Mass customization-oriented industrial production method, device, equipment and storage medium

Abstract

The embodiment of the application provides an industrial production method, device, equipment and storage medium for mass customization, and relates to the technical field of industrial artificial intelligence. The method verifies the candidate configuration combination of the product to be produced in multiple dimensions, and iteratively corrects the candidate configuration combination until a target configuration combination meeting the verification condition is obtained in the case that the verification fails. The target configuration combination and the corresponding target verification result are sent to a decision unit matched with the target configuration combination, so that the target configuration combination and the target verification result are sent to a production system after receiving the production instruction of the decision unit. The production system produces the product to be produced based on the target configuration combination and the target verification result. The method realizes the automatic correction of the customized production scheme, improves the correction accuracy and generation efficiency of the customized scheme, and thus helps to improve the product qualification rate.

Description

Technical Field

[0001] This application relates to the field of industrial artificial intelligence technology, and in particular to an industrial production method, apparatus, equipment and storage medium for mass customization. Background Technology

[0002] Against the backdrop of the manufacturing industry's transformation towards Massive Personalized Customization (MPC), customer demand for products is becoming increasingly diversified and fragmented.

[0003] Currently, product manufacturing solutions are mainly generated through single-round generative design tools, and the geometric manufacturability or process feasibility of candidate solutions are verified. If the verification fails, manual modification and redesign are carried out by human intervention.

[0004] However, relying on human experience to modify customized production plans makes it difficult to guarantee the efficiency of customized plan generation and product qualification rate. Summary of the Invention

[0005] This application provides an industrial production method, apparatus, equipment, and storage medium for mass customization, which solves the technical problem that existing technologies rely on manual experience to modify customized production plans, making it difficult to guarantee the efficiency of customized plan generation and product qualification rate.

[0006] In a first aspect, embodiments of this application provide an industrial production method for mass customization, comprising:

[0007] Obtain candidate configuration combinations for the products to be manufactured. These candidate configuration combinations are used to characterize the customized production schemes for the products to be manufactured.

[0008] The candidate configuration combinations were validated in production from multiple dimensions to obtain the validation results.

[0009] If the verification result indicates that the verification has failed, the candidate configuration combination is iteratively modified based on the verification result until the target verification result corresponding to the modified target configuration combination meets the verification conditions.

[0010] Send the target configuration combination and target verification results to the decision-making unit that matches the target configuration combination;

[0011] After receiving the production instruction from the decision-making unit, the target configuration combination and target verification results are sent to the production system so that the production system can produce the product to be produced based on the target configuration combination and target verification results.

[0012] In one possible implementation, obtaining candidate configuration combinations for the product to be manufactured includes:

[0013] Obtain the configuration condition parameters of the product to be produced. These configuration condition parameters are generated based on customer requirements. The configuration condition parameters include: social environment parameters, customer requirement parameters, service flow parameters, and failure dimension characteristics.

[0014] Input the configuration condition parameters into the base model to obtain candidate configuration combinations, which include: product specification parameters, embodied execution strategy, social environment parameters, customer demand parameters, service flow parameters, and value parameters.

[0015] In one possible implementation, multi-dimensional production verification is performed on candidate configuration combinations to obtain verification results, including:

[0016] The candidate configuration combinations are input into the digital twin simulation model, and the multi-dimensional production verification of the candidate configuration combinations is performed in parallel by multiple verification modules in the digital twin simulation model to obtain the verification results.

[0017] The multiple verification modules include at least two of the following: geometric manufacturability verification module, process feasibility verification module, embodied executability verification module, customer preference consistency verification module, and social acceptability verification module.

[0018] In one possible implementation, based on the verification results, the candidate configuration combinations are iteratively revised until the target verification result corresponding to the revised target configuration combination satisfies the verification conditions, including:

[0019] Feature extraction is performed on the failure dimension information in the verification results to obtain failure dimension features;

[0020] The failure dimension features are fed back to the base model, and the base model performs inference analysis on the failure dimension features and configuration condition parameters to obtain the corrected candidate configuration combination.

[0021] The revised candidate configuration combination is input into the digital twin simulation model to obtain the first verification result;

[0022] If the first verification result meets the verification conditions, the modified candidate configuration combination is output as the target configuration combination, and the first verification result is output as the target verification result.

[0023] If the first verification result does not meet the verification conditions and the preset termination conditions, the failure dimension features are obtained from the first verification result, and the above candidate configuration combination generation steps are re-executed; the termination conditions include at least: the maximum number of iterations reaches a preset value, and the number of consecutive verification failures in the same dimension reaches a preset threshold.

[0024] In one possible implementation, the method further includes:

[0025] If the preset termination conditions are met, an alarm message is output; the alarm message is generated based on the failure dimension information, the corrected candidate configuration combination, and the first verification result.

[0026] In one possible implementation, customer requirement parameters include customer group affiliation information, and the target configuration combination and target verification results are sent to a decision-making unit that matches the target configuration combination, including:

[0027] Retrieve decision units that match the customer group affiliation information from the preset decision unit configuration table;

[0028] The target configuration combination and target verification results are sent to the decision-making unit.

[0029] In one possible implementation, after obtaining the decision unit matching the customer group affiliation information from a preset decision unit configuration table, the method further includes:

[0030] Obtain the production line control system identifier corresponding to the decision-making unit from the predefined production line control system catalog;

[0031] Send the target configuration combination and target verification results to the production system, including:

[0032] Based on the production line control system identifier, target configuration combination, and target verification results, a target configuration table is generated and sent to the production system.

[0033] Among them, the target configuration combination and the target verification result are used to generate production work orders, and the production line control system identifier is used to instruct the production system to send the production work order to the corresponding production line control system.

[0034] Secondly, embodiments of this application provide an industrial production apparatus, comprising:

[0035] The acquisition module is used to acquire candidate configuration combinations for the products to be manufactured. These candidate configuration combinations represent customized production plans for the products to be manufactured.

[0036] The verification module is used to perform multi-dimensional production verification on candidate configuration combinations and obtain verification results.

[0037] The correction module is used to iteratively correct the candidate configuration combination based on the verification result when the verification result indicates that the verification has failed, until the target verification result corresponding to the corrected target configuration combination meets the verification conditions.

[0038] The sending module is used to send the target configuration combination and the target verification result to the decision unit that matches the target configuration combination;

[0039] The sending module is also used to send the target configuration combination and target verification results to the production system after receiving the production instructions from the decision unit, so that the production system can produce the products to be produced based on the target configuration combination and target verification results.

[0040] In one possible implementation, the above-mentioned device further includes: an input module;

[0041] The acquisition module is also used to acquire the configuration condition parameters of the product to be produced. These configuration condition parameters are generated based on customer requirements and include: social environment parameters, customer requirement parameters, service flow parameters, and failure dimension characteristics.

[0042] The input module is used to input configuration condition parameters into the base model to obtain candidate configuration combinations, which include: product specification parameters, embodied execution strategies, social environment parameters, customer demand parameters, service flow parameters, and value parameters.

[0043] In one possible implementation, the input module is also used to input candidate configuration combinations into the digital twin simulation model, and to obtain verification results by performing multi-dimensional production verification of the candidate configuration combinations in parallel through multiple verification modules in the digital twin simulation model.

[0044] The multiple verification modules include at least two of the following: geometric manufacturability verification module, process feasibility verification module, embodied executability verification module, customer preference consistency verification module, and social acceptability verification module.

[0045] In one possible implementation, the above-mentioned device further includes: a feature extraction module and an output module;

[0046] The feature extraction module is used to extract features from the failure dimension information in the verification results to obtain failure dimension features;

[0047] The sending module is also used to feed back the failure dimension features to the base model, and the base model performs reasoning analysis on the failure dimension features and configuration condition parameters to obtain the corrected candidate configuration combination.

