Intelligent scheduling and automatic planning system for tread extrusion process
The intelligent scheduling and automatic planning system solves the problem of low efficiency in traditional manual scheduling, and realizes efficient, reliable and flexible production management of the tire tread pressing process, thereby improving the stability and flexibility of the production line.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SAILUN GRP CO LTD
- Filing Date
- 2026-03-23
- Publication Date
- 2026-06-19
Smart Images
Figure CN122243089A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of tire manufacturing technology, and specifically to an intelligent scheduling and automatic planning system for the tread pressing process. Background Technology
[0002] In the tire manufacturing industry, tread extrusion is a crucial process that connects upstream and downstream processes. The rationality of its production plan and the efficiency of its execution directly affect the smoothness of the overall production line, the level of work-in-process inventory, and the on-time delivery of orders. Traditionally, the production planning and issuance for the tread extrusion process relies heavily on manual operation by planners, often referred to as "planning." Planners manually retrieve production orders from the Manufacturing Execution System (MES) or Enterprise Resource Planning (ERP) system, and based on their personal experience, consider multiple complex factors such as order specifications, delivery dates, equipment status, mold (die) availability, and material preparation. They manually split and merge orders, and assign specific production machines and start times to each order. This traditional model has several inherent drawbacks. First, it is extremely inefficient. With hundreds or thousands of orders, manual scheduling is time-consuming and prone to errors during calculation and transcription. Second, due to the limitations of manual processing, the scheduling solutions often fail to achieve optimal results. Planners often struggle to comprehensively consider and balance multiple conflicting objectives within a short period, such as minimizing mold changeover time, minimizing order delays, optimizing equipment energy consumption, and balancing the load across multiple production lines. This leads to frequent unplanned downtime, work-in-process inventory buildup, and order delays. Furthermore, this model is severely inadequate in responding to anomalies. When faced with urgent order insertions, sudden equipment failures, or material supply disruptions, manual rescheduling is slow, significantly impacting the flexibility and stability of the production system. Finally, scheduling quality relies excessively on the planner's individual experience and skills. This tacit knowledge is difficult to standardize, solidify, and pass on, and staff turnover will impact the stability of production plans. Therefore, the industry urgently needs a system solution that can automatically, intelligently, and efficiently schedule and issue plans for the tread extrusion process, and dynamically respond to production changes. This solution aims to overcome the bottlenecks of the traditional manual model and improve the intelligence and flexibility of tire manufacturing.
[0003] Therefore, existing technologies still need further development. Summary of the Invention
[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide an intelligent scheduling and automatic planning system for the tread extrusion process, so as to solve the problems existing in the prior art.
[0005] To achieve the above-mentioned technical objectives, the present invention provides an intelligent scheduling and automatic planning system for the tread extrusion process, comprising: The order parsing module is used to receive and process production orders from the upstream system and build a pool of orders that can be scheduled for production. The intelligent scheduling engine is connected to the order parsing module and is used to perform multi-level scheduling optimization on the pool of orders that can be scheduled to generate a detailed scheduling plan. The multi-level scheduling optimization includes initial screening of orders based on preset rules and multi-objective global optimization of the order sequence after initial screening. The work order issuance module is connected to the intelligent production scheduling engine and is used to convert the production plan into work order instructions that can be executed by the equipment, and automatically issue them to the corresponding tire tread extrusion production equipment. The monitoring and rescheduling module is used to collect production execution data in real time and compare it with the production schedule. When the preset rescheduling conditions are triggered, it drives the intelligent scheduling engine to dynamically reschedule the unexecuted plans.
[0006] Specifically, the intelligent scheduling engine includes a rule-based initial screening unit, a multi-objective optimization unit, and a simulation verification unit connected in sequence.
[0007] Specifically, the rule screening unit is configured to filter, merge, and initially sort orders based on at least one hard rule and clustering rule, including order specifications and mouthpiece matching, equipment capability range, and order similarity.
[0008] Specifically, the multi-objective optimization unit is configured to establish an optimization model with the objective function of minimizing the total mold change time, minimizing the total delay time, minimizing the total energy consumption, and maximizing the equipment load balance, and with the production line, die plate, and start time of the order allocation as decision variables, and to solve the model using an optimization algorithm.
[0009] Specifically, the optimization algorithm includes a genetic algorithm or a simulated annealing algorithm.
[0010] Specifically, the simulation verification unit is configured to perform time extrapolation and conflict detection on the production scheduling scheme output by the multi-objective optimization unit based on the discrete event simulation model, and automatically fine-tune the detected conflicts to output the final executable production scheduling instruction set.
[0011] Specifically, the monitoring and rescheduling module is configured such that the rescheduling conditions include at least one of equipment failure, material shortage, order change, and predicted schedule delay; the dynamic rescheduling follows the principle of minimum disturbance, prioritizing impact range assessment and local rescheduling.
[0012] Specifically, the system also includes integration interfaces with Manufacturing Execution System (MES) and Enterprise Resource Planning (ERP) systems.
[0013] Specifically, the work order issuance module is configured to automatically associate and load the standard process parameters corresponding to the tire tread specification when generating a work order instruction.
[0014] Specifically, the intelligent scheduling engine, work order issuance module, and monitoring and rescheduling module are integrated and deployed on the same server or server cluster.
