Filter processing control method and system

Abstract

The application belongs to the technical field of production scheduling and control, and discloses a filter processing control method and system, wherein the steps include: obtaining a production order list, current health degree evaluation values of each key device, and a product-device health degree influence matrix; based on the production order list, the current health degree evaluation values of each key device, and the product-device health degree influence matrix, an optimal processing scheduling scheme is determined in combination with optimization targets and constraint conditions; the optimal processing scheduling scheme is issued to a production site to perform processing production, and the device states of the key devices are continuously monitored, and the real-time health degree changes of the key devices are compared with planned prediction conditions; if the real-time health degree of any key device significantly decreases at a rate faster than the prediction result or approaches a warning threshold, processing scheduling scheme adjustment is performed for the remaining production tasks; thereby, the processing scheduling scheme of the filter can be effectively optimized.

Description

Technical Field

[0001] This application relates to the field of production scheduling and control technology, and more specifically, to a filter processing control method and system. Background Technology

[0002] Automotive filter manufacturing plants face the challenge of mixed-line production of multiple filter varieties. Different models are processed on shared production equipment, and their differences in structure, materials, and processes lead to varying cumulative effects on the equipment. For example, in key processes such as stamping, folding, and bonding, the production of specific models will affect key equipment performance indicators such as stamping die wear, folding machine tool precision, and bonding curing oven temperature uniformity at different rates and modes. These dynamic changes in equipment performance directly relate to product processing quality and production stability. Excessive wear of stamping dies may lead to defects in stamped parts; decreased precision of folding machine tools affects the folding quality of filter elements; and temperature fluctuations in the curing oven affect bonding strength. These quality issues not only increase costs but may also disrupt production rhythm.

[0003] Traditional production scheduling methods are typically based on equipment rated capacity, failing to fully consider the real-time performance status of equipment and the differentiated impact of different products on equipment status. This makes it difficult to predict when equipment will reach a critical point, thus hindering effective mitigation of quality or interruption risks. Equipment degradation can ultimately lead to unplanned downtime or adjustments, further disrupting production schedules. Existing scheduling systems struggle to proactively consider the cumulative impact of equipment status during planning and cannot dynamically adjust based on real-time status during production execution to address the risks posed by equipment performance degradation. Therefore, there is an urgent need for a method and system that integrates the dynamic performance status of equipment as a key input into production scheduling decisions to achieve more robust and efficient multi-product mixed-line production.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] The purpose of this application is to provide a filter processing control method and system that can effectively optimize the filter processing scheduling scheme.

[0006] In a first aspect, this application provides a filter processing control method for optimizing the processing scheduling scheme of filters in a mixed-line processing scenario where multiple types of filters share a production equipment. The steps of this method include:

[0007] A1. Obtain the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health impact matrix; the production order list includes product model, quantity, and delivery date; the product-equipment health impact matrix records the unit impact coefficient of each product model on the health of each key piece of equipment;

[0008] A2. Based on the production order list, the current health assessment values ​​of each key piece of equipment, and the product-equipment health impact matrix, combined with optimization objectives and constraints, determine the optimal processing scheduling scheme; the processing scheduling scheme includes the production sequence and the batch allocation scheme for each process; the optimization objectives include minimizing the total production time, minimizing the number of equipment switchovers, maximizing the overall health maintenance time of equipment, and balancing the rate of decline in the health of each piece of equipment; the constraints include all orders being completed before the delivery date, the production sequence conforming to the process flow, and the health assessment values ​​of all key pieces of equipment remaining above their preset health thresholds after all planned production tasks are completed;

[0009] A3. The optimal processing scheduling scheme is sent to the production site for processing and production, and the equipment status of each key equipment is continuously monitored. The real-time health status changes of each key equipment are compared with the planned forecast.

[0010] A4. If the real-time health rate of any critical equipment decreases significantly faster than the predicted result, or approaches the warning threshold, the processing scheduling plan for the remaining production tasks will be adjusted.

[0011] Secondly, this application provides a filter processing control system for optimizing the processing scheduling scheme of filters in mixed-line processing scenarios where multiple types of filters share production equipment. The system includes:

[0012] The information acquisition module is used to acquire a production order list, the current health assessment value of each key piece of equipment, and a product-equipment health impact matrix. The production order list includes product models, quantities, and delivery dates. The product-equipment health impact matrix records the unit impact coefficient of each product model on the health of each key piece of equipment.

[0013] The scheduling optimization module is used to determine the optimal processing scheduling scheme based on the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health impact matrix, combined with optimization objectives and constraints. The processing scheduling scheme includes the production sequence and the batch allocation scheme for each process. The optimization objectives include minimizing the total production time, minimizing the number of equipment switchovers, maximizing the overall health maintenance time of the equipment, and balancing the rate of decline in the health of each piece of equipment. The constraints include all orders being completed before the delivery date, the production sequence conforming to the process flow, and the health assessment value of all key equipment remaining above its preset health threshold after all planned production tasks are completed.

[0014] The execution and monitoring module is used to send the optimal processing scheduling plan to the production site for processing and production, and to continuously monitor the equipment status of each key piece of equipment, and compare the real-time health changes of each key piece of equipment with the planned forecast.

[0015] The scheduling adjustment module is used to adjust the processing scheduling scheme for the remaining production tasks when the real-time health rate of any critical equipment decreases significantly faster than the prediction result or approaches the warning threshold.

[0016] Beneficial effects: The filter processing control method and system provided in this application optimize the filter processing scheduling scheme by considering the real-time performance status of the equipment and the differentiated impact of different products on the equipment status, thereby achieving effective optimization of the filter processing scheduling scheme. Attached Figure Description

[0017] Figure 1 A flowchart of a filter manufacturing control method provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the filter processing control system provided in an embodiment of this application.

[0019] Labeling Explanation: 1. Information Acquisition Module; 2. Scheduling Optimization Module; 3. Execution and Monitoring Module; 4. Scheduling Adjustment Module. Detailed Implementation

[0020] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0021] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] refer to Figure 1 This application proposes a filter processing control method for optimizing the filter processing scheduling scheme in a mixed-line processing scenario where multiple types of filters share a production equipment. The method includes the following steps:

[0023] A1. Obtain the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health impact matrix; the production order list includes product model, quantity, and delivery date; the product-equipment health impact matrix records the unit impact coefficient of each product model on the health of each key piece of equipment;

[0024] A2. Based on the production order list, the current health assessment values ​​of each key piece of equipment, and the product-equipment health impact matrix, combined with optimization objectives and constraints, determine the optimal processing scheduling scheme; the processing scheduling scheme includes the production sequence and the batch allocation scheme for each process; the optimization objectives include minimizing the total production time, minimizing the number of equipment switchovers, maximizing the overall health maintenance time of equipment, and balancing the rate of decline in the health of each piece of equipment; the constraints include all orders being completed before the delivery date, the production sequence conforming to the process flow, and the health assessment values ​​of all key pieces of equipment remaining above their preset health thresholds after all planned production tasks are completed;

[0025] A3. The optimal processing scheduling scheme is sent to the production site for processing and production, and the equipment status of each key equipment is continuously monitored. The real-time health status changes of each key equipment are compared with the planned forecast.

[0026] A4. If the real-time health rate of any critical equipment decreases significantly faster than the predicted result, or approaches the warning threshold, the processing scheduling plan for the remaining production tasks will be adjusted.

[0027] In step A1, obtaining the current health assessment value of each key piece of equipment refers to quantifying the performance status of the equipment at a specific moment. This can be achieved by collecting the equipment's operating parameters, such as die wear of a stamping machine, tool precision of a folding machine, and temperature uniformity of a bonding and curing oven, and calculating based on the mapping relationship between these parameters and equipment health. The purpose is to provide an objective basis for scheduling decisions regarding the current health level of the equipment. The product-equipment health impact matrix records the predictive impact of the unit processing volume (e.g., per 1000 pieces) of different product models on the health of each key piece of equipment. This can be pre-built using historical data analysis, experiments, or expert experience and stored in a local database, retrieved when needed. Its main purpose is to quantify the differential impact of different products on equipment wear and tear, enabling scheduling to specifically consider these differences.

[0028] In step A2, the optimal processing scheduling scheme refers to a production plan that achieves the best balance among multiple optimization objectives while satisfying all preset constraints. This includes determining the production sequence of different product models and the batch size of each process. Its main purpose is to guide the production site in efficiently and reliably arranging production tasks. Optimization objectives include minimizing total production time, minimizing equipment changeovers, maximizing overall equipment health maintenance time, and balancing the rate of decline in the health of individual equipment. These objectives comprehensively consider production efficiency, auxiliary time, equipment lifespan, and load balancing. The aim is to comprehensively balance production demand and equipment status in scheduling decisions. Constraints include all orders being completed before the delivery date, the production sequence conforming to the process flow, and the health assessment values ​​of all critical equipment remaining above their preset health thresholds after all planned production tasks are completed. These constraints ensure the feasibility of the scheduling scheme, product quality, and equipment safety. Their main purpose is to guarantee the timely completion of production tasks, the correct execution of processes, and to avoid production interruptions or quality problems caused by low equipment status.