[0048] The input module is also used to input the modified candidate configuration combination into the digital twin simulation model to obtain the first verification result;

[0049] The output module is used to output the corrected candidate configuration combination as the target configuration combination when the first verification result meets the verification conditions, and to output the first verification result as the target verification result.

[0050] The acquisition module is also used to acquire the failure dimension features from the first verification result and re-execute the above-mentioned candidate configuration combination generation steps when the first verification result does not meet the verification conditions and the preset termination conditions; the termination conditions include at least: the maximum number of iterations reaches a preset value, and the number of consecutive verification failures in the same dimension reaches a preset threshold.

[0051] In one possible implementation, the output module is further configured to output an alarm message when a preset termination condition is met; the alarm message is generated based on failure dimension information, a corrected candidate configuration combination, and a first verification result.

[0052] In one possible implementation, the customer demand parameters include customer group affiliation information. The acquisition module is also used to obtain decision units that match the customer group affiliation information from a preset decision unit configuration table.

[0053] The sending module is also used to send the target configuration combination and target verification results to the decision unit.

[0054] In one possible implementation, the acquisition module is also used to acquire the production line control system identifier corresponding to the decision unit from a predefined production line control system catalog.

[0055] The sending module is also used to generate a target configuration table based on the production line control system identifier, target configuration combination, and target verification results, and send the target configuration table to the production system;

[0056] Among them, the target configuration combination and the target verification result are used to generate production work orders, and the production line control system identifier is used to instruct the production system to send the production work order to the corresponding production line control system.

[0057] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0058] The aforementioned memory stores instructions executed by the computer;

[0059] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0060] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0061] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0062] The industrial production method, apparatus, equipment, and storage medium for mass customization provided in this application obtains candidate configuration combinations for the product to be produced, performs multi-dimensional production verification on these combinations, obtains verification results, and then iteratively corrects the candidate configuration combinations based on the verification results when verification fails, until the target verification result corresponding to the corrected target configuration combination meets the verification conditions. The target configuration combination and the target verification result are then sent to a decision unit matching the target configuration combination. Upon receiving the production instruction from the decision unit, the target configuration combination and the target verification result are sent to the production system, enabling the production system to produce the product to be produced based on the target configuration combination and the target verification result. This method achieves automated correction of customized production schemes by performing multi-dimensional simulation verification on candidate configuration combinations and iteratively correcting them when verification fails, improving the accuracy and efficiency of customized scheme correction and generation, thereby contributing to a higher product qualification rate. Attached Figure Description

[0063] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0064] Figure 1 A flowchart illustrating an industrial production method for mass customization provided in this application embodiment. Figure 1 ;

[0065] Figure 2 A schematic diagram of the architecture of an industrial production system provided in this application embodiment;

[0066] Figure 3 A flowchart illustrating an industrial production method for mass customization provided in this application embodiment. Figure 2 ;

[0067] Figure 4 A schematic diagram of the structure of an industrial production device for mass customization provided in this application;

[0068] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in this application.

[0069] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0071] 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, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0072] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0073] First, let me explain the terms used in this application:

[0074] Trusted Execution Environment (TEE): A secure execution container based on hardware isolation that performs computations on sensitive inputs (customer body measurements, preference data, etc.) in an encrypted state.

[0075] In the field of mass personalized customization manufacturing, product solution formation and production execution are no longer limited to traditional standardized mass production, but are increasingly being developed for on-demand production scenarios such as clothing, home appliances, auto parts, and furniture components.

[0076] In such scenarios, customers typically raise differentiated requirements regarding dimensions such as size, structure, material, functional combination, and delivery cycle. The manufacturing system needs to quickly generate a configuration plan for the product to be produced after receiving the order, and complete a comprehensive judgment on manufacturability and feasibility before entering production.

[0077] To support this process, existing systems typically include a customer request acquisition terminal, an order management platform, a configuration rule engine or generative design module, a simulation verification module, and a production system connected to the shop floor manufacturing execution system. After customers submit their requirements through applications, web pages, or enterprise interfaces, the configuration rule engine or generative design module transforms the relevant requirement parameters into configuration combinations that can be used for design and production. Then, the simulation verification module verifies the geometric manufacturability or process feasibility of candidate solutions. If the verification fails, manual intervention is required for modification and redesign.

[0078] However, relying on human experience to modify customized production plans makes it difficult to guarantee the efficiency of customized plan generation and product qualification rate.

[0079] To address the aforementioned technical issues, this application provides an industrial production method for mass customization. This method involves multi-dimensional simulation verification of candidate configuration combinations for the product to be manufactured. If the verification fails, the method iteratively corrects the candidate configuration combination until a target configuration combination that meets the verification conditions is obtained, thus achieving automated correction of the customized production plan. Here, the configuration combination represents the customized production plan for the product to be manufactured. This method improves the accuracy and efficiency of customized plan correction and generation, thereby contributing to a higher product qualification rate.

[0080] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0081] Figure 1 A flowchart illustrating an industrial production method for mass customization provided in this application embodiment. Figure 1 The implementing entity of this application can be a central processing unit, cloud server, or other electronic device deployed in an industrial production system, or a processing architecture composed of the aforementioned devices working together. Figure 1 As shown, the method includes:

[0082] S101: Obtain candidate configuration combinations for the product to be manufactured.

[0083] The candidate configuration combination is used to characterize the customized production plan for the product to be produced. This candidate configuration combination carries the initial production plan for the product to be produced and is the core object for subsequent multi-dimensional production verification, iterative correction, and transmission to the decision-making unit and production system. The product to be produced can be, for example, clothing, home appliances, automotive parts, furniture components, or other on-demand customized products.

[0084] Specifically, the system obtains the customer's customized production requirements, analyzes these requirements, and combines them with a pre-built privacy database, carbon constraint database, etc., to generate multiple production configuration parameters. These parameters are then input into a pre-built inference model to obtain candidate configuration combinations.

[0085] In one possible implementation, Figure 2 This is a schematic diagram of the architecture of an industrial production system provided in an embodiment of this application. Figure 2 As shown, the industrial production system includes a client and a production server. The client includes a customer demand acquisition terminal 201 and an edge gateway 202 as client devices, used to collect customer demands, preprocess the customer demands, and send the preprocessed demand information to the production server. The production server includes a basic model inference cluster 203, a multi-dimensional digital twin simulation cluster 204, an MES (Manufacturing Execution System) interface, and a micro-enterprise unit production line control system 205.

[0086] The customer demand acquisition terminal 201 includes, but is not limited to, mobile applications, web front-ends, enterprise resource planning interfaces, customer relationship management interfaces, and information acquisition terminals, used to receive customized production requirements from customers. For example, the information acquisition terminal could be a scanner used for body measurement in the apparel industry.

[0087] For example, in the apparel scenario, customized production requirements include body measurement vectors, pattern preferences, and fabric requirements; in the home appliance scenario, customized production requirements include installation space dimensions, power requirements, and appearance style information; and in the automotive parts scenario, customized production requirements include compatible vehicle models, performance levels, and assembly interface parameters.

[0088] Edge gateway 202 is communicatively connected to customer requirement acquisition terminal 201. Edge gateway 202 includes a Trusted Execution Environment (TEE) and an authentication module. After receiving customer requirements, customer requirement acquisition terminal 201 transmits the requirements to edge gateway 202, where they are preprocessed to output structured configuration data. This preprocessing includes, but is not limited to, standardization and anonymization.

[0089] For example, after receiving a customized production request, the edge gateway 102 extracts the customer identity identifier from the request and then sends a verification request to the cloud-based identity management system through the identity authentication module to confirm the legitimacy, validity, and access permissions of the customer's identity. If the verification is successful, the identity authentication module further queries the privacy level, historical interaction permission scope, and operation permissions corresponding to the customer identity identifier, and sends the authentication result and the customer identity identifier back to the edge gateway.

[0090] In addition, the edge gateway generates carbon constraint limits based on product categories (such as refrigerators) and customer environmental preferences extracted from multimodal inputs, and generates ethical constraint labels according to preset ethical constraint rules.