[0015] Beneficial effects: Compared with existing technologies, the intelligent scheduling and automatic planning system for the tread extrusion process provided by this invention brings significant technological progress and beneficial effects by constructing a fully closed-loop automated architecture of "planning-execution-monitoring" and integrating multi-level intelligent decision-making methods.
[0016] First, it has achieved a leap forward in production efficiency and cost control. The system transforms production scheduling from hours of manual labor into automated calculations at the minute or even second level, greatly freeing up human resources. Through intelligent optimization algorithms, the system can automatically find the optimal production sequence to reduce the number of die plate changes and balance equipment load, thereby significantly reducing downtime and improving overall equipment utilization. Simultaneously, precise cycle time control ensures more synchronized integration between semi-finished tire treads and downstream molding processes, effectively reducing work-in-process inventory levels and related capital occupation and warehousing costs.
[0017] Secondly, improvements have been made in quality assurance and production stability. The system completely eliminates basic errors that may occur during manual production scheduling, such as incorrect specification settings and incorrect mold usage. More importantly, when automatically issuing production work orders, the system can simultaneously link and issue a set of verified standard process parameters for that tire tread specification, ensuring that each production run is executed according to the optimal and uniform process standards. This guarantees the consistency and stability of product quality and reduces quality fluctuations caused by improper parameter settings.
[0018] Third, the system endows the production line with unprecedented agility and production flexibility. Based on a real-time rescheduling mechanism triggered by dynamic events, the system can automatically and quickly respond to various production anomalies such as equipment failures, material shortages, and emergency order insertions. Following the principle of "minimal disturbance," the system performs local or global rescheduling, rapidly generating new feasible plans. This enhances the production system's anti-interference capabilities and robustness, making delivery commitments to customers more reliable and enabling the production line to better adapt to the trend of small-batch, multi-variety customized production.
[0019] Fourth, it has driven a fundamental shift in enterprise production decision-making from experience-driven to data- and algorithm-driven. The system's scheduling rules and optimization goals are quantifiable and configurable, forming a transparent and reproducible decision-making process. The cases accumulated during system operation, such as "which specification combination and which production sequence is most efficient," will become a valuable scheduling knowledge base for the enterprise, enabling the digital transfer of experience and continuous self-optimization of the system, laying a solid foundation for the enterprise's intelligent upgrade. Attached Figure Description
[0020] Figure 1 This is a schematic diagram of the intelligent scheduling and automatic planning system for the tread pressing process provided in a specific embodiment of the present invention. Detailed Implementation
[0021] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments in this application, other similar embodiments obtained by those skilled in the art without creative effort should all fall within the scope of protection of this application. Furthermore, directional terms mentioned in the following embodiments, such as "up," "down," "left," and "right," are only for reference to the directions in the accompanying drawings; therefore, the directional terms used are for illustrative purposes and not for limiting the invention.
[0022] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments.
[0023] Please see Figure 1 This invention provides an intelligent scheduling and automatic planning system for the tread extrusion process, comprising: The order parsing module is used to receive and process production orders from the upstream system and build a pool of orders that can be scheduled for production. The intelligent scheduling engine is connected to the order parsing module and is used to perform multi-level scheduling optimization on the pool of orders that can be scheduled to generate a detailed scheduling plan. The multi-level scheduling optimization includes initial screening of orders based on preset rules and multi-objective global optimization of the order sequence after initial screening. The work order issuance module is connected to the intelligent production scheduling engine and is used to convert the production plan into work order instructions that can be executed by the equipment, and automatically issue them to the corresponding tire tread extrusion production equipment. The monitoring and rescheduling module is used to collect production execution data in real time and compare it with the production schedule. When the preset rescheduling conditions are triggered, it drives the intelligent scheduling engine to dynamically reschedule the unexecuted plans.
[0024] It should be further explained that the order parsing module is used to receive and process production orders from Manufacturing Execution System (MES) or Enterprise Resource Planning (ERP) systems. Specifically, the intelligent scheduling engine is used to perform multi-level scheduling optimization, including rule initial screening, multi-objective global optimization, and micro-simulation verification. Specifically, the work order issuance module is used to convert the optimized and verified scheduling plan into equipment control instructions. Specifically, the monitoring and rescheduling module is used to achieve closed-loop feedback and dynamic adjustment of the plan execution process.
[0025] Furthermore, the specific design of the above scheme includes: (1) The production order data received by the order parsing module from the upstream MES / ERP system includes information such as specifications, quantity, priority, delivery date, customer level, and associated material codes. This module first preprocesses the orders, including data verification and format standardization, and then performs cluster analysis and conflict checking. For example, orders with identical specifications (such as all being 205 / 55R16 comfort pattern) are forcibly merged and treated as a single production batch to maximize production continuity. At the same time, material availability is checked to ensure that the required die plates, adhesives, etc., are ready for production. After preprocessing, a "schedulable order pool" is formed, containing all executable orders and their related constraints.
[0026] (2) The intelligent scheduling engine is the core computing unit of the system. Its workflow strictly follows a three-level progressive model of "rule initial screening, multi-objective global optimization, and micro-simulation verification". First, the rule engine quickly filters out obviously infeasible order combinations, generating a preliminary ordered candidate solution set, which "reduces the burden" for subsequent complex optimization. Then, the intelligent optimization algorithm is used to perform deep optimization on the candidate solution set to find the best scheduling scheme that balances multiple objectives. Finally, the optimized scheme is placed in a highly simulated virtual production environment for verification and fine-tuning to ensure the feasibility of the scheme.
