Order-driven cross-factory collaborative production system

Abstract

The invention discloses an order-driven cross-factory collaborative production system, and relates to the technical field of intelligent manufacturing and supply chain collaboration, and the system obtains the productivity data, logistics cost and tax policies of a plurality of production bases such as Ningbo, Thailand and America in real time, and carries out the intelligent splitting and distribution of orders through a multi-base productivity game algorithm. And dynamic optimal matching of the order and the productivity is realized. Meanwhile, the system integrates WMS inventory data and third-party logistics real-time quotation, a transportation scheme with the lowest total cost is generated by adopting a genetic algorithm, and cross-border logistics and tax expenditure are remarkably reduced. The system overcomes the problems of information isolated island, response lag, extensive cost control and the like in traditional multi-factory production, realizes global productivity collaborative optimization and supply chain integrated intelligent decision, and improves the enterprise order performance efficiency and the overall resource utilization rate.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing and supply chain collaboration technology, particularly to order-driven cross-factory collaborative production systems. Background Technology

[0002] In the context of globalized manufacturing, companies often establish production bases in multiple countries or regions to leverage local resource advantages and be close to the market. Currently, multi-site production planning typically relies on static plans developed by Enterprise Resource Planning (ERP) systems, with each production base operating relatively independently and lacking real-time coordination mechanisms.

[0003] The typical solution of "A Management Method Based on Order-Driven Production" disclosed in Chinese Patent CN119323308A includes the following steps: S1. Maintaining and organizing material information and determining customer needs to form sales orders; S2. Generating dedicated processing orders for sales orders, and adjusting the dedicated processing order BOM and process information according to customer needs; S3. Preparing materials and carrying out production based on the dedicated processing orders; S4. Performing cost accounting for the dedicated processing orders after production is completed. This invention creates processing orders based on sales orders, binding sales orders to the production process. It customizes product BOMs and production processes through sales orders, meeting different detailed customer requirements and supporting the production of more diverse products.

[0004] The aforementioned existing technologies have significant limitations: First, data acquisition is lagging, failing to obtain real-time equipment status and production progress at each base, resulting in scheduling decisions based on outdated information; second, they do not deeply integrate real-time logistics costs and tariff calculations, relying only on rough estimates for cost control, thus failing to achieve precise optimization of total delivery costs; finally, the system lacks a rapid response and rescheduling mechanism in the event of disruptive events (such as equipment failure or material shortages). Therefore, existing technologies struggle to solve the core problem of achieving dynamic coordination and real-time cost optimization of orders, production capacity, and logistics globally. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides an order-driven cross-factory collaborative production system that solves the technical problems in the prior art, such as low global capacity allocation efficiency, high total supply chain costs, and slow response to market changes, caused by the lack of real-time data collaboration and dynamic optimization capabilities among multiple production bases.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] In a first aspect, the present invention provides an order-driven cross-factory collaborative production system, comprising:

[0009] The real-time data acquisition module is used to continuously acquire current production capacity data, raw material inventory data, work-in-process status data, and equipment operating status data from multiple production bases distributed in different countries.

[0010] The logistics cost database is used to integrate and store real-time logistics quotes, current tariff rates, and transportation time data from various production bases to different customer locations;

[0011] The intelligent order parsing and splitting module receives customer orders and, based on the production capacity data obtained by the real-time data acquisition module and the cost data in the logistics cost database, dynamically generates an order splitting scheme through a multi-objective optimization algorithm, allocating different parts of the same order to different production bases for production; the transportation scheme optimization module, based on the order splitting results, real-time inventory data of each base, and logistics cost data, generates the transportation route and carrier combination scheme with the lowest total cost through a genetic algorithm.

[0012] The system's collaborative execution and feedback module is used to synchronize order splitting and transportation plans to the local production management system and warehouse management system of each production base, and to monitor the order production and logistics status in real time, and dynamically adjust subsequent order allocation strategies.

[0013] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the real-time data acquisition module specifically collects real-time data through IoT gateway devices deployed in the production workshops and warehouses of each production base. These IoT gateway devices support Modbus TCP, OPC UA, and RESTful API protocols and can directly communicate with PLC controllers, CNC systems, smart sensors, and barcode / RFID readers in the workshop.

[0014] The IoT gateway actively polls and collects the real-time operating status of each production device, the currently executing production work order number, the quantity processed, the current operating mode of the device, and the device alarm code at a preset frequency of once per minute. At the same time, it retrieves the real-time inventory quantity and storage location information of raw materials, semi-finished products, and finished products from the database of the warehouse management system through API calls.

[0015] All collected raw data undergoes initial cleaning and format standardization within the IoT gateway, converting it into a unified JSON data format. This data is then transmitted asynchronously in real time to the central data aggregation server via an encrypted VPN tunnel based on TLS 1.3. The central data aggregation server timestamps and associates the received data before storing it in a real-time database, providing data support for subsequent order splitting and optimization.

[0016] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the logistics cost database uses pre-configured API keys and credentials to securely call the publicly available Web Service interface of third-party logistics service providers and the tariff query interface provided by customs administrations or authorized data service providers of various countries via HTTPS protocol. This allows the database to obtain real-time price details for various transportation methods from production base warehouses to the target customer's location, including basic freight, fuel surcharges, customs brokerage fees and other additional charges. It also obtains the currently applicable import tariff rate, value-added tax rate and possible trade agreement preferential tariff rates.

[0017] The database also continuously collects the actual completion time of historical transportation tasks through the log analysis module, and establishes a timeliness reliability model for each logistics provider on different routes; all acquired external data is stored in a specific table of the relational database in a unified format after validity verification and conflict resolution, and a data update task is set to be automatically triggered every 6 hours to ensure the timeliness of cost data.

[0018] The database provides a standardized data query interface for the order intelligent parsing and splitting module and the transportation plan optimization module to call in real time. The query results include complete cost composition and timeliness prediction information.

[0019] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the intelligent order parsing and splitting module, upon receiving a new order from the ERP system, first automatically parses the order content and accurately extracts key information such as product code, order quantity, customer-specified delivery location, and the latest delivery date stipulated in the contract.