[0029] In step A3, continuously monitoring the equipment status of each key piece of equipment and comparing the real-time health changes of each key piece of equipment with the planned forecasts means acquiring the equipment's operating data in real time during production execution, calculating its health changes, and comparing them with the health change trends predicted when the scheduling plan was formulated. Its main purpose is to promptly detect abnormal fluctuations in equipment status.

[0030] In step A4, adjusting the processing schedule for the remaining production tasks means that when monitoring reveals that the rate of equipment health decline is significantly faster than the predicted rate (e.g., if the difference between the rate of health decline and the predicted rate of health decline exceeds a preset deviation threshold, then the actual rate of health decline is determined to be significantly faster than the predicted rate) or close to the warning threshold (e.g., if the deviation between the rate of health decline and the warning threshold falls within a preset tolerance range, then the actual rate of health decline is determined to be close to the warning threshold), the original scheduling plan is recalculated or modified based on the current equipment status and remaining tasks. (Specifically, an updated production order list can be generated based on the remaining tasks, and steps A1-A4 can be executed again for this updated production order list). This is mainly to cope with emergencies, reduce risks, and ensure production continuity.

[0031] The core innovation of this application lies in integrating the dynamic health status of equipment and its interaction with different products as key inputs into the multi-objective production scheduling optimization process. Combined with real-time monitoring and dynamic adjustment mechanisms in the production process, predictive scheduling and risk avoidance based on equipment health status are realized. This effectively solves the problem of neglecting equipment status changes in traditional methods, improves production stability and efficiency, and extends equipment lifespan.

[0032] Specifically, this method first acquires information on production task requirements, the current health status of equipment, and the impact patterns of different products on equipment health. Based on this information, combined with multiple preset optimization objectives and necessary constraints, a preliminary optimal production scheduling scheme is calculated, which determines the production sequence of products and the batch size of each process. Subsequently, this scheme is deployed to the production site for execution. During execution, the system continuously collects real-time equipment status data and converts it into real-time health changes, comparing it with the health change trend predicted during the initial scheme calculation. If the actual rate of decline in equipment health is found to be significantly faster than predicted, or if the equipment health is close to a preset warning threshold, it indicates that the equipment may face abnormal risks. At this point, the system triggers the rescheduling of unfinished production tasks, generating a new optimization scheme to address the current equipment status and avoid potential quality problems or downtime risks. The entire process forms a closed loop: data acquisition, optimization decision-making, execution monitoring, and anomaly adjustment, achieving the perception, prediction, optimization, and dynamic response to equipment health status.

[0033] By incorporating the dynamic health status of equipment into scheduling decisions, this application can predict potential changes in equipment status and prevent premature equipment degradation caused by continuous production of products with high wear and tear. By balancing the rate of health decline of each piece of equipment, it can prevent certain equipment from becoming bottlenecks and extend the overall service life of the equipment. Real-time monitoring and dynamic adjustment mechanisms can promptly respond to abnormal fluctuations in equipment status, reduce unplanned downtime and maintenance, and ensure the continuity and stability of production. Comprehensive multi-objective optimization can balance production efficiency and cost while ensuring equipment health. Ultimately, this improves the quality stability, production efficiency, and equipment utilization rate of filter processing in multi-product mixed-line production scenarios.

[0034] In some implementations, step A1, which involves obtaining the current health assessment value of each key device, includes:

[0035] Collect equipment status parameters for each key piece of equipment; the key equipment includes a stamping machine, a folding machine, and a bonding and curing oven; the equipment status parameters include the wear amount of the stamping die edge of the stamping machine, the tool contour accuracy deviation of the folding machine, and the internal temperature uniformity deviation of the bonding and curing oven.

[0036] Based on the equipment status parameters of each key device, and the mapping relationship between the equipment status parameters and the equipment health of each key device, the current health assessment value of each key device is determined.

[0037] Equipment status parameters refer to key physical quantities or indicators that directly reflect the operating performance, wear level, or working status of equipment. These parameters can be obtained through sensor acquisition, image recognition and analysis, signal processing, or manual inspection. The equipment status parameter-equipment health mapping relationship refers to the rules, models, or data structures that establish the correspondence between equipment status parameters and equipment health assessment values. This can be achieved using statistical models based on historical operating data and maintenance records, expert experience rule bases, or predictive models trained through machine learning algorithms. This equipment status parameter-equipment health mapping relationship is pre-established and stored in a local database, and is retrieved when needed.

[0038] The method provided in this application aims to accurately obtain the current health assessment values ​​of key equipment, thereby providing a more reliable data foundation for the overall filter processing control method. First, by collecting equipment status parameters of each key piece of equipment, the operating status of the equipment is quantified. For a stamping press, the wear of the stamping die cutting edge is monitored; for a folding machine, the deviation of the tool contour accuracy is monitored; and for a bonding and curing oven, the deviation of the internal temperature uniformity is monitored. These parameters directly reflect the key performance indicators of the equipment and can more accurately characterize the health status of the equipment. Second, based on the collected equipment status parameters and the corresponding equipment status parameter-equipment health assessment mapping relationship, the current health assessment value of each key piece of equipment is determined. This mapping relationship establishes a connection between equipment status parameters and health assessment values. Through a pre-set mapping relationship, specific equipment status parameters can be transformed into quantifiable health assessment values, thus providing a basis for subsequent scheduling optimization. In this way, the transformation from equipment status parameters to health assessment values ​​is realized, making the assessment of equipment health status more objective and accurate. Using this accurate health assessment value as input to the processing scheduling method allows subsequent scheduling optimizations to fully consider the actual carrying capacity and health status of the equipment. For example, when determining production sequences and batch allocations, it can avoid assigning high-load tasks to equipment with low health, or adjust batch sizes to balance the rate of decline in equipment health. This makes the overall scheduling scheme more forward-looking and robust, effectively avoiding production problems caused by equipment deterioration.

[0039] By collecting parameters that directly reflect key equipment performance, such as the wear of stamping die cutting edges, the accuracy deviation of folding machine tool contours, and the temperature uniformity deviation inside the bonding and curing oven, the actual operating status of the equipment can be captured more accurately. Based on the mapping relationship between equipment status parameters and equipment health, these parameters are transformed into quantified health assessment values, making the assessment of equipment health status more objective and standardized. This accurate health assessment provides a reliable data foundation for subsequent processing scheduling optimization, enabling scheduling schemes to fully consider the actual load-bearing capacity and potential risks of the equipment. This helps to formulate more reasonable production plans, avoid assigning high-load tasks to equipment with low health, thereby reducing excessive equipment wear and extending equipment life. At the same time, by avoiding quality problems that may be caused by equipment deterioration, the product qualification rate is improved and rework costs are reduced. In addition, accurate health assessment also makes the prediction of equipment maintenance needs more precise, which helps to arrange planned maintenance, reduce unplanned downtime, and improve the overall stability and efficiency of the production line.

[0040] It should be noted that key equipment is not limited to the equipment listed above, nor are equipment status parameters limited to those listed above; key equipment and equipment status parameters can be adjusted according to actual needs.

[0041] Furthermore, step A1, the step of obtaining the current health assessment value of each key device, may also include:

[0042] Collect environmental parameters of the environment where each key piece of equipment is located; these environmental parameters include workshop temperature, humidity, and air cleanliness.

[0043] Based on the collected environmental parameters, calculate the health impact correction coefficient for each key piece of equipment;

[0044] Based on the aforementioned health impact correction coefficient, the current health assessment value of each key device is corrected.

[0045] Specifically, collecting environmental parameters of the environment where key equipment is located can be achieved by deploying corresponding environmental sensors near the equipment. For example, temperature sensors, humidity sensors, and air quality sensors or dust concentration sensors can be deployed to acquire real-time or periodic data on workshop temperature, humidity, and air cleanliness in the equipment's operating environment. These environmental parameters are important external factors affecting equipment health. For example, high temperature and humidity may accelerate the corrosion and aging of metal parts, while particulate matter in the air may cause wear or blockage of precision components.