[0027] (3) The work order issuance module receives the final production scheduling instruction set from the intelligent production scheduling engine. This module transforms the abstract production scheduling plan (such as "production line L1 starts using die plate M1 to produce order A at 08:00") into specific instructions that can be recognized and executed by specific equipment. For example, it generates work orders containing information such as specific specification codes, process parameter sets, start / end times, and target output, and automatically issues them to the designated extruder, cutting machine controller, and material handling system through industrial communication protocols (such as OPCUA, ModbusTCP) or system interfaces.
[0028] (4) The monitoring and rescheduling module acquires information such as equipment status, production progress, and material consumption in real time through data acquisition terminals deployed on the production site. The system continuously compares the actual execution data with the production schedule. Once a deviation is detected or an abnormal event (such as an equipment failure alarm) is received, it determines whether to trigger rescheduling according to preset logic. If triggered, it calls the specific function of the intelligent scheduling engine to quickly reschedule the affected subsequent plans based on the principle of "minimum disturbance" and issues the new plan, thereby achieving agile response in the production process.
[0029] Understandably, this system constructs a complete closed loop from order receipt and instruction issuance to execution monitoring and dynamic adjustment. It fundamentally changes the traditional model of relying on planners for manual scheduling and planning, freeing planners from tedious, error-prone, and inefficient manual operations, reducing scheduling time from hours to minutes. Through algorithmic optimization, the system comprehensively considers multiple complex constraints such as mold changes, delivery dates, energy consumption, and load balancing to obtain better scheduling solutions, thereby reducing unplanned downtime, lowering work-in-process inventory, and improving overall equipment utilization. Simultaneously, its dynamic rescheduling capability gives the production line strong anti-interference and flexibility, enabling rapid response to abnormal situations such as order insertions and equipment failures, improving on-time order delivery rates.
[0030] Specifically, the intelligent scheduling engine includes a rule-based initial screening unit, a multi-objective optimization unit, and a simulation verification unit connected in sequence.
[0031] It should be further explained that the internal architecture of the intelligent scheduling engine includes a rule initial screening unit, a multi-objective optimization unit, and a simulation verification unit connected in sequence. The three units are executed sequentially, and the output of the previous unit serves as the input of the next unit.
[0032] Furthermore, the specific design of the above scheme includes: (1) The rule screening unit is the first level of the production scheduling process. It receives the "schedulable order pool" from the order parsing module and applies a series of preset rules with low computational complexity for rapid processing. Its goal is to quickly narrow down the solution space, provide a high-quality initial solution sequence for subsequent optimization algorithms, and avoid the optimization algorithm wasting computational resources on invalid solutions.
[0033] (2) The multi-objective optimization unit is the second level of the production scheduling process and also the core optimization layer. It receives the "order sequence to be optimized" output by the rule screening unit, and performs a deep search in a huge solution space containing multiple conflicting objectives based on this sequence. It adopts mathematical modeling and intelligent optimization algorithms to find a production scheduling scheme with the best comprehensive evaluation.
[0034] (3) The simulation verification unit is the third level of the production scheduling process, namely the feasibility verification layer. It receives the production scheduling plan that is "better" in mathematical model from the output of the multi-objective optimization unit. This plan may ignore some micro-sequence and logistics details on the production site. The simulation verification unit establishes a high-fidelity virtual production environment to dynamically extrapolate and stress test the plan, discover and correct potential timing conflicts, cycle mismatches and other problems, and output a final production scheduling instruction set that is truly executable and accurate to the minute or even the second.
[0035] Understandably, this multi-level architecture design of "rule filtering, global optimization, and micro-level verification" balances the speed, quality, and feasibility of production scheduling. Initial rule screening ensures efficiency, multi-objective optimization pursues overall optimality, and micro-level simulation ensures the reliability of the plan. The clear division of labor at each level, with progressive refinement, collectively constitutes an efficient and robust intelligent production scheduling decision-making process.
[0036] Specifically, the rule screening unit is configured to filter, merge, and initially sort orders based on at least one hard rule and clustering rule, including order specifications and mouthpiece matching, equipment capability range, and order similarity.
[0037] It should be further explained that the rule screening unit is configured to process orders based on multiple rules, including hard filtering rules, clustering and merging rules, and priority sorting rules.
[0038] Furthermore, the specific design of the above scheme includes: (1) Hard filtering rules are used to exclude orders that are technically infeasible. For example: a) Mouthplate matching rules: Filter out orders for which no mouthplate is currently available. The system will query the mouthplate inventory and status database to confirm whether the corresponding mouthplate is available and within its lifespan.
[0039] b) Equipment capacity rules: Filter out order specifications that exceed the maximum processing width, length, or speed limits of any available extrusion production line.
[0040] (2) Clustering and merging rules are used to reduce production batches. The main goal is to merge identical or highly similar orders to reduce the number of mold changes. For example: a) Complete merger: Orders with identical pattern codes, sizes, and rubber compound formulas are forcibly merged into one production batch.
[0041] b) Similar merging: Under certain strategies (such as to achieve the minimum economic production batch), orders with the same pattern series and similar sizes can be merged and produced using the same set of die plates, although minor adjustments to process parameters may be required.