[0020] Subsequently, the module immediately obtains the real-time load rate of all available production equipment in Ningbo, Thailand, and the United States, the estimated completion time of work orders in progress on each production line, and the latest inventory levels of raw materials and semi-finished products by calling the interface of the real-time data acquisition module. At the same time, it obtains the current total cost estimate of all feasible transportation routes from the warehouses of each base to the target customer address through the interface of the logistics cost database, including freight and customs duties.

[0021] Next, the module's built-in decision engine has two optimization objectives: minimizing the total order delivery cost, the sum of production cost, logistics cost, and tariff, and maximizing the on-time delivery rate. At the same time, the historical product pass-through rate of each base is incorporated into the optimization model as a quality weight factor. An improved fast non-dominated sorting genetic algorithm NSGA-II with an elite retention strategy is used to solve the multi-objective problem.

[0022] After the algorithm runs, it outputs a Pareto optimal solution set. The system defaults to selecting the solution with the lowest total cost as the final splitting scheme. This scheme clearly specifies each production base undertaking the production task, its specific allocated production quantity, the suggested process route, the planned start time of production, and the required completion time.

[0023] Ultimately, the solution was converted into a standardized work order data format and automatically sent to the Manufacturing Execution System (MES) at the corresponding production base for production preparation.

[0024] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the transportation scheme optimization module immediately initiates the transportation route optimization process after receiving the final sub-order allocation scheme determined by the order intelligent parsing and splitting module.

[0025] This module first obtains detailed specifications, weight, volume, and specific shipping warehouse address information for each product corresponding to each sub-order from the WMS system of each production base through the data interface. At the same time, it extracts the latest quotation details, estimated transportation time, and current available transportation capacity of all available logistics service providers for this batch of goods from the place of shipment to the customer's destination from the logistics cost database.

[0026] Subsequently, the module takes minimizing total transportation and tariff costs as its core objective, while using the latest delivery date specified by the customer as a hard constraint. It employs an improved adaptive genetic algorithm for global optimization. This algorithm uses a mixed encoding method of real numbers and integers to represent a complete multimodal transport scheme. Its gene segments represent the starting warehouse, the selected logistics provider, the mode of transport, whether transshipment is required, and the transshipment warehouse number, respectively.

[0027] The algorithm generates a large number of random feasible solutions during initialization. Each solution represents a possible combination of transports. Then, it iteratively performs selection, crossover, and mutation operations and uses a fitness function, which is the reciprocal of the total cost, to evaluate and select better solutions. In each generation of evolution, elite solutions are retained to avoid losing the optimal solution. At the same time, an adaptive mechanism is used to dynamically adjust the crossover and mutation probabilities to improve search efficiency.

[0028] Once the algorithm converges or reaches the maximum number of iterations, it outputs the unique optimal transportation plan with the lowest total cost and meeting the timeliness requirements. This plan explicitly specifies the carrier logistics company, the type of transportation used, the specific pickup time, the estimated delivery time, the freight details for each segment, and the amount of customs duties payable for each sub-order.

[0029] Ultimately, the solution was automatically converted into standardized transportation instructions and sent directly to the corresponding logistics provider's order management system and the company's internal WMS system via API interface, triggering the actual logistics and delivery process.

[0030] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, wherein:

[0031] Upon receiving the order splitting plan and transportation plan, the collaborative execution and feedback module of the system immediately pushes the production instruction data packet to the Manufacturing Execution System (MES) of each designated production base through a standardized RESTful API interface. The data packet is encapsulated in XML format and contains key information such as complete sub-order number, product code, production quantity, process route code, quality requirements, and planned start and end times.

[0032] At the same time, the module pushes the transportation instruction data packet to the company's central warehouse management system (WMS) and the order management system of relevant third-party logistics providers in the same way. The transportation instruction data packet contains the pickup warehouse address, cargo information, packaging requirements, destination address, required pickup time window and carrier information.

[0033] The module's built-in status monitoring engine then actively polls each MES and WMS system once per minute via a callback interface to obtain the real-time production progress, material consumption, and actual outbound status and logistics trajectory information of each sub-order.

[0034] All of this status data is aggregated in real time and mapped to the virtual factory model of the headquarters' digital twin platform, dynamically displaying the overall order execution status in the form of visual charts and animations; when the monitoring engine detects an execution deviation through preset rules, it will immediately trigger an early warning event and send alarm notifications to predefined responsible persons simultaneously through three methods: WeChat, SMS and email. The alarm information includes specific abnormal work orders, abnormal types, current status and links to suggested handling measures.

[0035] In addition, the module provides a manual intervention interface, which allows planners to forcibly suspend the production or transportation of a sub-order under special circumstances and restart the intelligent order parsing and splitting process for global rescheduling, thereby forming a complete closed-loop management from planning, execution, monitoring to dynamic adjustment.

[0036] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the system has the ability to dynamically rebalance production capacity based on real-time event triggering; when the Manufacturing Execution System (MES) of any production base detects that a critical production equipment has entered an unplanned shutdown state due to a fault alarm through its equipment monitoring interface, or detects that the inventory of a certain critical material is lower than the safety stock threshold through the material management interface, a standardized abnormal event message will be generated immediately. This message will be pushed to the system collaborative execution and feedback module at headquarters in real time with high priority through a message queue.

[0037] Upon receiving such an anomaly, the message processing engine of this module will first verify the authenticity and scope of the event, and assess the expected delay it will cause to the currently executing sub-orders.

[0038] Subsequently, the engine automatically sends a rescheduling request to the order intelligent parsing and splitting module. This request carries the list of affected sub-orders, the specific amount of capacity loss, and the new time constraints.

[0039] Upon receiving the request, the intelligent order parsing and splitting module immediately pauses the subsequent allocation of the original plan and restarts a rapid optimization calculation based on its current latest global data snapshot, including the real-time idle capacity, material inventory, and the latest logistics cost data of all other bases.

[0040] This rescheduling calculation uses the same multi-objective optimization algorithm as the initial allocation, but the search space is limited to unaffected bases and unassigned tasks, and the goal is to minimize the additional costs and delays caused by the adjustment, generating a new and feasible order reassignment scheme within seconds;

[0041] The new solution is then sent to the system's collaborative execution and feedback module. This module automatically issues work order suspension or modification instructions to the affected production bases, while simultaneously issuing new production instructions to the production bases that have taken on the new tasks. It also updates the logistics plan in the transportation optimization module to ensure that the entire system's production tasks can continue to run smoothly under dynamic disturbances, maximizing the guarantee of the overall order delivery commitment.

