Speed-first product production scheduling method and application thereof in garment production enterprise

Abstract

The invention relates to the technical field of garment production and manufacturing, and discloses a speed-first product production scheduling method and application thereof in a garment production enterprise, and the method comprises the steps: collecting real-time production data and multi-source order data of a product production line; based on the real-time production data and the multi-source order data, constructing a multi-target constraint optimization model with the shortest order delivery time as an optimization target; and solving the multi-target constraint optimization model by adopting a preset double-layer optimization strategy, outputting an optimal production scheduling plan, and dynamically adjusting the optimal production scheduling plan according to the real-time production state. The method focuses on minimizing order delivery time, gives consideration to multi-objective optimization of equipment utilization rate and production cost, and can carry out dynamic adjustment according to real-time production data.

Description

Technical Field

[0001] This invention relates to the field of garment manufacturing technology, specifically to a speed-priority product production scheduling method and its application in garment manufacturing enterprises. Background Technology

[0002] As competition in the apparel industry intensifies, the planned, standardized, and large-scale production model can no longer meet the diverse needs of consumers. Apparel companies are increasingly shifting towards small-batch, order-based production to adapt to changing demands, shorten product lifecycles, and improve inventory turnover. This places higher demands on the production management level of the apparel manufacturing industry.

[0003] Currently, when enterprises engage in large-scale personalized customization of clothing, they face problems such as uncertain delivery dates, slow order response times, and low workshop flexibility. Bottlenecks in the production process are not resolved promptly, and production progress is difficult to control, leading to failure to complete orders on time. Traditional clothing manufacturing production line scheduling methods lack scientific management. Personnel resources are locked up before the individual tasks of each process are completed, resulting in wasted resources and an inability to allocate them rationally and effectively, severely impacting work efficiency.

[0004] Existing supply chain scheduling systems often only consider a single objective, such as minimizing completion time or maximizing equipment utilization, failing to adequately address the urgency of order delivery. Furthermore, traditional algorithms lack the ability to adjust in real time to cope with dynamically changing production environments, resulting in poor performance.

[0005] In summary, existing technologies suffer from slow response times in garment production scheduling, untimely order delivery, and weak dynamic adjustment capabilities. Summary of the Invention

[0006] This invention provides a speed-priority product production scheduling method and its application in garment manufacturing enterprises, in order to solve the problems of slow response speed, untimely order delivery, and weak dynamic adjustment capability in garment production scheduling in the prior art.

[0007] In a first aspect, the present invention provides a speed-priority product production scheduling method, the method comprising: Collect real-time production data from the product production line and multi-source order data; Based on real-time production data and multi-source order data, a multi-objective constrained optimization model is constructed with the shortest order delivery time as the optimization objective. A pre-defined two-layer optimization strategy is used to solve the multi-objective constrained optimization model, output the optimal production scheduling plan, and dynamically adjust the optimal production scheduling plan according to the real-time production status.

[0008] This invention provides a speed-priority product production scheduling method. By comprehensively collecting real-time production and multi-source order data, it constructs a multi-objective constrained optimization model with the shortest order delivery time as the core objective, taking into account key factors such as equipment utilization and delayed orders to achieve a balance between speed and efficiency. Combined with a preset two-layer optimization strategy, it generates an initial optimal plan through global optimization and dynamically adjusts based on real-time production status. This effectively responds to dynamic disturbances such as order insertions and equipment failures, ultimately improving the accuracy, adaptability, and order delivery efficiency of production scheduling, meeting the needs of production scenarios requiring rapid market response. This invention focuses on minimizing order delivery time while considering multi-objective optimization of equipment utilization and production costs, and can dynamically adjust based on real-time production data. It solves the problems of slow response speed, untimely order delivery, and weak dynamic adjustment capabilities in existing garment production scheduling technologies.

[0009] In one optional implementation, real-time production data from the product production line and multi-source order data are collected, including: Multiple IoT sensors are used to collect real-time production data from the product production line. The real-time production data includes: status sensor data of garment hanging production equipment, RFID tag or unique code data, and work reporting information scanned by operators outside the hanging line. The system collects multi-source order data from supply chain management software, including order data, material bill of materials (BOM) data, inventory data, and process route data.

[0010] This invention provides a speed-priority product production scheduling method. It utilizes multiple IoT sensors to capture real-time dynamic data from the garment hanging production equipment, including status sensor data, RFID tag or unique code data, and work reporting information scanned by operators outside the hanging line. This ensures timely reflection of the actual operation of the production line and guarantees the real-time nature and on-site relevance of production data. Simultaneously, it accurately collects management-level data such as order details, material reserves, and process route planning from supply chain management software, achieving seamless integration of production execution data and enterprise management data. This avoids the limitations of a single data source and comprehensively covers key information across the entire chain from order demand to production completion. It provides a comprehensive, high-quality data foundation for subsequently building accurate multi-objective constraint optimization models, efficiently solving optimal scheduling plans, and dynamically adjusting the schedule. This effectively reduces scheduling deviations caused by missing, delayed, or fragmented data, improving the accuracy and reliability of subsequent scheduling decisions.

[0011] In one optional implementation, a multi-objective constrained optimization model is constructed based on real-time production data and multi-source order data, with the goal of minimizing order delivery time. This model includes: Based on real-time production data and multi-source order data, multiple constraints are determined. These constraints include: logical relationship constraints between processes, equipment resource capacity constraints, personnel skill matching constraints, material supply time and secondary processing constraints, and order priority constraints. Based on multiple constraints, a multi-objective constrained optimization model is constructed with the goal of minimizing order delivery time.

[0012] This invention provides a speed-priority product production scheduling method. Based on real-time and comprehensive production data and multi-source order data, it ensures that the model construction remains relevant to the actual production scenario, laying a data foundation for the feasibility of subsequent scheduling schemes. Simultaneously, by accurately identifying and incorporating five core constraints—process logic relationships (ensuring the rationality of the production process), equipment resource capabilities (avoiding equipment overload or idleness), personnel skill matching (preventing resource mismatch), material supply time and secondary material processing (eliminating material shortages in processes), and order priority (meeting high-value demands)—it comprehensively covers key constraints in product production scheduling from process and resource to demand, effectively avoiding scheduling problems caused by missing constraints. Furthermore, the model clearly defines the shortest order delivery time as the core optimization objective, while also considering key production indicators such as equipment utilization and personnel efficiency through multiple constraints, forming a speed-priority, multi-dimensional balanced optimization orientation. This provides a scientific and rigorous framework for solving the optimal production scheduling plan, ensuring that the final output scheduling scheme focuses on core delivery needs.

[0013] In one alternative implementation, the multi-objective constrained optimization model with the goal of minimizing order delivery time is expressed by the following formula: MinimizeF(X) =[f1(X),f2(X), f3(X)]; Where, Minimize F(X)=[f1(X),f2(X),f3(X)] represents minimizing the maximum completion time, minimizing the weighted delay time, and maximizing the equipment utilization; f1(X)=max(C i f1(X) is the maximum completion time, and f2(X) = Σw i T i f2(X) is the weighted delay time, f3(X) = -U, f3(X) is the maximum equipment utilization rate, and C i T is the completion time of order i. i w is the delay time of order i. i is the priority weight of order i, and U is the average equipment utilization rate.

[0014] In one optional implementation, the preset two-layer optimization strategy is a two-layer optimization strategy that combines an improved adaptive genetic algorithm with reinforcement learning components. A pre-defined two-layer optimization strategy is used to solve a multi-objective constrained optimization model, outputting an optimal production scheduling plan. This optimal production scheduling plan is then dynamically adjusted based on real-time production status, including: An improved adaptive genetic algorithm is used to solve the multi-objective constrained optimization model to generate an initial globally optimal production scheduling plan; Based on the initial globally optimal production scheduling plan, a reinforcement learning component is used in conjunction with the product production scenario to output the final optimal production scheduling plan; The optimal production scheduling plan is dynamically adjusted based on the deviation between the actual production progress and the planned production progress.

