Large language model driven flexible manufacturing system componentized intelligent programming method and system

Abstract

This invention discloses a modular intelligent programming method, device, and medium for flexible manufacturing systems driven by a large language model. The method includes a task decomposition step, a manufacturing subtask single-step description generation step, and a process generation step. The task decomposition step uses a large model to generate step-by-step tasks based on user-input task descriptions. The manufacturing subtask single-step description generation step uses the large model to generate single-step descriptions for each manufacturing subtask within the step-by-step tasks. The process generation step uses the large model to generate a complete description of the manufacturing process based on multiple manufacturing subtask single-step descriptions. This invention leverages the powerful natural language understanding and generation capabilities of large language models; by designing appropriate task prompts, the large language model can automatically plan the entire manufacturing process. The method of this invention can quickly generate detailed steps of the manufacturing process and allows for interactive iteration with engineers, improving the interpretability, flexibility, and adaptability of the planning process.

Description

Technical Field

[0001] This invention relates to the field of development process design technology, specifically to a modular intelligent programming method, device, and medium for flexible manufacturing systems driven by a large language model. Background Technology

[0002] In today's rapidly evolving Industry 5.0 era, the manufacturing industry is undergoing a profound transformation. At its core lies a shift from traditional standardized, large-scale production models to more flexible, personalized, and small-batch production. This transformation is not only a challenge to the existing manufacturing system but also a major test of future manufacturing capabilities. To meet the growing personalized demands of consumers while maintaining their competitiveness, manufacturers must build intelligent manufacturing systems capable of rapidly adjusting production line configurations and efficiently handling various customized orders.

[0003] In this context, automated planning in the manufacturing process becomes crucial. It not only impacts the delivery speed of the final product but also directly affects production efficiency, product quality, and cost control. An efficient automated planning system can significantly shorten the product cycle time from design to market, improve work efficiency by optimizing resource allocation, and ensure that every production stage meets the highest standards, thereby providing customers with high-quality products and services. Furthermore, such a system can help companies make more flexible cost management decisions when facing external factors such as fluctuations in raw material prices.

[0004] However, most manufacturing companies still primarily rely on the experience and expertise of human engineers for manufacturing process planning. While this method has played a crucial role in the past, its limitations are becoming increasingly apparent in today's volatile market environment: on the one hand, manual planning is time-consuming and inefficient; on the other hand, the error rate due to human factors is relatively high, making it difficult to guarantee consistent quality levels. More importantly, traditional manual planning methods often fail to respond promptly to drastic changes in market demand, undoubtedly posing significant challenges to enterprises. Therefore, developing and applying more intelligent and adaptive automated planning technologies has become key to driving the transformation and upgrading of the manufacturing industry. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a modular intelligent programming method, device, and medium for flexible manufacturing systems driven by a large language model.

[0006] The first aspect of this invention provides a modular intelligent programming method for flexible manufacturing systems driven by a large language model, including a task decomposition step, a manufacturing subtask single-step description generation step, and a process generation step.

[0007] The task decomposition step uses a large model to generate step-by-step tasks based on the user-input task description; each step-by-step task contains multiple manufacturing sub-tasks.

[0008] The manufacturing subtask single-step description generation step uses a large model to generate a manufacturing subtask single-step description for each manufacturing subtask in the step-by-step task; the manufacturing subtask single-step description is used to explain the manufacturing function of the manufacturing subtask associated with that manufacturing subtask;

[0009] The process generation step uses a large model to generate a complete description of the manufacturing process based on the step-by-step descriptions of multiple manufacturing sub-tasks.

[0010] Furthermore, the task decomposition step specifically includes the following steps:

[0011] Receive a user task description input by the user, and create a manufacturing task order based on the user task description; the manufacturing task order contains all the manufacturing functions that the task description may require.

[0012] The user task description is processed using a large language model to generate multiple manufacturing sub-tasks that form a step-by-step task; the manufacturing sub-tasks are used to describe the manufacturing functions included in the manufacturing task order and the manufacturing units that implement the manufacturing functions;

[0013] Receive user confirmation request for the step-by-step task; when the user accepts the step-by-step task, execute the step of generating a single-step description of the subtask and add the step-by-step task to the training dataset of the large language model; when the user rejects the step-by-step task, return to the step of processing the user task description using the large language model and regenerate the step-by-step task.

