problems and the target tool description information required by the target MCP server.

[0096] S360. Based on the target tool description information, call the target tool to the target MCP server, and return the target tool call result to the target language model for information integration to obtain the target response content corresponding to the mine scheduling problem.

[0097] Optionally, based on the target tool description information, the target tool is invoked to the target MCP server, including:

[0098] If the target tool description information has a configured token field, and the target user's authorized scheduling subsystem identifier information and the target user's authorized scheduling submodule identifier information contain the target MCP server's identifier information, then the target user's basic token is written into the target tool description information; based on the target tool description information after writing the basic token, the target tool is called to the target MCP server to obtain the target tool call result.

[0099] The target MCP server is the selected MCP server to which the call will be initiated. This MCP server corresponds to a calling subsystem or calling submodule. The token field is a parameter field pre-set in the target tool description information, representing the identity credentials of the target user that can be carried during the call. The identification information of the target MCP server serves as an index for permission verification. The system determines whether the user has permission to access the MCP server by checking whether this identification information is included in the cached authorization identification information.

[0100] Specifically, when initiating a target tool call to the target MCP server based on the target tool description information, the target MCP server's target tool description information includes a token field. The target MCP server's identification information is compared with the cached identification information of the target user's authorized scheduling subsystems and submodules. If the target MCP server's identification information is found in the target user's authorized calling subsystems and submodules, it indicates that the target MCP server is an authorized calling subsystem or submodule for the target user. This improves the accuracy of system call data access control and the compliance of the call. Subsequently, after confirming authorization, the target user's basic token is written to the token field of the target tool description information, providing authentication credentials for subsequent tool calls. By separating authorization verification from token injection, it ensures that identity credentials are only used in the authorized call stage, achieving security protection for user identity data.

[0101] S370. Return the target response content and the current session identifier information to the front-end page, so that when the front-end page determines that the target response content is the current session response content based on the returned current session identifier information, it displays the target response content on the front-end page.

[0102] Specifically, after the target language model generates the target response content, it packages the target response content and its corresponding current session identifier information together and returns it to the front-end page. Upon receiving the response, the front-end extracts the current session identifier information and compares it with the current session identifier information in the current chat window. If the extracted current session identifier information matches the current chat window's current session identifier information, the target response content is determined to be the response content corresponding to the mine scheduling question entered by the target user in this session. Furthermore, the target response content is displayed on the front-end page for the target user to view.

[0103] The technical solution provided in this invention obtains the target user's mining scheduling question, target token, and current session identifier information after the target user logs into the system and inputs the question on the front-end page. The target token is parsed to obtain the target user's basic token, the identifier information of the authorized scheduling subsystem, and the identifier information of the authorized scheduling submodule. Furthermore, the target user's basic token, the authorized scheduling subsystem identifier information, the authorized scheduling submodule identifier information, and the current session identifier information are cached. Based on the mining scheduling question, the configuration information of each MCP server among multiple MCP servers, and a preset prompt template, the target prompt text is determined. The target prompt text is input into a trained target language model for processing to obtain the target MCP server used to handle the mining scheduling question and the description information of the target tools required by the target MCP server. Based on the target tool description information, the target tool is invoked from the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration to generate the target response content corresponding to the mining scheduling question. Finally, the target response content and the cached current session identifier information are returned to the front-end page. Based on the returned current session identifier information, it is determined that the target response content corresponds to the mine scheduling question entered by the target user in this session. After confirming that the current session identifier information is consistent, the target response content is displayed on the front-end page. This ensures the security of user identity information, system data security, and access control when the target language model calls the target MCP server.

[0104] Figure 5 This is an overall framework diagram of a mine scheduling question-and-answer method provided in an embodiment of the present invention, combined with... Figure 5 Understand the technical solutions of the embodiments of the present invention.