[0091] Based on privacy levels, carbon constraint caps, and ethical constraint labels, a social and environmental parameter Ω is constructed.

[0092] Based on the parameters in the customized production requirements, and / or, the natural language requirements in the customized production requirements are parsed to obtain the customer requirement parameter C. Customized production requirements may include, for example, customer identity, body measurement data, preference data, and historical record summaries. Body measurement data may, for example, come from the customer requirement collection terminal 101, and preference data may be obtained from edge cache and / or through cloud collaboration. Furthermore, the historical record summary is stored and provided by the edge device by compressing past interaction records into feature vector form.

[0093] In addition, semantic analysis is performed on the parameters in the customized production requirements, and / or on the natural language requirements in the customized production requirements, combined with a rule engine or a lightweight classification model, to obtain service flow parameters S such as service flow type, delivery timeliness, and after-sales expectations.

[0094] Thus, we obtain the social environment parameter Ω, the customer demand parameter C, and the service flow parameter S. Based on the social environment parameter Ω, the customer demand parameter C, and the service flow parameter S, we construct configuration data <Ω, C, S>.

[0095] The structured configuration data is sent to the basic model inference cluster 203 for inference analysis, and the candidate configuration combinations are output to the multi-dimensional digital twin simulation cluster 204. The multi-dimensional digital twin simulation cluster 204 is deployed with a digital twin simulation model, which includes multiple verification modules. Each verification module is used to perform simulation verification of the candidate configuration combinations in different production dimensions.

[0096] Among them, the basic model inference cluster 203 carries large language or VLA (Vision-Language-Action) models. During the inference process, customer requirement parameters are first standardized, converting inputs from different sources, units, and formats into unified fields, and then mapped to corresponding product family templates based on product type. Afterwards, the basic model generates one or more candidate configuration combinations based on template constraints, historical order data, product configuration rules, and current manufacturing resource constraints.

[0097] In another possible implementation, candidate configuration combinations can be formed from manually input initial schemes, reused historical candidate schemes, results retrieved from a preset configuration library, or existing standard schemes adjusted parametrically. For example, designers input an initial structural scheme based on customer requirements in the configuration interface, and the production server automatically completes the process and material attributes based on built-in process rules and a material database. Furthermore, production schemes with a similarity exceeding a preset threshold to current customer requirements can be retrieved from a historical case library. Common configurations from these schemes can be migrated to the current order and rewritten to differentiate them based on new requirements. The similarity can be calculated comprehensively based on dimensional differences, functional overlap rates, material compatibility, and delivery time proximity. Based on this approach, even when facing long-tail requirements, an initial configuration input that can be subsequently verified can be formed, without relying entirely on pre-enumerated schemes.

[0098] To ensure that candidate configuration combinations can be stably processed in subsequent steps, this step can also add order unique identifier, version number, generation timestamp, source identifier, and data integrity verification fields to the candidate configuration combinations and store them in the configuration management database. If a customer request is found to be missing key parameters, the parameter completion process is triggered.

[0099] For example, the parameter completion process could involve sending a completion request to the client terminal, inferring default values ​​from historical customer profiles, or generating temporary configurations using default boundary values ​​based on product categories. Subsequently, risk items are further identified during the verification phase. This method allows candidate configuration combinations to not only meet clients' customized production needs but also to enter subsequent multi-dimensional production verification in a unified format, thereby improving solution execution efficiency.

[0100] S102: Perform multi-dimensional production verification on candidate configuration combinations to obtain verification results.

[0101] Multi-dimensional production validation is used to comprehensively examine candidate configuration combinations to determine whether they meet production conditions. The validation results characterize the pass or fail status of candidate configuration combinations after multi-dimensional production validation. These results may include, for example, an overall pass indicator, the pass status corresponding to each validation dimension, failure dimension information, failure reason code, risk level, quantitative score, and suggested correction directions.

[0102] In customized industrial production, factors that render a solution unfeasible often stem from multiple aspects, including structure, process, resources, delivery, and customer preferences. Therefore, inputting candidate configuration combinations into a pre-built digital twin simulation model, which is constructed based on a multi-source constraint environment, and simulating and verifying the candidate configuration combinations under this environment, helps to identify hidden conflicts before production and generate fine-grained feedback that can be used for correction.

[0103] Understandably, the digital twin simulation model constructs a multi-source constraint environment through multiple verification modules. Each verification module corresponds to a different production verification dimension, such as geometric manufacturability, process feasibility, and embodied executability.

[0104] Optionally, after receiving the candidate configuration combination, the digital twin simulation model determines the multiple verification modules to be activated based on the product type, and then inputs the candidate configuration combination into the multiple verification modules for simulation verification, obtaining multiple sub-verification results. Subsequently, the verification result of the candidate configuration combination is constructed based on the multiple sub-verification results.

[0105] For example, the correspondence between product types and verification dimensions is maintained in advance. Based on the product types parsed from the candidate configuration combinations, multiple dimensions to be verified are determined from this correspondence, and the verification module corresponding to each dimension to be verified is determined.

[0106] For example, customized clothing products require verification of geometric constraints, fabric and process matching, cutting path, and delivery timeliness; home appliances require verification of structural assembly interference, electrical safety clearance, equipment processing range, and logistics packaging compatibility; and automotive parts require verification of interface size consistency, processing accuracy attainability, assembly path, and production cycle time.

[0107] In one possible embodiment, such as Figure 2 As shown, the candidate configuration combinations are input into the multidimensional digital twin simulation cluster 204 to obtain the verification results. The digital twin simulation model deployed in the multidimensional digital twin simulation cluster 204 may include, for example, a geometric analysis module, a process simulation module, an equipment capacity matching module, a production cycle time assessment module, an assembly accessibility analysis module, a logistics and scheduling impact assessment module, and a customer preference consistency analysis module.

[0108] S103: If the verification result indicates that the verification has failed, based on the verification result, iteratively correct the candidate configuration combination until the target verification result corresponding to the corrected target configuration combination meets the verification conditions.

[0109] The verification results include, but are not limited to, verification identifiers and simulation results for each production dimension. The verification identifiers include a first identifier and a second identifier. The first identifier indicates that the verification has passed, and the second identifier indicates that the verification has failed.

[0110] If the verification fails, the verification result will also include verification failure information, which includes, but is not limited to, the verification failure reason entry, the verification failure dimension, and the verification score of the failure dimension.

[0111] Specifically, if the verification result indicates that the verification failed, the verification failure information is parsed and features are extracted to obtain the failure dimension features. Based on the failure dimension features and the candidate configuration combination, the above step S101 is re-executed to correct the candidate configuration combination and obtain the corrected candidate configuration combination.

[0112] For example, the failure dimension features are back-injected into the base model in step S101 above. The base model then re-executes the inference analysis operation using the failure dimension features and candidate configuration combinations as inference conditions, and outputs the corrected candidate configuration combinations.

[0113] Furthermore, after obtaining the modified candidate configuration combination, the above step S102 is executed to perform multi-dimensional simulation verification on the modified candidate configuration combination, obtain the verification result, and if the verification result meets the verification conditions, the modified candidate configuration combination is output as the target configuration combination, and the verification result is output as the target verification result.

[0114] The verification conditions include at least one of the following: the verification result indicates that the verification is passed; the preset core dimension is verified and the verification score of the non-core dimension is higher than the preset score; the number of iterations and corrections reaches the preset maximum number of iterations.

[0115] Understandably, different types of products require different verification conditions. For example, for safety-sensitive products, all four core dimensions—structure, safety, process, and equipment capability—may be met; for time-sensitive products, a production scheduling feasibility score higher than 0.85 may be required. This method, by setting different verification conditions, takes into account the quality, cost, and efficiency goals under different industrial scenarios.

