An intelligent production scheduling optimization system for a cigarette rolling plant

Abstract

The application relates to an intelligent production scheduling optimization system of a rolling workshop, and belongs to the technical field of production scheduling optimization systems.The system comprises a multi-source data fusion module, which is used for integrating order data of a manufacturing execution system, state data of a device Internet of Things platform and auxiliary material inventory data of a warehouse management system; a dynamic constraint modeling module, which is used for establishing a constraint model library comprising device capacity matrix constraints, card changing time rule constraints and auxiliary material coupling equation constraints; a double-chain genetic algorithm engine, which is used for adopting an improved genetic algorithm of process chain and device chain cooperation coding to perform multi-objective optimization of device resource allocation, material supply guarantee and order task scheduling under the condition of meeting the constraint model library, and generate a production scheduling scheme; and a closed-loop feedback executor, which is used for dynamically triggering the double-chain genetic algorithm engine to perform production scheduling scheme re-optimization when detecting that deviation from the current production scheduling scheme exceeds a preset threshold based on real-time collection of production performance data.

Description

Technical Field

[0001] This invention relates to the field of production scheduling optimization systems, and more particularly to an intelligent production scheduling optimization system for a roll packaging workshop. Background Technology

[0002] Production scheduling in tobacco packaging workshops is a highly complex optimization problem, characterized by multiple constraints, multiple objectives, strong coupling, and high dynamism. Traditional scheduling methods mainly rely on manual experience or simple rule-based systems, which have significant shortcomings: Data fragmentation: Key data such as order information (MES), real-time equipment status (IoT), and auxiliary material inventory (WMS) are scattered across different systems, making it difficult to effectively integrate and utilize them, resulting in incomplete and delayed basic production scheduling information.

[0003] The constraints are too simplistic: they mainly focus on single objectives such as order delivery time or equipment utilization, making it difficult to simultaneously optimize multi-dimensional objectives such as equipment efficiency, material availability, and order priority; and they model complex constraints (such as differences in equipment capabilities, mandatory cleaning time required for brand changes according to GMP standards, and strong coupling between specific brands and auxiliary materials / equipment) in a crude or ignored manner, resulting in poor feasibility of production scheduling plans.

[0004] Poor dynamic adaptability: When faced with disturbances such as emergency orders, sudden equipment failures, or material shortages, traditional methods respond slowly, often requiring manual rescheduling, which is time-consuming and makes it difficult to guarantee the optimization of the new solution.

[0005] Compliance Challenges: The tobacco industry's stringent GMP (Good Manufacturing Practice) requirements, especially the necessary cleaning processes when switching between different brands, necessitate precise adherence to time and procedural constraints during production scheduling. Manual scheduling is prone to errors, and existing systems often lack sufficient integration. Therefore, there is an urgent need for an intelligent production scheduling optimization system for cigarette packaging workshops that can deeply integrate multi-source heterogeneous data, accurately model complex dynamic constraints, achieve multi-objective global optimization, quickly respond to disturbances, and automatically ensure compliance. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent scheduling optimization system for a roll packaging workshop to solve the problems mentioned in the background art.

[0007] To achieve the above-mentioned objectives, this invention provides an intelligent scheduling optimization system for a roll packaging workshop, comprising: The multi-source data fusion module is used to integrate order data from the manufacturing execution system, status data from the equipment IoT platform, and auxiliary material inventory data from the warehouse management system. The dynamic constraint modeling module is used to build a constraint model library that includes equipment capacity matrix constraints, license plate change time rule constraints, and auxiliary material coupling equation constraints. A dual-chain genetic algorithm engine is used to perform multi-objective optimization of equipment resource allocation, material supply guarantee and order task scheduling under the condition of satisfying the constraint model library, and generate production scheduling plan; A closed-loop feedback actuator is used to dynamically trigger the dual-chain genetic algorithm engine to re-optimize the production schedule when the deviation from the current production schedule exceeds a preset threshold, based on real-time collected production performance data.