[0015] This invention provides a speed-priority product production scheduling method. It achieves complementary advantages through a two-layer collaborative architecture combining an improved adaptive genetic algorithm and a reinforcement learning component. On one hand, the improved adaptive genetic algorithm's global search capability effectively avoids local optima traps, generating a globally reasonable initial production scheduling plan for a multi-objective constrained optimization model, laying a high-efficiency foundation for subsequent scheduling optimization. On the other hand, the reinforcement learning component deeply integrates with the actual production scenario (such as real-time equipment status, order priority fluctuations, and material supply changes) to optimize the initial globally optimal plan in a scenario-specific way, compensating for the shortcomings of a single genetic algorithm in adapting to real-time production scenarios and outputting a final optimal plan that better meets on-site needs. Simultaneously, dynamic adjustments triggered by deviations between actual and planned production progress enable rapid responses to unexpected situations in production (such as equipment failures, emergency orders, and fluctuations in personnel efficiency), real-time correction of the scheduling plan, and ensuring that the optimal scheduling scheme is always synchronized with the actual production situation. This not only guarantees the core objective of minimizing order delivery time but also ensures efficient utilization of production resources, effectively solving the problems of insufficient global optimization or lagging dynamic response in traditional single-algorithm scheduling, and significantly improving the scientific nature, flexibility, and reliability of production scheduling.

[0016] In one optional implementation, an improved adaptive genetic algorithm is used to solve the multi-objective constrained optimization model to generate an initial globally optimal production scheduling plan, including: A two-layer coding system is determined based on the process characteristics and equipment type of product manufacturing. The two-layer coding system includes process coding and equipment coding. Based on a multi-objective constrained optimization model, the fitness value of each code is calculated; Based on the fitness value of each code, a new generation of population is generated through adaptive operations of selection, crossover, and mutation. A simulated annealing mechanism is used to perform global search optimization on the new generation population. When the preset iteration termination condition is met, the initial global optimal production scheduling plan is output.

[0017] This invention provides a speed-priority product production scheduling method. Based on the characteristics of product production processes and equipment types, it designs a two-layer coding structure of process codes and equipment codes. This structure accurately matches production process logic (such as process sequence dependencies) with equipment capability adaptability (such as matching specialized equipment with corresponding processes), breaking through the search space limitations of single codes. It provides a coding carrier that fits the actual production scenario for subsequent optimization, avoiding invalid solutions caused by codes deviating from production realities. Based on a multi-objective constraint optimization model, the fitness value of each code is calculated, quantifying the degree to which the code fits the core objective of shortest order delivery time and secondary objectives such as equipment utilization and delayed orders. This ensures that the selection direction is highly consistent with scheduling requirements, resulting in high-quality codes. The preservation of genetic algorithms provides a scientific basis; by using fitness values ​​to drive adaptive operations of selection, crossover, and mutation to generate a new generation of population, it can retain high-quality coding genes based on the elite preservation strategy, and dynamically adjust the crossover rate and mutation rate to balance the breadth of the algorithm's exploration of new solutions and the speed of convergence to optimal solutions, effectively avoiding the premature convergence problem that is prone to occur in traditional fixed-parameter genetic algorithms; a simulated annealing mechanism is introduced to perform global search optimization on the new generation of population, and the probability of accepting inferior solutions is dynamically controlled by temperature decay, breaking the local optimum trap, and ensuring that when the iteration reaches the preset termination condition, the output of the initial global optimal production scheduling plan has both global rationality and practical feasibility, laying a high-quality and highly feasible foundation for the scenario-based optimization of subsequent reinforcement learning components.

[0018] In one alternative implementation, the reinforcement learning component employs a dual deep Q-learning network architecture, the reinforcement learning component including a state space, an action space, and a reward function; Based on the initial globally optimal production scheduling plan, a reinforcement learning component is used in conjunction with the product production scenario to output the final optimal production scheduling plan, including: The state space, action space, and reward function of the reinforcement learning component are constructed based on real-time production data and multi-source order data. Based on the state space, a dual-deep Q-learning network architecture is used to output the optimal adjustment action; Based on the optimal adjustment action and reward function, the reward value of the multi-objective constrained optimization model is calculated, and the optimal adjustment action is optimized based on the reward value. The optimized optimal adjustment action is then used as the final optimal production scheduling plan.

[0019] This invention provides a speed-priority product production scheduling method, employing a dual-deep Q-learning network architecture. This effectively avoids the overestimation of Q-values ​​common in traditional Q-learning, providing a more accurate value assessment basis for action decisions and ensuring the scientific nature of adjustments. The state space is constructed based on real-time production data and multi-source order data, comprehensively covering dynamic aspects of the production site such as equipment operating status, order progress, and material supply. The action space focuses on core production scheduling needs such as order allocation and equipment scheduling, ensuring that both state perception and action output deeply align with the actual product production scenario, avoiding ineffective decisions detached from the actual situation. Furthermore, the reward function and multi-objective constraints... By optimizing the linkage between model objectives, the impact of actions on key indicators such as order delivery time, equipment utilization, and the number of delayed orders can be quantitatively assessed. This guides adjustments to converge towards the core objective of shortest order delivery time and multi-dimensional production needs. Through a closed-loop mechanism of state perception, action output, reward evaluation, and action optimization, actions are continuously iterated and optimized. This ensures that the final optimal production scheduling plan inherits the rationality of the initial globally optimal plan while accurately adapting to the real-time differences and dynamic changes in product production scenarios. This significantly improves the actual executability and multi-objective optimization effect of the scheduling plan, effectively compensating for the shortcomings of single global optimization in terms of on-site adaptability.

[0020] Secondly, the present invention provides a production scheduling method for garment manufacturing enterprises that applies the speed-priority product production scheduling method of the first aspect described above, the method comprising: Receive multi-source order data from garment manufacturers; Based on multi-source order data from garment manufacturing enterprises, the speed-first product scheduling method described in the first aspect above is run to output the optimal production scheduling plan. A visual scheduling scheme is generated based on the optimal production scheduling plan, and the execution status of garment manufacturing enterprises is monitored to dynamically adjust the optimal production scheduling plan.

[0021] The application of the speed-priority garment production scheduling method provided by this invention in the garment supply chain management system ensures the timely delivery of garment orders (especially peak season best-selling items and VIP customized orders), effectively improves the flexibility and resource utilization efficiency of garment production lines, and fully meets the core demands of garment enterprises to cope with rapid market changes and personalized customization needs.

[0022] Thirdly, the present invention provides a speed-priority product production scheduling device, the device comprising: The real-time production data and multi-source order data acquisition module is used to collect real-time production data and multi-source order data from the product production line. The multi-objective constrained optimization model construction module is used to build a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective, based on real-time production data and multi-source order data. The multi-objective constrained optimization model solving and dynamic adjustment module is used to solve the multi-objective constrained optimization model using a preset two-level optimization strategy, output the optimal production scheduling plan, and dynamically adjust the optimal production scheduling plan according to the real-time production status.

[0023] Fourthly, the present invention provides a production scheduling device for a garment manufacturing enterprise that applies the speed-priority product production scheduling method of the first aspect described above. The device includes: The multi-source order data receiving module is used to receive multi-source order data from garment manufacturing enterprises; The speed-first product scheduling module is used to run the speed-first product scheduling method mentioned above based on multi-source order data from garment manufacturing enterprises, and output the optimal production scheduling plan. The production plan dynamic adjustment module is used to generate a visual scheduling scheme based on the optimal production scheduling plan, and to monitor the execution of the garment manufacturing enterprise and dynamically adjust the optimal production scheduling plan.

[0024] Fifthly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the speed-priority product production scheduling method of the first aspect or any corresponding embodiment described above, and a production scheduling method for a garment manufacturing enterprise applying the speed-priority product production scheduling method of the first aspect described above.

[0025] In a sixth aspect, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the speed-priority product production scheduling method of the first aspect or any corresponding embodiment thereof, and a production scheduling method for a garment manufacturing enterprise applying the speed-priority product production scheduling method of the first aspect.