[0046] Based on collected environmental parameters, correction coefficients are calculated to assess the impact of current environmental conditions on equipment health. These correction coefficients quantify the degree of influence of environmental conditions on equipment health. The calculation of these correction coefficients can be based on a pre-established model of the relationship between environmental parameters and equipment health. This model can be constructed through analyzing historical data, conducting experimental tests, or combining expert experience with lookup tables, mathematical formulas, or machine learning models. Different equipment types (such as stamping presses, folding machines, and bonding curing ovens) have varying sensitivities to environmental parameters; therefore, corresponding correction coefficient calculation methods need to be established for different equipment. For example, correction factors for the corrosion or wear rate of stamping dies or folding machine tools can be calculated based on workshop temperature and humidity, or correction factors for the performance degradation of precision components inside the equipment can be calculated based on air cleanliness.

[0047] Subsequently, based on the health impact correction coefficient, the current health assessment values ​​of each key piece of equipment are corrected. The correction process involves combining the preliminary health assessment values ​​obtained from equipment status parameters with the calculated health impact correction coefficient to obtain a more accurate assessment value that reflects the actual health status of the equipment in the current environment. The correction method can be either multiplying the preliminary assessment value by the correction coefficient or adding the correction coefficient as an offset to the preliminary assessment value; the specific method depends on the definition of the correction coefficient and the established model.

[0048] The working principle of this solution is to combine the impact of the equipment's inherent state parameters with external environmental parameters on its health. First, by collecting equipment state parameters, the health status caused by the equipment's own use and wear can be assessed. Simultaneously, by collecting environmental parameters and calculating correction coefficients, the additional impact of the current environment on the equipment's health can be quantified. Combining these two pieces of information, through a correction process, ensures that the final health assessment value not only reflects the equipment's inherent state but also takes into account the accelerating or mitigating effects of the external environment. For example, even if the equipment state parameters indicate acceptable health, if the environment is characterized by high temperature, high humidity, or high pollution, the correction coefficient will lower the health assessment value, thus more accurately reflecting the equipment's actual health level and potential risks under the current harsh environment. This comprehensive assessment method, which considers both internal and external factors, provides more comprehensive and accurate equipment health information.

[0049] In some implementations, step A2 includes:

[0050] A201. For the production order list, based on the product model and the product-equipment health influence matrix, the products are divided into multiple product families using the K-means algorithm, so that the products in the same product family have similar health influence on the key equipment, and a product family list is obtained.

[0051] A202. Based on the product family list and the current health assessment value of each key equipment, a genetic algorithm is used to generate multiple initial production sequences. Combined with the product-equipment health influence matrix, the equivalent health decline rate of each key equipment under each initial production sequence is calculated to screen candidate production sequences from multiple initial production sequences.

[0052] A203. For each candidate production sequence, construct a batch allocation model with the objective of minimizing the variance of the health decline rate among equipment. Use dynamic programming algorithm to determine the optimal batch allocation scheme for each process. Under the constraints of order delivery period and process flow, balance the health decline rate of each equipment and calculate the total production time, number of equipment changeovers and overall equipment health maintenance time under the scheme.

[0053] A204. Based on the total production time, number of equipment changeovers, overall equipment health maintenance time, and variance of the rate of decline in health between equipment, the TOPSIS method is used to comprehensively evaluate each candidate production sequence and its batch allocation scheme, and the scheme with the highest comprehensive evaluation score is selected as the optimal processing scheduling scheme.

[0054] In step A201, the K-means algorithm is an iterative clustering analysis algorithm that divides a dataset into a predetermined number of clusters, such that the distance between data points within a cluster is less than the distance between data points between clusters. It can use Euclidean distance or Manhattan distance as the distance metric. A product family refers to a set of products with similar characteristics grouped together by a clustering algorithm. In this scheme, products within a product family have a similar impact on the health of key equipment.

[0055] In step A202, the genetic algorithm refers to a random search algorithm that simulates the biological evolution process in nature. It iteratively optimizes the population through operations such as selection, crossover, and mutation to find the optimal solution to the problem. It can be implemented using binary encoding or real-number encoding. The equivalent health rate of decline is an indicator that measures the degree of health decline of equipment per unit time due to production tasks. This indicator comprehensively considers the specific impact of different products on the equipment and the duration of the production process.

[0056] In step A203, the batch allocation model refers to a mathematical model describing how to decompose the total production task into multiple production batches and allocate them to different processes or equipment. This model includes an objective function and a series of constraints. Dynamic programming is a method for optimizing multi-stage decision-making processes. It obtains the global optimum by decomposing the problem into interrelated subproblems and solving them sequentially. It can be implemented using a bottom-up or top-down approach. The variance of the rate of health decline among equipment measures the dispersion of the rate of health decline of different key equipment relative to the average rate of decline after completing the production task. A smaller variance indicates a more balanced health consumption among the equipment.

[0057] In step A204, the TOPSIS method refers to a multi-attribute decision analysis method that evaluates the relative merits of each alternative solution by calculating the distance between each alternative solution and the ideal optimal and ideal worst solutions. It can be implemented using Euclidean distance or Chebyshev distance as the distance calculation method.

[0058] This solution addresses the balance between equipment health and production efficiency in multi-product mixed-line production through a series of collaborative steps. First, the received production order list is categorized using the K-means algorithm to create a product family list. This categorization is based on product model and its impact on equipment health, grouping products with similar impacts together, effectively simplifying subsequent scheduling optimization and laying the foundation for sequence generation based on product families. Next, based on the obtained product family list and the current health assessment values ​​of each key piece of equipment, a genetic algorithm generates multiple initial production sequences. The global search capability of the genetic algorithm helps explore a broad production sequence space and find potential optimization orders. Simultaneously, combined with the product-equipment health impact matrix, the equivalent health decline rate of each key piece of equipment under each initial production sequence is calculated. By evaluating the impact of different sequences on equipment health, candidate production sequences with relatively small impacts on equipment health can be preliminarily screened, thus considering equipment health maintenance at the sequence level. Finally, for each selected candidate production sequence, a batch allocation model is constructed with the objective of minimizing the variance of the health decline rate among equipment. This model seeks the optimal batch allocation scheme under constraints such as order delivery time and process flow. A dynamic programming algorithm is used to solve the model, decomposing the complex batch allocation problem into multiple sub-problems. By progressively optimizing the batch allocation decisions at each stage, the globally optimal batch allocation scheme that balances the rate of health degradation across all equipment is ultimately determined. This balanced batch allocation strategy helps prevent premature degradation of some equipment due to continuous high-load production, achieving a smooth depletion of equipment health. Finally, for each candidate production sequence and its corresponding optimal batch allocation scheme, multiple evaluation indicators are calculated, including total production time, number of equipment changeovers, overall equipment health maintenance time, and variance of the rate of health degradation among equipment. The TOPSIS method is used to comprehensively evaluate these schemes. This method can balance multiple conflicting optimization objectives and select the optimal scheme that performs balancedly across all indicators. Through the organic combination of the above steps, this scheme comprehensively considers equipment health and production efficiency from product classification and production sequence generation to batch allocation optimization, generating an optimal processing scheduling scheme that better meets actual production needs.

[0059] Consider a specific implementation scenario, such as a filter manufacturing plant that produces various models of air filters and oil filters, sharing key equipment such as stamping presses, folding machines, and bonding and curing ovens. First, obtain the current production order list, which includes the quantity and delivery date of different filter models. Simultaneously, obtain the current health assessment values ​​of the stamping press, folding machine, and bonding and curing oven, as well as a product-equipment health influence matrix recording the unit influence coefficient of each filter model on the health of these machines. Next, for each product model in the production order, based on the product-equipment health influence matrix, use the K-means algorithm to divide these products into several product families. For example, products that significantly affect the wear of stamping press dies can be grouped into one category, and products that significantly affect the precision of folding machine cutting tools can be grouped into another. This yields a list of product families. Then, based on this list of product families and the current health assessment values ​​of the equipment, a genetic algorithm is used to generate multiple possible production sequences. For example, if product family 1 contains products A and B, and product family 2 contains product C, an initial sequence could be to produce product family 1 first, then product family 2. Within product family 1, the order could be A then B, or B then A. For each generated initial sequence, the equivalent health decline rate of the stamping machine, folding machine, and bonding curing oven under that sequence is calculated using the product-equipment health influence matrix. Based on these decline rates, several production sequences with relatively low or balanced equivalent health decline rates are selected as candidates. For each candidate production sequence, a batch allocation model is constructed. The goal of this model is to minimize the variance of the health decline rate of the stamping machine, folding machine, and bonding curing oven, while also satisfying constraints such as all orders being completed before the delivery date and the production sequence conforming to the process flow. A dynamic programming algorithm is used to solve this batch allocation model to determine the optimal batch size and batch order for producing different products in each process (e.g., stamping, folding, bonding). After determining the optimal batch allocation scheme, the total production time, number of equipment changeovers, and overall equipment health maintenance time under this scheme are calculated. Finally, based on the calculated total production time, number of equipment changeovers, overall equipment health maintenance time, and variance of the rate of decline in health among equipment, the TOPSIS method is used to comprehensively evaluate all candidate production sequences and their corresponding optimal batch allocation schemes. The TOPSIS method calculates the distance between each scheme and the ideal optimal and ideal worst-case schemes, and calculates a comprehensive evaluation score based on the distance. The scheme with the highest comprehensive evaluation score is selected as the final optimal processing scheduling scheme.