[0042] (3) Priority sorting rules are used to give the filtered and merged order sequence a preliminary processing order. The sorting criteria usually include: a) Order delivery time: The more urgent the delivery time, the higher the priority.
[0043] b) Customer Tier: Orders from strategic or important customers have higher priority.
[0044] c) Order size: Under certain strategies, small-batch orders will be prioritized to quickly release production capacity.
[0045] Understandably, through initial rule screening, the system can quickly eliminate obvious conflicts in the order pool, simplifying hundreds or thousands of original orders into a smaller, preliminarily ordered "order sequence to be optimized." This significantly reduces the difficulty and computational burden of subsequent multi-objective optimization algorithms, improving the response speed of the entire production scheduling system. Simultaneously, simple clustering and merging immediately brings the direct benefit of reducing mold changeovers, raising the baseline of production scheduling efficiency.
[0046] Specifically, the multi-objective optimization unit is configured to establish an optimization model with the objective function of minimizing the total mold change time, minimizing the total delay time, minimizing the total energy consumption, and maximizing the equipment load balance, and with the production line, die plate, and start time of the order allocation as decision variables, and to solve the model using an optimization algorithm.
[0047] It should be further explained that the multi-objective optimization unit is configured to establish a multi-objective optimization mathematical model and solve it through an optimization algorithm to obtain the comprehensive optimal production scheduling plan.
[0048] Furthermore, the specific design of the above scheme includes: (1) The optimization model describes a production scheduling plan by defining decision variables. Specific decision variables include: a) Binary decision variables : indicates an order Whether to allocate to the production line Production begins. If allocation is required, then... Otherwise, it is 0. Wherein, Represents the order index. Represents the production line index.
[0049] b) Continuous decision variables : indicates an order The start time of production.
[0050] c) Integer decision variables : indicates an order The number of the lip-shaped plate used.
[0051] (2) The objective of the optimization model is to find a set of decision variable values that minimizes (or maximizes) the value of an objective function that integrates multiple sub-objectives. Specifically, the objective function... It is usually expressed as a weighted sum of multiple sub-objectives to achieve trade-offs among the multiple objectives. A preferred functional form is as follows: in: This represents the total mold changeover time. It calculates the sum of the mold changeover preparation times required for two adjacent orders using different die plates on the same production line. Reducing this value improves equipment time utilization.
[0052] This represents the total delay time. It calculates the sum of the actual completion time for each order exceeding its promised delivery date, i.e. ,in For orders Completion time, For orders Delivery time. Minimizing this will improve customer satisfaction.
[0053] This represents total energy consumption. It is calculated based on the rated power of each production line and the planned operating time for each order (including necessary warm-up time). Minimizing this will reduce production costs.
[0054] This indicates the degree of equipment load balance. It measures the degree of balance in workload among production lines and can be expressed as the reciprocal of the variance of the total working hours of each production line, or directly as a negative number of "(maximum load - minimum load)". Since the objective is to minimize... ,and The preceding character is negative, therefore the optimization process tends to maximize it. Even with a more balanced equipment load, a balanced load contributes to production stability and full utilization of resources.
[0055] (3) Weighting coefficients , , , This represents the different emphases that enterprises place on four dimensions: efficiency, delivery, cost, and stability. The optimal values can be determined based on historical production data, using expert experience or more advanced preference learning algorithms. A set of exemplary optimal values is as follows: , , , The reason for choosing this set of values is that, in a typical tire manufacturing scenario, reducing mold changes improves equipment operating efficiency. ) and ensuring on-time delivery of orders ( ) is usually given the highest priority; reducing energy consumption ( This is a crucial cost control aspect; while ensuring the first three points are met, efforts should be made to balance the load on each production line. This is to facilitate production management.
[0056] (4) The optimization model also includes a series of constraints to ensure the feasibility of the production scheduling plan. The main constraints include: a) Order uniqueness constraint: An order can only be produced on one and only one piece of equipment, i.e. .
[0057] b) Resource exclusivity constraint: Only one order can be produced by the same equipment at any given time. This is determined by the order's start time. Processing time and decision variables This is implicitly guaranteed during algorithm iteration.
[0058] c) Material readiness constraint: The order start time must be after the readiness time of its required materials (mainly rubber).
[0059] d) Mouth-shaped plate matching constraints: assigned to orders Lip shaper It must belong to the set of mouthpieces allowed for use in this order.
[0060] Understandably, by establishing a mathematical model that includes decision variables, multi-objective functions, and complex constraints, and by transforming the unique requirements of tire production such as mold changing and die plate matching into mathematical constraints, the system can automatically search for the optimal production scheduling scheme under given preferences within a vast, multi-dimensional solution space. This surpasses the traditional production scheduling method that relies on human experience and can only consider a few objectives, achieving the quantification, scientification, and optimization of the production scheduling process.
[0061] It should be further explained that after the multi-objective optimization unit applies the optimization algorithm, it will output a production scheduling scheme containing detailed timing arrangements. To illustrate its output results and optimization effect, a convergent scheme with given weight coefficients will be used as an example. Assume the system's configured optimization objective weights are... , , , This weighting combination reflects a typical emphasis placed by enterprises on production efficiency, order fulfillment, energy costs, and equipment stability. Optimization algorithms (such as genetic algorithms) iteratively calculate under this weighting, eventually converging and outputting a production scheduling plan. The key results can be summarized as follows: (1) Mold change arrangement: A total of 3 die plate change operations are required throughout the entire production cycle. Specifically, on the L1 production line, after the production of order A is completed, the die plate needs to be changed from M1 to M3 to produce subsequent orders; on the L2 production line, the die plate needs to be changed from M2 to M4.