[0026] In a seventh aspect, the present invention provides a computer program product, including computer instructions for causing a computer to execute the speed-priority product production scheduling method of the first aspect or any corresponding embodiment thereof, and a production scheduling method for a garment manufacturing enterprise applying the speed-priority product production scheduling method of the first aspect. Attached Figure Description

[0027] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0028] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first type of speed-priority product production scheduling method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the second process of the speed-priority product production scheduling method according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the third process of the speed-priority product production scheduling method according to an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the execution of the reinforcement learning component in the speed-priority product production scheduling method according to an embodiment of the present invention; Figure 6 This is a flowchart illustrating the application of the speed-priority product production scheduling method according to an embodiment of the present invention in a garment manufacturing enterprise. Figure 7 This is a structural block diagram of a speed-priority product production scheduling device according to an embodiment of the present invention; Figure 8 This is a structural block diagram of a production scheduling device for a garment manufacturing enterprise according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0031] As an optional application scenario of this invention, such as Figure 1 As shown, application 101 is installed in terminal device 110, and user 130 can interact with application 101 through terminal device 110 and / or access device of terminal device 110.

[0032] For example, application 101 can be any application that provides question-and-answer related services. For instance, application 101 could be a question-and-answer interactive application, such as a text-to-text application, an image-to-text application, etc. Figure 1 In the application scenario shown, if application 101 is active, the terminal device 110 can display the interface 102 of application 101. The interface 102 may include various pages that application 101 can provide, such as interactive pages, settings pages, query pages, etc.

[0033] In some embodiments, terminal device 110 is communicatively connected to server 120 to provide services to application 101. Terminal device 110 may be a mobile terminal, fixed terminal, or portable terminal, etc., including but not limited to mobile phones, desktop computers, laptop computers, multimedia tablets, e-book devices, gaming devices, or any combination thereof, including accessories and peripherals of these devices or any combination thereof. In some embodiments, terminal device 110 may also support any type of interface, and server 120 may be various types of computing systems or servers capable of providing computing power, including but not limited to mainframes, edge computing nodes, computing devices in cloud environments, etc.

[0034] It should be noted that, Figure 1 This is merely an example of an application scenario and does not limit the scope of protection of this invention.

[0035] The embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the pages shown in the drawings are merely examples, and various page designs are possible in practice. The various graphic elements on the page may have different arrangements and different visual representations; one or more elements may be omitted or replaced, and one or more other elements may also be present, without any limitation in the embodiments of the present invention. Furthermore, the embodiments described below primarily pertain to terminal device 110. It should be understood that the actions described relative to terminal device 110 can be performed by application 101 on terminal device 110, or can be performed by application 101 in conjunction with its server (e.g., server 120).

[0036] Traditional production scheduling algorithms lack the ability to dynamically adjust when faced with industry pain points unique to garment production, such as the coupling of multiple processes (e.g., sewing depends on cutting), the time-consuming fabric switching (e.g., switching between denim and knitted fabrics requires cleaning equipment), and frequent emergency orders.

[0037] This invention provides a speed-priority product production scheduling method that focuses on minimizing order delivery time while taking into account multiple objectives such as equipment utilization and production costs. It can also dynamically adjust based on real-time production data, thus solving the problems of slow response speed, untimely order delivery, and weak dynamic adjustment capability in the prior art of garment production scheduling.

[0038] According to an embodiment of the present invention, a speed-priority product production scheduling method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0039] This embodiment provides a speed-priority product production scheduling method, which can be used in product production lines. Figure 2 This is a flowchart of a speed-priority product production scheduling method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Collect real-time production data and multi-source order data from the product production line.

[0040] Specifically, a product production line refers to a continuous production activity unit that is organically combined with a series of specialized equipment, orderly processes, suitable personnel, material flow systems, and other elements to realize the transformation from raw materials to semi-finished products to finished products. Its core function is to efficiently complete the entire production cycle of a specific product according to a preset process. It is the core carrier for manufacturing enterprises to achieve large-scale and standardized production.

[0041] For example, in the context of garment production, the specific form of a garment production line can be further visualized: with fabric → finished garment as the transformation goal, core processes such as fabric pretreatment → cutting (automatic cutting bed) → sewing (high-speed sewing machine) → ironing (steam ironing machine) → quality inspection (intelligent quality inspection equipment) → packaging are connected in series. Skilled personnel such as cutting workers, sewing workers, and quality inspectors are equipped, and the flow management of fabric, cut pieces, and finished garments is realized through material tracking systems such as RFID tags or unique code data, forming a complete garment production operation chain.

[0042] In this step, by adapting to the product production scenario, we comprehensively acquire real-time production-related data during the production line operation, as well as multi-source order data from different channels that cover core product order information, providing basic data support for the subsequent construction and solution of the scheduling model.

[0043] Real-time production-related data refers to various types of data that are generated and dynamically updated in real time during the operation of the product production line, and are directly related to the production execution status, resource utilization, and process progress rhythm. It covers the core elements of the entire production chain and needs to be collected in combination with specific manufacturing scenarios.

[0044] Multi-source order data refers to a collection of various information related to the entire lifecycle of a product order, obtained from multiple different channels, systems, or participants. Its core is to cover key information across the entire chain from demand initiation to production delivery, avoiding information bias or omissions caused by a single data source, and providing comprehensive and accurate demand basis for production scheduling.

[0045] Step S202: Based on real-time production data and multi-source order data, construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective.

[0046] Specifically, based on the collected real-time production data and multi-source order data, focusing on the primary optimization direction of minimizing order delivery time, while also taking into account the multi-dimensional needs in production scheduling, a multi-objective optimization model with multiple constraints is built to clarify the core direction and boundaries of scheduling optimization.

[0047] Multiple constraints refer to the multi-dimensional limiting factors that must be included when building a product production scheduling optimization model to ensure that the scheduling plan conforms to actual production rules, resource capabilities, and demand priorities. These factors cover core aspects such as production process logic, resource supply capacity, personnel skill matching, material support rhythm, and order demand sequencing. They are the key boundaries to prevent the scheduling plan from deviating from reality and becoming unimplementable, and are especially suitable for the multi-process and multi-resource collaborative production characteristics of manufacturing industries such as clothing.

[0048] Step S203: The multi-objective constraint optimization model is solved using a preset two-layer optimization strategy, the optimal production scheduling plan is output, and the optimal production scheduling plan is dynamically adjusted according to the real-time production status.

[0049] Specifically, a pre-defined two-layer optimization strategy is used to calculate and solve the constructed multi-objective constrained optimization model to obtain the optimal production scheduling plan. At the same time, the real-time production status is continuously monitored, and the optimal production scheduling plan is flexibly adjusted according to the difference between the actual production situation and the plan to ensure that the scheduling plan adapts to the production dynamics.

[0050] The speed-priority product production scheduling method provided in this embodiment comprehensively collects real-time production and multi-source order data, constructs a multi-objective constrained optimization model with the shortest order delivery time as the core objective, and takes into account key factors such as equipment utilization and delayed orders to achieve a balance between speed and efficiency. Combined with a preset two-layer optimization strategy, it generates an initial optimal plan through global optimization and dynamically adjusts based on real-time production status, effectively responding to dynamic disturbances such as order insertions and equipment failures. Ultimately, it improves the accuracy, adaptability, and order delivery efficiency of production scheduling, meeting the needs of production scenarios requiring rapid market response. This invention focuses on minimizing order delivery time while considering multi-objective optimization of equipment utilization and production costs, and can dynamically adjust based on real-time production data, solving the problems of slow response speed, untimely order delivery, and weak dynamic adjustment capabilities in existing garment production scheduling technologies.

[0051] This embodiment provides a speed-priority product production scheduling method, which can be used in product production lines. Figure 3 This is a flowchart of a speed-priority product production scheduling method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S301: Collect real-time production data and multi-source order data from the product production line.

[0052] Specifically, step S301 includes: Step S3011: Use a variety of IoT sensors to collect real-time production data of the product production line. The real-time production data includes: status sensor data of garment hanging production equipment, RFID tag or unique code data, and work reporting information scanned by operators outside the hanging line.

[0053] Among them: 1) Equipment status sensor: collects information on the operating status, working efficiency and fault information of production equipment.

[0054] 2) RFID (Radio Frequency Identification) Tags: Tracking the flow of materials and work-in-process. The core function of RFID tags is to track the flow of materials and work-in-process in real time. For example, in garment production, RFID tags can be attached to fabric rolls, cut pieces, and finished garments. RFID readers in the workshop collect tag data in real time to obtain information such as the current process of the fabric (e.g., cutting / sewing area), the processing progress of the cut pieces (e.g., sewn completed / awaiting ironing), and the flow path of the finished garment (e.g., after quality inspection and awaiting packaging). This data is synchronized to the production scheduling system, providing accurate data support for material supply time and secondary processing constraints, as well as process progress monitoring, avoiding process stagnation or scheduling deviations caused by opaque material flow information.