[0008] Furthermore, the bistranded encoding in the bistranded genetic algorithm engine includes: A process chain, whose elements are a brand-process combination, is jointly identified by the brand ID and the process type code; The equipment chain consists of physical equipment numbers. Each equipment chain element is associated with a set of process chain elements that it can execute, i.e., a process capability set, through a preset mapping table.

[0009] Furthermore, the process type coding includes at least feeding, coiling, packaging, and label changing / cleaning.

[0010] Furthermore, when performing re-optimization, the dual-chain genetic algorithm engine adopts a device chain local mutation strategy, mainly adjusting the gene segments associated with the device that has undergone a state change.

[0011] Furthermore, the label-changing time rule constraint is based on tobacco industry GMP standards, accurately modeling the mandatory cleaning time interval required for switching between different label combinations.

[0012] Furthermore, the conditions for triggering the closed-loop feedback actuator include one or more of the following: unplanned equipment downtime exceeding a first time threshold, changes in order demand information exceeding a second threshold, and expected shortage lead time of key auxiliary materials being less than a third time threshold.

[0013] Furthermore, the multi-source data fusion module cleans, transforms, and standardizes the integrated data.

[0014] Compared with the prior art, the beneficial effects of the present invention are: Global optimization capabilities are significantly enhanced: Through a dual-chain genetic algorithm engine, the three key dimensions of equipment allocation, material matching, and order scheduling are innovatively optimized collaboratively within a unified framework.

[0015] It overcomes the limitations of traditional single-objective (such as only considering delivery time or equipment load) or simple linear weighting methods when dealing with complex coupling relationships.

[0016] It can find a better overall production scheduling solution while meeting all complex constraints, and comprehensively improve equipment utilization, on-time order delivery rate and material utilization efficiency.

[0017] Dynamic adaptability has been greatly improved: A closed-loop feedback actuator based on real-time production performance data enables the system to quickly detect disturbances.

[0018] In particular, the dual-chain encoding mechanism, combined with a local mutation strategy for the equipment chain (when equipment status changes abruptly or tasks change, the gene segments involving the equipment and its associated tasks are mainly adjusted, rather than starting from scratch), greatly improves the response speed of production rescheduling. Practical application verification shows that when typical disturbances such as order changes occur, the time for the system to regenerate the optimized solution is reduced to 1 / 6 of the original manual scheduling method, significantly reducing production downtime.

[0019] Compliance assurance is embedded automatically: The dynamic constraint modeling module incorporates GMP standard constraints for the tobacco industry, particularly for the accurate modeling of label changing and cleaning times and procedures.

[0020] During the generation and re-optimization of production scheduling plans, the system automatically ensures that the time interval between grade switching meets the mandatory cleaning requirements, and the cleaning task is accurately arranged in the equipment chain as a specific "process".

[0021] It completely eliminates the risk of violating GMP standards due to human error or deficiencies in traditional systems, ensuring product quality and production compliance. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are some embodiments of the invention, but not all embodiments.

[0023] A roll packaging workshop intelligent scheduling optimization system includes the following interrelated functional modules: Multi-source data fusion module: This module serves as the system's data foundation, responsible for collecting key production data from the workshop's underlying systems in real-time or near real-time. Its core function is to integrate: order data from the Manufacturing Execution System (MES), such as order data including brand, quantity, priority, delivery date, etc.; equipment status data from the Internet of Things (IoT) platform, including equipment start-up, shutdown, operation, fault, efficiency, and current production tasks; and auxiliary material inventory data from the Warehouse Management System (WMS), including the availability of auxiliary materials strongly related to the brand, such as filter rods, cigarette paper, tipping paper, ink, small boxes, and cartons. This module cleans, transforms, and standardizes heterogeneous data, providing accurate, consistent, and comprehensive data input for subsequent optimization.