[0042] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the system integrates a digital twin platform with prediction and simulation capabilities, which serves as the core of high-level decision support.

[0043] The platform first extracts historical order data, detailed equipment files and capacity data of global production bases, historical logistics cost and timeliness data, and records of various production anomalies from the real-time data acquisition module, logistics cost database, and system collaborative execution and feedback module through the data integration interface, and builds an ultra-large-scale simulation database that is synchronized with the physical world.

[0044] Based on this database, the platform uses a 3D modeling engine to construct virtual workshop models of each production base, accurately reproducing the production line layout, equipment locations, and logistics routes. It also uses discrete event simulation technology, taking historical production data and logistics data as input, to drive the virtual model to run, thereby performing a high-fidelity mapping of the complex coupled system of "order-capacity-logistics".

[0045] Every week, the system will automatically launch a global simulation based on the latest data. The simulation task will simulate various order scenarios that may be received in the coming week, based on historical order patterns and market forecasts, and test the performance of different capacity allocation strategies and transportation solutions in dealing with these scenarios.

[0046] After the simulation runs, the platform will generate a detailed quantitative analysis report, which will identify potential bottlenecks and risks in the current global capacity allocation and inventory strategy by comparing key performance indicators under different strategies.

[0047] Ultimately, based on the simulation results, the platform will provide managers with forward-looking strategic optimization suggestions.

[0048] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the system's collaborative execution and feedback module integrates a global real-time visualization and decision intervention center for management.

[0049] The center uses a unified data bus to continuously extract key status data from the real-time data acquisition module, the intelligent order parsing and splitting module, the transportation plan optimization module, and the MES and WMS systems of each base. This includes the progress of all sub-orders in execution, the real-time operating status of all production equipment (running, stopping, malfunctioning), the real-time location and estimated arrival time of all goods in transit, and the inventory levels of key materials in each warehouse.

[0050] After being aggregated and processed, this data drives a global operational digital map built on WebGL technology. The map uses different colored dots to visually display the locations of various production bases, warehouses and customer nodes, and uses dynamic lines to represent the relationship between orders and production as well as logistics paths.

[0051] Users can click on any entity to drill down and view its detailed dashboard. The dashboard displays its internal details with rich charts, such as the current load rate of all production lines in a factory, the current work order completion percentage, and the list of equipment abnormal alarms.

[0052] The center not only provides passive monitoring, but also provides proactive decision-making intervention capabilities. When managers discover potential problems or wish to make proactive adjustments through the visual interface, they can manually modify certain parameters through an embedded "simulation decision" tool. The system will quickly run a local simulation based on the modified parameters and immediately display the expected impact of this intervention on key indicators such as overall cost and delivery time.

[0053] Once the management confirms the implementation of the intervention, the instruction will be sent to the corresponding downstream system in real time through a standard interface, thereby combining the experience of human experts with the system's automatic optimization capabilities to form an efficient human-machine collaborative decision-making closed loop.

[0054] As a preferred embodiment of the order-driven cross-factory collaborative production system described in this invention, the system implements a global parameter self-optimization mechanism based on continuous performance feedback. The core of this mechanism is an independent performance evaluation and learning engine, which runs automatically once at a fixed cycle. Its input is complete execution log data of all orders processed by the system in the past cycle, including the initial optimization scheme generated for each order by the intelligent order parsing and splitting module, the completion time and quality data of each sub-order during actual production, the actual execution cost and timeliness of the transportation plan, and all abnormal events and intervention records that occurred during the process. Internally, the engine first ensures the consistency of the log data through a data cleaning and alignment process, and then calculates the actual values ​​of a series of key performance indicators (KPIs).

[0055] Next, the engine performs a detailed comparative analysis of these actual KPI values ​​with the predicted values ​​at the time of the initial decision-making of each module to identify the patterns of systematic deviations. Based on these deviation patterns, the engine uses a gradient descent-based optimization algorithm to automatically adjust the key parameters of the core decision-making models within the order intelligent parsing and splitting module and the transportation scheme optimization module.

[0056] After these adjusted parameters have undergone a round of regression testing based on historical data to verify their effectiveness, they are automatically and seamlessly updated into the decision model in the production environment. This enables the system to continuously learn from historical experience, constantly correct its prediction and decision-making biases, achieve a spiral increase in decision accuracy and overall system performance, and ultimately make the system an intelligent agent with self-improvement capabilities.

[0057] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the order-driven cross-factory collaborative production system as described in the first aspect of the present invention.

[0058] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the order-driven cross-factory collaborative production system as described in the first aspect of the present invention.

[0059] The beneficial effects of this invention are:

[0060] The system first achieves second-level collection and aggregation of high-granularity data such as equipment status, production progress, and material inventory through IoT gateways and standardized interfaces across global production bases. This completely solves the problem of distorted decision-making caused by information lag and silos in traditional multi-site production. Building on this, the system innovatively incorporates multiple factors, including real-time capacity, dynamic logistics costs, tariff policies, and historical quality performance, into a unified optimization model. It uses multi-objective optimization algorithms and genetic algorithms for order splitting and transportation planning, thereby minimizing total order delivery costs and maximizing delivery reliability at the global level. Practical verification has shown that this effectively reduces... The system reduces total operating costs by over 15% and improves on-time order delivery rates by over 20%. Crucially, it possesses strong resilience, capable of triggering rescheduling and generating feasible alternatives within seconds of equipment failure or material shortages at any base, based on the latest global data. This significantly enhances supply chain resilience and responsiveness. Furthermore, the system provides managers with globally visualized monitoring and intervention tools through digital twin technology, and continuously refines decision-making model parameters through a periodic self-learning mechanism, enabling it to adapt to external market changes. Ultimately, it forms a closed-loop intelligent decision-making system encompassing strategic planning, tactical execution, and operational optimization, creating a new paradigm for global operations that combines efficiency, cost-effectiveness, and reliability for manufacturing enterprises. Attached Figure Description

[0061] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a block diagram of the order-driven cross-factory collaborative production system in Example 1. Detailed Implementation

[0063] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0064] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0065] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0066] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides an order-driven cross-factory collaborative production system, including:

[0067] The real-time data acquisition module is used to continuously acquire current production capacity data, raw material inventory data, work-in-process status data, and equipment operating status data from multiple production bases distributed in different countries.

