Four-line production scheduling optimization system and method based on thick line

Abstract

The invention relates to the technical field of thick line four-line scheduling optimization, and discloses a thick line four-line production scheduling optimization system and method. The method comprises the steps of generating a scheduling configuration data set by collecting initial scheduling parameters and real-time task trigger signals; performing dynamic priority evaluation on the data set, and determining whether a multilayer optimization analysis mechanism is started or not; after starting, calculating a structure efficiency index, synchronously capturing a material flow dynamic state, integrating the two to intelligently partition, and identifying a key evaluation unit; the operation oscillation index and the output rate are monitored, the influence coefficient on scheduling disturbance is analyzed, and the performance is quantified in combination with the task execution duration; extracting an energy consumption utilization rate and a maximum throughput capacity derivation unit weight factor; and fusing the performance quantification result and the weight factor, and outputting a comprehensive optimization index of the current scheduling scene. The method is adaptive to complex scheduling requirements of thick four-line production, and the accuracy and comprehensiveness of scheduling optimization are improved.

Description

Technical Field

[0001] This invention relates to the field of coarse-line four-line scheduling optimization technology, specifically to a coarse-line four-line production scheduling optimization system and method. Background Technology

[0002] As a crucial link in industrial manufacturing, coarse-line production, particularly the four-line parallel production mode, is widely used in various coarse-line processing scenarios due to its ability to increase output scale and production flexibility. Such production systems involve complex processes such as multi-line coordination, dynamic material flow, and parallel execution of multiple tasks. The scheduling process requires comprehensive consideration of multiple factors, including initial parameter configuration, real-time task changes, equipment operating status, and material supply efficiency, making scheduling optimization significantly more difficult than for a single production line.

[0003] Currently, coarse-grained four-line production scheduling mostly adopts traditional fixed rules or static optimization models, which are difficult to adapt to dynamic changes in the production process. In the data acquisition stage, existing methods often only focus on initial scheduling parameters, and the capture of real-time task trigger signals is not timely and comprehensive enough. This results in the scheduling configuration dataset lacking dynamic adaptability and failing to provide basic data that fits the actual production status for subsequent optimization analysis.

[0004] In terms of priority assessment, traditional methods often employ a single evaluation dimension, failing to dynamically adjust based on factors such as the urgency of production tasks and system load. This frequently leads to over-optimization or under-optimization, wasting computational resources and failing to guarantee scheduling effectiveness. Furthermore, the lack of scientific basis for incorporating multi-layered optimization analysis prevents the optimization process from specifically focusing on core issues.

[0005] In the production line analysis and zoning process, existing technologies often sever the connection between structural performance and material flow dynamics, relying solely on a single indicator for zoning. This leads to inaccurate identification of key evaluation units and an inability to accurately locate the core links that play a decisive role in scheduling effectiveness. Furthermore, the monitoring of the operational status of key units is not comprehensive enough, failing to fully consider the synergistic impact of operational oscillation indicators and output rates. The analysis of the impact coefficients of scheduling disturbances lacks depth, making it difficult for performance quantification results to reflect the actual operational level of the units.

[0006] In terms of weight factor derivation and comprehensive optimization, existing methods often neglect the balance between energy consumption utilization and maximum throughput capacity, and the weight allocation lacks rationality. This results in the final optimization results failing to fully cover the multi-dimensional needs of production efficiency, resource utilization, and other aspects, and thus cannot provide a scientific and effective decision-making reference for coarse-line four-line production scheduling. The existence of these problems leads to limited scheduling efficiency and inadequate resource allocation in coarse-line four-line production systems, making it difficult to fully leverage the advantages of four-line parallel production and hindering the improvement of the overall operational efficiency of the production system. Summary of the Invention

[0007] The purpose of this invention is to provide a coarse-line four-line production scheduling optimization system and method to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, this invention provides a coarse-line four-line production scheduling optimization method, the method comprising: Collect the initial scheduling parameters and real-time task trigger signals of the coarse-line four-line production system to generate a scheduling configuration dataset; Dynamically prioritize the scheduling configuration dataset to determine whether to enable a multi-level optimization analysis mechanism; When the multi-layer optimization analysis mechanism is enabled, the structural efficiency index of the production line is calculated, and the dynamic status of material flow is captured simultaneously. By integrating structural performance indicators and material flow dynamics, the production line is intelligently partitioned to identify key evaluation units; Continuously monitor the operational oscillation indicators and output rates of key evaluation units, and analyze the impact coefficients of each unit on scheduling disturbances; The performance of key evaluation units is quantified based on the impact coefficient and task execution time. Extract the energy consumption utilization rate and maximum throughput capacity of key evaluation units, and derive the unit weighting factors; By integrating the performance quantification results and unit weight factors, a comprehensive optimization index for the coarse-line four-line production system under the current scheduling scenario is output.