[0024] Dynamic constraint modeling module: Based on the information provided by the multi-source data fusion module, this module constructs a mathematical model library that describes the complex rules and constraints of production scheduling in the roll packaging workshop; The core constraint models included in the constraint library are: Equipment Capability Matrix Constraints: Precisely describe the range of grades that each winding and sealing machine and packaging unit (physical equipment) can produce, as well as its theoretical / actual processing capacity and efficiency parameters in each process (such as feeding, winding and sealing, packaging, and boxing).

[0025] Brand switching time rules constraints: Accurately model the mandatory time interval (cleaning time) caused by GMP cleaning requirements when switching between different brands; the rules are refined into specific cleaning durations and procedures for different brand combinations (such as from brand A to brand B); Constraints on the coupling equation of auxiliary materials: Establish a strong correlation equation between the grade and the specific auxiliary materials required (filter rod type, cigarette paper, packaging materials, etc.); ensure that in the production scheduling plan, during the time period allocated to the equipment to produce a certain grade, the inventory of all auxiliary materials required meets the completeness requirements; This module supports dynamic updates of constraints, such as changes in equipment status, inventory consumption, and new orders that trigger changes in constraint conditions.

[0026] Bis-stranded genetic algorithm engine: This module is the core optimizer of this invention, responsible for solving the optimal or near-optimal (meeting business needs) production scheduling scheme under the constraints defined by the dynamic constraint modeling module.

[0027] Its innovation lies in the use of an improved genetic algorithm (double-chain genetic algorithm) that uses collaborative coding of process chain and equipment chain. Coding mechanism: A chromosome contains two gene chains that map to each other.

[0028] Process chain: Its element is "brand name-process combination", which is jointly identified by brand name ID and process type code (such as feeding, coiling, packaging, brand changing and cleaning, etc.) to identify an atomic production task unit.

[0029] Equipment chain: Its elements are physical equipment numbers (such as winding and sealing machine 01, packaging machine 03), which identify the equipment that performs the tasks in the corresponding process chain.

[0030] Mapping Association: Equipment chains and process chains are associated through a mapping table. This mapping table defines which "brand-process combinations" (process chain elements) each physical equipment number (equipment chain element) can perform, i.e. the process capability set of the equipment, ensuring that the equipment allocation scheme is within the equipment capability range.

[0031] This algorithm engine simultaneously optimizes three dimensions of objectives: equipment resource allocation, material (auxiliary material) supply guarantee, and order task (delivery date / priority) achievement. Specifically, it aims to minimize total completion time, maximize equipment utilization, minimize the number of rework orders / time, ensure timely delivery of high-priority orders, and ensure material supply matching.

[0032] Closed-loop feedback actuator: This module enables closed-loop control and dynamic optimization of production scheduling plans.

[0033] Its core function is to monitor production performance data from the multi-source data fusion module in real time, such as actual task start / completion time, actual equipment status changes, actual material consumption / warehousing, and actual order progress.

[0034] When the deviation between actual production and the original production schedule is detected to exceed a preset threshold (such as equipment failure causing downtime, delays in the arrival of key auxiliary materials, emergency orders, or excessive cleaning time for actual product rebranding), this module automatically triggers a re-optimization process.

[0035] Re-optimization process: The current actual state (as a new starting point) and the latest data from the multi-source data fusion module are input into the dynamic constraint modeling module to update the constraint model. Then, the dual-chain genetic algorithm engine quickly generates a new optimized production scheduling plan to achieve dynamic adjustment of production scheduling.

[0036] The process chain element is a combination of grade and process (grade ID + process type code), which is the smallest unit that accurately describes the production task.

[0037] The equipment chain element is the physical equipment number, which is associated with the process capability set through a mapping table to ensure the feasibility of the equipment allocation scheme and the accurate modeling of equipment capabilities.

[0038] The intelligent scheduling optimization system for roll packaging workshops provided by this invention has the following significant advantages compared with existing technologies: Global optimization capabilities are significantly enhanced: Through a dual-chain genetic algorithm engine, the three key dimensions of equipment allocation, material matching, and order scheduling are innovatively optimized collaboratively within a unified framework.