[0014] Furthermore, each of the manufacturing units is associated with one or more manufacturing subtasks; each of the manufacturing subtasks is associated with one or more manufacturing functions.

[0015] Furthermore, the single-step description generation step of the manufacturing subtask specifically includes the following steps:

[0016] A single-step description of each manufacturing subtask is generated using a large language model; each manufacturing subtask includes the manufacturing function in the corresponding manufacturing subtask, the dependencies required by the manufacturing function, and the output content of the manufacturing function.

[0017] Receive user confirmation request for the step-by-step description of the manufacturing subtask; when the user accepts the step-by-step description of the manufacturing subtask, execute the process generation step and add the step-by-step description of the manufacturing subtask to the training dataset of the large language model; when the user rejects the step-by-step description of the manufacturing subtask, return to the step of generating the step-by-step description of each manufacturing subtask using the large language model, and regenerate the step-by-step task.

[0018] Furthermore, the dependencies required by the manufacturing function refer to the parameters of the manufacturing order that the manufacturing function is associated with, as well as the output content of other manufacturing functions.

[0019] Furthermore, the process generation step specifically includes the following steps:

[0020] Using a large language model, interface relationships and content interaction relationships are established between manufacturing functions associated with multiple manufacturing subtasks based on the single-step description of the manufacturing subtasks; the interface relationships are used to describe the interfaces that enable data interaction between manufacturing functions; the content interaction relationships are used to describe the specific output content of data interaction between manufacturing functions.

[0021] A complete description of the manufacturing process is generated using a large language model based on the interface relationships and content interaction relationships.

[0022] Receive user confirmation request for the complete description; when the user accepts the complete description, use the generated complete description as the manufacturing process and add the complete description to the training dataset of the large language model; when the user rejects the complete description, return to the step of generating a complete description of the manufacturing process using the large language model based on the interface relationship and content interaction relationship, and regenerate the complete description.

[0023] Furthermore, before generating a complete description of the manufacturing process using a large language model based on the interface relationships and content interaction relationships, the following steps are also included:

[0024] The system provides a pre-defined manufacturing process template to the large language model, enabling the model to generate a complete description of the manufacturing process based on the template.

[0025] Furthermore, it also includes the following steps:

[0026] Encapsulate a finite state machine within each manufacturing subtask;

[0027] When executing the complete description of the manufacturing process, the execution state of each manufacturing function is obtained through the finite state machine encapsulated in the manufacturing subtask;

[0028] Based on the execution state obtained from the finite state machine, determine whether there are any errors in the complete description of the current execution.

[0029] A second aspect of the present invention provides an electronic device, including a processor and a memory;

[0030] The memory is used to store programs;

[0031] The processor executes the program to implement the modular intelligent programming method for a flexible manufacturing system driven by a large language model.

[0032] A third aspect of the present invention provides a computer-readable storage medium storing a program that is executed by a processor to implement the aforementioned modular intelligent programming method for a flexible manufacturing system driven by a large language model.

[0033] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned method.

[0034] The embodiments of the present invention have the following beneficial effects: The present invention provides a modular intelligent programming method, device, and medium for flexible manufacturing systems driven by a large language model. Utilizing the powerful natural language understanding and generation capabilities of a large language model, the entire manufacturing process can be automatically planned by the large language model simply by designing appropriate task prompts, thus solving the aforementioned problems. The method of the present invention can quickly generate detailed steps of the manufacturing process and can interactively iterate with engineers, improving the interpretability, flexibility, and adaptability of the planning process, thereby achieving automatic conversion from requirement description to an executable manufacturing process. This invention is widely applicable to manufacturing process planning in the Industry 5.0 era.

[0035] Additional aspects and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description or may be learned by practice of the invention. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the basic process of a modular intelligent programming method for a flexible manufacturing system driven by a large language model, according to the present invention.