[0116] Optionally, iterative correction can follow a hierarchical strategy of parameter adjustment, local reconstruction, and global regeneration. When the failure involves only a few consecutive parameters exceeding limits, parameter adjustment can be performed, such as adjusting dimensions, tolerances, material thickness, process sequence, or equipment assignment identifiers. When the failure involves conflicts between modules, local reconstruction is performed, such as replacing local components, changing connection methods, or rearranging assembly paths. When the failure involves multi-dimensional systemic conflicts, global regeneration is performed, that is, retaining customer requirement parameters and key business boundaries, and re-invoking the basic model to generate new candidate configuration combinations. Through this shallow-to-deep correction mechanism, the computational waste caused by completely rebuilding the solution after each failure can be avoided, while improving the success rate of correction in complex scenarios.

[0117] This step can be, for example, by... Figure 2 The basic model inference cluster 203 is implemented.

[0118] This step is crucial for the convergence of this solution. Its core lies in not treating verification failure as the end point of the process, but rather transforming the failure information contained in the verification results into a correction driver. This allows candidate configuration combinations to continuously evolve within an automated closed loop until a target configuration combination that meets manufacturing conditions and business constraints is obtained. By transforming failed verification results into closed-loop feedback for configuration generation, this step achieves continuous correction and convergence of candidate configuration combinations. This overcomes the problems of fragmented generation and verification in existing technologies, and the reliance on repeated manual modifications after failures, thereby improving the initial success rate of the target configuration combination.

[0119] S104: Send the target configuration combination and target verification results to the decision unit that matches the target configuration combination.

[0120] The decision-making unit is used to receive the target configuration combination and target verification results, and output the execution entity of the production instructions accordingly.

[0121] Specifically, after obtaining the target configuration combination and target verification results, customer demand parameters can be extracted from the target configuration combination. These parameters include, but are not limited to, customer group affiliation, product category, geographic attributes, service contract attributes, delivery level, green manufacturing level requirements, and cost sensitivity. Subsequently, based on the extracted customer demand parameters, decision units that match these parameters are determined from a pre-defined decision unit configuration table. This table includes the business domain, applicable product type, service area, resource coverage, priority rules, and load status of each decision unit. These decision units can be, for example, a factory-level scheduling center, a business unit-level order distribution node, a regional manufacturing service unit, a micro-enterprise unit, or a service execution node corresponding to a specific customer group.

[0122] For example, when the customer's demand parameters correspond to high-end customized clothing, the delivery location is in the East China region, and the customer has a VIP (Very Important Person) service contract, the customer is matched to the corresponding high-end customized regional decision-making unit; when the customer's demand parameters correspond to partial customization of standard home appliances and emphasize rapid delivery, the customer is matched to a regional decision-making unit that is closer to the customer and has the corresponding assembly capabilities.

[0123] After the decision-making unit is determined, the target configuration combination and target verification results are constructed into a structured message, and this structured message is sent to the aforementioned decision-making unit. The structured message can be encapsulated using JSON (JavaScript Object Notation), XML (Extensible Markup Language), binary messages, or data formats compatible with industry protocols.

[0124] The structured message mentioned above may include fields such as order identifier, customer requirement parameter summary, target configuration combination details, target verification result, version number, timestamp, priority, and security check code. This structured message can be sent to the matching decision-making unit via message queue, service interface call, enterprise service bus, or industrial internet communication link.

[0125] After receiving the structured message, the decision-making unit verifies message integrity and version consistency. If the verification passes, the structured message is registered as a task pending decision. Subsequently, relevant decision-makers in the decision-making unit confirm whether to execute the task. After receiving confirmation from the relevant decision-makers, the decision-making unit generates a production instruction based on the target configuration and sends the production instruction to the production server.

[0126] For example, after generating a task to be decided, the decision-making unit displays the task in the task list and sends a prompt message to the relevant decision-makers. This prompt message includes, but is not limited to, the product name and production deadline in the target configuration combination. The prompt message is used to remind the relevant decision-makers to decide whether to execute the task as soon as possible within the production deadline.

[0127] Relevant decision-makers click on a task from the list of tasks to be decided, entering the decision-making interface. This interface includes a target configuration combination, a confirmation component, a rejection component, and a remarks component. In response to the decision-maker's click on the confirmation component, a production instruction is sent to the production server.

[0128] Understandably, if multiple candidate decision-making units meet the matching conditions, a secondary screening is conducted by combining real-time resource status, historical performance, transportation timeliness, and work-in-process load to determine the final receiving entity.

[0129] In one possible embodiment, the production instruction is generated by combining the resource status of the production line control system under the decision-making unit, and / or the service flow information in the target configuration combination. The resource status may include, for example, equipment availability, workstation load, personnel skill coverage, material inventory, energy window, and maintenance plan; the service flow information may include, for example, order priority, customer service level, order consolidation opportunities, logistics coordination status, and cross-plant collaboration constraints.

[0130] S105: After receiving the production instruction from the decision-making unit, the target configuration combination and target verification results are sent to the production system so that the production system can produce the product to be produced based on the target configuration combination and target verification results.

[0131] The production system can be, for example, a MES system.

[0132] Specifically, after receiving the production instruction returned by the decision-making unit, the production server performs a consistency verification to confirm whether the version number of the target configuration combination in the production instruction is consistent with the version number of the current target configuration combination. If they are consistent, the target configuration combination and the target verification result are associated and encapsulated with the production instruction and sent to the production system. The production system then generates a production work order based on the target configuration combination and the target verification result, and sends the production work order to the production line control system under the decision-making unit in step S104 above.

[0133] For example, a work order is written through the Manufacturing Execution System (MES) interface and then sent to the specific production line control system equipment control unit via the industrial control gateway. The production line control system parses the production work order to obtain data such as the process sequence, production parameters corresponding to each process, and inspection requirements. The parsed data is then sent to the process controllers at each workstation, which in turn issues process execution instructions to produce the products to be manufactured.

[0134] like Figure 2 As shown, the MES (Manufacturing Execution System) interface and the micro-enterprise unit production line control system 205 are used to receive target configuration combinations and target verification results, generate production work orders, and execute specific production operations.

[0135] In one possible embodiment, after receiving the target configuration combination and target verification results, the production system first generates the corresponding production task instance and maps the configuration parameters to executable instructions for the equipment. For example, dimensional and structural parameters are mapped to CNC machining programs, cutting paths, or welding trajectories; process parameters are mapped to temperature, pressure, speed, stitch spacing, coating thickness, or assembly torque settings; and quality requirements are mapped to online inspection thresholds and sampling rules. The production system then initiates the manufacturing process according to the time window and resource binding relationships determined by the production instructions, sequentially completing processes such as processing, assembly, inspection, and packaging, and recording actual execution data at key nodes. The actual execution data may include equipment operating status, measured values ​​of process parameters, process completion time, quality inspection results, and abnormal alarm information.

[0136] In another possible embodiment, the production system can also send the actual execution data back to the industrial production server after production is completed or during the execution process, for subsequent iterative optimization of the basic model, multidimensional data twin model, etc.

[0137] This step, by sending the target configuration combination and target verification results to the production system after receiving the production instruction, ensures that the execution side obtains the final production data that has undergone decision matching and resource organization, thus avoiding invalid issuance and duplicate scheduling.

[0138] The industrial production method for mass customization provided in this application involves acquiring candidate configuration combinations for the product to be manufactured, performing multi-dimensional production verification on these combinations, obtaining verification results, and then iteratively correcting the candidate configuration combinations based on the verification results when verification fails, until the target configuration combination obtained after correction meets the verification conditions. The target configuration combination and the target verification results are then sent to a decision unit matching the target configuration combination. Upon receiving the production instruction from the decision unit, the decision unit sends the target configuration combination and the target verification results to the production system, enabling the production system to produce the product to be manufactured based on the target configuration combination and the target verification results. Here, the configuration combination represents the customized production scheme for the product to be manufactured. This method, by performing multi-dimensional simulation verification on candidate configuration combinations and iteratively correcting them when verification fails until a target configuration combination that meets the verification conditions is obtained, achieves automated correction of customized production schemes, improves the accuracy and efficiency of customized scheme correction, and thus helps to improve the product qualification rate.