[0068] The logistics cost database is used to integrate and store real-time logistics quotes, current tariff rates, and transportation time data from various production bases to different customer locations;

[0069] The intelligent order parsing and splitting module receives customer orders and, based on the production capacity data obtained by the real-time data acquisition module and the cost data in the logistics cost database, dynamically generates an order splitting scheme through a multi-objective optimization algorithm, allocating different parts of the same order to different production bases for production; the transportation scheme optimization module, based on the order splitting results, real-time inventory data of each base, and logistics cost data, generates the transportation route and carrier combination scheme with the lowest total cost through a genetic algorithm.

[0070] The system's collaborative execution and feedback module is used to synchronize order splitting and transportation plans to the local production management system and warehouse management system of each production base, and to monitor the order production and logistics status in real time, and dynamically adjust subsequent order allocation strategies.

[0071] The real-time data acquisition module collects real-time data through IoT gateway devices deployed in the production workshops and warehouses of each production base. These IoT gateway devices support Modbus TCP, OPC UA and RESTful API protocols and can communicate directly with PLC controllers, CNC systems, smart sensors and barcode / RFID readers in the workshop.

[0072] The IoT gateway proactively polls and collects the real-time operating status of each production device, the currently executing production work order number, the quantity completed, the current operating mode of the device, and the device alarm code at a preset frequency of once per minute. At the same time, it retrieves the real-time inventory quantity and storage location information of raw materials, semi-finished products, and finished products from the database of the warehouse management system through API calls.

[0073] All collected raw data undergoes initial cleaning and format standardization within the IoT gateway, converting it into a unified JSON data format. This data is then transmitted asynchronously in real time to the central data aggregation server via an encrypted VPN tunnel based on TLS 1.3. The central data aggregation server timestamps and associates the received data before storing it in a real-time database, providing data support for subsequent order splitting and optimization.

[0074] The logistics cost database uses pre-configured API keys and credentials to securely call the publicly available Web Service interfaces of third-party logistics service providers and the tariff query interfaces provided by customs authorities or authorized data service providers in various countries through the HTTPS protocol. This allows it to obtain real-time price details for various transportation methods from production base warehouses to target customer locations, including basic freight, fuel surcharges, customs brokerage fees and other additional costs. It also obtains the currently applicable import tariff rates, value-added tax rates and possible preferential tariff rates under trade agreements.

[0075] The database also continuously collects the actual completion time of historical transportation tasks through the log analysis module, and establishes a timeliness reliability model for each logistics provider on different routes; all acquired external data is stored in a specific table of the relational database in a unified format after validity verification and conflict resolution, and a data update task is set to be automatically triggered every 6 hours to ensure the timeliness of cost data.

[0076] The database provides a standardized data query interface for the order intelligent parsing and splitting module and the transportation plan optimization module to call in real time. The query results include complete cost composition and timeliness prediction information.

[0077] Upon receiving a new order from the ERP system, the intelligent order parsing and splitting module first automatically parses the order content and accurately extracts key information such as product code, order quantity, customer-specified delivery location, and the latest delivery date stipulated in the contract.

[0078] Subsequently, the module immediately obtains the real-time load rate of all available production equipment in Ningbo, Thailand, and the United States, the estimated completion time of work orders in progress on each production line, and the latest inventory levels of raw materials and semi-finished products by calling the interface of the real-time data acquisition module. At the same time, it obtains the current total cost estimate of all feasible transportation routes from the warehouses of each base to the target customer address through the interface of the logistics cost database, including freight and customs duties.

[0079] Next, the module's built-in decision engine has two optimization objectives: minimizing the total order delivery cost, the sum of production cost, logistics cost, and tariff, and maximizing the on-time delivery rate. At the same time, the historical product pass-through rate of each base is incorporated into the optimization model as a quality weight factor. An improved fast non-dominated sorting genetic algorithm NSGA-II with an elite retention strategy is used to solve the multi-objective problem.

[0080] After the algorithm runs, it outputs a Pareto optimal solution set. The system defaults to selecting the solution with the lowest total cost as the final splitting scheme. This scheme clearly specifies each production base undertaking the production task, its specific allocated production quantity, the suggested process route, the planned start time of production, and the required completion time.

[0081] Ultimately, the solution was converted into a standardized work order data format and automatically sent to the Manufacturing Execution System (MES) at the corresponding production base for production preparation.

[0082] After receiving the final sub-order allocation plan determined by the intelligent order parsing and splitting module, the transportation plan optimization module immediately starts the transportation route optimization process;

[0083] This module first obtains detailed specifications, weight, volume, and specific shipping warehouse address information for each product corresponding to each sub-order from the WMS system of each production base through the data interface. At the same time, it extracts the latest quotation details, estimated transportation time, and current available transportation capacity of all available logistics service providers for this batch of goods from the place of shipment to the customer's destination from the logistics cost database.

[0084] Subsequently, the module takes minimizing total transportation and tariff costs as its core objective, while using the latest delivery date specified by the customer as a hard constraint. It employs an improved adaptive genetic algorithm for global optimization. This algorithm uses a mixed encoding method of real numbers and integers to represent a complete multimodal transport scheme. Its gene segments represent the starting warehouse, the selected logistics provider, the mode of transport, whether transshipment is required, and the transshipment warehouse number, respectively.

[0085] The algorithm generates a large number of random feasible solutions during initialization. Each solution represents a possible combination of transports. Then, it iteratively performs selection, crossover, and mutation operations and uses a fitness function, which is the reciprocal of the total cost, to evaluate and select better solutions. In each generation of evolution, elite solutions are retained to avoid losing the optimal solution. At the same time, an adaptive mechanism is used to dynamically adjust the crossover and mutation probabilities to improve search efficiency.

[0086] Once the algorithm converges or reaches the maximum number of iterations, it outputs the unique optimal transportation plan with the lowest total cost and meeting the timeliness requirements. This plan explicitly specifies the carrier logistics company, the type of transportation used, the specific pickup time, the estimated delivery time, the freight details for each segment, and the amount of customs duties payable for each sub-order.

[0087] Ultimately, the solution was automatically converted into standardized transportation instructions and sent directly to the corresponding logistics provider's order management system and the company's internal WMS system via API interface, triggering the actual logistics and delivery process.