[0038] Figure 2 This is a detailed flowchart illustrating a modular intelligent programming method for a flexible manufacturing system driven by a large language model, as described in this invention.

[0039] Figure 3 This is a schematic diagram illustrating the relationship between manufacturing units, manufacturing sub-tasks, and manufacturing functions in the task decomposition steps of this invention.

[0040] Figure 4 This is a schematic diagram illustrating the specific execution relationship between the manufacturing subtasks of this invention.

[0041] Figure 5 This is a schematic diagram of the functional code block generated in the single-step description generation step of the manufacturing subtask of the present invention.

[0042] Figure 6 This is a schematic diagram of a single-step description function code block for the manufacturing subtask of a specific robotic arm moving component.

[0043] Figure 7 This is a schematic diagram illustrating possible errors encountered in the job planning method of this invention.

[0044] Figure 8 This is a schematic diagram of the structure of an electronic device according to the present invention.

[0045] Figure 9 This is a schematic diagram of a computer-readable storage medium structure according to the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0047] In today's rapidly evolving Industry 5.0 era, intelligent manufacturing systems are facing a shift from standardized, large-scale production to personalized, small-batch production. This shift demands high flexibility and adaptability in manufacturing processes to quickly respond to rapidly changing market demands. Against this backdrop, automated planning of the manufacturing process becomes particularly important, as it directly impacts production efficiency, product quality, and cost control. Currently, manufacturing process planning relies primarily on the experience and knowledge of human engineers, a method that is inefficient, error-prone, and ill-suited to adapting to rapid changes in market demands.

[0048] While machine learning and artificial intelligence technologies have made some progress in manufacturing process planning in recent years, these methods typically require large amounts of training data and scenario-specific knowledge, resulting in poor generalizability. Furthermore, the manufacturing processes generated by existing methods often lack interpretability, require manual post-processing, and struggle to handle complex multi-device collaborative planning problems. These issues severely restrict the widespread application of automated manufacturing process planning technology in industry and fail to meet the demands for flexibility and adaptability in the Industry 5.0 era.

[0049] To address the aforementioned problems, the first embodiment of this invention provides a modular intelligent programming method for flexible manufacturing systems driven by a large language model. For example... Figure 1 As shown, the method of this embodiment mainly includes a task decomposition step, a manufacturing subtask single-step description generation step, and a process generation step.

[0050] The task decomposition step generates step-by-step tasks based on the user-input task description. The manufacturing subtask single-step description generation step generates a single-step description for each manufacturing subtask within the step-by-step tasks. The process generation step uses a large model to generate a complete description of the manufacturing process based on the multiple manufacturing subtask single-step descriptions.

[0051] The detailed process of the embodiments of the present invention is as follows: Figure 2 As shown. The method proposed in this embodiment of the invention can automatically parse the manufacturing task workflow generation framework of user manufacturing order requirements through a large language model, and can simultaneously call heterogeneous industrial instances to integrate into the workflow.

[0052] The implementation process of each step of this invention is described in detail below.

[0053] Task decomposition steps: The task decomposition steps in this embodiment of the invention specifically include the following steps:

[0054] S101. Receive user task description input by the user, and create a manufacturing task order based on the user task description.

[0055] The task description expressed in natural language by the user cannot be directly used as a manufacturing process; therefore, it needs to be understood by a large language model. The large language model used in this invention refers to a class of large models in the field of Natural Language Processing (NLP), typically used for processing text data and understanding natural language. The main characteristic of these large language models is that they are trained on large-scale corpora to learn various grammatical, semantic, and contextual rules of natural language. Specific large language models include the GPT series (OpenAI), Bard (Google), and Wenxin Yiyan (Baidu).

[0056] In some embodiments, the user input task requirements may have a specific format that is the same as the training set format used to train the large language model, making it easier for the large language model to understand the user task requirements.

[0057] While understanding the task description, the large language model extracts keywords for manufacturing functions based on the keywords in the task description, generating a manufacturing task order. The manufacturing task order contains descriptions of various manufacturing functions to match manufacturing units provided by the manufacturing service provider. To fulfill the task requirements, this embodiment of the invention establishes a manufacturing unit library provided by the manufacturing service provider, encompassing manufacturing units that implement various manufacturing functions, using these manufacturing units as the basic tools for processing task requirements.