[0139] Figure 3 A flowchart illustrating an industrial production method for mass customization provided in this application embodiment. Figure 2 ,like Figure 3 As shown, in this embodiment... Figure 1-2 Based on the examples, a possible industrial production method for mass customization is described in detail, the method comprising:

[0140] S301: Obtain the configuration parameters of the product to be manufactured.

[0141] The system receives configuration condition parameters sent by the edge gateway. These configuration condition parameters are generated based on customer requirements and include social environment parameters, customer requirement parameters, service flow parameters, and failure dimension features. They are an input set used to constrain the direction of basic model generation.

[0142] Social environment parameters are used to characterize external compliance conditions, environmental requirements, or carbon emission restrictions related to the product to be manufactured; customer demand parameters are used to characterize customer requirements for size, function, material, appearance, or delivery cycle; service flow parameters are used to characterize conditions such as order flow, service type, and delivery time.

[0143] Failure dimension features are used to characterize the dimensions that failed in the preceding validation and the reasons for their failure, so that the base model can correct the generation direction accordingly.

[0144] It should be noted that the failure dimension feature is a zero vector in the first round of inference, meaning that the configuration condition parameters in the first round of inference are <Ω0, C0, S0, 0>. During the k-th iteration to correct the configuration combination, the configuration condition parameters become <Ω0, C0, S0, ΔF_{k-1}>. Here, ΔF_{k-1} represents the failure dimension feature corresponding to the candidate configuration combination obtained in the (k-1)-th iteration of inference. ΔF_{k-1} can be, for example, a feature vector.

[0145] S302: Input the configuration condition parameters into the basic model to obtain candidate configuration combinations.

[0146] The candidate configuration combinations include product specification parameters, embodied execution strategies, social environment parameters, customer demand parameters, service flow parameters, and value parameters.

[0147] The candidate configuration condition parameters are input into the base model, which then performs inference analysis on these parameters to obtain the inference result. This inference result includes combinations of candidate configurations.

[0148] In some embodiments, the inference results described above may also include confidence scores for candidate configuration conditions.

[0149] The base model can be a generative network or a multimodal inference model trained for customized manufacturing scenarios. In some embodiments, the base model includes an encoding network, a feature extraction network, and a decoding network. The encoding network maps configuration condition parameters to latent space condition vectors. The feature extraction network includes a learnable adjacency matrix that characterizes the spatial relationships between social environment parameters, customer demand parameters, service flow parameters, and failure dimension features.

[0150] The feature extraction network performs graph convolution operations on the latent space conditional vectors based on the adjacency matrix, resulting in an aggregated feature matrix that integrates the spatial influence relationships across various dimensions. For example, multiplying the adjacency matrix by the latent space vectors yields an aggregated feature matrix A that encodes all the deep features required to generate product specification parameters, embodied execution strategies, and value parameters.

[0151] The decoding network includes at least a first output head, a second output head, and a third output head. The decoding network performs branch decoding on the above-mentioned aggregated feature matrix, mapping it to product specification parameters via the first output head, to an embodied execution strategy via the second output head, and to a value parameter via the third output head.

[0152] Understandably, the training set of the base model includes multiple historical configuration condition parameters, each of which carries an output annotation set, which includes product specification parameters, embodied execution strategies, and value parameters.

[0153] S303: Input the candidate configuration combination into the digital twin simulation model, and perform multi-dimensional production verification of the candidate configuration combination in parallel through multiple verification modules in the digital twin simulation model to obtain the verification result.

[0154] The multiple verification modules include at least two of the following: geometric manufacturability verification module, process feasibility verification module, embodied executability verification module, customer preference consistency verification module, and social acceptability verification module.

[0155] Digital twin simulation models are used to construct virtual verification environments corresponding to actual production environments, capable of receiving candidate configuration combinations and performing parallel simulations. The geometric manufacturability verification module verifies in the virtual environment whether the product structure, assembly clearances, interference relationships, and key contours meet manufacturing requirements; the process feasibility verification module determines whether candidate configuration combinations can be implemented by existing processes based on equipment capabilities, process windows, and material constraints; the embodied executability verification module simulates the motion paths of robots, fixtures, or production line actuators to assess whether grasping, alignment, and assembly actions are reachable and achievable; the customer preference consistency verification module compares the matching degree of candidate configuration combinations with the style, function, and attributes corresponding to customer requirement parameters; and the social acceptability verification module assesses the compliance of candidate configuration combinations based on environmental protection, carbon emissions, privacy, or ethical constraints.

[0156] Understandably, the various verification dimensions can be executed in parallel or sequentially according to a preset order. When a scenario has high efficiency requirements, a parallel mode can be used to shorten the total latency. When a scenario has high cost control requirements, a low-cost, rapid screening can be performed first, followed by a high-precision simulation.

[0157] In practical implementation, a multidimensional digital twin simulation cluster can consist of a simulation server, a task scheduling unit, and multiple parallel verification engines. The digital twin simulation model is deployed within this cluster. Upon receiving candidate configuration combinations, the digital twin simulation model decomposes these combinations into geometric parameters, process parameters, execution parameters, and preference parameters, which are then input into the corresponding verification modules within the digital twin simulation model. Each verification module calls a verification engine to independently calculate and output local verification conclusions based on a preset model and constraint library. Subsequently, the multiple local verification conclusions are summarized to obtain the final verification result. Furthermore, the task scheduling unit is used for task distribution, scheduling, and resource allocation.

[0158] The aforementioned local validation conclusions include, but are not limited to, Boolean values ​​and validation scores. In some embodiments, the results of each local validation are summarized to form a Boolean pass matrix and a comprehensive score vector, and the validation result is constructed based on the Boolean pass matrix and the comprehensive score vector. The Boolean pass matrix can be, for example, (1, 1, 1, 1, 1), used to represent that all five dimensions of production validation have passed.

[0159] It should be noted that if any dimension fails validation, the validation result will also include information about the failed dimension. This information may include, for example, a description of the constraint conflict, the location of the out-of-limit parameter, the deviation amount, and the identifier of the conflicting object.

[0160] For example, in a clothing industry scenario where production is based on body measurements, customers submit body scan data and style preferences through an application. The body scan data can be, for example, a 26-dimensional anthropometric vector. After the edge network acquires the body scan data and style preferences, it analyzes and encodes them to obtain configuration condition parameters <Ω0, C0, S0>, where Ω0 is a social environment parameter, C0 is a customer demand parameter, and S0 is a service flow parameter.

[0161] Furthermore, the edge gateway sends the configuration condition parameters <Ω0, C0, S0> to the base model. Through inference analysis by the base model, candidate configuration combinations <Ω0, C0, P_k, S0, I_k, V_k, ΔF_{k-1}> are obtained. Here, P_k represents the candidate clothing pattern obtained in the Kth round of inference, I_k represents the sewing process obtained in the Kth round of inference, V_k represents the value parameter obtained in the Kth round of inference, and ΔF_{k-1} is the failure dimension feature of the (k-1)th round. If k=1, then ΔF_{k-1}=0.

[0162] The above candidate configuration combination <Ω0, C0, P_k, S0, I_k, V_k, ΔF_{k-1}> is input into the digital twin simulation model, which includes a geometric manufacturability verification module D1, a process feasibility verification module D2, an embodied executability verification module D3, a customer preference consistency verification module D4, and a social acceptability verification module D5.

[0163] The simulation verification content of the geometric manufacturability verification module D1 includes, but is not limited to, the consistency of the total seam length of multi-faceted garments. The simulation verification content of the process feasibility verification module D2 includes, but is not limited to, the ability of the selected fabric to be processed on current sewing equipment. The simulation verification content of the embodied executability verification module D3 includes, but is not limited to, simulating the gripping and / or alignment of irregularly shaped facets by a robotic arm. The simulation verification content of the customer preference consistency verification module D4 includes, but is not limited to, the matching between the preference vector generated based on candidate configuration combinations and the customer preference vector. The simulation verification content of the social acceptability verification module D5 includes, but is not limited to, carbon footprint accounting; for example, if recycled fabrics are used, the verification score output by module D5 will be higher.