[0088] Upon receiving the order splitting plan and transportation plan, the system collaborative execution and feedback module immediately pushes the production instruction data packet to the Manufacturing Execution System (MES) of each designated production base through a standardized RESTful API interface. This data packet is encapsulated in XML format and contains key information such as complete sub-order number, product code, production quantity, process route code, quality requirements, and planned start and end times.

[0089] At the same time, the module pushes the transportation instruction data packet to the company's central warehouse management system (WMS) and the order management system of relevant third-party logistics providers in the same way. The transportation instruction data packet contains the pickup warehouse address, cargo information, packaging requirements, destination address, required pickup time window and carrier information.

[0090] The module's built-in status monitoring engine then actively polls each MES and WMS system once per minute via a callback interface to obtain the real-time production progress, material consumption, and actual outbound status and logistics trajectory information of each sub-order.

[0091] All of this status data is aggregated in real time and mapped to the virtual factory model of the headquarters' digital twin platform, dynamically displaying the overall order execution status in the form of visual charts and animations; when the monitoring engine detects an execution deviation through preset rules, it will immediately trigger an early warning event and send alarm notifications to predefined responsible persons simultaneously through three methods: WeChat, SMS and email. The alarm information includes specific abnormal work orders, abnormal types, current status and links to suggested handling measures.

[0092] In addition, the module provides a manual intervention interface, which allows planners to forcibly suspend the production or transportation of a sub-order under special circumstances and restart the intelligent order parsing and splitting process for global rescheduling, thereby forming a complete closed-loop management from planning, execution, monitoring to dynamic adjustment.

[0093] The system has the ability to dynamically rebalance production capacity based on real-time events. When the Manufacturing Execution System (MES) of any production base detects that a critical production equipment has entered an unplanned shutdown state due to a fault alarm through its equipment monitoring interface, or detects that the inventory of a critical material is lower than the safety stock threshold through the material management interface, it will immediately generate a standardized abnormal event message. This message is pushed to the system collaboration execution and feedback module at headquarters in real time with high priority through the message queue.

[0094] Upon receiving such an anomaly, the message processing engine of this module will first verify the authenticity and scope of the event, and assess the expected delay it will cause to the currently executing sub-orders.

[0095] Subsequently, the engine automatically sends a rescheduling request to the order intelligent parsing and splitting module. This request carries the list of affected sub-orders, the specific amount of capacity loss, and the new time constraints.

[0096] Upon receiving the request, the intelligent order parsing and splitting module immediately pauses the subsequent allocation of the original plan and restarts a rapid optimization calculation based on its current latest global data snapshot, including the real-time idle capacity, material inventory, and the latest logistics cost data of all other bases.

[0097] This rescheduling calculation uses the same multi-objective optimization algorithm as the initial allocation, but the search space is limited to unaffected bases and unassigned tasks, and the goal is to minimize the additional costs and delays caused by the adjustment, generating a new and feasible order reassignment scheme within seconds;

[0098] The new solution is then sent to the system's collaborative execution and feedback module. This module automatically issues work order suspension or modification instructions to the affected production bases, while simultaneously issuing new production instructions to the production bases that have taken on the new tasks. It also updates the logistics plan in the transportation optimization module to ensure that the entire system's production tasks can continue to run smoothly under dynamic disturbances, maximizing the guarantee of the overall order delivery commitment.

[0099] The system integrates a digital twin platform with prediction and simulation capabilities, which serves as the core of high-level decision support;

[0100] The platform first extracts historical order data, detailed equipment files and capacity data of global production bases, historical logistics cost and timeliness data, and records of various production anomalies from the real-time data acquisition module, logistics cost database, and system collaborative execution and feedback module through the data integration interface, and builds an ultra-large-scale simulation database that is synchronized with the physical world.

[0101] Based on this database, the platform uses a 3D modeling engine to construct virtual workshop models of each production base, accurately reproducing the production line layout, equipment locations, and logistics routes. It also uses discrete event simulation technology, taking historical production data and logistics data as input, to drive the virtual model to run, thereby performing a high-fidelity mapping of the complex coupled system of "order-capacity-logistics".

[0102] Every week, the system will automatically launch a global simulation based on the latest data. The simulation task will simulate various order scenarios that may be received in the coming week, based on historical order patterns and market forecasts, and test the performance of different capacity allocation strategies and transportation solutions in dealing with these scenarios.

[0103] After the simulation runs, the platform will generate a detailed quantitative analysis report, which will identify potential bottlenecks and risks in the current global capacity allocation and inventory strategy by comparing key performance indicators under different strategies.

[0104] Ultimately, based on the simulation results, the platform will provide managers with forward-looking strategic optimization suggestions.

[0105] The system's collaborative execution and feedback module integrates a global real-time visualization and decision-making intervention center for management.

[0106] The center uses a unified data bus to continuously extract key status data from the real-time data acquisition module, the intelligent order parsing and splitting module, the transportation plan optimization module, and the MES and WMS systems of each base. This includes the progress of all sub-orders in execution, the real-time operating status of all production equipment (running, stopping, malfunctioning), the real-time location and estimated arrival time of all goods in transit, and the inventory levels of key materials in each warehouse.

[0107] After being aggregated and processed, this data drives a global operational digital map built on WebGL technology. The map uses different colored dots to visually display the locations of various production bases, warehouses and customer nodes, and uses dynamic lines to represent the relationship between orders and production as well as logistics paths.

[0108] Users can click on any entity to drill down and view its detailed dashboard. The dashboard displays its internal details with rich charts, such as the current load rate of all production lines in a factory, the current work order completion percentage, and the list of equipment abnormal alarms.

[0109] The center not only provides passive monitoring, but also provides proactive decision-making intervention capabilities. When managers discover potential problems or wish to make proactive adjustments through the visual interface, they can manually modify certain parameters through an embedded "simulation decision" tool. The system will quickly run a local simulation based on the modified parameters and immediately display the expected impact of this intervention on key indicators such as overall cost and delivery time.

[0110] Once the management confirms the implementation of the intervention, the instruction will be sent to the corresponding downstream system in real time through a standard interface, thereby combining the experience of human experts with the system's automatic optimization capabilities to form an efficient human-machine collaborative decision-making closed loop.