[0055] 3) Intelligent personnel terminal: collects personnel location, skill level, and work efficiency information.

[0056] Step S3012: Collect multi-source order data from the supply chain management software. The multi-source order data includes order data, material BOM data, inventory data, and process route data.

[0057] Supply chain management software includes ERP (Enterprise Resource Planning) systems, MES (Manufacturing Execution System) systems, and WSM (Warehouse Management System) systems. Order data, BOM (Bill of Materials) data, inventory data, and process route data are obtained from supply chain management software (ERP, MES, WMS, etc.).

[0058] It should be noted that all collected real-time production data and multi-source order data are cleaned, denoised, and standardized before being stored in a time-series database for use by the algorithm.

[0059] Step S302: Based on real-time production data and multi-source order data, construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective.

[0060] Specifically, step S302 includes: Step S3021: Based on real-time production data and multi-source order data, determine multiple constraints; the multiple constraints include: logical relationship constraints between processes, equipment resource capacity constraints, personnel skill matching constraints, material supply time and secondary processing constraints, and order priority constraints.

[0061] in: 1) Logical relationship constraints between processes: S ij ≥C i (h)+T_transport / / The start time of process j is greater than or equal to the completion time of the preceding process h plus the material transfer time; 2) Equipment resource capacity constraint: Σxijk≤1 / / One piece of equipment can only process one process at a time.

[0062] 3) Personnel skill matching constraint: If operator k does not have the skill to operate equipment j, then xijk=0.

[0063] 4) Material supply time and secondary processing constraints: S ij≥T_material_ready / / The process start time must be later than the material preparation time.

[0064] 5) Order priority constraint: If the priority of order i is higher than that of order j, then Ci ≤ Cj.

[0065] Among them, S ij C represents the start time of process j in order i. i (h) represents the completion time of the preceding process h of order i, T_transport represents the material transfer time, T_material_ready represents the material preparation time, and xijk is a 0-1 variable indicating whether process j of order i is assigned to equipment k.

[0066] Step S3022: Based on multiple constraints, construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective.

[0067] Specifically, the multi-objective constrained optimization model with the goal of minimizing order delivery time is expressed by the following formula: MinimizeF(X)=[f1(X),f2(X),f3(X)](1); Where, MinimizeF(X)=[f1(X),f2(X),f3(X)] represents minimizing the maximum completion time, minimizing the weighted delay time, and maximizing the equipment utilization; f1(X)=max(C i f1(X) is the maximum completion time, and f2(X) = Σw i T i f2(X) is the weighted delay time, f3(X) = -U, f3(X) is the maximum equipment utilization rate, and C i T is the completion time of order i. i w is the delay time of order i. i is the priority weight of order i, and U is the average equipment utilization rate.

[0068] Step S303 involves solving the multi-objective constrained optimization model using a pre-defined two-layer optimization strategy, outputting the optimal production scheduling plan, and dynamically adjusting the optimal production scheduling plan based on real-time production status. For details, please refer to [link to relevant documentation]. Figure 2 Step S203 of the illustrated embodiment will not be described again here.

[0069] The speed-priority product production scheduling method provided in this embodiment is based on real and comprehensive real-time production data and multi-source order data, ensuring that the model construction does not deviate from the actual production scenario, and laying a data foundation for the feasibility of subsequent scheduling schemes. In addition, the model clearly defines the shortest order delivery time as the core optimization objective, and at the same time takes into account key production indicators such as equipment utilization and personnel efficiency through multiple constraints, forming a speed-priority, multi-dimensional balanced optimization orientation, providing a scientific and rigorous framework for solving the optimal production scheduling plan, and ensuring that the final output scheduling scheme can focus on core delivery requirements.

[0070] This embodiment provides a speed-priority product production scheduling method, which can be used in product production lines. Figure 4 This is a flowchart of a speed-priority product production scheduling method according to an embodiment of the present invention, such as... Figure 4 As shown, the process includes the following steps: Step S401: Collect real-time production data and multi-source order data from the product production line. For details, please refer to [link / reference]. Figure 3 Step S301 of the illustrated embodiment will not be described again here.

[0071] Step S402: Based on real-time production data and multi-source order data, construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective. For details, please refer to [link to details]. Figure 3 Step S302 of the illustrated embodiment will not be described again here.

[0072] Step S403: The multi-objective constraint optimization model is solved using a preset two-layer optimization strategy, the optimal production scheduling plan is output, and the optimal production scheduling plan is dynamically adjusted according to the real-time production status.

[0073] Specifically, the preset two-layer optimization strategy is a two-layer optimization strategy that combines an improved adaptive genetic algorithm with reinforcement learning components; the above step S403 includes: Step S4031: The improved adaptive genetic algorithm is used to solve the multi-objective constraint optimization model to generate an initial global optimal production scheduling plan.

[0074] In some optional implementations, step S4031 above includes: Step a1: Determine the dual-layer coding based on the process characteristics and equipment type of product manufacturing. The dual-layer coding includes process coding and equipment coding.

[0075] Coding Design: A two-tiered coding system using process coding and equipment coding is adopted. Process coding uses real-number coding based on process sequence, while equipment coding uses integer coding based on equipment allocation.

[0076] 1) Process coding: The sequence of processes is represented by a real number sequence (e.g., [2-4-1-3] represents the order process execution sequence), and the garment order is broken down into process sequences such as cutting, sewing, ironing, and quality inspection.

[0077] Example: A shirt order code [2-4-1-3] means: 2 (cutting) → 4 (embroidery) → 1 (sewing) → 3 (buttoning).

[0078] 2) Equipment coding: Integer sequences represent the mapping between processes and equipment (e.g., [3-1-2] indicates that process 1 is processed by equipment 3). The dual coding evolves independently, avoiding the search space limitations of a single coding.

[0079] Associate with garment-specific equipment (such as automatic cutting bed code 1, high-speed sewing machine code 3).

[0080] Example: Equipment code [3-1-2] means: the cutting process is performed by automatic cutting bed No. 1, and the sewing is performed by high-speed sewing machine No. 3.

[0081] Step a2: Calculate the fitness value for each code based on the multi-objective constrained optimization model.

[0082] Specifically, the following fitness function is designed, which is linked to production indicators, to evaluate the performance of individuals: Fitness=1 / [α1×f1(X)+α2×f2(X)-α3×f3(X)](2); Where α1, α2, and α3 are weighting coefficients, and the garment order characteristics are adjusted according to the actual needs of the enterprise, as shown in Table 1 below. f1(X)=max(C i f1(X) is the maximum completion time, and f2(X) = Σw i T i f2(X) is the weighted delay time, f3(X) = -U, f3(X) is the maximum equipment utilization rate, and C i T is the completion time of order i. i w is the delay time of order i. i is the priority weight of order i, and U is the average equipment utilization rate.

[0083] Table 1 Weighting Coefficients

[0084] For example: Detailed regulations for the regulation of the apparel industry: 1) α1 (Delivery time core coefficient): Technical effect: When f1(X) (maximum completion time) increases, the fitness value decreases significantly, forcing the algorithm to prioritize compressing the critical path time.

[0085] Industry logic: During peak seasons / promotional seasons for apparel, improve α1 to ensure the delivery of best-selling products.

[0086] #Peak season parameter setting rules are as follows: if is_peak_season(order_data): α1=0.7 # This is a 16.7% increase from the base value of 0.6.

[0087] Among them, peak_season represents the peak season.

[0088] Example as follows: An e-commerce apparel company set α1=0.75 during the "Double 11" shopping festival. The longest delivery time for coat orders has been reduced from 72 hours to 58 hours, and the out-of-stock rate for popular items has decreased by 12.3%.

[0089] 2) α2 (delay penalty coefficient): Technical function: To impose stronger delay penalties on high-priority orders (w) i (With nested weights), VIP order delays will significantly reduce adaptability.

[0090] Industry logic: High-end customer orders require special protection, for example: #VIP order weight amplification mechanism: if order.priority=='VIP': wi=base_weight×3.0 # Regular orders have a weight of 3 times.