[0058] S102. Use a large language model to process the user task description and generate multiple manufacturing sub-tasks to form a step-by-step task.

[0059] In this embodiment of the invention, the specific relationship diagram between the manufacturing unit, the manufacturing subtask, and the manufacturing unit is as follows: Figure 3 As shown. In this embodiment of the invention, each manufacturing unit is associated with one or more manufacturing subtasks; each manufacturing subtask is associated with one or more manufacturing functions.

[0060] Specifically, for each manufacturing subtask t in the entire step-by-step task T i Its properties are defined by the following formula:

[0061] t i ={desc,paratrans}

[0062] Here, `desc` describes the manufacturing function to be performed by defining a single-step description of the manufacturing subtask, while `paratrans` represents the parameters received from adjacent manufacturing subtasks and the parameters passed out to other subtasks after processing by the internal algorithm. In `t`... i and its adjacent subtasks t i+1 ,t i-1 In this process, subtasks are interconnected through executable parameters in paratrans. Since manufacturing subtasks consist of multiple manufacturing functions—for example, a subtask of robotic object grasping requires a combination of servo motor and gripper control—and manufacturing units can perform various subtasks, embodiments of this invention require arranging and combining multiple subtasks in a consumer goods manufacturing order to generate a task workflow capable of completing the order. The manufacturing service provider must coordinate the execution logic of the manufacturing functions to derive a complete sequence of step-by-step tasks.

[0063] since Figure 3It can be observed that, since each manufacturing subtask in this embodiment of the invention is associated with only one manufacturing unit, each manufacturing subtask will only call one manufacturing unit for manufacturing during execution, avoiding the situation where multiple manufacturing units interfere with each other, causing the manufacturing subtask to fail. The specific granularity of the manufacturing subtask division is determined based on the level of detail of the task steps stored in the preset knowledge base. By providing manufacturing subtasks with the same level of granularity during the training phase of the large language model, the level of detail of the manufacturing subtasks output by the large language model can be maintained at the same level.

[0064] For example, when a user inputs a task description such as "using a depth camera and a robotic arm to perform a block grasping process," the task decomposition steps can yield the following manufacturing sub-tasks: "Step 1: To facilitate block identification, a depth camera is needed to locate the block. Step 2:...". It is evident that the task decomposition steps in this embodiment of the invention extrapolate the user's concise input to task steps with the same level of detail as those in the knowledge base, generating multiple manufacturing sub-tasks, thereby facilitating the execution of subsequent steps.

[0065] S103. Receive the user's confirmation request for the step-by-step task; when the user accepts the step-by-step task, execute the step of generating a single-step description of the subtask and add the step-by-step task to the training dataset of the large language model; when the user rejects the step-by-step task, return to the step of processing the user's task description using the large language model and regenerate the step-by-step task.

[0066] After the large language model outputs the step-by-step tasks, this embodiment of the invention can display the model-generated step-by-step tasks to the user. If the user believes that the manufacturing function implemented in the task is incomplete or conflicting, they can reject the distributed tasks output by the large language model. The large language model will then establish a multi-turn dialogue with the user, gradually decomposing the task process and generating manufacturing sub-tasks, ultimately obtaining the step-by-step task output accepted by the user. Simultaneously, the step-by-step task output accepted by the user is added to the large language model's training dataset, allowing the model to optimize subsequent task description decomposition and step-by-step task generation.

[0067] Manufacturing Subtask Single-Step Description Generation Steps: In this embodiment of the invention, the manufacturing subtask single-step description generation steps specifically include the following steps:

[0068] S201. Use a large language model to generate a step-by-step description of each manufacturing subtask.