[0062] (2) Order delivery status: Under this scheme, most orders can be completed on time, but order B will be delayed by 2 hours. This delay is the result of a global trade-off made by the algorithm under the given objective function: in order to enable order B to be produced on the L1 production line using the same M1 die plate as the earlier order A, thereby avoiding an additional die change operation (which helps to reduce the total die change time). The system "accepted" a partial delay in order B, which reflects the reduction in mold changes (target). ) and reducing delays (target) A compromise between these two approaches.
[0063] (3) Equipment load balancing: The planned total production loads for each production line are as follows: L1 production line 15 hours, L2 production line 14 hours, and L3 production line 10 hours. This load distribution presents a relatively balanced situation, avoiding excessive concentration of production tasks on a single production line. This is conducive to the stable utilization of production resources and the stability of production organization, which is precisely the equipment load balancing target in the optimization objective. The desired effect.
[0064] (4) Comprehensive evaluation of the scheme: based on the aforementioned objective function Calculate the evaluation value of this plan. The value is 42.7. To illustrate the optimization effect of this system, a set of comparative data is provided for reference: Under the same order and constraint conditions, if production scheduling is carried out in a completely random manner, its... The value is typically around 65 to 80; if only rule-based initial screening is applied without multi-objective optimization, its... The value is approximately 55. The optimized solution in this example... The value (42.7) is significantly lower than these two comparison scenarios, which quantitatively proves that the "rule-based initial screening combined with multi-objective global optimization" method adopted by this system can effectively search for and output a production schedule that is better in terms of overall cost.
[0065] Understandably, this example concretely demonstrates the working results of the intelligent scheduling engine. It doesn't generate an "ideal" plan that excels in all single metrics, but rather, through algorithmic search and trade-offs across a multi-dimensional objective space, arrives at a practical solution that best performs overall given business preferences (weighting coefficients). This solution includes specific equipment assignments, die-cutting plate usage sequences, job start and end times, and the expected completion status of orders. This information collectively forms a detailed and quantifiable scheduling Gantt chart, serving as the direct basis for subsequent work order issuance and execution monitoring.
[0066] Specifically, the optimization algorithm includes a genetic algorithm or a simulated annealing algorithm.
[0067] It should be further noted that the multi-objective optimization unit is configured to use a genetic algorithm or a simulated annealing algorithm to solve the multi-objective optimization model.
[0068] Furthermore, the specific design of the above scheme includes: Taking the genetic algorithm as an example, the specific steps for its application in this system are as follows: (1) Encoding: A hybrid encoding method is used to represent a gestational age (chromosome). For example, a chromosome can be divided into three parts: a) The first part is the order sequence gene, which represents an order of all orders to be scheduled, such as [A,C,B,E,D].
[0069] b) The second part is the equipment allocation gene, which corresponds to the order of the first part and indicates the production line to which each order is allocated, such as [L1,L2,L1,L3,L2].
[0070] c) The third part is the mouth-shaped plate allocation gene, which corresponds to the order in the first part and indicates the mouth-shaped plate number used for each order, for example, [M1,M3,M1,M4,M2]. This chromosome fully expresses a production scheduling scheme: order A is produced on line L1 using mouth-shaped plate M1, order C is produced on line L2 using mouth-shaped plate M3, and so on.
[0071] (2) Initialize the population: generate randomly A feasible scheduling scheme that satisfies basic hard constraints (such as die plate matching) constitutes the initial population. Population size The optimal value is 100. The reason for choosing this value is that a population that is too small (e.g., 20) lacks diversity and is prone to getting trapped in local optima; a population that is too large (e.g., 500) will significantly increase the computation time per iteration. 100 is an empirical value that strikes a good balance between search capability and computational efficiency.
[0072] (3) Fitness assessment: For each chromosome in the population, decode its specific production scheduling plan and calculate the start and end times of each order. Then, substitute the plan into the objective function. Calculate its total cost Since the goal is to minimize Define fitness ,in This is a small positive number to prevent division by zero errors. A higher fitness value indicates a better overall performance of the production scheduling plan.
[0073] (4) Selection: Individuals are selected from the current population to enter the mating pool using a tournament selection method. Random selection is performed each time. Individuals ( The optimal value is 3), and the individual with the highest fitness is selected. This process is repeated until the size of the mating pool is the same as the population size. Tournament selection ensures that excellent individuals have a higher probability of being selected while maintaining a certain level of selection pressure.
[0074] (5) Crossover: Individuals in the mating pool are paired up with a certain crossover probability. Perform a crossover operation to generate offspring individuals. For the order sequence portion, sequential crossover (OX) is used to preserve the relative order in the parent sequence. For the equipment and mouthpiece allocation portions, two-point crossover is used. Crossover probability. The optimal value is 0.8. A high crossover probability is beneficial for the rapid spread of superior genes and population evolution.
[0075] (6) Mutation: For offspring individuals produced by crossover, with a small probability of mutation. Perform mutation operations. Mutation operations include randomly swapping the positions of two orders, or randomly changing the production line or die plate assigned to an order (constraints must be satisfied). Mutation probability. The optimal value is 0.05. A lower mutation probability can introduce new genes and increase population diversity while avoiding excessive damage to good solutions.