[0164] S304: If the verification result indicates that the verification failed, extract features from the failure dimension information in the verification result to obtain the failure dimension features.

[0165] The failure dimension information is as explained in step S303 above, and will not be repeated here.

[0166] Specifically, after obtaining the verification results corresponding to the candidate configuration combinations, the dimensions that failed verification are first analyzed, the corresponding failure dimension information is extracted, and it is encoded as failure dimension features. These failure dimension features include, but are not limited to, constraint conflict descriptions, verification scores of failure dimensions, damaged geometric regions, and embodied failure joints. Among them, constraint conflict descriptions are, for example, "Interference detection failed: Part A and Part B collide at coordinates (10, 20, 30)".

[0167] S305: Feed the failure dimension features back to the base model. The base model then performs reasoning analysis on the failure dimension features and configuration condition parameters to obtain the corrected candidate configuration combination.

[0168] After receiving the failure dimension features, the base model updates the failure dimension features to the configuration condition parameters to obtain the first configuration condition parameters for this round of inference. Then, it performs inference analysis on the first configuration condition parameters to obtain the corrected candidate configuration combination.

[0169] S306: Input the modified candidate configuration combination into the digital twin simulation model to obtain the first verification result.

[0170] This step is similar to the explanation of step S303 above, and will not be repeated here.

[0171] S307: If the first verification result meets the verification conditions, the modified candidate configuration combination is output as the target configuration combination, and the first verification result is output as the target verification result.

[0172] When the first verification result meets the preset verification conditions, the modified candidate configuration combination is determined as the target configuration combination, and the corresponding first verification result is determined as the target verification result, which serves as the basis for subsequent decision-making and production execution.

[0173] S308: If the first verification result does not meet the verification conditions and the preset termination conditions, obtain the failure dimension features from the first verification result and execute the above step S302.

[0174] The preset termination condition is used to control the number of iterations or the failure threshold to avoid infinite loops caused by repeated corrections. This termination condition includes at least the maximum number of iterations reaching a preset value, and the number of consecutive verification failures in the same dimension reaching a preset threshold.

[0175] Understandably, the maximum number of iterations is jointly determined by production line cycle time constraints and customer waiting tolerance. The tighter the cycle time and the lower the waiting tolerance, the smaller the maximum number of iterations. This maximum number of iterations can be, for example, between 3 and 10. Furthermore, the aforementioned preset threshold prematurely terminates the current synthesis round when a single dimension becomes a bottleneck. This preset threshold can be, for example, between 2 and 4. In practical applications, this threshold can also be adjusted according to different product lines, and this application embodiment does not limit this.

[0176] If the first verification result still does not meet the verification conditions, it is further determined whether the preset termination condition is triggered. If the determination result indicates that the preset termination condition is not triggered, the failure dimension feature is extracted from the first verification result, and the basic model is returned to perform the above step S302 to continue the iterative correction operation.

[0177] This method uses a closed-loop feedback mechanism to enable the basic model to continuously absorb verification failure information and converge round by round until it outputs a target configuration combination that can be used for production.

[0178] S309: Output alarm information when the preset termination conditions are met.

[0179] The alarm information is generated based on the failure dimension information, the corrected candidate configuration combination, and the first verification result. It is used to send a prompt to the production server or manual maintenance terminal when the iterative correction process cannot continue to converge, so that maintenance personnel can quickly locate the constraint that caused the verification failure.

[0180] Specifically, after each verification, the failed dimensions are recorded and the number of iterations is accumulated. When the accumulated number reaches the preset maximum number of iterations, the termination condition is determined to have been met. In addition, the number of times each dimension fails verification is counted separately, and the termination condition is determined to have been met when the number of consecutive verification failures reaches a preset threshold.

[0181] Understandably, alarm information can be output to the decision-making unit, manual review terminal, or operation and maintenance platform through message queues, interface pop-ups, log records, or interface pushes to indicate that there is a persistent constraint conflict in the current configuration scheme.

[0182] S310: Obtain the decision unit that matches the customer group affiliation information from the preset decision unit configuration table; send the target configuration combination and target verification results to the decision unit.

[0183] Customer requirement parameters include customer group affiliation information. This information indicates the customer's geographic location, market segment, or service contract tag, and serves as the search key for selecting a decision unit. A pre-defined decision unit configuration table stores the mapping relationship between each decision unit and the customer group affiliation information.

[0184] Specifically, after generating the target configuration combination and target verification results, customer group affiliation information is parsed from the customer requirement parameters. This customer group affiliation information is then used as a query condition to determine the decision unit matching the customer group affiliation information from the decision unit configuration table. Subsequently, the target configuration combination and target verification results are encapsulated into a unified data message, which is then sent to the decision unit. This data message includes, but is not limited to, configuration version number, verification dimension summary, failure dimension characteristics, timestamp, and order identifier.

[0185] Understandably, the decision unit configuration table can be deployed in a database or cache service. For example, the decision unit configuration table may include a decision unit identifier, applicable customer group tags, manageable product types, and corresponding permission information. This decision unit configuration table can be, for example, a micro-enterprise unit registry, which records the registration information, capability attributes, status, and interfaces of all available micro-enterprise units (M_C) to quickly and accurately filter out micro-enterprise units capable of executing the target configuration table.

[0186] This step selects a decision-making unit from a preset configuration table based on customer group affiliation information, so that the target configuration combination is confirmed by the production decision-making body that is compatible with the customer group before execution, thereby avoiding mismatch problems caused by fixed routing.

[0187] This step significantly improves resource utilization and response efficiency in production by dynamically adapting to micro-enterprise units. For example, in automotive parts manufacturing, when a customer requests non-standard sized parts, the system matches them with micro-enterprise units that have flexible processing capabilities, avoiding idleness on fixed production lines due to equipment incompatibility.

[0188] S311: Obtain the production line control system identifier corresponding to the decision unit from the predefined production line control system catalog; generate a target configuration table based on the production line control system identifier, target configuration combination, and target verification results, and send the target configuration table to the production system.

[0189] Among them, the target configuration combination and the target verification result are used to generate production work orders, and the production line control system identifier is used to instruct the production system to send the production work order to the corresponding production line control system.

[0190] The production line control system catalog stores the correspondence between decision-making units and production line control system identifiers. This catalog can be pre-established by the enterprise when modeling the production line organizational structure and updated synchronously during production line adjustments, equipment relocations, or control system replacements. The production line control system identifier uniquely identifies a specific control system, enabling the production system to route work orders and issue control commands according to this identifier.

[0191] Specifically, after determining the decision unit that matches the target configuration combination, the corresponding production line control system identifier is queried from the production line control system catalog according to the decision unit, and the identifier, the target configuration combination, and the target verification results are encapsulated together into a target configuration table.

[0192] The target configuration table can be expressed as a structured data object, database record, or message payload. For example, the target configuration table may include configuration number, verification status, adapted production line code, and related fields required for work order generation.

[0193] After receiving the target configuration table, the decision-making unit generates a production work order based on the target configuration combination and target verification results. It then sends the work order to the corresponding production line control system in conjunction with the production line control system identifier, thereby enabling the control system to organize production resources and execute manufacturing tasks according to the target configuration combination.

[0194] In this step, the target configuration combination, target verification results, and production line control system identifier are combined to form a target configuration table with routing attributes, enabling the decision-making unit to directly complete work order construction and production line mapping based on a unified carrier, thereby helping to improve the speed of customized production.