[0069] In this embodiment of the invention, the manufacturing subtasks generated by the large language model have specific execution relationships. For example... Figure 4As shown, the execution relationships between manufacturing subtasks include sequential execution, conditional execution, and parallel execution. Sequential execution means starting with the first manufacturing subtask and executing each subtask sequentially until completion, without any intermediate conditional or parallel execution flows. Conditional execution means starting with the first manufacturing subtask and executing one or more branch tasks based on conditional judgments until completion. Parallel execution means starting with the first manufacturing subtask and executing sequentially up to the parallel judgment-related manufacturing subtasks, synchronously executing the subsequent parts of all parallel branches, and ending the task after all parallel branches are completed.

[0070] since Figure 3 It can be observed that the execution of each manufacturing function also depends on external dependencies. These dependencies may originate from the results of other functions or be based on internal parameter injection from consumer manufacturing orders. Therefore, upon receiving the step-by-step tasks from the previous step, this embodiment of the invention formally describes the parameter transmission for each step using a large language model. These parameter transmissions encapsulate the dependencies required to execute the manufacturing functions of the sub-tasks, as well as the outputs generated by executing these tasks, ultimately resulting in a... Figure 5 The shown function code block.

[0071] Specifically, the functional code block includes the dependent parameters involved in the manufacturing subtask and the internal algorithm functions responsible for its implementation. For example, calling the robotic arm movement function requires inputting the six-dimensional coordinate data specifying the position to be moved to. It also requires executing functions that sequentially call the robotic arm movement function.

[0072] In this embodiment of the invention, the generated function code block Paratrans consists of an API and a DATA interface. The API interface is the method for calling a specific function of the function code block, and the DATA interface is used for data transfer between function code blocks. For example, for a task requiring a robotic arm to perform, the robot arm's end effector is moved to a specified position by using the moveline linear movement API of the robotic arm function code block and passing in 6DoF coordinates. A key aspect of this stage is the effective processing of input parameters and the generation of parameters required by other function code blocks after internal operations. A specific example of this robotic arm function code block is shown below. Figure 6 As shown.

[0073] S202. Receive the user's confirmation request for the single-step description of the manufacturing subtask; when the user accepts the single-step description of the manufacturing subtask, execute the process generation step and add the single-step description of the manufacturing subtask to the training dataset of the large language model; when the user rejects the single-step description of the manufacturing subtask, return to the step of generating the single-step description of each manufacturing subtask using the large language model, and regenerate the step-by-step task.

[0074] After the large language model outputs a single-step description of the manufacturing subtask, this embodiment of the invention can display the model-generated single-step description to the user. If the user believes that the manufacturing function implemented in the single-step description is incomplete or conflicting, they can reject the single-step description output by the large language model. The large language model will then establish a multi-turn dialogue with the user, allowing the user to manually adjust the code block generated by the model, ultimately obtaining a single-step description of the manufacturing subtask that the user accepts. Simultaneously, the single-step description of the manufacturing subtask that the user accepts is added to the training dataset of the large language model, allowing the model to optimize subsequent single-step description generation for manufacturing subtasks.

[0075] Process generation steps: In this embodiment of the invention, the process generation steps specifically include the following steps:

[0076] S301. Use a large language model to establish interface relationships and content interaction relationships between manufacturing functions associated with multiple manufacturing subtasks based on the single-step description of the manufacturing subtasks.

[0077] The key focus of the workflow generation step in this embodiment is ensuring information interaction between the current functional code block and other functional code blocks in each subtask step. Paratrans is divided into two parts: [API, Data]; where the interface relationship API describes the interface for data interaction between manufacturing functions; and the content interaction relationship DATA describes the specific output content of data interaction between manufacturing functions. By establishing seamless parameter interaction between manufacturing subtasks and other manufacturing subtasks, the entire task is guaranteed to execute smoothly and without errors, assuming it has been divided into subtasks. During the workflow generation phase, the large language model not only considers parameter transfer between subtasks but also understands the specific requirements in the task description to establish parameter transfer relationships between all subtasks. This involves API interaction and DATA interaction between adjacent functional code blocks to ensure that task execution meets expected results.

[0078] S302. Use a large language model to generate a complete description of the manufacturing process based on interface relationships and content interaction relationships.