[0076] (7) Iteration: Replace the old population with a new generation of population generated through selection, crossover, and mutation, and repeat steps (3) to (6). The termination condition for the algorithm iteration can be reaching the preset maximum number of iterations. (The preferred value is 200 generations), or the optimal fitness no longer significantly improves over multiple generations. When the iteration terminates, the chromosome with the highest fitness in each generation is output, and after decoding, it becomes a relatively optimal scheduling scheme on the "Pareto front" found by the system.
[0077] Understandably, intelligent optimization algorithms such as genetic algorithms, by simulating the "selection-crossover-mutation" evolutionary mechanism in nature, can effectively perform global searches in complex, nonlinear, and multimodal solution spaces, avoiding getting trapped in local optima. They do not depend on the convexity or differentiability of the problem, making them particularly suitable for solving combinatorial optimization problems like production scheduling. By adjusting the algorithm's parameters (such as population size and crossover / mutation probabilities), a flexible trade-off can be struck between solution quality and computation time.
[0078] Specifically, the simulation verification unit is configured to perform time extrapolation and conflict detection on the production scheduling scheme output by the multi-objective optimization unit based on the discrete event simulation model, and automatically fine-tune the detected conflicts to output the final executable production scheduling instruction set.
[0079] It should be further explained that the simulation verification unit is configured to establish a discrete event simulation environment for virtual execution and verification of the optimized production scheduling plan.
[0080] Furthermore, the specific design of the above scheme includes: (1) Establishment of a discrete event simulation model. This model establishes an accurate, event-driven mathematical model of each entity and its interaction logic in the production system. Specifically, it includes: a) Equipment Model: Build a model for each extrusion production line and each cutting machine, including its status (idle, production, mold change, fault), extrusion speed (e.g., 60 meters / minute), mold change time (e.g., 30 minutes to change the die plate), preheating time and other parameters.
[0081] b) Material flow model: Model materials such as rubber and tire strips, and define their flow logic, transmission time (e.g., the transmission from the press outlet to the cutting machine takes 2 minutes) and buffering rules between processes.
[0082] c) Process constraint model: Transforms industry-specific process requirements into time or logical constraints. For example, the tread must be cooled for at least 5 minutes after being pressed out (cooling time constraint) before cutting; some patterns need to be produced symmetrically, and the left and right treads need to be pressed out continuously (pattern symmetry constraint).
[0083] (2) Dynamic simulation and conflict detection. The simulation verification unit takes the production scheduling plan (usually in Gantt chart form) output by the multi-objective optimization unit as input and triggers events strictly according to the time points of the plan in the simulation environment, such as "08:00, L1 line starts to replace M1 die board" and "08:30, L1 line starts to produce order A". The simulation clock advances step by step, and the system monitors and records all events and entity statuses in real time.
[0084] (3) Conflict Detection and Automatic Fine-tuning. During the dynamic simulation, the system will detect micro-conflicts that were not considered in the optimization scheme, mainly including two types: a) Time Conflict: For example, simulation analysis reveals that, according to the original plan, the tire tread of order A arrives at the cutting machine at 08:35, but its extrusion completion time is 08:30, which does not meet the process constraint of "at least 5 minutes of cooling". After detecting this conflict, the system will automatically insert the necessary waiting time (such as 5 minutes) before the cutting process, and thus adjust the start time of the upstream extrusion process or the arrangement of subsequent orders on the same production line in reverse.
[0085] b) Resource Conflicts: For example, simulation reveals that two extruders simultaneously request the same AGV to transport rubber material at 09:00. The system will adjust the request time of one of the devices according to preset priority rules (such as proximity service, first-come, first-served) and adjust its production cycle accordingly. The fine-tuning process is iterative; after correcting one conflict, the system will continue to simulate and check for new conflicts until a complete production scheduling instruction set that is conflict-free and can be executed smoothly in the simulation environment is obtained.
[0086] Understandably, micro-simulation verification serves as a crucial bridge connecting the "ideal mathematical optimal solution" with the "actually executable production plan." Optimization algorithms focus on the mathematical optimality of the global objective, potentially neglecting dynamic details and rigid process constraints on the production floor. Simulation verification, through a "simulate first, then execute" approach, exposes and resolves these potential problems in advance, such as cycle time mismatch, logistical congestion, and process violations. This significantly improves the reliability and executability of the production scheduling plan, avoiding the risk of the plan "crashing" in actual production after it has been issued.
[0087] Specifically, the monitoring and rescheduling module is configured such that the rescheduling conditions include at least one of equipment failure, material shortage, order change, and predicted schedule delay; the dynamic rescheduling follows the principle of minimum disturbance, prioritizing impact range assessment and local rescheduling.
[0088] It should be further noted that the monitoring and rearrangement module is configured with preset rearrangement trigger conditions and performs dynamic rearrangement based on the principle of minimum disturbance.
[0089] Furthermore, the specific design of the above scheme includes: (1) The rearrangement trigger condition is a configurable set, mainly including: a) Passive triggering condition: Reordering directly triggered by external systems or events.