[0105] like Figure 5 As shown, this embodiment explains the overall implementation of the solution based on the above optional implementation methods, specifically including:

[0106] The mine scheduling system adopts a layered architecture, consisting of an infrastructure layer, a communication layer, a service layer, and an application layer from top to bottom. These layers work together to achieve the entire mine scheduling process. The infrastructure layer integrates hardware resources such as a computing power service machine, a large-scale model operation and maintenance toolchain, and a model factory. It constructs a labeled question-and-answer dataset for mine scheduling to train and fine-tune models, providing computing power support for the system. The communication layer ensures real-time transmission of scheduling commands through broadcasting, LEDs, and other methods. The service layer integrates the mine dataset (labeled question-and-answer dataset for mine scheduling) and the model factory, and relies on the vector database and relational database in the data service layer for data storage and retrieval. The application layer integrates the MCP capability factory (MCP server) and basic component library, encapsulating various calling tools and components. Leveraging voice intelligent interaction services and AI core services, it implements functions such as domain knowledge question answering, real-time data dashboards, converged communication, page navigation, and element location. Finally, it integrates with various mine scheduling subsystems or submodules through interfaces to complete mine scheduling.

[0107] This invention, in its embodiments, acquires a mine scheduling question input by a target user on a front-end page. Based on the mine scheduling question, configuration information of multiple MCP servers, and a preset prompt template, a target prompt text is determined. Further, the target prompt text is input into a target language model for processing, obtaining descriptions of the target MCP server used to handle the mine scheduling question and the target tools required by the target MCP server. Based on the target tool descriptions, the target tool is invoked from the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration to obtain the target response content corresponding to the mine scheduling question. Finally, the target response content is returned to the front-end page for display. This achieves mine scheduling responses, improves the efficiency of the integration between natural language and the scheduling system, and enhances the accuracy of front-end information display.

[0108] Figure 6 This is a schematic diagram of the structure of a mine dispatching question-and-answer device provided in an embodiment of the present invention, as shown below. Figure 6 As shown, the device includes: a mine scheduling problem acquisition module 410, a target prompt text determination module 420, a target tool description information acquisition module 430, a target response content acquisition module 440, and a target response content display module 450.

[0109] The system includes the following modules: a mine scheduling problem acquisition module 410, used to acquire the mine scheduling problem input by the target user on the front-end page; a target prompt text determination module 420, used to determine the target prompt text based on the mine scheduling problem, the configuration information of each MCP server among multiple MCP servers, and a preset prompt template, wherein the MCP server is a server that follows the Model Context Protocol, and the MCP server corresponds one-to-one with the scheduling subsystem or scheduling submodule in the mine scheduling system; a target tool description information acquisition module 430, used to input the target prompt text into the target language model for processing to obtain the target MCP server used to handle the mine scheduling problem and the target tool description information required by the target MCP server; a target response content acquisition module 440, used to call the target tool to the target MCP server based on the target tool description information, and return the target tool call result to the target language model for information integration to obtain the target response content corresponding to the mine scheduling problem; and a target response content display module 450, used to return the target response content to the front-end page for display.

[0110] This invention, in its embodiments, acquires a mine scheduling question input by a target user on a front-end page. Based on the mine scheduling question, configuration information of multiple MCP servers, and a preset prompt template, a target prompt text is determined. Further, the target prompt text is input into a target language model for processing, obtaining descriptions of the target MCP server used to handle the mine scheduling question and the target tools required by the target MCP server. Based on the target tool descriptions, the target tool is invoked from the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration to obtain the target response content corresponding to the mine scheduling question. Finally, the target response content is returned to the front-end page for display. This achieves mine scheduling responses, improves the efficiency of the integration between natural language and the scheduling system, and enhances the accuracy of front-end information display.

[0111] Based on the above technical solutions, the device may optionally include:

[0112] The dataset construction module is used to build a labeled question-and-answer dataset for mine scheduling based on the mine scheduling information database;

[0113] The target language model acquisition module is used to fine-tune the pre-trained language model based on the mine scheduling labeled question-and-answer dataset and the low-rank adaptation fine-tuning method to obtain the target language model.

[0114] Based on the above technical solutions, the dataset construction module includes: each sample question in the mine scheduling labeled question-and-answer dataset includes: the questioning time of the sample question and the chain-thinking prompt information; the label content corresponding to each sample question and the output of the pre-trained language model are both structured data.

[0115] Based on the above technical solutions, the dataset construction module includes: structured data including multiple fields and field information corresponding to each field; wherein, the fields include: the type of operation requested by the user, the time when the operation occurred, the target object of the operation, the priority field, and the location field.

[0116] Based on the above technical solutions, the device may optionally include:

[0117] The business scheduling execution module is used to parse the target response content to obtain business scheduling instructions, and execute the corresponding business scheduling based on the business scheduling instructions; wherein, the business scheduling instructions include page jump instructions and / or element location instructions.