[0079] After the relationships between manufacturing subtasks are established, the execution results of each functional code block are transmitted to the API interfaces of subsequent code blocks via API interfaces. Each code block obtains the task execution results from the preceding manufacturing subtasks and processes them using the necessary functions encapsulated within the code block to obtain further execution results. The large language model can generate a complete description of the manufacturing process based on the interface relationships between various manufacturing subtasks, for use by front-line manufacturing personnel.

[0080] In some embodiments, before the large language model generates a complete description of the manufacturing process, a preset manufacturing process template can be imported first, allowing the large language model to generate a complete description of the manufacturing process based on the template. The manufacturing process template is a manufacturing process description format determined by the manufacturer, making it easy for the manufacturer to understand and interpret. This step effectively utilizes existing knowledge and experience to generate a coherent and practical task workflow, thereby improving the efficiency and accuracy of task execution.

[0081] S303. Receive the user's confirmation request for the complete description; when the user accepts the complete description, use the generated complete description as the manufacturing process and add the complete description to the training dataset of the large language model; when the user rejects the complete description, return to the step of generating a complete description of the manufacturing process using the large language model based on the interface relationship and content interaction relationship, and regenerate the complete description.

[0082] After the large language model outputs a complete description of the manufacturing process, this embodiment of the invention can display the model-generated complete description to the user. If the user believes that the complete description does not fully implement the manufacturing function or that there are conflicts, they can reject the complete description output by the large language model. The large language model will then establish a multi-turn dialogue with the user and adjust the output of a new complete description based on the user's feedback, ultimately obtaining a complete description output acceptable to the user. Simultaneously, the complete description output accepted by the user is added to the large language model's training dataset to help the model optimize subsequent complete description generation.

[0083] In practice, it has been found that the step-by-step tasks, single-step descriptions, and complete descriptions generated by large language models are not always executable, and various errors may exist. Most of these errors can be summarized as follows: Figure 7 The type shown:

[0084] Type (1) error: A deadlock may occur in the full description. After the executable loop condition of the manufacturing subtask is met, the current manufacturing subtask execution loop should be exited. However, the erroneous process will continue to iterate the current cycle until a timeout occurs, instead of continuing to execute subsequent processes.

[0085] Type (2) errors: Data transmission anomalies manifest in the execution sequence, with discrepancies observed in the information exchanged between functions. These errors include incorrect data input identifiers and mismatched data formats. For example, a robotic arm designed to manipulate six-dimensional coordinate data may provide image data from a camera that the camera cannot interpret.

[0086] Type (3) errors: Errors in sequence API inputs and outputs, especially in conditional and parallel processes. The corresponding sequence APIs for conditional steps and parallel processing must be invoked to execute the appropriate procedures. Directly transitioning from standard functionality to conditional or parallel processes is not feasible and would disrupt the expected process flow.

[0087] Type (4) error: Failure to execute a step in an executable sequence results in incomplete process execution. Such omissions may truncate the workflow, thereby compromising the overall integrity and functionality of the system.

[0088] To address this, embodiments of the present invention use a finite state machine to monitor the execution process of each manufacturing subtask, specifically including the following steps:

[0089] S401. Encapsulate the finite state machine in each manufacturing subtask;

[0090] S402. When executing a complete description of the manufacturing process, obtain the execution state of each manufacturing function through the finite state machine encapsulated in the manufacturing subtask;

[0091] S403. Based on the execution state obtained from the finite state machine, determine whether there are any errors in the complete description of the current execution.

[0092] For the step-by-step task T = {t1, t2, ... t} n In this embodiment of the invention, finite state machines are injected into the functional code blocks corresponding to each manufacturing subtask. In each finite state machine, 0 indicates that the current functional code block has not been executed, and 1 indicates that the current functional code block has been executed. By deploying finite state machines in the manufacturing subtasks, this embodiment of the invention only needs to maintain one state machine controller, which can execute the corresponding finite state machines according to the progress of the step-by-step tasks. When the controller executes a certain finite state machine, it triggers the state transition condition, S: 0 → 1; when the manufacturing subtask is completed, S: 1 → 0. Simultaneously, the executed finite state machines are listed to record the executed finite state machines and the execution count of each finite state machine.