[0060] The above technical solutions effectively categorize products with similar impacts on equipment health, simplifying subsequent scheduling optimization. Production sequences generated based on product families and equipment health can initially consider equipment health maintenance, selecting production sequences with relatively minor impacts on equipment health. By constructing a batch allocation model with the objective of minimizing the variance of health decline rates among equipment and employing dynamic programming, a balanced health consumption of key equipment can be achieved while meeting production constraints, avoiding excessive wear and tear on localized equipment. Finally, the TOPSIS method is used for comprehensive evaluation of multiple objectives, balancing factors such as production efficiency, changeover costs, and equipment health to select the optimal scheduling scheme that balances performance across all aspects. These measures work synergistically, enabling the generated processing scheduling scheme to more effectively balance equipment health and production efficiency, extend equipment lifespan, reduce the risk of unplanned downtime, and improve production stability.

[0061] Preferably, step A202 may include:

[0062] B1. Based on the current health assessment values ​​of each key device, set the initial population size, selection probability, crossover probability, and mutation probability for the genetic algorithm;

[0063] B2. Based on the product family list, and according to the initial population size, multiple initial production sequences are generated using a segmented coding method; where each product family corresponds to a segment in the production sequence, and the product order within the segment is randomly generated;

[0064] B3. For each initial production sequence, based on the product-equipment health impact matrix, calculate the predicted decay value of the health of each key piece of equipment and the corresponding production time after completing the processing of each reference batch of products under the initial production sequence;

[0065] B4. Based on the predicted decay value and the corresponding production time, calculate the equivalent health rate of each key equipment, and calculate the fitness function value of the genetic algorithm with the goal of minimizing the maximum equivalent health rate of decay.

[0066] B5. Based on the selection probability, crossover probability, and mutation probability, perform selection, crossover, and mutation operations of the genetic algorithm to generate a new initial production sequence, and repeat steps B3 and B4 until the preset number of iterations is met or the convergence condition is reached.

[0067] B6. Based on the fitness function value, select the multiple production sequences with the highest fitness from the final set of initial production sequences as candidate production sequences.

[0068] In step B1, the operating parameters of the genetic algorithm are dynamically adjusted. For example, when the health of the devices is generally low, the population size or mutation probability can be increased to enhance the global search capability; when the health of the devices is high, the mutation probability can be appropriately reduced to speed up the convergence.

[0069] Segmented coding refers to representing a complete production sequence as a structure composed of multiple segments, where each segment corresponds to a product family. The order of products within a segment can be randomly arranged. This coding method can ensure that products of the same product family are adjacent in the sequence, while allowing changes in the order within the family.

[0070] The equivalent health rate of decline characterizes the average decline in equipment health within a unit production time. It comprehensively considers the absolute decline in health and the production time required to achieve that decline, and can be used to measure the long-term impact of different production sequences on equipment health. Minimizing the maximum equivalent health rate of decline means seeking, during optimization, the production sequence that minimizes the decline rate of the fastest-decreasing equipment among all critical equipment. This helps balance the health losses of various equipment and prevents any single equipment from deteriorating too quickly and becoming a bottleneck. The fitness function value is a numerical value used to evaluate the quality of each production sequence in the genetic algorithm. This value is related to the optimization objective (minimizing the maximum equivalent health rate of decline). A production sequence with a higher fitness value indicates a lower maximum equivalent health rate of decline, making it a better solution. The fitness function can be set according to actual needs, for example, Fitness = K / (1 + max(Variance)), where Fitness is the fitness function value, max(Variance) represents the maximum equivalent health rate of decline, and K is a constant.

[0071] Among them, the candidate production sequence refers to a set of production sequences with high fitness selected from the final population after iterative optimization by the genetic algorithm. These sequences represent the better production order options considering the influence of equipment health and will be used as input for subsequent batch allocation steps.

[0072] This scheme optimizes the performance of the genetic algorithm in generating candidate production sequences through a series of steps. First, based on the current health assessment values ​​of key equipment, the key parameters of the genetic algorithm are dynamically adjusted, allowing the search strategy to adapt to the actual state of the equipment. Next, based on the product family list, an initial production sequence population is constructed using a segmented encoding method. This encoding method integrates the product family structure into the sequence representation, helping to reduce equipment switching while maintaining the diversity of orders within the family. Then, using the product-equipment health influence matrix, the equipment health decay and production time under each production sequence are predicted, and the equivalent health decay rate is calculated. This rate index comprehensively reflects the long-term impact of the production sequence on equipment health. Minimizing the maximum equivalent health decay rate is used as the optimization objective to guide the search direction of the genetic algorithm, making it tend to generate production sequences that balance the health losses of various equipment. Through selection, crossover, and mutation operations of the genetic algorithm, the population continuously evolves, and sequences with high fitness are retained and optimized. After a preset number of iterations or reaching the convergence condition, the multiple sequences with the highest fitness are selected from the final population as candidate production sequences. These candidate sequences are the preferred options selected after fully considering the impact of equipment health. This approach, combined with steps such as dividing products into product families, constructing a batch allocation model, and using the TOPSIS method for comprehensive evaluation, forms a complete processing scheduling optimization process. By providing candidate production sequences with high adaptability that take equipment health into account, this approach lays the foundation for subsequent batch allocation and overall scheme evaluation, enabling the final determined processing scheduling scheme to more effectively balance production efficiency, equipment changeover, and equipment health, thereby solving the problem of existing technologies struggling to cope with dynamic changes in equipment status.

[0073] For example, in practical implementation, the parameters of the genetic algorithm can be dynamically adjusted based on the current health assessment values ​​of the stamping machine, folding machine, and bonding curing oven. If the health of the equipment is generally at a moderate level, the initial population size can be set to a moderate value, such as 100 production sequences; the selection probability can be set to 0.8, the crossover probability to 0.9, and the mutation probability to 0.05. Based on the product family list partitioned by the K-means algorithm, for example, containing product family A (product models X, Y) and product family B (product models Z, W), the initial production sequences are generated using a segmented encoding method. An initial sequence can be represented as [segment of product family A, segment of product family B], where the segment of product family A can be [X,Y] or [Y,X], and the segment of product family B can be [Z,W] or [W,Z]. For example, an initial sequence can be [X,Y,Z,W]. For this sequence, based on the product-equipment health influence matrix, the predicted decay values ​​and corresponding production times for the health of the stamping press, folding machine, and curing oven for producing a reference batch (i.e., a unit production task, e.g., 1000 units) of products X, Y, Z, and W are calculated. For example, product X has a greater impact on the health of the stamping press and a smaller impact on the folding machine; product Y has a greater impact on the folding machine and a smaller impact on the stamping press; products Z and W have a greater impact on the curing oven. The total predicted health decay value and total production time for each piece of equipment under the sequence [X,Y,Z,W] are accumulated and calculated. Then, based on these values, the equivalent health decay rate for each piece of equipment is calculated. For example, the equivalent decay rate for the stamping press is obtained by dividing the total decay by the total time; similar calculations are performed for the folding machine and the curing oven. The fitness function value for this sequence is calculated with the objective of minimizing the largest equivalent health decay rate among these three pieces of equipment. Next, based on the set selection probability, crossover probability, and mutation probability, genetic operations are performed on the current population to generate a new set of production sequences, and the fitness is calculated again. This process can be iterated, for example, repeated 1000 times, or until the average fitness of the population no longer shows a significant improvement (i.e., the convergence condition is met). Finally, based on the calculated fitness function value, several sequences with the highest fitness are selected from the set of production sequences generated in the last iteration, for example, the top 10 sequences in terms of fitness, as candidate production sequences for use in subsequent steps.

[0074] Preferably, step A203 may include:

[0075] C1. For each candidate production sequence, based on the product model of each product family in the product family list and combined with the product-equipment health influence matrix, calculate the unit health influence coefficient of each product family on each key equipment, and construct the equipment health influence coefficient matrix.

[0076] C2. Based on the equipment health impact coefficient matrix and the current health assessment value of each key equipment, a batch allocation model is constructed with the objective of minimizing the variance of the health decline rate among equipment. The constraints of the batch allocation model include: ensuring that all orders are completed before the delivery period, ensuring that the production sequence conforms to the process flow, and ensuring that the health assessment value of all key equipment remains above its preset health threshold after all planned production tasks are completed.