[0118] Based on the above technical solutions, the mine scheduling problem acquisition module includes:

[0119] The information acquisition unit is used to acquire the mine scheduling question entered by the target user on the front-end page, the target token corresponding to the target user, and the session identifier information for this round.

[0120] The identification information acquisition unit is used to parse the target token to obtain the target user's basic token, the target user's authorized scheduling subsystem identification information, and the target user's authorized scheduling submodule identification information.

[0121] The information caching unit is used to cache the target user's basic token, the target user's authorized scheduling subsystem identification information, the target user's authorized scheduling submodule identification information, and the current session identification information.

[0122] Based on the above technical solutions, the target response content acquisition module also includes:

[0123] The basic token writing unit is used to write the target user's basic token into the target tool description information if a token field has been configured in the target tool description information, and the target user's authorized scheduling subsystem identifier information and the target user's authorized scheduling submodule identifier information contain the identifier information of the target MCP server.

[0124] The target tool call result acquisition unit is used to call the target tool to the target MCP server based on the target tool description information written to the base token, so as to obtain the target tool call result.

[0125] Based on the above technical solutions, the target response content display module also includes:

[0126] The target response content display unit is used to return the target response content and the current round of conversation identification information to the front-end page, so that when the front-end page determines that the target response content is the current round of conversation response content based on the returned current round of conversation identification information, it will display the target response content on the front-end page.

[0127] The mine scheduling question-and-answer device provided in this embodiment of the invention can execute a mine scheduling question-and-answer method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0128] Figure 7 This is a schematic diagram of a mine dispatching question-and-answer system provided in an embodiment of the present invention. This embodiment is applicable to situations involving mine dispatching responses. Figure 7 As shown, the system specifically includes: a front-end page 510, an MCP client 520, and multiple MCP servers 530;

[0129] The front-end page 510 is used to implement the mine scheduling question-and-answer method in the above embodiments. The MCP client 520 is used to implement the mine scheduling question-and-answer method in the above embodiments. Each MCP server 530 is used to implement the mine scheduling question-and-answer method in the above embodiments.

[0130] The mine scheduling question-and-answer system in this embodiment of the invention obtains the mine scheduling question input by the target user on the front-end page. Based on the mine scheduling question, the configuration information of multiple MCP servers, and the preset prompt template, the target prompt text is determined. Further, the target prompt text is input into the target language model for processing, and the target MCP server used to process the mine scheduling question and the target tool description information required by the target MCP server are obtained. Based on the target tool description information, the target tool is invoked from the target MCP server, and the result of the target tool invocation is returned to the target language model for information integration to obtain the target answer content corresponding to the mine scheduling question. Finally, the target answer content is returned to the front-end page for display. By encapsulating the scheduling subsystem and scheduling submodule in the mine scheduling system into an MCP server that conforms to the Model Context Protocol, intelligent docking between the target language model and the mine scheduling system is achieved based on multiple MCP servers. This enables intelligent question-and-answering in mine scheduling using the target language model, improving the efficiency and accuracy of mine scheduling.

[0131] Figure 8A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0132] like Figure 8 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0133] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0134] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a mine scheduling question-and-answer method.

[0135] In some embodiments, a mine scheduling question-and-answer method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the mine scheduling question-and-answer method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform a mine scheduling question-and-answer method by any other suitable means (e.g., by means of firmware).

[0136] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0137] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0138] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0139] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0140] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0141] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0142] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and no limitation is imposed herein.