[0090] ① Equipment failure: Triggered when a fault alarm occurs on a certain extrusion production line or a key piece of equipment (such as a cutting machine); ② Material shortage: Triggered when the system detects that the required rubber material for a certain production batch is insufficient in stock or cannot be delivered on time; ③ Quality anomaly: Triggered when the online inspection system detects a continuous stream of non-conforming products, requiring a shutdown for adjustment or investigation; ④ Order Change / Installation: Triggered when MES / ERP issues a new urgent order, or cancels or modifies an existing order.
[0091] b) Active triggering conditions: Rearrangement triggered by the system based on real-time data prediction and analysis.
[0092] ① Key order progress delays: The system compares the planned and actual progress and triggers a mechanism when it predicts that a high-priority order will not be completed on time. ② Potential conflict warning: Based on the current execution rhythm, the simulation predicts that a resource conflict will occur at some point in the future (such as two devices simultaneously requiring AGVs).
[0093] (2) Minimum Disturbance Rescheduling Principle and Process. When the rescheduling condition is triggered, the system does not simply rerun the global scheduling, but follows the "minimum disturbance" principle, as follows: a) Impact Assessment: The system first analyzes the impact of the abnormal event. For example, if a failure occurs on production line L1 at 10:00, and repair is expected to take 2 hours, the system will lock all unexecuted orders assigned to line L1 with a scheduled start time after 10:00. These orders constitute the "affected order set." Orders scheduled to start production before 10:00 will generally remain unchanged. b) Partial Rescheduling Attempt: The system attempts to reschedule only the "affected order set" while keeping the plans of other production lines unchanged. It considers transferring these orders to other available production lines or postponing their start time on line L1. The goal is to minimize changes to the original global plan (especially for orders already in execution) while satisfying order constraints. c) Global Rescheduling Backup: If a local rescheduling fails to find a feasible solution (e.g., affected orders have special die requirements and can only be produced by the faulty production line), the system initiates a global rescheduling to optimize the production schedule for all unexecuted orders. Even so, the system can still set constraints to try to keep the plans of executed orders unchanged.
[0094] Understandably, the real-time rescheduling mechanism triggered by dynamic events is key to the system's "intelligence" and "adaptive" capabilities. It transforms production planning from rigid, one-off instructions into flexible guidance that can dynamically adjust to the production environment. Compared to traditional, slow-responding manual rescheduling, this system can respond to various uncertainties in production in real time and automatically, quickly generating new feasible plans. This significantly improves the robustness of the production system and the reliability of order delivery, meeting the flexible production needs of small batches and diverse product varieties.
[0095] Specifically, the system also includes integration interfaces with Manufacturing Execution System (MES) and Enterprise Resource Planning (ERP) systems.
[0096] It should be further noted that the intelligent scheduling and automatic planning system for the tread pressing process also includes an integrated interface module for data exchange with external information systems.
[0097] Furthermore, this integrated interface module typically includes interfaces with Manufacturing Execution System (MES) and Enterprise Resource Planning (ERP) systems, as specifically designed below: (1) Interface with the ERP system: This interface is mainly used to receive master data for production orders. Through this interface, the system periodically or in real time retrieves or receives production order information issued by the ERP system. This information includes at least: order number, product specification code, production quantity, customer code, promised delivery date, priority identifier, etc. This is the original input for the system to perform production scheduling calculations; (2) Interface with MES system: This interface is responsible for bidirectional data communication.
[0098] a) Upstream direction (data acquisition): This system acquires real-time dynamic data from the production site through MES, including: the current operating status (running, stopped, faulty) of each piece of equipment (extruder, cutting machine), current production tasks, completed output, material consumption, quality inspection results, etc. This data is the basis for the monitoring and rescheduling module to compare the plan with the actual situation; b) Downward Direction (Instruction Issuance): This system will issue the final work order instructions generated through production scheduling and verification via the MES interface. The MES will then further decompose these instructions and issue them to the underlying equipment control systems (such as the extruder PLC), material handling systems (such as the AGV scheduling system), and operator terminals. At the same time, the standard process parameters associated with the work order (such as temperature, pressure, speed, etc.) will also be issued to ensure the consistency of production quality.
[0099] Understandably, by seamlessly integrating with ERP and MES systems through standard integration interfaces (such as APIs, Web Services, and middleware databases), this intelligent scheduling system ensures its integration into the enterprise's existing IT architecture, forming a closed-loop data flow from business orders to production execution. It avoids information silos, enabling production plans to be formulated based on accurate orders and real-time operating conditions, and allowing plans to be automatically and accurately transmitted to the execution layer, truly achieving the integration and automation of "planning-execution-monitoring".
[0100] Specifically, the work order issuance module is configured to automatically associate and load the standard process parameters corresponding to the tire tread specification when generating a work order instruction.
[0101] It should be further explained that the work order issuance module is configured to automatically associate and load the standard process parameters that match the tire tread product specifications corresponding to the work order during the process of converting the production schedule into specific work order instructions.
[0102] Furthermore, the system maintains a "product specification-process parameter" mapping database internally or through integration with the upper-level PLM (Product Lifecycle Management) system. This database stores a verified "golden set of process parameters" corresponding to each tread specification (uniquely determined by tread code, size, ply grade, rubber compound formula, etc.). When the work order issuance module generates a specific work order based on the production schedule (e.g., "Produce order A on line L1, specification 205 / 55R16 comfort type"), it automatically queries the database using the specification code of the work order as an index to obtain the corresponding standard process parameters. These parameters typically include, but are not limited to: temperature setpoints for each section of the barrel, die plate temperature, screw speed, traction speed, cooling water temperature, and cutting length. Then, the work order issuance module encapsulates these process parameters as necessary components of the work order instruction, along with the production instruction (start time, quantity, etc.), and sends them to the corresponding extrusion production line control system through the MES interface. Upon receiving the work order, the equipment control system can automatically call these parameters for production preparation and settings, thereby achieving "one-click specification change".