[0093] In summary, the embodiments of this invention realize the transformation of manufacturing process planning from manual experience to automated and intelligent processes, improving planning efficiency; enhancing the flexibility and adaptability of manufacturing process planning, enabling it to quickly respond to rapid changes in market demands; reducing errors that may occur in manually designing manufacturing process steps, and improving the quality of manufacturing process planning. Furthermore, the embodiments of this invention enhance user participation in manufacturing process planning through interactive design, increasing user satisfaction; reducing the reliance on professional knowledge and skills in manufacturing process planning, and lowering the technical threshold. Finally, the embodiments of this invention enable rapid response to market demands, enhance the market competitiveness of enterprises, and can be widely applied in enterprise manufacturing process design.

[0094] Figure 8 This is a schematic diagram of the electronic device proposed in the second embodiment of the present invention. In this embodiment, the memory stores program instructions for implementing the modular intelligent programming method for a flexible manufacturing system driven by a large language model, as described in any of the above embodiments. The processor executes the program instructions stored in the memory to perform modular intelligent programming of the flexible manufacturing system driven by a large language model. The processor can also be referred to as a CPU (Central Processing Unit). The processor may be an integrated circuit chip with signal processing capabilities. The processor can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0095] The methods described in the first embodiment of the present invention are applicable to the embodiments of the present electronic device. The specific functions implemented by the embodiments of the present electronic device are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above methods.

[0096] Figure 9 This is a schematic diagram of the structure of a computer-readable storage medium according to the third embodiment of the present invention. The computer-readable storage medium of the fourth embodiment of the present invention stores program instructions capable of implementing the above-described large language model-driven modular intelligent programming method for flexible manufacturing systems. These program instructions can be stored in the storage medium in the form of a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of the present invention. The aforementioned computer-readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0097] The methods described in the first embodiment of the present invention are applicable to the computer-readable storage medium embodiment. The specific functions implemented by the computer-readable storage medium embodiment are the same as those in the above method embodiment, and the beneficial effects achieved are also the same as those achieved by the above method.

[0098] This embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned related steps to realize the modular intelligent programming method for flexible manufacturing systems driven by a large language model provided in the above embodiment.

[0099] It should be noted that 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 the present invention 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.

[0100] Those skilled in the art will understand that modules in the device of the embodiments of the present invention can be adaptively modified and placed in one or more devices different from those embodiments. Modules, units, or components in the embodiments of the present invention can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the corresponding claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the corresponding claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0101] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0102] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0103] Furthermore, the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. In particular, for embodiments such as apparatus and devices, since they are basically similar to the method embodiments, the relevant parts can be referred to the description of the method embodiments. The apparatus, devices, and other embodiments described above are merely illustrative, and the modules, units, etc., described as separate components may or may not be physically separate, that is, they may be located in one place or distributed in multiple places, such as nodes in a system network. Specifically, some or all of the modules and units can be selected according to actual needs to achieve the purpose of the above-described embodiment solutions. Those skilled in the art can understand and implement this without creative effort.

[0104] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0105] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0106] Furthermore, the terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying relative importance, or implicitly specifying the number of technical features indicated in this embodiment. Therefore, features defined with terms such as "first" and "second" in the embodiments of this invention can explicitly or implicitly indicate that the embodiment includes at least one of those features. In the description of this invention, the word "multiple" means at least two or more, such as two, three, four, etc., unless otherwise explicitly specified in the embodiments.

[0107] In embodiments of the present invention, the terms "comprising," "including," or any other variations thereof are intended to cover a 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 a process, method, article, or apparatus. Without further limitations, 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. Furthermore, components, features, and elements with the same names in different embodiments of the present invention may have the same meaning or different meanings, the specific meaning of which must be determined by its interpretation in that specific embodiment or further in conjunction with the context of that specific embodiment.