[0143] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A mine dispatching question-and-answer method, characterized in that, include: Obtain the mine scheduling question entered by the target user on the front-end page; Based on the aforementioned mine scheduling problem, the configuration information of each MCP server among multiple MCP servers, and the preset prompt template, the target prompt text is determined. The MCP server is a server that follows the Model Context Protocol, and the MCP server corresponds one-to-one with the scheduling subsystem or scheduling submodule in the mine scheduling system. The target prompt text is input into the target language model for processing to obtain the target MCP server for handling the mine scheduling problem and the target tool description information required by the target MCP server. Based on the target tool description information, the target tool is invoked to the target MCP server, and the target tool invocation result is returned to the target language model for information integration to obtain the target response content corresponding to the mine scheduling problem. The target response content is returned to the front-end page for display. The step of obtaining the mine scheduling question input by the target user on the front-end page includes: Obtain the mine scheduling question entered by the target user on the front-end page, the target token corresponding to the target user, and the session identifier information for this round; The target token is parsed to obtain the target user's base token, the target user's authorized scheduling subsystem identifier information, and the target user's authorized scheduling submodule identifier information; The target user's basic token, the target user's authorized scheduling subsystem identifier, the target user's authorized scheduling submodule identifier, and the current session identifier are cached. The step of invoking the target tool to the target MCP server based on the target tool description information includes: If a token field is configured in the target tool description information, and the target user's authorized scheduling subsystem identifier information and the target user's authorized scheduling submodule identifier information contain the identifier information of the target MCP server, then the target user's basic token is written into the target tool description information. Based on the target tool description information written to the base token, the target tool is invoked to the target MCP server to obtain the target tool invocation result.

2. The method according to claim 1, characterized in that, The method further includes: Based on the mine scheduling information database, a labeled question-and-answer dataset for mine scheduling is constructed. Based on the labeled question-and-answer dataset for mine scheduling and the low-rank adaptation fine-tuning method, the pre-trained language model is fine-tuned to obtain the target language model.

3. The method according to claim 2, characterized in that, Each sample question in the labeled question-and-answer dataset for mine scheduling includes: the question's questioning time and chained thinking prompts; The label content corresponding to each sample question and the output of the pre-trained language model are both structured data.

4. The method according to claim 3, characterized in that, The structured data includes multiple fields and field information corresponding to each field; The fields include: the type of operation requested by the user, the time when the operation occurred, the target object of the operation, the priority field, and the location field.

5. The method according to claim 1, characterized in that, The method further includes: The target response content is parsed to obtain a service scheduling instruction, and the corresponding service scheduling is executed based on the service scheduling instruction; The service scheduling instructions include page navigation instructions and / or element location instructions.

6. The method according to claim 1, characterized in that, The step of returning the target response content to the front-end page for display on the front-end page includes: The target response content and the current session identifier information are returned to the front-end page, so that when the front-end page determines that the target response content is the current session response content based on the returned current session identifier information, the front-end page displays the target response content.

7. A mine dispatching question-and-answer device, characterized in that, include: The mine scheduling problem acquisition module is used to acquire mine scheduling problems entered by target users on the front-end page; The target prompt text determination module is used to determine the target prompt text based on the mine scheduling problem, the configuration information of each MCP server among multiple MCP servers, and the preset prompt template. The MCP server is a server that follows the Model Context Protocol, and the MCP server corresponds one-to-one with the scheduling subsystem or scheduling submodule in the mine scheduling system. The target tool description information acquisition module is used to input the target prompt text into the target language model for processing, so as to obtain the target MCP server for processing the mine scheduling problem and the target tool description information required by the target MCP server. The target response content acquisition module is used to call the target tool to the target MCP server based on the target tool description information, and return the target tool call result to the target language model for information integration, so as to obtain the target response content corresponding to the mine scheduling problem; The target response content display module is used to return the target response content to the front-end page so as to display the target response content on the front-end page; The mine scheduling problem acquisition module includes: The information acquisition unit is used to acquire the mine scheduling question entered by the target user on the front-end page, the target token corresponding to the target user, and the current session identifier information; The identification information acquisition unit is used to parse the target token to obtain the target user's basic token, the target user's authorized scheduling subsystem identification information, and the target user's authorized scheduling submodule identification information. The information caching unit is used to cache the target user's basic token, the target user's authorized scheduling subsystem identifier information, the target user's authorized scheduling submodule identifier information, and the current session identifier information; The target response content acquisition module includes: The basic token writing unit is used to write the basic token of the target user into the target tool description information if a token field has been configured in the target tool description information, and the target user's authorized scheduling subsystem identifier information and the target user's authorized scheduling submodule identifier information contain the identifier information of the target MCP server. The target tool call result acquisition unit is used to call the target tool to the target MCP server based on the target tool description information written to the base token, so as to obtain the target tool call result.

8. A mine dispatching question-and-answer system, characterized in that, The system includes: a front-end page, an MCP client, and multiple MCP servers; The MCP client is used to implement the mine scheduling question-and-answer method as described in any one of claims 1-6.

Images