[0195] Optionally, during the production line execution system's execution of production orders, actual execution metrics are fed back to the digital twin simulation model to update the parameters of each verification module. These actual execution metrics indicate data such as cycle time and yield during production line execution, for example, the success rate of robotic arm grasping and the cycle time of a sewing machine. Verification modules, such as geometric verification models, can optimize their interference detection algorithms using actual cycle time data.

[0196] After the execution entity completes production, this method feeds back actual execution metrics (such as cycle time and yield) to the digital twin simulation model to update the verification module. For example, the geometric verification module optimizes the interference detection algorithm using actual cycle time data, making subsequent verification results more consistent with actual production line conditions, forming a "production-feedback-optimization" closed loop.

[0197] This method continuously optimizes the predictive capabilities of the verification module through an online learning mechanism, ensuring that subsequent configurations better match actual production line conditions. For example, in garment customization, actual cycle time data is used to adjust the parameters of the geometric verification model, making the subsequently generated dress design more compatible with the robotic arm's grasping capabilities, significantly improving the production line execution success rate.

[0198] The industrial production method for mass customization provided in this application involves obtaining the configuration parameters of the product to be produced, inputting these parameters into a base model to obtain candidate configuration combinations, and then inputting these candidate configuration combinations into a digital twin simulation model. Multiple verification modules within the digital twin simulation model perform multi-dimensional production verification of the candidate configuration combinations in parallel to obtain verification results. These multiple verification modules include at least two of the following: a geometric manufacturability verification module, a process feasibility verification module, an embodied executability verification module, a customer preference consistency verification module, and a social acceptability verification module.

[0199] This method, through multi-dimensional digital twin simulation verification, can identify issues such as geometric interference, process conflicts, and preference mismatches in candidate configuration combinations in advance. This helps to obtain target configuration combinations that meet customer production needs and are compliant and highly executable, thereby reducing scrap rates and ensuring production compliance and social acceptability. Furthermore, when the verification result indicates a failure, the failure dimension features are fed back to the base model, triggering iterative optimization. This reduces manual intervention in correcting configuration combinations and helps improve the efficiency of customized solution generation and product qualification rate.

[0200] If the verification result indicates that the verification failed, feature extraction is performed on the failure dimension information in the verification result to obtain failure dimension features. These failure dimension features are then fed back to the base model. The base model performs reasoning analysis on the failure dimension features and configuration condition parameters to obtain a corrected candidate configuration combination. This corrected candidate configuration combination is then input into the digital twin simulation model to obtain the first verification result. If the first verification result does not meet the verification conditions and the preset termination conditions, the failure dimension features are retrieved from the first verification result. Based on these failure dimension features and the aforementioned configuration condition parameters, the candidate configuration combination generation operation is re-executed. If the first verification result meets the verification conditions, the corrected candidate configuration combination is output as the target configuration combination, and the first verification result is output as the target verification result.

[0201] This method, when a candidate configuration combination fails simulation verification, injects the failure dimension features back into the base model. This allows the base model to directionally correct the candidate configuration combination based on the failure dimension information, thereby improving the convergence efficiency of complex personalized configurations. Furthermore, a digital twin simulation model is introduced for re-verification, enabling the corrected candidate configuration combination to be checked in a timely manner in a virtual environment, reducing the risk of conflicts in actual production.

[0202] In addition, when the preset termination conditions are met, an alarm message is output, which is generated based on the failure dimension information, the corrected candidate configuration combination, and the first verification result.

[0203] This method employs a termination control mechanism to promptly halt invalid iterations when candidate configuration combinations, despite repeated modifications, still fail to meet verification conditions. This prevents iterative modifications from consuming excessive computing power and reducing order response efficiency. Furthermore, failure dimension information is output along with the current solution as an alarm message, facilitating subsequent targeted adjustments to the basic model, verification rules, or configuration constraints. This improves the convergence and troubleshooting efficiency of personalized customization production.

[0204] After obtaining the target configuration combination and target verification results, the decision unit matching the customer group affiliation information is obtained from the preset decision unit configuration table, and the target configuration combination and target verification results are sent to the decision unit. This allows the decision unit to directly know the solution content and verification conclusions, reducing repeated verification and manual relay, improving decision response efficiency, and enhancing the accuracy and consistency of solution distribution in customized production.

[0205] After determining the decision-making unit, the corresponding production line control system identifier is retrieved from a predefined production line control system catalog. Then, based on the production line control system identifier, the target configuration combination, and the target verification results, a target configuration table is generated and sent to the production system. In this method, because the production system can accurately send production work orders to the corresponding control system based on the production line control system identifier, this method helps reduce the probability of incorrect work order dispatch, routing confusion, and manual reassignment, and improves the efficiency of the connection between target configuration combination verification and production execution, thereby improving the response speed and execution stability of personalized customized production.

[0206] Figure 4 This application provides a structural schematic diagram of an industrial production device designed for mass customization. Figure 4 As shown, the industrial production apparatus 40 may include:

[0207] The acquisition module 401 is used to acquire candidate configuration combinations of the product to be produced, which are used to characterize the customized production scheme of the product to be produced.

[0208] Verification module 402 is used to perform multi-dimensional production verification on candidate configuration combinations and obtain verification results;

[0209] The correction module 403 is used to iteratively correct the candidate configuration combination based on the verification result when the verification result indicates that the verification has failed, until the target verification result corresponding to the corrected target configuration combination meets the verification conditions.

[0210] The sending module 404 is used to send the target configuration combination and the target verification result to the decision unit that matches the target configuration combination;

[0211] The sending module 404 is also used to send the target configuration combination and target verification result to the production system after receiving the production instruction from the decision unit, so that the production system can produce the product to be produced based on the target configuration combination and target verification result.

[0212] In one possible implementation, the above-mentioned device further includes: an input module 405;

[0213] The acquisition module 401 is also used to acquire the configuration condition parameters of the product to be produced. These configuration condition parameters are generated based on customer requirements. The configuration condition parameters include: social environment parameters, customer requirement parameters, service flow parameters, and failure dimension characteristics.

[0214] Input module 405 is used to input configuration condition parameters into the basic model to obtain candidate configuration combinations, which include: product specification parameters, embodied execution strategy, social environment parameters, customer demand parameters, service flow parameters, and value parameters.

[0215] In one possible implementation, the input module 405 is further configured to input the candidate configuration combination into the digital twin simulation model, and perform multi-dimensional production verification of the candidate configuration combination in parallel through multiple verification modules in the digital twin simulation model to obtain the verification result.

[0216] The multiple verification modules include at least two of the following: geometric manufacturability verification module, process feasibility verification module, embodied executability verification module, customer preference consistency verification module, and social acceptability verification module.

[0217] In one possible implementation, the above-mentioned device further includes: a feature extraction module 406 and an output module 407;

[0218] The feature extraction module 406 is used to extract features from the failure dimension information in the verification results to obtain failure dimension features;

[0219] The sending module 404 is also used to feed back the failure dimension features to the base model, and through the base model, to perform reasoning analysis on the failure dimension features and configuration condition parameters to obtain the corrected candidate configuration combination.

[0220] The input module 405 is also used to input the modified candidate configuration combination into the digital twin simulation model to obtain the first verification result;

[0221] The output module 407 is used to output the modified candidate configuration combination as the target configuration combination when the first verification result meets the verification conditions, and to output the first verification result as the target verification result.

[0222] The acquisition module 401 is further configured to, when the first verification result does not meet the verification conditions and the preset termination conditions, acquire the failure dimension features from the first verification result and re-execute the above-mentioned candidate configuration combination generation steps; the termination conditions include at least: the maximum number of iterations reaches a preset value, and the number of consecutive verification failures in the same dimension reaches a preset threshold.

[0223] In one possible implementation, the output module 407 is further configured to output alarm information when a preset termination condition is met; the alarm information is generated based on failure dimension information, the corrected candidate configuration combination, and the first verification result.

[0224] In one possible implementation, the customer demand parameters include customer group affiliation information. The acquisition module 401 is also used to obtain decision units that match the customer group affiliation information from a preset decision unit configuration table.