[0077] C3. Using a dynamic programming algorithm, the batch allocation problem is decomposed into multiple sub-problems based on the equipment health influence coefficient matrix. Each stage corresponds to the batch allocation decision of a process. The optimal batch allocation scheme for each stage is calculated iteratively until the globally optimal batch allocation scheme that satisfies the constraints and minimizes the variance of the rate of decline in health among equipment is obtained.

[0078] C4. Based on the optimal batch allocation scheme determined by the dynamic programming algorithm, calculate the total production time, number of equipment changeovers, and overall equipment health maintenance time under this optimal batch allocation scheme.

[0079] The product family list refers to a list of products within the same product family. The equipment health impact coefficient matrix is ​​a matrix calculated for a specific candidate production sequence, based on the product models of each product family in the product family list and their positions in the sequence, combined with the product-equipment health impact matrix, reflecting the cumulative health impact of each product family on each key piece of equipment.

[0080] In the batch allocation problem, the batch allocation decision for each process can be considered as a stage. By solving the optimal decision at each stage, the overall optimal solution is gradually approximated. Decomposing the batch allocation problem into multiple stages means dividing the entire batch allocation decision process of the production task into a series of interrelated decision steps according to the process sequence or product family sequence. Each step solves a portion of the batch allocation problem. Iteratively calculating the optimal batch allocation scheme for each stage means, in the dynamic programming process, starting from the initial stage, calculating the value of all possible decision schemes in the current stage based on the optimal solution of the previous stage, selecting the optimal scheme, and then proceeding to the next stage, repeating this process until all stages are completed. The globally optimal batch allocation scheme refers to the specific batch allocation scheme that achieves the optimal value of the optimization objective of the batch allocation model while satisfying all constraints.

[0081] The overall working principle of this scheme is as follows: After obtaining candidate production sequences, the impact of each product family on the health of key equipment under a specific production sequence is first quantified by calculating the equipment health impact coefficient matrix. Then, based on this impact matrix and the current health of the equipment, a batch allocation model is constructed with the objective of minimizing the variance of the rate of health decline among equipment. This model not only pursues balanced use of equipment health but also strictly incorporates constraints that production must meet, such as order delivery time, process flow, and equipment health thresholds. To solve this complex model, a dynamic programming algorithm is employed. Dynamic programming decomposes the entire batch allocation problem into multiple interrelated stage decisions, each stage corresponding to the batch allocation of a process or product family. By iteratively calculating the optimal decision for each stage and using the optimal solutions to subproblems to construct the optimal solution for the larger problem, a globally optimal batch allocation scheme that satisfies all constraints and minimizes the variance of the rate of health decline is finally obtained. Finally, the performance of this optimal scheme is evaluated, calculating key indicators such as total production time, number of equipment changeovers, and overall equipment health maintenance time. This solution, combined with the aforementioned steps for determining candidate production sequences, further refines the management of equipment health at the batch allocation level, building upon the initial screening of production sequences that take health into account in the aforementioned steps. By balancing the rate of decline in the health of each piece of equipment, it effectively addresses the challenges brought about by the dynamic changes in equipment status during multi-variety mixed-line production, thereby improving the robustness and efficiency of production scheduling.

[0082] The following is a detailed explanation of this solution using a specific embodiment. Consider a specific production scenario where multiple models of filters need to be processed sequentially on key equipment such as a stamping press, a folding machine, and a bonding and curing oven. Assume a candidate production sequence has been determined through the aforementioned steps. This solution first consults the product-equipment health impact matrix based on the product models included in the sequence, calculates the unit health impact coefficient of each product on each key piece of equipment, and constructs an equipment health impact coefficient matrix. For example, if the sequence includes models A and B, and the stamping press is a key piece of equipment, then the unit health impact coefficients of models A and B on the stamping press are calculated. Next, a batch allocation model is constructed based on the current health assessment value of the stamping press. The goal of this model is to determine the production batches of models A and B on the stamping press such that the variance of the rate of health decline of the stamping press and other key equipment (such as the folding machine and the curing oven) is minimized. The constraints of the model include ensuring that the total production volume of models A and B meets the order requirements, ensuring that model A is produced before model B (if the sequence is arranged this way), and ensuring that the health of the stamping press is higher than a preset threshold after production is completed. The model is solved using dynamic programming. The batch allocation problem can be decomposed into multiple stages; for example, the first stage determines the batch allocation of model A, and the second stage determines the batch allocation of model B. In each stage, different batch allocation schemes are considered (e.g., model A is produced entirely on press 1, or partially on press 1 and partially on press 2, if multiple similar machines exist), and the impact of each scheme on equipment health and the required time are calculated. Through iterative calculation and comparison, the batch allocation scheme that minimizes the variance of the final equipment health decline rate is found. Finally, based on the determined optimal batch allocation scheme, the total production time required to complete all tasks, the number of machine changeovers, and the expected duration of overall equipment health maintenance are calculated.

[0083] Preferably, step C3 may include:

[0084] C301. For each stage of batch allocation decision, calculate the rate of decline of health of each key device under different batch allocation schemes in the current stage based on the equipment health impact coefficient matrix;

[0085] C302. Based on the calculated health rate decline, an optimization function is constructed with the objective of minimizing the variance of the health rate decline among devices, and a penalty term is introduced to penalize batch allocation schemes that do not meet the constraints of order delivery period, process flow, or device health; wherein, the magnitude of the penalty term is proportional to the degree of constraint violation;

[0086] C303. An improved dynamic programming algorithm is adopted. By introducing a heuristic search strategy, the algorithm prioritizes searching for batch allocation schemes that meet the constraints and have a balanced rate of equipment health decline. Combined with a pruning strategy, the search space is dynamically adjusted according to the remaining equipment health and the remaining production tasks.

[0087] C304. Iteratively calculate the optimal batch allocation scheme for each stage until a globally optimal batch allocation scheme that satisfies the constraints and minimizes the variance of the rate of decline in health among devices is obtained, and record the optimal batch allocation scheme for each stage.

[0088] Batch allocation decision refers to determining the production batch allocated to different equipment for each process during production scheduling. The calculation method for the equipment health degradation rate is described above.

[0089] The optimization function is a mathematical expression used to quantify the merits of batch allocation schemes, with the goal of minimizing the variance of the rate of decline in health among equipment. A penalty term is added to the optimization function; this value increases when a batch allocation scheme violates constraints, thereby lowering the scheme's evaluation score. The degree of constraint violation refers to the severity to which the batch allocation scheme fails to meet order delivery deadlines, process flow, or equipment health constraints.

[0090] The improved dynamic programming algorithm, based on the standard dynamic programming algorithm, introduces additional strategies to enhance search efficiency and optimization performance. Heuristic search strategy is a search method that utilizes experiential knowledge or heuristic information from the problem domain to guide the search direction, prioritizing paths more likely to lead to the optimal solution. Pruning strategy, during the search process, evaluates the potential of the current path or state and terminates branches unlikely to lead to the optimal solution early, thereby reducing the search space. Equipment remaining health refers to the assessed health value of the equipment after completing a portion of its production tasks. Remaining production tasks refer to the total number of products from all production orders that still need to be completed after the current stage. The search space refers to the set of all possible batch allocation schemes that the dynamic programming algorithm needs to consider when searching for the optimal solution.

[0091] Among them, the globally optimal batch allocation scheme refers to the scheme that can satisfy all constraints and minimize the variance of the rate of decline in health among devices among all possible batch allocation schemes.

[0092] This solution addresses the shortcomings of standard dynamic programming algorithms in batch allocation optimization by improving the dynamic programming algorithm. First, based on the equipment health influence coefficient matrix, the rate of health decline for each key piece of equipment under different batch allocation schemes at the current stage is calculated. Quantifying the impact of different allocation schemes on equipment health provides a basis for subsequent optimization. Second, an optimization function is constructed with the objective of minimizing the variance of the rate of health decline among equipment, and a penalty term is introduced. This approach not only considers the balance of equipment health but also ensures the feasibility of the scheme through the penalty term, i.e., satisfying the constraints of order delivery time, process flow, and equipment health. Introducing the penalty term effectively prevents the algorithm from getting trapped in infeasible solutions, ensuring the practicality of the final solution. Then, an improved dynamic programming algorithm is employed. By introducing a heuristic search strategy, it prioritizes searching for batch allocation schemes that satisfy the constraints and have a balanced rate of equipment health decline. This strategy significantly reduces the search space and improves the algorithm's efficiency. Simultaneously, combined with a pruning strategy, the search space is dynamically adjusted based on the remaining equipment health and remaining production tasks, further improving search efficiency and avoiding ineffective searches. Finally, the optimal batch allocation scheme for each stage is iteratively calculated until the global optimal solution is obtained, and the optimal scheme for each stage is recorded, ensuring the global optimality of the final scheme and providing a basis for subsequent scheduling adjustments. The entire scheme, together with the constructed batch allocation model and the determined candidate production sequences, enables the entire scheduling scheme to comprehensively consider multiple factors such as production sequence, batch allocation, equipment health, delivery time, and process flow, thereby finding a feasible scheduling scheme that can balance the rate of equipment health decline in complex multi-product mixed-line production scenarios.