[0091] Here, order.priority represents the order priority.

[0092] Taking a custom suit company as an example, the weight of VIP orders was set to wi=3.0. During the genetic algorithm iteration, a temporary inferior solution was accepted through simulated annealing mechanism (probability of 0.22). In the end, the order delay rate of gold card customers decreased from 15.2% to 3.7%, and the equipment utilization rate increased to 89%.

[0093] 3) α3 (Equipment Utilization Coefficient): Technical function: The negative sign increases adaptability when equipment utilization is improved, but we should be wary of over-optimization leading to a deterioration in delivery time.

[0094] Industry logic: Proactive optimization for high-depreciation equipment: #Dynamically adjust based on equipment type if equipment_type=='Automatic Cutting Bed': α3=0.25 #Enhance the utilization of high-cost equipment.

[0095] elif equipment_type=='Ordinary sewing machine': α3=0.08#Inexpensive equipment weakens constraints.

[0096] Example as follows: A denim clothing factory sets α3=0.28 for its laser cutting machine (¥1.5 million / unit). Utilization rate increased from 68% to 87%; depreciation cost per unit decreased by ¥0.38.

[0097] The weighting coefficient adjustment mechanism and examples are shown in Table 2 below: Table 2 Weighting coefficient adjustment mechanism and cases

[0098] For example, the dynamic adjustment mechanism of the α coefficient is designed specifically for the needs of the apparel industry: 1) Equipment differences: Woven equipment α3=0.18 (reduced downtime for cleaning), knitting equipment α3=0.22 (circular knitting machine operates continuously).

[0099] 2) Seasonal strategy library: Establish a rule library to map production seasons to coefficients (adjust α1 during peak season and α3 during off-season).

[0100] Step a3: Based on the fitness value of each code, a new generation of population is generated through adaptive operations of selection, crossover, and mutation.

[0101] Specifically, the selection process employs a tournament selection method and an elite retention strategy to ensure that outstanding individuals are preserved for the next generation. The detailed process can be found in relevant technical documents and will not be elaborated upon here.

[0102] Crossover operation: Using a method similar to OX crossover, an adaptive crossover rate is designed. Pc=Pcmax-(Pcmax-Pcmin)×(g / G)(3); Where Pcmax and Pcmin are the maximum and minimum crossover rates, g is the current iteration number, and G is the total number of iterations.

[0103] Crossover rate: Pc = 0.9 - 0.5 × g / G (high initial exploration, convergence later).

[0104] Mutation rate: Pm = 0.1 + 0.3 × g / G (prevents premature maturity in the early stage and maintains diversity in the later stage).

[0105] This allows us to break free from the limitations of fixed parameters and achieve a dynamic balance between exploration and development.

[0106] Mutation operation: An adaptive mutation rate is designed by combining exchange mutation and inversion mutation. Pm=Pmmin+(Pmmax-Pmmin)×(g / G) (4).

[0107] Step a4: Use simulated annealing to perform global search optimization on the new generation population. When the preset iteration termination condition is met, output the initial global optimal production scheduling plan.

[0108] In simulated annealing, when the objective function value (fitness) of the new solution is worse than that of the old solution (i.e., f... new >f old The algorithm accepts the inferior solution with a certain probability to avoid getting trapped in a local optimum. The probability of accepting an inferior solution is... Calculated using the following formula: (5); Where: f new f represents the objective function value (fitness value) of the new solution. old This represents the objective function value of the current solution (old solution); T is the current temperature parameter. This formula shows that when the temperature T is high, the probability of accepting a suboptimal solution is greater, and as the temperature decreases, the probability of accepting a suboptimal solution gradually decreases.

[0109] Temperature decay formula: The temperature parameter T decays (cools down) after each generation of evolution in the genetic algorithm according to the following formula: T= × (6); in: is the initial temperature; g is the current iteration number (g-th generation); 0.95 is the decay coefficient (i.e., a 5% temperature decrease per generation). This formula indicates that the temperature decays exponentially with the number of iterations, and with each iteration, the temperature drops to 95% of the previous generation's temperature.

[0110] In simulated annealing, the physical meaning of temperature T is the tolerance for accepting suboptimal solutions: at high temperatures (when g is small, T is close to T0), the algorithm is more likely to accept worse solutions, thus escaping local optima and achieving global exploration; as the number of iterations g increases, temperature T increases by 0.95. g As the exponential growth rate gradually decreases, the probability of accepting poor solutions decreases, and the algorithm gradually converges to a better local solution, eventually stabilizing near the globally better solution, thus avoiding the local optimum trap of traditional genetic algorithms.

[0111] Step S4032: Based on the initial global optimal production scheduling plan, a reinforcement learning component is used in conjunction with the product production scenario to output the final optimal production scheduling plan.

[0112] In some optional implementations, the reinforcement learning component employs a dual deep Q-learning network architecture, the reinforcement learning component including a state space, an action space, and a reward function; step S4032 above includes: Step b1: Construct the state space, action space, and reward function of the reinforcement learning component based on real-time production data and multi-source order data.

[0113] Specifically, to enhance the algorithm's dynamic adjustment capability, this embodiment integrates a deep reinforcement learning component, the structure of which is as follows: 1. Status space: including equipment status (idle, busy, faulty), order progress (percentage completed, remaining processes), and personnel status (work efficiency, fatigue level).

[0114] Action space: including order allocation (which equipment to assign an order to), equipment scheduling (equipment work plan), and personnel allocation (personnel allocation and adjustment).

[0115] Reward Function: Design a multi-objective reward function, as shown in the following formula: (7); Where β1=0.5, β2=0.3, β3=0.2, and β4=0.2. Cmax_prev and Cmax_curr are the maximum completion times before and after adjustment, ΣwT_prev and ΣwT_curr are the weighted delay times before and after adjustment, and Uprev and Ucurr are the average equipment utilization rates before and after adjustment. N delay is the number of newly generated delayed orders, and β1, β2, β3, and β4 are all weighting coefficients.

[0116] Network structure: A Double Deep Q-Network (DDQN) structure is adopted, which includes a main network and a target network to reduce the problem of overestimation of Q value.

[0117] Implementation details of reinforcement learning components: 1) Definition of state space: Equipment status: {Idle, In Process, Faulty} + Remaining working hours; Order progress: {Number of completed processes, estimated time for remaining processes}; Personnel Status: {Skill Level, Real-time Efficiency Coefficient}; The state space uses multi-dimensional state modeling to cover production dynamics.

[0118] To enhance the forward-looking nature of state awareness, this embodiment integrates an LSTM-based device idle prediction model into the state space of the reinforcement learning component. This model dynamically predicts the future idle time of each device by analyzing the current task progress, historical performance data, and material flow. This upgrades the state space from a static "device busy" to a dynamic "device expected to be idle in X minutes," enabling the reinforcement learning component to perform forward-looking scheduling (e.g., pre-allocating urgent orders before devices become idle). This achieves a leap from passive response to proactive optimization, significantly reducing inter-process waiting time and further ensuring order delivery speed.

[0119] 2) Motion space design: Order allocation: Select orders and target devices from the pool of orders to be allocated; Equipment scheduling: Adjusting equipment work plans (such as inserting orders or rearranging work processes); Personnel allocation: Allocate suitable skilled personnel across work processes; The motion space uses three layers of motion linkage to achieve resource synergy optimization.

[0120] 3) Reward function design: (8); The reward function adopts a multi-objective quantitative reward, namely shortening the longest construction period, reducing the weighted delay time, improving equipment utilization, and penalizing delayed orders, driving the strategy to converge toward the core indicators.

[0121] Step b2: Based on the state space, a dual-deep Q-learning network architecture is used to output the optimal adjustment action.

[0122] Specifically, based on a main network + target network architecture using a Double Deep Q-Network (DDQN), the optimal adjustment action is output according to the real-time state of the state space (covering triple linkage of orders, equipment, and personnel): Order allocation: Select orders from the pool of pending orders (such as urgent children's clothing orders) and match them to idle devices (such as an automatic cutting bed that will be idle in 8.2 minutes). Equipment scheduling: Adjust the equipment work plan (e.g., reassign the work processes originally assigned to the faulty equipment to the standby high-speed sewing machine). Personnel allocation: Allocate suitable skilled personnel across processes (e.g., temporarily assign A-level sewing workers to bottleneck processes to improve sewing efficiency).