[0225] The sending module 404 is also used to send the target configuration combination and target verification results to the decision unit.

[0226] In one possible implementation, the acquisition module 401 is further configured to acquire the production line control system identifier corresponding to the decision unit from a predefined production line control system catalog.

[0227] The sending module 404 is also used to generate a target configuration table based on the production line control system identifier, target configuration combination and target verification results, and send the target configuration table to the production system;

[0228] Among them, the target configuration combination and the target verification result are used to generate production work orders, and the production line control system identifier is used to instruct the production system to send the production work order to the corresponding production line control system.

[0229] The industrial production apparatus for mass customization provided in this application is used to execute the aforementioned industrial production method embodiment for mass customization. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0230] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in this application. Figure 5 The illustrated electronic device 50 includes a memory 51, a processor 52, and a communication interface 53. The memory 51, processor 52, and communication interface 53 are communicatively connected to each other. For example, the memory 51, processor 52, and communication interface 53 can be connected via a network. Alternatively, the electronic device 50 may also include a bus 54. The memory 51, processor 52, and communication interface 53 are communicatively connected to each other via the bus 54. Figure 5 It is an electronic device 50 in which the memory 51, processor 52, and communication interface 53 are connected to each other via bus 54.

[0231] The memory 51 can be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 51 can store programs, and when the program stored in the memory 51 is executed by the processor 52, the processor 52 and the communication interface 53 are used to execute the industrial production method for mass customization described in any of the foregoing embodiments. The memory can also store data required by this industrial production method for mass customization.

[0232] The processor 52 can be a general-purpose CPU, microprocessor, application-specific integrated circuit (ASIC), graphics processing unit (GPU), or one or more integrated circuits.

[0233] Processor 52 can also be an integrated circuit chip with signal processing capabilities. In implementation, the industrial production method for large-scale customization described in this application can be completed through integrated logic circuits in the hardware of processor 52 or instructions in software form. The aforementioned processor 52 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, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the following embodiments of this application. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the following embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 51. Processor 52 reads the information in memory 51 and, in conjunction with its hardware, completes the industrial production method for mass customization proposed in this application.

[0234] Communication interface 53 uses transceiver modules, such as, but not limited to, transceivers, to enable communication between electronic device 50 and other devices or communication networks. For example, data sets can be acquired through communication interface 53.

[0235] When the aforementioned electronic device 50 includes a bus 54, the bus 54 may include a path for transmitting information between various components of the electronic device 50 (e.g., memory 51, processor 52, communication interface 53).

[0236] This application also provides a computer-readable storage medium, which may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. Specifically, the computer-readable storage medium stores program instructions, which are used in the methods described in the above embodiments.

[0237] This application also provides a program product including executable instructions stored in a readable storage medium. At least one processor of an electronic device can read the executable instructions from the readable storage medium, and the execution of the executable instructions by the at least one processor causes the electronic device to implement the industrial production methods for mass customization provided by the various embodiments described above.

[0238] The term "multiple" in this document refers to two or more. The term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Furthermore, the character " / " in this document generally indicates an "or" relationship between the preceding and following related objects; in formulas, " / " indicates a "division" relationship. Additionally, it should be understood that in the description of this application, words such as "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or order.

[0239] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application.

[0240] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. An industrial production method for mass customization, characterized in that, include: Obtain candidate configuration combinations for the products to be manufactured, wherein the candidate configuration combinations are used to characterize the customized production scheme for the products to be manufactured; The candidate configuration combinations were subjected to multi-dimensional production verification to obtain verification results; If the verification result indicates that the verification has failed, the candidate configuration combination is iteratively modified based on the verification result until the target verification result corresponding to the modified target configuration combination satisfies the verification conditions. The target configuration combination and the target verification result are sent to the decision unit that matches the target configuration combination; After receiving the production instruction from the decision unit, the target configuration combination and the target verification result are sent to the production system so that the production system can produce the product to be produced based on the target configuration combination and the target verification result.

2. The method according to claim 1, characterized in that, The process of obtaining candidate configuration combinations for the products to be manufactured includes: Obtain the configuration condition parameters of the product to be produced. The configuration condition parameters are generated based on customer requirements and include: social environment parameters, customer requirement parameters, service flow parameters, and failure dimension characteristics. The configuration condition parameters are input into the base model to obtain candidate configuration combinations, which include: product specification parameters, embodied execution strategies, social environment parameters, customer demand parameters, service flow parameters, and value parameters.

3. The method according to claim 1, characterized in that, The multi-dimensional production verification of the candidate configuration combinations, to obtain the verification results, includes: The candidate configuration combinations are input into the digital twin simulation model, and the multi-dimensional production verification of the candidate configuration combinations is performed in parallel by multiple verification modules in the digital twin simulation model to obtain the verification results. The plurality of verification modules include at least two of the following: geometric manufacturability verification module, process feasibility verification module, embodied executability verification module, customer preference consistency verification module, and social acceptability verification module.

4. The method according to claim 2, characterized in that, The step of iteratively refining the candidate configuration combinations based on the verification results until the target verification result corresponding to the refined target configuration combination satisfies the verification conditions includes: Feature extraction is performed on the failure dimension information in the verification results to obtain failure dimension features; The failure dimension features are fed back to the base model, and the base model performs inference analysis on the failure dimension features and the configuration condition parameters to obtain the corrected candidate configuration combination. The modified candidate configuration combination is input into the digital twin simulation model to obtain the first verification result; If the first verification result satisfies the verification condition, the modified candidate configuration combination is output as the target configuration combination, and the first verification result is output as the target verification result. If the first verification result does not meet the verification conditions and the preset termination conditions, the failure dimension features are obtained from the first verification result, and the above-mentioned candidate configuration combination generation steps are re-executed; the termination conditions include at least: the maximum number of iterations reaches a preset value, and the number of consecutive verification failures in the same dimension reaches a preset threshold.

5. The method according to claim 4, characterized in that, The method further includes: If the preset termination conditions are met, an alarm message is output; the alarm message is generated based on the failure dimension information, the corrected candidate configuration combination, and the first verification result.

6. The method according to any one of claims 1-5, characterized in that, The customer demand parameters in the candidate configuration combination include customer group affiliation information. Sending the target configuration combination and the target verification result to the decision unit matching the target configuration combination includes: Obtain the decision unit that matches the customer group affiliation information from the preset decision unit configuration table; The target configuration combination and the target verification result are sent to the decision unit.

7. The method according to claim 6, characterized in that, After obtaining the decision unit matching the customer group affiliation information from the preset decision unit configuration table, the method further includes: Obtain the production line control system identifier corresponding to the decision-making unit from the predefined production line control system catalog; Sending the target configuration combination and the target verification result to the production system includes: Based on the production line control system identifier, the target configuration combination, and the target verification result, a target configuration table is generated and sent to the production system; The target configuration combination and the target verification result are used to generate a production work order, and the production line control system identifier is used to instruct the production system to send the production work order to the corresponding production line control system.

8. An industrial production device designed for mass customization, characterized in that, include: The acquisition module is used to acquire candidate configuration combinations of the product to be manufactured, wherein the candidate configuration combinations are used to characterize the customized production scheme of the product to be manufactured. The verification module is used to perform multi-dimensional production verification on the candidate configuration combinations and obtain verification results. The correction module is used to iteratively correct the candidate configuration combination based on the verification result when the verification result indicates that the verification has failed, until the target verification result corresponding to the corrected target configuration combination meets the verification conditions. A sending module is used to send the target configuration combination and the target verification result to a decision unit that matches the target configuration combination; The sending module is further configured to send the target configuration combination and the target verification result to the production system after receiving the production instruction from the decision unit, so that the production system can produce the product to be produced based on the target configuration combination and the target verification result.

9. An electronic device, characterized in that, include: Processor and memory; The processor is communicatively connected to the memory; The memory stores computer instructions; The processor executes computer instructions stored in the memory to implement the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

Images