[0093] In one embodiment, for batch allocation decisions at each stage, the rate of health decline of each key piece of equipment under different batch allocation schemes at the current stage can be calculated based on the equipment health influence coefficient matrix. For example, for a certain process, a batch of products needs to be processed, and this batch of products can be allocated to multiple pieces of equipment with the same function. For each possible allocation combination, the total expected decline in health of the equipment after completing the allocated task can be calculated based on the product's influence coefficient on the unit health of the equipment, and then divided by the processing time to obtain the decline rate. Based on the calculated health decline rate, an optimization function can be constructed with the objective of minimizing the variance of the health decline rate among equipment. For example, the optimization function can be expressed as the sum of the squares of the differences between the health decline rate of each piece of equipment and the average decline rate. Simultaneously, a penalty term can be introduced. For example, if a scheme causes an order to be unable to be completed before the delivery date, the penalty term can be set as the number of days exceeding the deadline multiplied by a large weight coefficient; if the process flow is violated, the penalty term can be set to a fixed large value; if it is predicted that the health of a piece of equipment will be below a threshold after the task is completed, the penalty term can be set as the degree to which it is below the threshold multiplied by another weight coefficient. The optimization objective becomes minimizing (variance + penalty term). An improved dynamic programming algorithm is employed. Heuristic search can be introduced into the state transition process of dynamic programming. For example, when considering transitioning from the current state to the next state, schemes that minimize the difference in the rate of health decline among devices can be prioritized. Specifically, when generating candidate transition states, the variance of their corresponding health decline rate and penalty term can be calculated first, and the candidate states can be sorted according to a heuristic function, prioritizing the exploration of states with smaller heuristic function values. Simultaneously, pruning strategies can be incorporated. For example, if a batch allocation scheme causes a device's health to fall below a preset threshold after completing the remaining tasks, or significantly exceed the order delivery period, the exploration of that scheme and its subsequent states can be immediately stopped. The remaining health of the device and the remaining production tasks can serve as the basis for determining whether to perform pruning. The optimal batch allocation scheme for each stage is iteratively calculated. Starting from the first process, the state value corresponding to all possible batch allocation schemes can be calculated, and the optimal path to reach that state can be recorded. Then, based on the state of the first stage, all possible batch allocation schemes for the second stage can be calculated, and the state value and optimal path can be updated. This process is repeated until the last process stage. Finally, in the state of the last stage, the complete batch allocation scheme corresponding to the state with the best value is selected as the globally optimal scheme. During the iteration process, the optimal decision at each stage can be recorded for later backtracking or analysis.

[0094] Preferably, step A204 may include:

[0095] D1. Based on each candidate production sequence and its batch allocation scheme, a normalized decision matrix is ​​constructed using the range standardization method according to the total production time, number of equipment switching times, overall equipment health maintenance time, and variance of the rate of decline of health between equipment. Among them, the total production time and number of equipment switching times are cost indicators, the overall equipment health maintenance time is a benefit indicator, and the variance of the rate of decline of health between equipment is a cost indicator.

[0096] D2. Based on the normalized decision matrix, determine the ideal solution and the negative ideal solution; the ideal solution is the maximum value of each benefit-type indicator and the minimum value of each cost-type indicator, and the negative ideal solution is the minimum value of each benefit-type indicator and the maximum value of each cost-type indicator.

[0097] D3. For each candidate production sequence and its batch allocation scheme, calculate the weighted Euclidean distance between its corresponding index and the ideal solution and the negative ideal solution;

[0098] D4. Based on the weighted Euclidean distance, calculate the relative proximity of each candidate production sequence and its batch allocation scheme as a comprehensive evaluation score, and select the scheme with the highest comprehensive evaluation score as the optimal processing scheduling scheme.

[0099] The range standardization method is a data standardization technique that linearly scales the original data to a fixed interval (usually [0,1] or [-1,1]) to eliminate the influence of different indicator units and numerical ranges, making different indicators comparable. Its specific calculation method is existing technology and will not be detailed here. The normalized decision matrix is ​​a standardized decision matrix where the elements are the standardized values ​​of each candidate solution for each indicator. Cost-related indicators are those where a smaller value is better, such as total production time and equipment changeover times. Benefit-related indicators are those where a larger value is better, such as the overall equipment health maintenance time.

[0100] An ideal solution is a virtual optimal solution where all benefit indicators reach their maximum values ​​and all cost indicators reach their minimum values ​​among all candidate solutions. A negative ideal solution is a virtual worst solution where all benefit indicators reach their minimum values ​​and all cost indicators reach their maximum values ​​among all candidate solutions.

[0101] Weighted Euclidean distance refers to the Euclidean distance calculated after considering the weights of each indicator. It is used to measure the distance between a candidate solution and the ideal solution or the negative ideal solution. It can be implemented by using the standard Euclidean distance formula and multiplying it by the corresponding weights.

[0102] Among them, relative proximity refers to a comprehensive index that measures how close a candidate solution is to the ideal solution and how far it is from the negative ideal solution. It is usually obtained by calculating the ratio of the distance of the candidate solution to the negative ideal solution to the sum of the distances to the ideal solution and the negative ideal solution.

[0103] This application provides a comprehensive evaluation method based on TOPSIS to select the optimal processing scheduling scheme from multiple candidate production sequences and their batch allocation schemes. This method comprehensively considers multiple optimization objectives, such as total production time, number of equipment changeovers, overall equipment health maintenance time, and variance of the rate of decline in health among equipment, thus ensuring both production efficiency and equipment health maintenance. First, based on each candidate production sequence and its batch allocation scheme, and according to the total production time, number of equipment changeovers, overall equipment health maintenance time, and variance of the rate of decline in health among equipment, a normalized decision matrix is ​​constructed using the range standardization method. Range standardization eliminates the influence of different indicator dimensions and numerical ranges, making different indicators comparable and laying the foundation for subsequent comprehensive evaluation. Total production time and number of equipment changeovers are defined as cost-type indicators, overall equipment health maintenance time as benefit-type indicators, and variance of the rate of decline in health among equipment as cost-type indicators, accurately reflecting the optimization direction of each indicator. Then, based on the normalized decision matrix, the ideal solution and the negative ideal solution are determined. The ideal solution represents the optimal value of all indicators, and the negative ideal solution represents the worst value of all indicators. By defining the ideal solution and the negative ideal solution, a benchmark can be provided for subsequent calculations of the closeness of candidate solutions to the ideal solution. Next, for each candidate production sequence and its batch allocation scheme, the weighted Euclidean distance between its corresponding indicators and the ideal and negative ideal solutions is calculated. The weighted Euclidean distance comprehensively considers the differences between each indicator and the ideal and negative ideal solutions and adjusts according to the weight of the indicator. By calculating the weighted Euclidean distance, the merits of each candidate solution can be quantified. Finally, based on the weighted Euclidean distance, the relative closeness of each candidate production sequence and its batch allocation scheme is calculated as a comprehensive evaluation score, and the scheme with the highest comprehensive evaluation score is selected as the optimal processing scheduling scheme. The relative closeness comprehensively considers the closeness of the candidate solution to the ideal solution and the distance from the negative ideal solution, and can comprehensively reflect the overall performance of the candidate solution. Selecting the scheme with the highest relative closeness as the optimal scheme can achieve a balance among multiple optimization objectives, thereby obtaining the best processing scheduling scheme.

[0104] This evaluation method, combined with the candidate production sequences and batch allocation schemes generated in the aforementioned steps, forms a complete process from scheme generation to scheme selection. This ensures that the final selected processing scheduling scheme not only performs well on a single objective but also achieves an effective balance among multiple conflicting objectives, thereby solving the problem of how to select the optimal scheduling scheme while taking into account both production efficiency and equipment health.