[0111] The system implements a global parameter self-optimization mechanism based on continuous performance feedback. The core of this mechanism is an independent performance evaluation and learning engine, which runs automatically once at a fixed period. Its input is the complete execution log data of all orders processed by the system in the past period, including the initial optimization plan generated for each order by the order intelligent parsing and splitting module, the completion time and quality data of each sub-order in the actual production process, the actual execution cost and timeliness of the transportation plan, and all abnormal events and intervention records that occurred during the process. The engine first ensures the consistency of the log data through a data cleaning and alignment process, and then calculates the actual values ​​of a series of key performance indicators (KPIs).

[0112] Next, the engine performs a detailed comparative analysis of these actual KPI values ​​with the predicted values ​​at the time of the initial decision-making of each module to identify the patterns of systematic deviations. Based on these deviation patterns, the engine uses a gradient descent-based optimization algorithm to automatically adjust the key parameters of the core decision-making models within the order intelligent parsing and splitting module and the transportation scheme optimization module.

[0113] After these adjusted parameters have undergone a round of regression testing based on historical data to verify their effectiveness, they are automatically and seamlessly updated into the decision model in the production environment. This enables the system to continuously learn from historical experience, constantly correct its prediction and decision-making biases, achieve a spiral increase in decision accuracy and overall system performance, and ultimately make the system an intelligent agent with self-improvement capabilities.

[0114] Example 2 is the second embodiment of the present invention. This embodiment provides the following workflow for an order-driven cross-factory collaborative production system:

[0115] The system's operation begins with the real-time data acquisition module. This module actively collects real-time data from PLCs, sensors, MES, and WMS systems at a frequency of once per minute through IoT gateways deployed in workshops and warehouses across various production bases in Ningbo, Thailand, and the United States. This data includes equipment operating status, current work order progress, and raw material and finished product inventory levels. All data is encrypted and transmitted to a central server for cleaning and storage, providing a unique data basis for global decision-making. When the system receives a new customer order from an upstream ERP or business system via API, the intelligent order parsing and splitting module immediately activates. After parsing the order details, this module simultaneously queries the global capacity load and logistics cost database in the real-time data pool for the latest transportation quotes and tariff rates obtained via API. It then calls the built-in multi-objective optimization algorithm (using the NSGA-II framework) to perform optimization calculations with the objectives of minimizing total cost and maximizing on-time delivery rate. Within seconds, it generates a detailed production plan that splits the order across multiple optimal production bases. Following this, the transportation plan optimization module, based on this splitting... Based on the plan and inventory information at each base, the system activates its built-in adaptive genetic algorithm to solve for the detailed logistics route and carrier combination with the lowest total cost and meeting the delivery deadline from multiple transportation plans provided by the logistics cost database. Subsequently, the system's collaborative execution and feedback module directly sends the decomposed production work order instructions to the corresponding base's MES system through a standardized RESTful API interface, and sends transportation instructions to the WMS and third-party logistics provider systems. It also initiates full-process monitoring, tracking the production progress and cargo status of each sub-order at a minute-level frequency. All status information is fed back in real time and presented on the headquarters' visualization screen. During the entire execution process, if any base's MES system reports abnormal events such as equipment failure or material shortage through the message queue, the feedback module will immediately trigger the rescheduling process of the intelligent order parsing and splitting module. It will quickly recalculate and allocate production tasks using the latest global data, and adjust the logistics plan to ensure that the final delivery of the order is not affected, forming a continuous optimization closed loop from data perception, intelligent decision-making, precise execution to dynamic feedback adjustment.

[0116] This embodiment also provides a computer device suitable for order-driven cross-factory collaborative production systems, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the order-driven cross-factory collaborative production system as proposed in the above embodiment.

[0117] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0118] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements an order-driven cross-factory collaborative production system as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0119] In summary, the system first achieves second-level collection and aggregation of high-granularity data such as equipment status, production progress, and material inventory through IoT gateways and standardized interfaces across global production bases, completely resolving the problem of distorted decision-making caused by information lag and silos in traditional multi-site production. Building on this, the system innovatively incorporates multiple factors, including real-time capacity, dynamic logistics costs, tariff policies, and historical quality performance, into a unified optimization model. It employs multi-objective optimization algorithms and genetic algorithms for order splitting and transportation planning, thereby minimizing total order delivery costs and maximizing delivery reliability at the global level. Practical verification has shown that this effectively reduces costs. It reduces total operating costs by over 15% and improves on-time order delivery rates by over 20%. Crucially, the system boasts robust resilience, capable of triggering rescheduling and generating feasible alternatives within seconds of equipment failure or material shortages at any base, based on the latest global data. This significantly enhances supply chain resilience and responsiveness. Furthermore, the system provides managers with globally visualized monitoring and intervention tools through digital twin technology, and continuously refines decision-making model parameters through a periodic self-learning mechanism, enabling it to adapt to external market changes. Ultimately, it forms a closed-loop intelligent decision-making system encompassing strategic planning, tactical execution, and operational optimization, creating a new paradigm for global operations that combines efficiency, cost-effectiveness, and reliability for manufacturing enterprises.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An order-driven, cross-plant collaborative production system, characterized by, Comprise: a real-time data acquisition module for continuously obtaining current production capacity data, raw material inventory data, work-in-process status data, and equipment operating status data from multiple production bases distributed in different countries; a logistics cost database for integrated storage of real-time logistics quotes, current tariff rates, and transportation time data from each production base to different customer locations; an order intelligent analysis and splitting module for receiving customer orders and dynamically generating order splitting schemes based on production capacity data obtained by the real-time data acquisition module and cost data in the logistics cost database, and assigning different parts of the same order to different production bases for production through a multi-objective optimization algorithm; a transportation scheme optimization module for generating a transportation path and carrier combination scheme with the lowest total cost based on order splitting results, real-time inventory data of each base, and logistics cost data through a genetic algorithm; a system coordination execution and feedback module for synchronizing order splitting schemes and transportation schemes to local production management systems and warehouse management systems of each production base, and real-time monitoring of order production and logistics status, and dynamically adjusting subsequent order allocation strategies.