[0103] Understandably, this feature deeply integrates production scheduling with process execution, eliminating the risks of delays and errors inherent in the traditional model where planners issue paper work orders and operators manually set equipment parameters based on experience or by consulting manuals. It ensures that every production run adheres to optimal and consistent process standards, significantly improving product quality consistency and stability. Simultaneously, it further reduces reliance on skilled operators and shortens preparation time during specification changes.
[0104] Specifically, the intelligent scheduling engine, work order issuance module, and monitoring and rescheduling module are integrated and deployed on the same server or server cluster.
[0105] It should be further explained that the intelligent scheduling engine, work order issuance module, and monitoring and rescheduling module are integrated and deployed as software services on the same physical server or in a cluster environment consisting of multiple servers.
[0106] Furthermore, the system's software architecture typically employs a layered or microservice design, but all core services (order parsing, rule initial screening, multi-objective optimization, simulation verification, work order assembly and distribution, data monitoring, reordering triggers, etc.) are deployed on the same high-performance, high-reliability server or server cluster. These services exchange data and collaborate through efficient internal communication mechanisms (such as message queues and RPC calls). The database (storing orders, equipment, process parameters, historical data, etc.) can be deployed within the same cluster or in an independent storage system accessible via a high-speed network. This centralized deployment approach helps ensure low latency and high throughput for data interaction between modules, meets the intensive computing resource requirements of scheduling calculations (especially optimization algorithm iterations), and facilitates unified system maintenance, upgrades, and expansion. The preferred server operating system is Linux, with a runtime environment including Java or Python. The database can be a relational database such as PostgreSQL or MySQL for storing business data, and can also be paired with a time-series database for storing high-frequency real-time monitoring data.
[0107] Understandably, integrating and deploying core functional modules avoids the network latency, data inconsistencies, and complex operation and maintenance issues that may arise from distributed deployment. This allows the entire intelligent scheduling system to operate as a unified, highly cohesive "intelligent brain," capable of rapidly processing orders, executing complex optimization algorithms, and responding in real-time to changes in the production environment, ensuring the overall system's responsiveness, stability, and reliability.
[0108] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.
[0109] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.
Claims
1. An intelligent scheduling and automatic planning system for the tread pressing process, characterized in that, include: The order parsing module is used to receive and process production orders from the upstream system and build a pool of orders that can be scheduled for production. The intelligent scheduling engine is connected to the order parsing module and is used to perform multi-level scheduling optimization on the pool of orders that can be scheduled to generate a detailed scheduling plan. The multi-level scheduling optimization includes initial screening of orders based on preset rules and multi-objective global optimization of the order sequence after initial screening. The work order issuance module is connected to the intelligent production scheduling engine and is used to convert the production plan into work order instructions that can be executed by the equipment, and automatically issue them to the corresponding tire tread extrusion production equipment. The monitoring and rescheduling module is used to collect production execution data in real time and compare it with the production schedule. When the preset rescheduling conditions are triggered, it drives the intelligent scheduling engine to dynamically reschedule the unexecuted plans.
2. The system according to claim 1, characterized in that, The intelligent scheduling engine includes a rule-based initial screening unit, a multi-objective optimization unit, and a simulation verification unit connected in sequence.
3. The system according to claim 2, characterized in that, The rule screening unit is configured to filter, merge, and initially sort orders based on at least one hard rule and clustering rule, including order specifications and mouthpiece matching, equipment capability range, and order similarity.
4. The system according to claim 2 or 3, characterized in that, The multi-objective optimization unit is configured to establish an optimization model with the objective function of minimizing the total mold change time, minimizing the total delay time, minimizing the total energy consumption, and maximizing the equipment load balance, and with the production line, die plate, and start time of the order allocation as decision variables, and to solve the model using an optimization algorithm.
5. The system according to claim 4, characterized in that, The optimization algorithm includes a genetic algorithm or a simulated annealing algorithm.
6. The system according to claim 2, characterized in that, The simulation verification unit is configured to perform time extrapolation and conflict detection on the production scheduling scheme output by the multi-objective optimization unit based on the discrete event simulation model, and automatically fine-tune the detected conflicts to output the final executable production scheduling instruction set.
7. The system according to claim 1, characterized in that, The monitoring and rescheduling module is configured such that the rescheduling conditions include at least one of equipment failure, material shortage, order change, and predicted schedule delay; the dynamic rescheduling follows the principle of minimum disturbance, prioritizing impact range assessment and local rescheduling.
8. The system according to claim 1, characterized in that, The system also includes integration interfaces with Manufacturing Execution System (MES) and Enterprise Resource Planning (ERP) systems.
9. The system according to claim 1 or 8, characterized in that, The work order issuance module is configured to automatically associate and load the standard process parameters corresponding to the tire tread specification when generating a work order instruction.
10. The system according to claim 1, characterized in that, The intelligent scheduling engine, work order issuance module, and monitoring and rescheduling module are integrated and deployed on the same server or server cluster.