[0039] It overcomes the limitations of traditional single-objective (such as only considering delivery time or equipment load) or simple linear weighting methods when dealing with complex coupling relationships.

[0040] It can find a better overall production scheduling solution while meeting all complex constraints, and comprehensively improve equipment utilization, on-time order delivery rate and material utilization efficiency.

[0041] Dynamic adaptability has been greatly improved: A closed-loop feedback actuator based on real-time production performance data enables the system to quickly detect disturbances.

[0042] In particular, the dual-chain encoding mechanism, combined with a local mutation strategy for the equipment chain (when equipment status changes abruptly or tasks change, the gene segments involving the equipment and its associated tasks are mainly adjusted, rather than starting from scratch), greatly improves the response speed of production rescheduling. Practical application verification shows that when typical disturbances such as order changes occur, the time for the system to regenerate the optimized solution is reduced to 1 / 6 of the original manual scheduling method, significantly reducing production downtime.

[0043] Compliance assurance is embedded automatically: The dynamic constraint modeling module incorporates GMP standard constraints for the tobacco industry, particularly for the accurate modeling of label changing and cleaning times and procedures.

[0044] During the generation and re-optimization of production scheduling plans, the system automatically ensures that the time interval between grade switching meets the mandatory cleaning requirements, and the cleaning task is accurately arranged in the equipment chain as a specific "process".

[0045] It completely eliminates the risk of violating GMP standards due to human error or deficiencies in traditional systems, ensuring product quality and production compliance.

[0046] The technical solution of the present invention has been described above in conjunction with specific embodiments. However, it should be noted that the above descriptions are only for explaining the solution of the present invention and should not be construed as a specific limitation on the scope of protection of the invention in any way. Based on this explanation, those skilled in the art can conceive of other specific embodiments or equivalent substitutions of the present invention without creative effort, and all such embodiments or substitutions will fall within the scope of protection of the present invention.

Claims

1. An intelligent scheduling optimization system for a roll packaging workshop, characterized in that, include: The multi-source data fusion module is used to integrate order data from the manufacturing execution system, status data from the equipment IoT platform, and auxiliary material inventory data from the warehouse management system. The dynamic constraint modeling module is used to build a constraint model library that includes equipment capacity matrix constraints, license plate change time rule constraints, and auxiliary material coupling equation constraints. A dual-chain genetic algorithm engine is used to perform multi-objective optimization of equipment resource allocation, material supply guarantee and order task scheduling under the condition of satisfying the constraint model library, and generate production scheduling plan; A closed-loop feedback actuator is used to dynamically trigger the dual-chain genetic algorithm engine to re-optimize the production schedule when the deviation from the current production schedule exceeds a preset threshold, based on real-time collected production performance data.

2. As described in claim 1, characterized in that, The bichain encoding in the bichain genetic algorithm engine includes: A process chain, whose elements are a brand-process combination, is jointly identified by the brand ID and the process type code; The equipment chain consists of physical equipment numbers. Each equipment chain element is associated with a set of process chain elements that it can execute, i.e., a process capability set, through a preset mapping table.

3. As described in claim 2, characterized in that, The process type codes include at least feeding, coiling, packaging, and cleaning.

4. As described in claim 1, characterized in that, When performing re-optimization, the dual-chain genetic algorithm engine adopts a device chain local mutation strategy, which mainly adjusts the gene segments associated with the devices that have undergone state changes.

5. As described in claim 1, characterized in that, The brand switching time rules are based on tobacco industry GMP standards and accurately model the mandatory cleaning time intervals required for switching between different brand combinations.

6. As described in claim 1, characterized in that, The conditions for triggering the closed-loop feedback actuator include one or more of the following: unplanned equipment downtime exceeding a first time threshold, changes in order demand information exceeding a second threshold, and expected shortage lead time of key auxiliary materials being less than a third time threshold.

7. As described in claim 1, characterized in that, The multi-source data fusion module cleans, transforms, and standardizes the integrated data.