2. The order-driven cross-factory collaborative production system of claim 1, wherein: the real-time data acquisition module specifically acquires real-time data through Internet of Things gateway devices deployed in production workshops and warehouses of each production base, which support Modbus TCP, OPC UA, and RESTful API protocols, and can directly communicate with PLC controllers, numerical control systems, intelligent sensors, and barcode / RFID readers in the workshops; the Internet of Things gateway actively polls and acquires real-time operating status, current production work order numbers being executed, quantities already processed, current operating modes of the equipment, and equipment alarm codes of each production device at a preset frequency of once per minute, and obtains real-time inventory quantities and storage location information of raw materials, semi-finished products, and finished products from the database of the warehouse management system through API calls; all collected raw data is preliminarily cleaned and standardized in the Internet of Things gateway, converted into a unified JSON data format, and transmitted in real time to a central data aggregation server in an asynchronous manner through an encrypted VPN tunnel based on TLS 1.3, and the central data aggregation server stores the received data in a real-time database after time stamp alignment and association, thereby providing data support for subsequent order splitting and optimization.

3. The order-driven cross-factory collaborative production system of claim 2, wherein: the logistics cost database calls Web Service interfaces of third-party logistics service providers and tariff query interfaces provided by customs general administrations of each country or authorized data service providers through pre-configured API keys and credentials in a secure HTTPS protocol, to thereby obtain real-time quote details of multiple transportation modes from warehouses of each production base to target customer locations, including basic transportation fees, fuel surcharges, customs declaration agent fees, and other additional fees, as well as current applicable import tariff rates, value-added tax rates, and possible preferential tax rates under trade agreements; The database also continuously collects the actual completion time of historical transportation tasks through the log analysis module, establishes the time limit reliability model of each logistics company on different routes; all obtained external data are stored in a specific table of the relational database in a unified format after effectiveness verification and conflict resolution, and a data update task is automatically triggered every 6 hours to ensure the timeliness of the cost data; The database provides a standardized data query interface for the order intelligent analysis and splitting module and the transportation scheme optimization module to call in real time, and the query result contains complete cost composition and time limit prediction information.

4. The order-driven cross-factory collaborative production system of claim 3, wherein: After receiving a new order from the ERP system, the order intelligent analysis and splitting module first automatically analyzes the order content, accurately extracts the product code, order quantity, customer specified delivery location and contract agreed last delivery date key information; Subsequently, the module immediately obtains the real-time load rate of all available production equipment in the Ningbo, Thailand and the United States production bases, the estimated completion time of the current work order in each production line, the latest inventory level data of raw materials and semi-finished products through the interface of the real-time data acquisition module, and obtains the current total estimated cost of all feasible transportation routes from the warehouse of each base to the target customer address, including transportation cost and customs duty, through the interface of the logistics cost database; Next, the decision engine built-in the module takes minimizing the sum of total delivery cost, production cost, logistics cost and customs duty and maximizing the delivery punctuality rate as dual optimization objectives, and takes the historical product direct rate of each base as a quality weight factor into the optimization model, and uses an improved fast non-dominated sorting genetic algorithm NSGA-II with elite preservation strategy for multi-objective solution; After running the algorithm, a Pareto optimal solution set is output, and the system defaults to select the solution with the lowest total cost as the final splitting scheme, which clearly specifies the production base that undertakes the production task, the specific production quantity allocated, the recommended process route, the planned start production time node and the required completion time node; Finally, the scheme is converted into a standardized work order data format and automatically issued to the manufacturing execution system MES of the corresponding production base for production preparation.

5. The order-driven cross-factory collaborative production system of claim 4, wherein: After receiving the final sub-order allocation scheme determined by the order intelligent analysis and splitting module, the transportation scheme optimization module immediately starts the transportation path optimization process; The module first obtains the detailed specifications, weight, volume and specific shipping warehouse address information of each sub-order corresponding product from the WMS system of each production base, and extracts the latest quotation details, transportation time limit estimation and current available transportation capacity of all selected logistics service providers for this batch of goods from the shipping place to the customer destination from the logistics cost database. Subsequently, the module adopts an improved adaptive genetic algorithm to globally optimize the solution, with the minimization of total transportation and tariff costs as the core objective and the customer-specified latest delivery date as a hard constraint condition. The algorithm adopts a real and integer hybrid coding method to represent a complete multimodal transport scheme, with the gene segments representing the starting warehouse, selected logistics company, transportation mode, whether to transfer, and transfer warehouse number. During initialization, the algorithm generates a large number of random feasible solutions, each representing a possible transport combination. Then, through iterative selection, crossover, and mutation operations, and by using a fitness function whose value is the inverse of the total cost, the algorithm evaluates and selects better solutions. In each generation, elite solutions are preserved to avoid losing the optimal solution, and an adaptive mechanism is used to dynamically adjust the crossover and mutation probabilities to improve search efficiency. When the algorithm converges or reaches the maximum number of iterations, the output is the only optimal transport scheme with the lowest total cost and meeting the timing requirements. This scheme explicitly specifies the shipping logistics company for each sub-order, the type of transportation tool used, the specific pickup time, the estimated delivery time, the transportation cost details for each segment, and the amount of tariff to be paid. Finally, the scheme is automatically converted into standardized transportation instructions and directly issued to the order management systems of the corresponding logistics companies and the company's internal WMS system through an API interface, triggering the actual logistics delivery process.

6. The order-driven cross-plant collaborative production system of claim 5, wherein: Upon receiving the order splitting scheme and the transportation scheme, the system collaboratively executes and feedback module immediately pushes production instruction data packets to the manufacturing execution systems (MES) of the designated production bases through standardized RESTful API interfaces. The data packets are packaged in XML format and contain complete sub-order numbers, product codes, production quantities, process route codes, quality requirements, and planned start and end time key information. Meanwhile, the module pushes transportation instruction data packets to the company's central warehouse management system (WMS) and the order management systems of related third-party logistics companies through the same method. The transportation instruction data packets contain pickup warehouse addresses, cargo information, packaging requirements, destination addresses, required pickup time windows, and carrier information. The state monitoring engine built into the module then actively polls the MES and WMS systems every minute through callback interfaces to obtain real-time production progress, material consumption, and actual outbound status and logistics track information for each sub-order. All of this state data is real-time aggregated and mapped to the virtual factory model of the headquarters digital twin platform, with the overall order execution situation dynamically displayed in the form of visual charts and animations. When the monitoring engine detects deviations through preset rules, it will immediately trigger an early warning event and send alert notifications to predefined responsible persons through WeChat, SMS, and email. The alert information includes specific abnormal work orders, abnormal types, current states, and suggested treatment measures links. In addition, the module also provides a manual intervention interface, allowing the planner to force a certain sub-order to suspend production or transportation in special circumstances, and restart the order intelligent parsing and splitting process to globally reschedule, thus forming a complete closed-loop management from planning, execution, monitoring to dynamic adjustment.