[0123] Step b3: Based on the optimal adjustment action and reward function, calculate the reward value of the multi-objective constrained optimization model, optimize the optimal adjustment action based on the reward value, and take the optimized optimal adjustment action as the final optimal production scheduling plan.

[0124] Specifically, the optimal adjustment action is substituted into the multi-objective reward function to calculate the reward value, and the DDQN (Double Deep Q-Network) network parameters are updated in reverse to optimize subsequent adjustment strategies. Strategy update: If the reward value is positive (e.g., delivery time is shortened by 2 hours, utilization rate is increased by 5%), the current action strategy is strengthened; if the reward value is negative (e.g., 3 new delayed orders are added), the network parameters are corrected to avoid similar adjustments.

[0125] Step S4033: Based on the deviation between the actual production progress and the planned production progress, dynamically adjust the optimal production scheduling plan.

[0126] Specifically, through real-time data monitoring, a rescheduling mechanism is triggered when the deviation between actual production progress and planned progress exceeds a threshold. For example: By collecting real-time actual production progress data (such as the actual completion rate of each process, the time spent, and the status of work-in-process), and comparing it with the planned progress in the preset optimal production scheduling plan (such as the planned start / end time of the process and the planned capacity), the deviation value is calculated (such as the actual progress of a process lagging behind the plan by 2 hours, the actual load of a certain equipment exceeding the plan by 10%), and the deviation type is identified (such as process delay, sudden equipment failure, material supply delay, personnel efficiency fluctuation, etc.).

[0127] Secondly, based on the degree (minor / serious) and cause of the deviation, and in conjunction with production constraints (such as the logical sequence of processes, equipment capacity limits, order priorities, and material availability), targeted adjustment strategies should be formulated: If it is a minor deviation (such as a process being 30 minutes behind schedule), the deviation can be reduced by fine-tuning the start time of subsequent processes and optimizing personnel allocation (such as allocating idle personnel to provide support); if it is a serious deviation (such as a process being halted for more than 2 hours due to a critical equipment failure), in-depth adjustments need to be initiated, such as using backup equipment, adjusting the order production sequence (prioritizing high-priority orders), coordinating expedited material supply, and extending effective working time (such as temporary overtime), to prevent the deviation from being transmitted to subsequent processes.

[0128] Finally, the adjusted scheduling plan is synchronized to the production execution system, and the adjustment effect is continuously monitored: if the deviation is reduced to an acceptable range (e.g., lag time less than 15 minutes), the adjusted plan is maintained; if there is still a significant deviation after adjustment, the above perception-analysis-adjustment process is repeated until the production schedule returns to the planned track or a new optimal plan is formed. The entire process relies on real-time data feedback and constraint verification to ensure that the adjustment actions conform to the actual production rules and can quickly respond to emergencies, avoiding order delivery delays caused by the disconnect between static planning and dynamic production.

[0129] This embodiment also provides a device idle prediction method: namely, the prediction model based on an LSTM network described above, with input dimensions including: #Feature Engineering Input Dimensions (Apparel Industry Specific): features=[current_stitch / total_stitch, #percentage of remaining stitches in the current process; fabric_type, # Fabric type (0: knitted, 1: woven); equipment_maintenance_countdown # Equipment maintenance countdown (hours) # Data source: MES system real-time messages.

[0130] Prediction example: The sewing machine is predicted to complete the current 2000 stitches of jeans in 15.3 minutes.

[0131] Real-time order insertion action execution: When the DDQN network detects that the device idle window is greater than 5 minutes: Action instruction: Insert the urgent order into the idle device.

[0132] Case study in the apparel industry: A fast fashion company's order placement scenario, such as... Figure 5 As shown.

[0133] Status: The automatic cutting bed has been detected to complete its current task in 8.2 minutes.

[0134] Action: Insert 50 pieces of printed T-shirt fabric that urgently need to be reworked.

[0135] Result: The delay time for this batch of orders was reduced by 3.2 hours.

[0136] The speed-priority product production scheduling method provided in this embodiment achieves complementary advantages through a two-layer collaborative architecture of improved adaptive genetic algorithm + reinforcement learning component. It effectively solves the problems of insufficient global optimization or lag in dynamic response in traditional single algorithm scheduling, and significantly improves the scientificity, flexibility and reliability of production scheduling.

[0137] As one or more specific application embodiments of the present invention, the speed-priority product production scheduling method provided by the present invention will be further described in detail, specifically including the following steps: (1) Data collection and processing: Real-time production data and multi-source order data of the garment production line are collected through IoT sensors and supply chain management software, including information such as equipment status, personnel skill level, process time, order urgency, and material supply. The data is then cleaned and standardized.

[0138] This includes: collecting real-time data from the production line through various IoT sensors, including: 1) Equipment status sensors: collect equipment operating status, working efficiency, and fault information; 2) RFID tags: used to track the movement of materials and work-in-process; 3) Intelligent personnel terminal: Collects personnel location, skill level, and work efficiency information; Simultaneously, order data, material data, and process route data are obtained from supply chain management software (ERP, MES, WMS, etc.). All collected data is cleaned, denoised, and standardized before being stored in a time-series database for algorithm use.

[0139] (2) Modeling: Construct a multi-objective constrained optimization model with the shortest order delivery time as the primary optimization objective. The model considers the following constraints: logical relationship constraints between processes, equipment resource capacity constraints, personnel skill matching constraints, material supply time and secondary processing constraints, and order priority constraints.

[0140] in: 1) Logical relationship constraints between processes: S ij ≥C i (h)+T_transport / / The start time of process j is greater than or equal to the completion time of the preceding process h plus the material transfer time; 2) Equipment resource capacity constraint: Σxijk≤1 / / One piece of equipment can only process one process at a time.

[0141] 3) Personnel skill matching constraint: If operator k does not have the skill to operate equipment j, then xijk=0.

[0142] 4) Material supply time and secondary processing constraints: S ij ≥T_material_ready / / The process start time must be later than the material preparation time.

[0143] 5) Order priority constraint: If the priority of order i is higher than that of order j, then Ci ≤ Cj.

[0144] Among them, S ij C represents the start time of process j in order i. i (h) represents the completion time of the preceding process h of order i, T_transport represents the material transfer time, T_material_ready represents the material preparation time, and xijk is a 0-1 variable indicating whether process j of order i is assigned to equipment k.

[0145] The multi-objective constrained optimization model with the goal of minimizing order delivery time is expressed by the following formula: MinimizeF(X)=[f1(X),f2(X),f3(X)](1); Where, MinimizeF(X)=[f1(X),f2(X),f3(X)] represents minimizing the maximum completion time, minimizing the weighted delay time, and maximizing the equipment utilization; f1(X)=max(C i f1(X) is the maximum completion time, and f2(X) = Σw i T i f2(X) is the weighted delay time, f3(X) = -U, f3(X) is the maximum equipment utilization rate, and C i T is the completion time of order i. i w is the delay time of order i. i is the priority weight of order i, and U is the average equipment utilization rate.

[0146] (3) Algorithm Solution: A two-layer optimization strategy combining an improved adaptive genetic algorithm and a reinforcement learning component is employed for the solution. The improved adaptive genetic algorithm includes: a two-layer encoding method using process encoding and equipment encoding; an adaptive crossover rate and mutation rate adjustment mechanism; the introduction of simulated annealing to avoid local optima; and the use of an elite retention strategy to ensure algorithm convergence. The reinforcement learning component uses a deep Q-learning network. The state space includes equipment state, order progress, and personnel state, while the action space includes order allocation, equipment scheduling, and personnel deployment. The reward function is designed based on order completion time, equipment utilization rate, and the number of delayed orders.

[0147] (4) Dynamic adjustment: Through real-time data monitoring, when the deviation between the actual production progress and the plan exceeds the threshold, a rescheduling mechanism is triggered.