[0108] Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention. Other embodiments of the present invention will readily conceive of by considering the specification and practicing the invention. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

Claims

1. A modular intelligent programming method for flexible manufacturing systems driven by a large language model, characterized in that, This includes task decomposition steps, single-step description generation steps for manufacturing subtasks, and process generation steps. The task decomposition step uses a large model to generate step-by-step tasks based on the user-input task description; each step-by-step task contains multiple manufacturing sub-tasks. The manufacturing subtask single-step description generation step uses a large model to generate a manufacturing subtask single-step description for each manufacturing subtask in the step-by-step task; the manufacturing subtask single-step description is used to explain the manufacturing function associated with the manufacturing subtask; The process generation step uses a large model to generate a complete description of the manufacturing process based on the single-step description of multiple manufacturing sub-tasks. The task decomposition step specifically includes the following steps: Receive a user task description input by the user, and create a manufacturing task order based on the user task description; the manufacturing task order contains all the manufacturing functions that the task description may require. The user task description is processed using a large language model to generate multiple manufacturing sub-tasks that form a step-by-step task; the manufacturing sub-tasks are used to describe the manufacturing functions included in the manufacturing task order and the manufacturing units that implement the manufacturing functions; Receive user confirmation request for the step-by-step task; when the user accepts the step-by-step task, execute the step of generating a single-step description of the subtask and add the step-by-step task to the training dataset of the large language model; when the user rejects the step-by-step task, return to the step of processing the user task description using the large language model and regenerate the step-by-step task. The single-step description generation step of the manufacturing subtask specifically includes the following steps: A single-step description of each manufacturing subtask is generated using a large language model; each manufacturing subtask includes the manufacturing function in the corresponding manufacturing subtask, the dependencies required by the manufacturing function, and the output content of the manufacturing function. The system receives a confirmation request from the user for the step-by-step description of the manufacturing subtask; when the user accepts the step-by-step description, the system executes the process generation step and adds the step-by-step description of the manufacturing subtask to the training dataset of the large language model; when the user rejects the step-by-step description of the manufacturing subtask, the system returns to the step of generating step-by-step descriptions of each manufacturing subtask using the large language model, and regenerates the step-by-step tasks. The process generation steps specifically include the following steps: Using a large language model, interface relationships and content interaction relationships are established between manufacturing functions associated with multiple manufacturing subtasks based on the single-step description of the manufacturing subtasks; the interface relationships are used to describe the interfaces that enable data interaction between manufacturing functions; the content interaction relationships are used to describe the specific output content of data interaction between manufacturing functions. A complete description of the manufacturing process is generated using a large language model based on the interface relationships and content interaction relationships. Receive user confirmation request for the complete description; when the user accepts the complete description, use the generated complete description as the manufacturing process and add the complete description to the training dataset of the large language model; when the user rejects the complete description, return to the step of generating a complete description of the manufacturing process using the large language model based on the interface relationship and content interaction relationship, and regenerate the complete description.

2. The modular intelligent programming method for flexible manufacturing systems driven by a large language model according to claim 1, characterized in that, Each manufacturing unit is associated with one or more manufacturing subtasks; each manufacturing subtask is associated with one or more manufacturing functions.

3. The modular intelligent programming method for a flexible manufacturing system driven by a large language model according to claim 1, characterized in that, The dependencies required for this manufacturing function refer to the parameters associated with the manufacturing order and the output content associated with other manufacturing functions.

4. The modular intelligent programming method for a flexible manufacturing system driven by a large language model according to claim 1, characterized in that, Before generating a complete description of the manufacturing process using a large language model based on the interface relationships and content interaction relationships, the following steps are also included: The system provides a pre-defined manufacturing process template to the large language model, enabling the model to generate a complete description of the manufacturing process based on the template.

5. The modular intelligent programming method for a flexible manufacturing system driven by a large language model according to claim 1, characterized in that, It also includes the following steps: Encapsulate a finite state machine within each manufacturing subtask; When executing the complete description of the manufacturing process, the execution state of each manufacturing function is obtained through the finite state machine encapsulated in the manufacturing subtask; Based on the execution state obtained from the finite state machine, determine whether there are any errors in the complete description of the current execution.

6. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement a modular intelligent programming method for flexible manufacturing systems driven by a large language model, as described in any one of claims 1-5.

Images