[0105] In one embodiment, multiple candidate production sequences generated by previous steps can be obtained first. Each sequence corresponds to an optimal batch allocation scheme determined by a dynamic programming algorithm, along with the calculated results of total production time, number of equipment changesovers, overall equipment health maintenance time, and variance of the rate of decline in health among equipment. This data can be organized into a decision matrix, where rows represent different candidate schemes and columns represent different evaluation indicators. Next, the data in the decision matrix is ​​subjected to range standardization, scaling all indicator values ​​to the [0,1] interval to obtain a normalized decision matrix. For cost-related indicators (total production time, number of equipment changesovers, variance of the rate of decline in health among equipment), the standardization formula can be (max-x) / (max-min), where x represents the original data value to be standardized, min represents the minimum value in the original dataset, and max represents the maximum value in the original dataset. For benefit-related indicators (overall equipment health maintenance time), the standardization formula can be (x-min) / (max-min). Then, based on the normalized decision matrix, the ideal solution vector (the maximum value in each column) and the negative ideal solution vector (the minimum value in each column) are determined. Subsequently, for each candidate solution, the weighted Euclidean distance between its row vector in the normalized decision matrix and the ideal solution vector and the negative ideal solution vector is calculated. The weights of each indicator can be preset or determined through other methods; for example, higher weights can be assigned to equipment health-related indicators based on actual production needs. Finally, based on the calculated distances to the ideal solution and the negative ideal solution, the relative proximity of each candidate solution is calculated. A higher relative proximity value indicates that the solution is closer to the ideal solution and further away from the negative ideal solution, resulting in better overall performance. The solution with the highest relative proximity is selected as the final processing scheduling solution.

[0106] Preferably, step D4 may include:

[0107] D401. For each candidate production sequence and its batch allocation scheme, calculate the predicted decrease in the health assessment value of each key piece of equipment after completing the production task under the current scheme, based on the equipment health impact coefficient matrix.

[0108] D402. Based on the predicted decrease, determine whether the health of any key equipment decreases below its corresponding health threshold. If so, calculate the degree to which it falls below the health threshold, and penalize the relative closeness of the candidate production sequence and its batch allocation scheme according to the degree to which it falls below the health threshold, to obtain the comprehensive evaluation score after penalty; where the greater the degree to which it falls below the health threshold, the greater the penalty.

[0109] D403. Select the scheme with the highest comprehensive evaluation score after penalty as the optimal processing scheduling scheme.

[0110] The predicted decrease refers to the amount by which the health assessment value of each key piece of equipment will decrease after the completion of the plan, relative to the amount before the start of the task, based on the equipment health impact coefficient matrix and the production task volume under the current plan. It can be obtained by cumulatively calculating the sum of the health impact of each product family on each piece of equipment.

[0111] The health threshold refers to the minimum acceptable health assessment value preset for each key piece of equipment. A value below this threshold may indicate a significant decline in equipment performance or a potential risk of failure. It can be set based on the equipment manufacturer's recommendations, historical maintenance records, or production quality requirements. The degree to which the health threshold is below the predicted health assessment value refers to the difference between the predicted equipment health assessment value and its corresponding health threshold, or a relative measure calculated based on this difference. This can be determined by directly calculating the difference or by normalizing the difference. The penalty intensity refers to the deduction or adjustment factor applied to the overall evaluation score of the solution based on the degree to which the health threshold is below the threshold. This can be determined using a preset penalty function or a lookup table, which reflects the relationship between the degree to which the threshold is below the threshold and the penalty intensity. Relative proximity, in the TOPSIS evaluation method, measures the closeness between a candidate solution and the ideal optimal solution. It can be obtained by calculating the ratio of the distance between the solution and the ideal solution and the negative ideal solution. The comprehensive evaluation score after penalty refers to the final evaluation score after adjusting the relative closeness of the scheme based on whether the predicted equipment health is below the threshold and to what extent. It can be obtained by subtracting the penalty intensity from the relative closeness.

[0112] In this application, the scheme comprehensively evaluates candidate production sequences and their batch allocation schemes. First, for each candidate scheme, using the equipment health impact coefficient matrix, the predicted decrease in the health assessment value of each key piece of equipment after completing the production task under that scheme is calculated. This predicted decrease reflects the potential cumulative impact of the scheme on equipment health. Next, based on the calculated predicted decrease, it is determined whether the health assessment value of any key piece of equipment will drop below its preset health threshold. If such a situation is predicted, the specific degree to which the equipment health falls below the threshold is further calculated. Based on this degree of below-threshold, a penalized adjustment is made to the relative closeness of the candidate scheme obtained in the previous TOPSIS evaluation, resulting in a penalized comprehensive evaluation score. The greater the degree below the health threshold, the greater the penalty, leading to a lower penalized score. Finally, among all candidate schemes, the scheme with the highest penalized comprehensive evaluation score is selected as the final optimal processing scheduling scheme. In this way, the proposed solution incorporates equipment health thresholds as a hard or heavily penalized constraint into the solution evaluation and selection process. This effectively avoids selecting solutions that, while performing well in other indicators, may lead to critical equipment prematurely entering a dangerous health zone. Compared to solutions that rely solely on TOPSIS comprehensive scores for selection, this approach more reliably ensures equipment health, reduces the risk of unplanned downtime, and thus improves the robustness and reliability of production scheduling.

[0113] By introducing a device health threshold penalty mechanism based on the TOPSIS comprehensive evaluation, the proposed solution can effectively identify and avoid scheduling schemes that may cause the health of critical equipment to drop to a dangerous level. This ensures that the final selected optimal processing scheduling scheme not only performs well in terms of total production time and number of equipment changeovers, but also ensures that the device health is maintained within a safe range. This significantly reduces the risk of unplanned downtime and production interruptions caused by device health issues, and improves the stability and continuity of the production process.

[0114] refer to Figure 2 This application provides a filter processing control system for optimizing the processing scheduling scheme of filters in mixed-line processing scenarios where multiple types of filters share production equipment. The system includes:

[0115] Information acquisition module 1 is used to acquire the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health influence matrix; the production order list includes product model, quantity, and delivery date; the product-equipment health influence matrix records the unit influence coefficient of each product model on the health of each key piece of equipment (for details, refer to step A1 above);

[0116] The scheduling optimization module 2 is used to determine the optimal processing scheduling scheme based on the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health impact matrix, combined with optimization objectives and constraints. The processing scheduling scheme includes the production sequence and the batch allocation scheme for each process. The optimization objectives include minimizing the total production time, minimizing the number of equipment switching, maximizing the overall health maintenance time of the equipment, and balancing the rate of decline of the health of each piece of equipment. The constraints include that all orders are completed before the delivery date, the production sequence conforms to the process flow, and the health assessment value of all key pieces of equipment remains above its preset health threshold after all planned production tasks are completed (refer to step A2 above for details).

[0117] The execution and monitoring module 3 is used to send the optimal processing scheduling plan to the production site for processing and production, and to continuously monitor the equipment status of each key equipment, and compare the real-time health changes of each key equipment with the planned forecast (for details, refer to step A3 above).

[0118] The scheduling adjustment module 4 is used to adjust the processing scheduling scheme for the remaining production tasks when the real-time health rate of any critical equipment decreases significantly faster than the prediction result or approaches the warning threshold (for details, refer to step A4 above).

[0119] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A filter processing control method, used to optimize the filter processing scheduling scheme in a mixed-line processing scenario where multiple types of filters share production equipment, characterized in that, The steps of this method include: A1. Obtain the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health impact matrix; the production order list includes product model, quantity, and delivery date; the product-equipment health impact matrix records the unit impact coefficient of each product model on the health of each key piece of equipment; A2. Based on the production order list, the current health assessment values ​​of each key piece of equipment, and the product-equipment health impact matrix, combined with optimization objectives and constraints, determine the optimal processing scheduling scheme; the processing scheduling scheme includes the production sequence and the batch allocation scheme for each process; the optimization objectives include minimizing the total production time, minimizing the number of equipment switchovers, maximizing the overall health maintenance time of equipment, and balancing the rate of decline in the health of each piece of equipment; the constraints include all orders being completed before the delivery date, the production sequence conforming to the process flow, and the health assessment values ​​of all key pieces of equipment remaining above their preset health thresholds after all planned production tasks are completed; A3. The optimal processing scheduling scheme is sent to the production site for processing and production, and the equipment status of each key equipment is continuously monitored. The real-time health status changes of each key equipment are compared with the planned forecast. A4. If the real-time health rate of any critical equipment decreases significantly faster than the predicted result, or approaches the warning threshold, the processing scheduling plan for the remaining production tasks will be adjusted.

2. The filter processing control method according to claim 1, characterized in that, Step A1, which involves obtaining the current health assessment values ​​of each key device, includes: Collect equipment status parameters for each key piece of equipment; the key equipment includes a stamping machine, a folding machine, and a bonding and curing oven; the equipment status parameters include the wear amount of the stamping die edge of the stamping machine, the tool contour accuracy deviation of the folding machine, and the internal temperature uniformity deviation of the bonding and curing oven. Based on the equipment status parameters of each key device, and the mapping relationship between the equipment status parameters and the equipment health of each key device, the current health assessment value of each key device is determined.