[0148] This invention also provides a system applying the aforementioned speed-priority product production scheduling method, comprising: a data acquisition module, an algorithm engine module, a visualization module, and a system integration interface. The system integration interface provides RESTful API services, enabling data exchange and integration with enterprise ERP (Enterprise Resource Planning), MES (Manufacturing Execution System), and WMS (Warehouse Management System) systems. Wherein: 1) Data acquisition module: Collects production line data in real time through Internet of Things (IoT) technology.

[0149] 2) Algorithm Engine Module: The core algorithm module, which implements the above-mentioned optimization algorithms.

[0150] 3) Visualization module: Displays scheduling results and production progress using Gantt charts, dashboards, and other methods.

[0151] 4) System integration interface: Provides RESTful API services for integration with existing enterprise systems.

[0152] The system adopts a microservice architecture, where each module can be deployed and expanded independently, improving system flexibility and maintainability.

[0153] This invention also provides a production scheduling method for garment manufacturing enterprises that applies a speed-priority product production scheduling method, such as... Figure 6 As shown, the method includes: Step S601: Receive multi-source order data from garment manufacturing enterprises.

[0154] Step S602: Based on the multi-source order data of garment manufacturing enterprises, run the speed-first product scheduling method described above and output the optimal production scheduling plan.

[0155] Step S603: Generate a visual scheduling scheme based on the optimal production scheduling plan, and dynamically adjust the optimal production scheduling plan by monitoring the execution status of garment manufacturing enterprises.

[0156] The specific process of the above embodiments is the same as that of the speed-priority product production scheduling method, and will not be described again here.

[0157] To verify the effectiveness of the method of this invention, a simulation experiment was conducted on actual production data from a garment enterprise. The experimental results show that, compared with the method of this invention, by improving the double-layer encoding of the genetic algorithm to expand the search space and combining adaptive parameters to find a better process sequence; the real-time order insertion mechanism of reinforcement learning compresses equipment idle time, reducing order completion time by 18.61%; and by using the reinforcement learning component to dynamically trigger equipment rescheduling through state awareness (such as equipment idle prediction), blocking waiting time is reduced by 30.97%. The reward function includes a penalty term for delayed orders (…). 0.2 N The delay-based forced strategy prioritizes urgent orders. The following table (Table 3) shows a performance comparison of the algorithms obtained through experiments: Table 3 Algorithm Performance Comparison Table 1

[0158] Note: The test environment in Table 3 above is: Intel Xeon Gold 6248R / 256GB RAM, which complies with GB / T39848-2021 standard.

[0159] In summary, the performance comparison between the speed-priority product production scheduling method provided by this invention and the traditional genetic algorithm is shown in Table 4 below: Table 4 Algorithm Performance Comparison Table 2

[0160] The speed-priority product production scheduling method provided in this embodiment first collects real-time production data and multi-source order data from the garment production line; then, it constructs a multi-objective constrained optimization model with the shortest order delivery time as the primary optimization objective; the improved adaptive genetic algorithm includes a simulated annealing mechanism, where the temperature decay formula is T=T0×0.95. g (g is the number of iterations), the probability of accepting a suboptimal solution is The reinforcement learning method employs a dual-depth Q-network, with an action space encompassing triple linkages of order allocation, equipment scheduling, and personnel deployment. It ultimately outputs an optimal production scheduling plan, dynamically adjusted based on real-time production status. This method can be widely applied to production scheduling systems in various garment manufacturing enterprises, and is particularly suitable for personalized custom garment production environments characterized by small batches, diverse varieties, and short lead times. After implementing this invention, enterprises can significantly improve on-time order delivery rates, reduce production cycles, and increase equipment utilization and personnel productivity.

[0161] This invention has been piloted in apparel companies, and the results show that the average order completion time has been shortened by 25.8%, the proportion of delayed orders has decreased by 60.3%, and the equipment utilization rate has increased by 15.6%, achieving significant economic benefits. It can effectively shorten order completion time, reduce the proportion of delayed orders, and improve equipment utilization rate, and is particularly suitable for the production environment of personalized custom-made apparel with small batches, multiple varieties, and short delivery times.

[0162] This embodiment also provides a speed-priority product production scheduling device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0163] This embodiment provides a speed-priority product production scheduling device, such as... Figure 7 As shown, it includes: The real-time production data and multi-source order data acquisition module 701 is used to collect real-time production data and multi-source order data from the product production line.

[0164] The multi-objective constrained optimization model construction module 702 is used to construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective, based on real-time production data and multi-source order data.

[0165] The multi-objective constrained optimization model solving and dynamic adjustment module 703 is used to solve the multi-objective constrained optimization model using a preset two-layer optimization strategy, output the optimal production scheduling plan, and dynamically adjust the optimal production scheduling plan according to the real-time production status.

[0166] In some alternative implementations, the real-time production data and multi-source order data acquisition module 701 includes: The real-time production data acquisition unit is used to collect real-time production data of the product production line using a variety of IoT sensors. The real-time production data includes: status sensor data of garment hanging production equipment, RFID tag or unique code data, and work reporting information scanned by operators outside the hanging line.

[0167] The multi-source order data acquisition unit is used to collect multi-source order data from supply chain management software. Multi-source order data includes order data, material BOM data, inventory data, and process route data.

[0168] In some optional implementations, the multi-objective constrained optimization model construction module 702 includes: The multi-constraint determination unit is used to determine multiple constraints based on real-time production data and multi-source order data. The multiple constraints include: logical relationship constraints between processes, equipment resource capacity constraints, personnel skill matching constraints, material supply time and secondary processing constraints, and order priority constraints.

[0169] The multi-objective constrained optimization model building unit is used to construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective, based on multiple constraints.

[0170] In one alternative implementation, the multi-objective constrained optimization model with the goal of minimizing order delivery time is expressed by the following formula: MinimizeF(X)=[f1(X),f2(X),f3(X)]; Where, MinimizeF(X)=[f1(X),f2(X),f3(X)] represents minimizing the maximum completion time, minimizing the weighted delay time, and maximizing the equipment utilization; f1(X)=max(C i f1(X) is the maximum completion time, and f2(X) = Σw i T i f2(X) is the weighted delay time, f3(X) = -U, f3(X) is the maximum equipment utilization rate, and C i T is the completion time of order i. i w is the delay time of order i. i is the priority weight of order i, and U is the average equipment utilization rate.

[0171] In one optional implementation, the preset two-layer optimization strategy is a two-layer optimization strategy combining an improved adaptive genetic algorithm and a reinforcement learning component; the multi-objective constrained optimization model solving and dynamic adjustment module 703 includes: The model solving unit is used to solve the multi-objective constrained optimization model using an improved adaptive genetic algorithm to generate an initial global optimal production scheduling plan.

[0172] The reinforcement learning unit is used to output the final optimal production scheduling plan based on the initial global optimal production scheduling plan and by combining reinforcement learning components with the product production scenario.

[0173] The dynamic adjustment unit is used to dynamically adjust the optimal production scheduling plan based on the deviation between the actual production progress and the planned production progress.

[0174] In one optional implementation, the model solving unit includes: The dual-layer coding determines the sub-unit and is used to determine the dual-layer coding based on the process characteristics and equipment type of product manufacturing. The dual-layer coding includes process coding and equipment coding.

[0175] The fitness value calculation subunit is used to calculate the fitness value of each code based on the multi-objective constrained optimization model.

[0176] The adaptive operation subunit is used to generate a new generation of population based on the fitness value of each encoding through adaptive operations of selection, crossover, and mutation.

[0177] The global search subunit is used to perform global search optimization on the new generation population using simulated annealing. When the preset iteration termination condition is met, it outputs the initial global optimal production scheduling plan.

[0178] In one optional implementation, the reinforcement learning component employs a dual deep Q-learning network architecture, the reinforcement learning component including a state space, an action space, and a reward function; the reinforcement learning unit includes: The reinforcement learning component construction subunit is used to construct the state space, action space, and reward function of the reinforcement learning component based on real-time production data and multi-source order data.

[0179] The optimal adjustment action output subunit is used to output the optimal adjustment action based on the state space and employing a dual-deep Q-learning network architecture.

[0180] The optimal adjustment action optimization subunit is used to calculate the reward value of the multi-objective constrained optimization model based on the optimal adjustment action and the reward function, and optimize the optimal adjustment action based on the reward value, and use the optimized optimal adjustment action as the final optimal production scheduling plan.