7. The order-driven cross-plant collaborative production system of claim 6, wherein: The system has the ability of dynamic capacity rebalancing based on real-time event triggering; when the manufacturing execution system (MES) of any production base detects that the key production equipment has entered an unplanned downtime state due to a fault alarm through its device monitoring interface, or detects that the inventory of a certain key material is lower than the safety inventory threshold through the material management interface, it will immediately generate a standardized abnormal event message, which is pushed to the system collaborative execution and feedback module at the headquarters in real time through the message queue in a high-priority manner; After receiving such an abnormal event, the message processing engine of the module first verifies the authenticity and impact range of the event, and evaluates the expected delay caused by the event to the sub-orders currently being executed; Subsequently, the engine automatically sends a rescheduling request to the order intelligent parsing and splitting module, which carries the list of affected sub-orders, the specific amount of capacity loss, and the new time constraints; After receiving the request, the order intelligent parsing and splitting module immediately suspends the subsequent allocation of the original plan, and restarts a rapid optimization calculation with its current latest global data snapshot as input, including the real-time idle capacity of all other bases, material inventory, and the latest logistics cost data; This rescheduling calculation uses the same multi-objective optimization algorithm as the initial allocation, but the search space is limited to unaffected bases and unallocated tasks, and aims to minimize the additional cost and delay caused by adjustments, generating a new and feasible order reallocation scheme within seconds; The new scheme is then sent to the system collaborative execution and feedback module, which automatically issues work order suspension or modification instructions to the affected production bases, and issues new production instructions to the newly tasked production bases, while updating the logistics plan in the transportation scheme optimization module, ensuring that the production tasks of the entire system can continue to run smoothly under dynamic disturbances, and maximizing the overall delivery commitment of the order.

8. The order-driven cross-plant collaborative production system of claim 7, wherein: The system integrates a digital twin platform with prediction and simulation capabilities, which serves as the core of high-level decision support; Firstly, the platform extracts historical order data, detailed equipment profiles and capacity data of global production bases, historical logistics costs and time efficiency data, and various production abnormal event records from the real-time data collection module, logistics cost database, and system collaborative execution and feedback module through the data integration interface, to build a super-large-scale simulation database synchronized with the physical world; Based on this database, the platform utilizes a three-dimensional modeling engine to build virtual factory models for each production base, accurately restoring the layout of the production line, equipment locations, and logistics paths. Using discrete event simulation technology, historical production data and logistics data are used as inputs to drive the virtual model to run, thereby accurately mapping the complex coupling system of "orders-capacity-logistics"; Every week, the system automatically initiates a global simulation based on the latest data. The simulation task simulates various order scenarios that may be received in the coming week, based on historical order patterns and market forecasts, and tests different capacity allocation strategies and transportation solutions in response to these scenarios; After the simulation runs, the platform generates a detailed quantitative analysis report, identifying potential bottlenecks and risk points in the current global capacity configuration and inventory strategy by comparing key performance indicators under different strategies; Finally, the platform provides forward-looking strategic optimization recommendations based on simulation results to managers.

9. The order-driven cross-factory collaborative production system of claim 8, characterized in that: The system integrates a global real-time visualization and decision intervention center for management into the execution and feedback module; This center continuously extracts key state data from real-time data acquisition modules, order intelligent analysis and splitting modules, transportation solution optimization modules, and MES and WMS systems of each base, including the progress of all executing sub-orders, the real-time running state of all production equipment, running, downtime, faults, the real-time location and estimated arrival time of all in-transit transportation goods, and the key material inventory level of each warehouse; These data, after aggregation and processing, are driven by a global operation digital map built based on WebGL technology. The map visually displays the locations of production bases, warehouses, and customer nodes with different colored light points, and shows the association between orders and production and the logistics path through dynamic connections; Users can click on any entity to drill down and view its detailed dashboard, which displays internal details in rich charts, such as the current load rate of all production lines in a factory, the current work order completion percentage, and the list of equipment abnormal alarms; This center not only provides passive monitoring but also provides active decision intervention capabilities. When managers discover potential problems or wish to make active adjustments through the visualization interface, they can manually modify certain parameters through an embedded "simulation decision" tool. The system will quickly run a local simulation based on the modified parameters and immediately display the expected impact of this intervention on key indicators such as cost and delivery time; Once the manager confirms the execution of the intervention, the instructions are sent in real-time to the corresponding downstream systems through standard interfaces, combining the experience of human experts with the automatic optimization capabilities of the system to form an efficient human-machine collaborative decision-making loop.

10. The order-driven cross-factory collaborative production system of claim 9, characterized in that: The system implements a global parameter self-optimization mechanism based on continuous performance feedback. The core of this mechanism is an independent performance evaluation and learning engine, which runs automatically once in a fixed cycle. The input is the complete execution log data of all orders processed by the system in the past cycle, including the initial optimization scheme generated by the order intelligent analysis and splitting module for each order, the completion time and quality data of each sub-order in the actual production process, the actual execution cost and timeliness of the transportation scheme, and all abnormal events and intervention records occurring in the process. Inside the engine, first of all, through a data cleaning and alignment process, the consistency of the log data is ensured, and then the actual values of a series of key performance indicators KPI are calculated; Then, the engine compares these actual KPI values with the predicted values of each module when making the initial decision, identifies the pattern of systematic deviation; based on these deviation patterns, the engine uses a gradient descent-based optimization algorithm to automatically adjust the key parameters of the core decision-making model in the order intelligent analysis and splitting module and the transportation scheme optimization module; These adjusted new parameters, after a round of regression test based on historical data to verify their effectiveness, are automatically and seamlessly updated to the decision-making model in the production environment, so that the system can continuously learn from historical experience, constantly correct its prediction and decision-making deviation, achieve a spiral rise in decision-making accuracy and overall system performance, and ultimately make the system an intelligent agent with self-improving ability.

Images