3. The filter processing control method according to claim 2, characterized in that, Step A1, the step of obtaining the current health assessment value of each key device, also includes: Collect environmental parameters of the environment where each key piece of equipment is located; these environmental parameters include workshop temperature, humidity, and air cleanliness. Based on the collected environmental parameters, calculate the health impact correction coefficient for each key piece of equipment; Based on the aforementioned health impact correction coefficient, the current health assessment value of each key device is corrected.

4. The filter processing control method according to claim 1, characterized in that, Step A2 includes: A201. For the production order list, based on the product model and the product-equipment health influence matrix, the products are divided into multiple product families using the K-means algorithm, so that the products in the same product family have similar health influence on the key equipment, and a product family list is obtained. A202. Based on the product family list and the current health assessment value of each key equipment, a genetic algorithm is used to generate multiple initial production sequences. Combined with the product-equipment health influence matrix, the equivalent health decline rate of each key equipment under each initial production sequence is calculated to screen candidate production sequences from multiple initial production sequences. A203. For each candidate production sequence, construct a batch allocation model with the objective of minimizing the variance of the health decline rate among equipment. Use dynamic programming algorithm to determine the optimal batch allocation scheme for each process. Under the constraints of order delivery period and process flow, balance the health decline rate of each equipment and calculate the total production time, number of equipment changeovers and overall equipment health maintenance time under the scheme. A204. Based on the total production time, number of equipment changeovers, overall equipment health maintenance time, and variance of the rate of decline in health between equipment, the TOPSIS method is used to comprehensively evaluate each candidate production sequence and its batch allocation scheme, and the scheme with the highest comprehensive evaluation score is selected as the optimal processing scheduling scheme.

5. The filter processing control method according to claim 4, characterized in that, Step A202 includes: B1. Based on the current health assessment values ​​of each key device, set the initial population size, selection probability, crossover probability, and mutation probability for the genetic algorithm; B2. Based on the product family list, and according to the initial population size, multiple initial production sequences are generated using a segmented coding method; where each product family corresponds to a segment in the production sequence, and the product order within the segment is randomly generated; B3. For each initial production sequence, based on the product-equipment health impact matrix, calculate the predicted decay value of the health of each key piece of equipment and the corresponding production time after completing the processing of each reference batch of products under the initial production sequence; B4. Based on the predicted decay value and the corresponding production time, calculate the equivalent health rate of each key equipment, and calculate the fitness function value of the genetic algorithm with the goal of minimizing the maximum equivalent health rate of decay. B5. Based on the selection probability, crossover probability, and mutation probability, perform selection, crossover, and mutation operations of the genetic algorithm to generate a new initial production sequence, and repeat steps B3 and B4 until the preset number of iterations is met or the convergence condition is reached. B6. Based on the fitness function value, select the multiple production sequences with the highest fitness from the final set of initial production sequences as candidate production sequences.

6. The filter processing control method according to claim 4, characterized in that, Step A203 includes: C1. For each candidate production sequence, based on the product model of each product family in the product family list and combined with the product-equipment health influence matrix, calculate the unit health influence coefficient of each product family on each key equipment, and construct the equipment health influence coefficient matrix. C2. Based on the equipment health impact coefficient matrix and the current health assessment value of each key equipment, a batch allocation model is constructed with the objective of minimizing the variance of the health decline rate among equipment. The constraints of the batch allocation model include: ensuring that all orders are completed before the delivery period, ensuring that the production sequence conforms to the process flow, and ensuring that the health assessment value of all key equipment remains above its preset health threshold after all planned production tasks are completed. C3. Using a dynamic programming algorithm, the batch allocation problem is decomposed into multiple sub-problems based on the equipment health influence coefficient matrix. Each stage corresponds to the batch allocation decision of a process. The optimal batch allocation scheme for each stage is calculated iteratively until the globally optimal batch allocation scheme that satisfies the constraints and minimizes the variance of the rate of decline in health among equipment is obtained. C4. Based on the optimal batch allocation scheme determined by the dynamic programming algorithm, calculate the total production time, number of equipment changeovers, and overall equipment health maintenance time under this optimal batch allocation scheme.

7. The filter processing control method according to claim 6, characterized in that, Step C3 includes: C301. For each stage of batch allocation decision, calculate the rate of decline of health of each key device under different batch allocation schemes in the current stage based on the equipment health impact coefficient matrix; C302. Based on the calculated health rate decline, an optimization function is constructed with the objective of minimizing the variance of the health rate decline among devices, and a penalty term is introduced to penalize batch allocation schemes that do not meet the constraints of order delivery period, process flow, or device health; wherein, the magnitude of the penalty term is proportional to the degree of constraint violation; C303. An improved dynamic programming algorithm is adopted. By introducing a heuristic search strategy, the algorithm prioritizes searching for batch allocation schemes that meet the constraints and have a balanced rate of equipment health decline. Combined with a pruning strategy, the search space is dynamically adjusted according to the remaining equipment health and the remaining production tasks. C304. Iteratively calculate the optimal batch allocation scheme for each stage until a globally optimal batch allocation scheme that satisfies the constraints and minimizes the variance of the rate of decline in health among devices is obtained, and record the optimal batch allocation scheme for each stage.

8. The filter processing control method according to claim 6, characterized in that, Step A204 includes: D1. Based on each candidate production sequence and its batch allocation scheme, a normalized decision matrix is ​​constructed using the range standardization method according to the total production time, number of equipment switching times, overall equipment health maintenance time, and variance of the rate of decline of health between equipment. Among them, the total production time and number of equipment switching times are cost indicators, the overall equipment health maintenance time is a benefit indicator, and the variance of the rate of decline of health between equipment is a cost indicator. D2. Based on the normalized decision matrix, determine the ideal solution and the negative ideal solution; the ideal solution is the maximum value of each benefit-type indicator and the minimum value of each cost-type indicator, and the negative ideal solution is the minimum value of each benefit-type indicator and the maximum value of each cost-type indicator. D3. For each candidate production sequence and its batch allocation scheme, calculate the weighted Euclidean distance between its corresponding index and the ideal solution and the negative ideal solution; D4. Based on the weighted Euclidean distance, calculate the relative proximity of each candidate production sequence and its batch allocation scheme as a comprehensive evaluation score, and select the scheme with the highest comprehensive evaluation score as the optimal processing scheduling scheme.

9. A filter processing control method according to claim 8, characterized in that, Step D4 includes: D401. For each candidate production sequence and its batch allocation scheme, calculate the predicted decrease in the health assessment value of each key piece of equipment after completing the production task under the current scheme, based on the equipment health impact coefficient matrix. D402. Based on the predicted decrease, determine whether the health of any key equipment decreases below its corresponding health threshold. If so, calculate the degree to which it falls below the health threshold, and penalize the relative closeness of the candidate production sequence and its batch allocation scheme according to the degree to which it falls below the health threshold, to obtain the comprehensive evaluation score after penalty; where the greater the degree to which it falls below the health threshold, the greater the penalty. D403. Select the scheme with the highest comprehensive evaluation score after penalty as the optimal processing scheduling scheme.

10. A filter processing control system, used to optimize the filter processing scheduling scheme in a mixed-line processing scenario where multiple types of filters share production equipment, characterized in that, The system includes: The information acquisition module is used to acquire a production order list, the current health assessment value of each key piece of equipment, and a product-equipment health impact matrix. The production order list includes product models, quantities, and delivery dates. The product-equipment health impact matrix records the unit impact coefficient of each product model on the health of each key piece of equipment. The scheduling optimization module is used to determine the optimal processing scheduling scheme based on the production order list, the current health assessment value of each key piece of equipment, and the product-equipment health impact matrix, combined with optimization objectives and constraints. The processing scheduling scheme includes the production sequence and the batch allocation scheme for each process. The optimization objectives include minimizing the total production time, minimizing the number of equipment switchovers, maximizing the overall health maintenance time of the equipment, and balancing the rate of decline in the health of each piece of equipment. The constraints include all orders being completed before the delivery date, the production sequence conforming to the process flow, and the health assessment value of all key equipment remaining above its preset health threshold after all planned production tasks are completed. The execution and monitoring module is used to send the optimal processing scheduling plan to the production site for processing and production, and to continuously monitor the equipment status of each key piece of equipment, and compare the real-time health changes of each key piece of equipment with the planned forecast. The scheduling adjustment module is used to adjust the processing scheduling scheme for the remaining production tasks when the real-time health rate of any critical equipment decreases significantly faster than the prediction result or approaches the warning threshold.

Images