[0181] This embodiment also provides a production scheduling device for a garment manufacturing enterprise that applies a speed-priority product production scheduling method. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0182] This embodiment provides a production scheduling device for a garment manufacturing enterprise that applies a speed-priority product production scheduling method, such as... Figure 8 As shown, it includes: The multi-source order data receiving module 801 is used to receive multi-source order data from garment manufacturing enterprises. The speed-first product scheduling module 802 is used to run the speed-first product scheduling method based on multi-source order data of garment manufacturing enterprises and output the optimal production scheduling plan.

[0183] The production plan dynamic adjustment module 803 is used to generate a visual scheduling scheme based on the optimal production scheduling plan and to monitor the execution of the garment manufacturing enterprise to dynamically adjust the optimal production scheduling plan.

[0184] The speed-priority product production scheduling device provided in this embodiment of the invention can execute the speed-priority product production scheduling method provided in any embodiment of the invention. The garment manufacturing enterprise production scheduling device provided in this embodiment of the invention can execute the garment manufacturing enterprise production scheduling method applying the speed-priority product production scheduling method provided in any embodiment of the invention, possessing the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0185] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0186] The following is a detailed reference. Figure 9 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 901, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 902 or a program loaded from memory 908 into random access memory (RAM) 903. RAM 903 also stores various programs and data required for the operation of the electronic device. The processor 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0187] Typically, the following devices can be connected to I / O interface 905: input devices 906 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 907 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 908 including, for example, magnetic tapes, hard disks, etc.; and communication devices 909. Communication device 909 allows electronic devices to exchange data via wireless or wired communication with other devices. Although Figure 9 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0188] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 909, or installed from a memory 908, or installed from a ROM 902. When the computer program is executed by the processor 901, it performs the functions defined in the speed-priority product production scheduling method of the embodiments of the present invention and the above-described production scheduling method for garment manufacturing enterprises applying the speed-priority product production scheduling method.

[0189] Figure 9 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0190] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the speed-priority product production scheduling method shown in the above embodiments and the garment manufacturing enterprise production scheduling method applying the speed-priority product production scheduling method described above are implemented.

[0191] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0192] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A speed-priority product production scheduling method, characterized in that, The method includes: Collect real-time production data from the product production line and multi-source order data; Based on the real-time production data and multi-source order data, a multi-objective constrained optimization model is constructed with the shortest order delivery time as the optimization objective. The multi-objective constraint optimization model is solved using a preset two-layer optimization strategy, and the optimal production scheduling plan is output. The optimal production scheduling plan is then dynamically adjusted based on the real-time production status.

2. The speed-priority product production scheduling method according to claim 1, characterized in that, The collected real-time production data and multi-source order data from the product production line include: Multiple IoT sensors are used to collect real-time production data from the product production line. The real-time production data includes: status sensor data of garment hanging production equipment, RFID tag or unique code data, and work reporting information scanned by operators outside the hanging line. Multi-source order data is collected from enterprise supply chain management software. The multi-source order data includes order data, material BOM data, inventory data, and process route data.

3. The speed-priority product production scheduling method according to claim 1, characterized in that, Based on the aforementioned real-time production data and multi-source order data, a multi-objective constrained optimization model is constructed with the goal of minimizing order delivery time, including: Based on the real-time production data and multi-source order data, multiple constraints are determined; these multiple constraints include: logical relationship constraints between processes, equipment resource capacity constraints, personnel skill matching constraints, material supply time and secondary processing constraints, and order priority constraints. Based on the aforementioned multiple constraints, a multi-objective constrained optimization model is constructed with the goal of minimizing order delivery time.

4. The speed-priority product production scheduling method according to claim 3, characterized in that, The multi-objective constrained optimization model with the goal of minimizing order delivery time is expressed by the following formula: MinimizeF(X)=[f1(X),f2(X),f3(X)]; Where, MinimizeF(X)=[f1(X),f2(X),f3(X)] represents minimizing the maximum completion time, minimizing the weighted delay time, and maximizing the equipment utilization; f1(X)=max(C i f1(X) is the maximum completion time, and f2(X) = Σw i T i f2(X) is the weighted delay time, f3(X) = -U, f3(X) is the maximum equipment utilization rate, and C i T is the completion time of order i. i It is the delay time of order i, w i is the priority weight of order i, and U is the average equipment utilization rate.

5. The speed-priority product production scheduling method according to claim 1, characterized in that, The preset two-layer optimization strategy is a two-layer optimization strategy that combines an improved adaptive genetic algorithm with reinforcement learning components; The process of solving the multi-objective constrained optimization model using a preset two-layer optimization strategy, outputting an optimal production scheduling plan, and dynamically adjusting the optimal production scheduling plan based on real-time production status includes: An improved adaptive genetic algorithm is used to solve the multi-objective constrained optimization model to generate an initial globally optimal production scheduling plan; Based on the initial global optimal production scheduling plan, a reinforcement learning component is used in conjunction with the product production scenario to output the final optimal production scheduling plan. The optimal production scheduling plan is dynamically adjusted based on the deviation between the actual production progress and the planned production progress.

6. The speed-priority product production scheduling method according to claim 5, characterized in that, The step of solving the multi-objective constrained optimization model using an improved adaptive genetic algorithm to generate an initial globally optimal production scheduling plan includes: A two-layer coding system is determined based on the process characteristics and equipment type of product manufacturing, wherein the two-layer coding system includes process coding and equipment coding; Based on the multi-objective constrained optimization model, the fitness value of each code is calculated; Based on the fitness value of each encoding, a new generation of population is generated through adaptive operations of selection, crossover, and mutation; A simulated annealing mechanism is used to perform global search optimization on the new generation population. When the preset iteration termination condition is met, the initial global optimal production scheduling plan is output.

7. The speed-priority product production scheduling method according to claim 5, characterized in that, The reinforcement learning component adopts a dual deep Q-learning network architecture, and the reinforcement learning component includes a state space, an action space, and a reward function. Based on the initial globally optimal production scheduling plan, a reinforcement learning component is used in conjunction with the product production scenario to output the final optimal production scheduling plan, including: The state space, action space, and reward function of the reinforcement learning component are constructed based on real-time production data and multi-source order data. Based on the state space, a dual-deep Q-learning network architecture is used to output the optimal adjustment action; Based on the optimal adjustment action and reward function, the reward value of the multi-objective constrained optimization model is calculated, and the optimal adjustment action is optimized based on the reward value. The optimized optimal adjustment action is then used as the final optimal production scheduling plan.

8. A production scheduling method for a garment manufacturing enterprise applying the speed-priority product production scheduling method according to any one of claims 1 to 7, characterized in that, The method includes: Receive multi-source order data from garment manufacturers; Based on the multi-source order data of the garment manufacturing enterprise, the speed-first product scheduling method according to any one of claims 1 to 7 is run to output the optimal production scheduling plan; A visual scheduling scheme is generated based on the optimal production scheduling plan, and the execution status of garment manufacturing enterprises is monitored to dynamically adjust the optimal production scheduling plan.

9. A speed-priority product production scheduling device, characterized in that, The device includes: The real-time production data and multi-source order data acquisition module is used to collect real-time production data and multi-source order data from the product production line. The multi-objective constrained optimization model construction module is used to construct a multi-objective constrained optimization model with the shortest order delivery time as the optimization objective based on the real-time production data and multi-source order data. The multi-objective constrained optimization model solving and dynamic adjustment module is used to solve the multi-objective constrained optimization model using a preset two-layer optimization strategy, output the optimal production scheduling plan, and dynamically adjust the optimal production scheduling plan according to the real-time production status.

10. A production scheduling device for a garment manufacturing enterprise applying the speed-priority product production scheduling method according to any one of claims 1 to 7, characterized in that, The device includes: The multi-source order data receiving module is used to receive multi-source order data from garment manufacturing enterprises; The speed-priority product scheduling operation module is used to run the speed-priority product scheduling method according to any one of claims 1 to 7 based on the multi-source order data of the garment manufacturing enterprise, and output the optimal production scheduling plan. The production plan dynamic adjustment module is used to generate a visual scheduling scheme based on the optimal production scheduling plan, and to monitor the execution of the garment manufacturing enterprise to dynamically adjust the optimal production scheduling plan.

Images