[0009] Preferably, dynamic priority evaluation is performed on the scheduling configuration dataset to determine whether to enable a multi-level optimization analysis mechanism, specifically including: Extract device operating limits and scheduling frequency features from the scheduling configuration dataset; By comparing the upper limit of equipment operation with the preset capacity threshold, it can be determined whether the production line load exceeds the limit; Analyze whether the scheduling frequency characteristics exceed the normal fluctuation range; If the upper limit of equipment operation is exceeded or the scheduling frequency characteristics are abnormal, the multi-level optimization analysis mechanism is skipped and a low optimization level is directly output; otherwise, the multi-level optimization analysis mechanism is activated.

[0010] Preferably, the structural efficiency index of the production line is calculated, and the dynamic status of material flow is captured simultaneously, specifically including: The upper limit of equipment operation and scheduling frequency characteristics are normalized, and a weighted fusion algorithm is applied to generate structural performance indicators. Deploy flow sensing devices along the production line transmission path to measure material flow rate and flow fluctuations in real time; The fluid continuity model is used to calculate the material throughput per unit time, which is used as the dynamic state of the material flow.

[0011] Preferably, by integrating structural performance indicators and material flow dynamics, the production line is intelligently partitioned to identify key evaluation units, specifically including: Input structural performance indicators and material flow dynamics into the cluster analysis model to generate partition reference values; The production line is divided into logical segments based on the partition reference values, with each segment serving as an evaluation unit. For each assessment unit, calculate its material demand intensity and failure history frequency; Material demand intensity is derived by analyzing unit equipment load and material flow models. The historical frequency of failures is obtained by statistically analyzing the ratio of the number of failures to the runtime of the unit within a historical period. Key evaluation units were selected based on a comprehensive score of material demand intensity and failure history frequency.

[0012] Preferably, the operational oscillation indicators and output rates of key evaluation units are continuously monitored, and the impact coefficients of each unit on scheduling disturbances are analyzed, specifically including: High-precision sensor arrays are used to collect production sequences and operating parameters of key evaluation units in real time; A sliding window variance calculation is performed on the production series to obtain the operating oscillation index; The output rate is determined by the slope of a linear regression of the change in output per unit time. The model will run the oscillation index and output rate input impact prediction model and output the impact coefficient of each key assessment unit on scheduling disturbances.

[0013] Preferably, the performance of key evaluation units is quantified based on the impact coefficient and task execution time, specifically including: Record the time interval between the start and end of the key evaluation units as the task execution duration; Standardize and convert the impact coefficient and task execution time. A nonlinear mapping function is applied to combine the standardized values ​​into a performance quantization score.

[0014] Preferably, the energy consumption utilization rate and maximum throughput capacity of key evaluation units are extracted, and unit weighting factors are derived, specifically including: Apply standard test loads to key evaluation units and measure their energy consumption to output ratio to obtain energy utilization rate; By gradually increasing the load until the system reaches its stable limit, the maximum output value is recorded as the maximum throughput capacity. The entropy method is used to calculate the energy consumption efficiency and maximum throughput capacity by weighting, and a unit weight factor is generated.

[0015] Preferably, by integrating performance quantification results and unit weighting factors, a comprehensive optimization index for the coarse-line four-line production system under the current scheduling scenario is output, specifically including: The performance quantification score of each key evaluation unit is multiplied by the corresponding unit weight factor to obtain the weighted score; The weighted scores of all key assessment units are summed. The summation is converted into a comprehensive optimization index between zero and one hundred by applying normalization.

[0016] Preferably, the initial scheduling parameters and real-time task trigger signals of the coarse-line four-line production system are collected to generate a scheduling configuration dataset, specifically including: Read equipment configuration parameters and historical scheduling records from the production management database; The IoT gateway captures the timestamps and task types of production task trigger events in real time. The device configuration parameters, historical scheduling records, and real-time task data are integrated into a structured scheduling configuration dataset.

[0017] Preferably, the present invention also includes a coarse-wire four-line production scheduling optimization system, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of the coarse-wire four-line production scheduling optimization method described above.

[0018] Compared with the prior art, the beneficial effects of the present invention are: By synchronously collecting initial scheduling parameters and real-time task trigger signals, the generated scheduling configuration dataset contains both basic system operating information and dynamic changing elements in the production process. This makes the basic data for optimization analysis more comprehensive and timely, accurately mapping the real-time operating status of the production system. The introduction of a dynamic priority evaluation mechanism makes the activation of multi-layer optimization analysis mechanisms more targeted. Optimization strategies can be flexibly adjusted according to the actual situation of the scheduling configuration dataset, avoiding resource waste or inadequate optimization in the traditional static evaluation mode, and making the optimization process more aligned with actual production needs.

[0019] In the multi-layer optimization analysis process, the structural efficiency index is calculated simultaneously with the dynamic state of material flow, achieving an organic combination of the hardware structural characteristics of the production system and the dynamic characteristics of material flow. This breaks the limitation of the separation between the two in traditional analysis and provides more comprehensive decision support for subsequent intelligent partitioning. Intelligent partitioning based on the integrated results of the two can accurately identify key evaluation units, focusing the optimization focus on the links that play a core role in scheduling effectiveness, avoiding the dispersed investment of optimization resources, and improving the targeting of scheduling optimization.

[0020] By continuously monitoring the operational oscillation indicators and output rates of key evaluation units, we can comprehensively capture the dynamic changes during unit operation, accurately analyze the impact coefficients of each unit on scheduling disturbances, and ensure that the performance quantification process fully integrates the correlation between the actual operating performance of the units and scheduling disturbances, making the quantification results more reflective of the true operating level of the units. The derivation of the unit weight factors comprehensively considers energy consumption utilization and maximum throughput capacity, focusing on both production efficiency and rational resource utilization, making the weight allocation more scientific and balanced, and avoiding optimization biases caused by single-dimensional considerations.

[0021] The integration of performance quantification results with unit weighting factors enables the final comprehensive optimization index to fully cover the multi-dimensional characteristics of the production system, including structural efficiency, material flow, unit operation, and resource utilization, thus comprehensively presenting the system's operating status and optimization direction under the current scheduling scenario. The entire methodology closely revolves around the core characteristics of coarse-line four-line production, such as multi-line collaboration and dynamic changes. Each step is designed to meet actual production needs, forming a closed-loop system from data collection, analysis, and partitioning to quantification and optimization. This effectively adapts to the complex scheduling requirements of coarse-line four-line production, making scheduling optimization more accurate, comprehensive, and practical. It helps the production system fully leverage the advantages of four-line parallelism, achieving a synergistic improvement in overall operational efficiency. Attached Figure Description

[0022] Figure 1 This is a schematic diagram illustrating the working principle of the coarse-line four-line production scheduling optimization method described in this invention. Figure 2 A flowchart for calculating structural performance and material flow state; Figure 3 A flowchart for analyzing oscillations and influence coefficients; Figure 4 A comprehensive optimization analysis diagram of energy efficiency and production capacity of key units in the production system; Figure 5 This is a chart analyzing the execution of production task scheduling. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1This invention provides a coarse-coarse four-line production scheduling optimization method. The method includes: collecting initial scheduling parameters and real-time task triggering signals of the coarse-coarse four-line production system to generate a scheduling configuration dataset; dynamically prioritizing the scheduling configuration dataset to determine whether to enable a multi-layer optimization analysis mechanism; when the multi-layer optimization analysis mechanism is enabled, calculating the structural efficiency index of the production line and simultaneously capturing the dynamic state of material flow; integrating the structural efficiency index and the dynamic state of material flow to intelligently partition the production line and identify key evaluation units; continuously monitoring the operational oscillation index and output rate of the key evaluation units and analyzing the influence coefficient of each unit on scheduling disturbances; quantifying the performance of the key evaluation units based on the influence coefficient and task execution time; extracting the energy consumption utilization rate and maximum throughput capacity of the key evaluation units to derive unit weight factors; and fusing the performance quantification results and unit weight factors to output a comprehensive optimization index of the coarse-coarse four-line production system under the current scheduling scenario. This method adaptively optimizes the scheduling strategy through a multi-layer analysis mechanism, improving production efficiency and resource utilization.

[0025] Example 1: In specific implementation, the upper limit of equipment operation and scheduling frequency characteristics are extracted from the scheduling configuration dataset. The upper limit of equipment operation refers to the maximum number of tasks or materials that the production equipment can process per unit time. This value is usually derived from the technical specifications of the equipment manufacturer or calibrated through historical operating data. The scheduling frequency characteristics reflect the distribution and regularity of the time intervals in which production tasks are triggered. They are calculated by analyzing the timestamp sequence in the scheduling log. In some embodiments, the extraction of the upper limit of equipment operation involves querying the equipment performance table in the production database, while the scheduling frequency characteristics are quantified by statistically analyzing the variance and mean of task triggering events in the most recent period. It can be understood that this extraction process relies on the real-time connection between the data interface and the production management system to ensure the accuracy and timeliness of the information. In practical implementation, the overload of the production line is determined by comparing the upper limit of equipment operation with a preset capacity threshold. The preset capacity threshold is a fixed or dynamically adjusted value pre-set according to the production line's design capacity and safe operation specifications, such as a percentage of historical peak load. The comparison operation uses a numerical comparison algorithm. If the upper limit of equipment operation is greater than or equal to the preset capacity threshold, it is determined to be an overload state. This state indicates that the production line may face overload risk, affecting system stability. It is understandable that the setting of the preset capacity threshold needs to consider equipment aging factors and capacity elasticity to maintain the rationality of the judgment. In practical implementation, the scheduling frequency characteristics are analyzed to see if they exceed the normal fluctuation range. The normal fluctuation range is a statistical model established through long-term monitoring of scheduling data, such as using the three Sigma rule to calculate the upper and lower limits. If the value of the scheduling frequency characteristics exceeds this range, it is marked as abnormal. The analysis process includes calculating the deviation between the current scheduling frequency and the historical baseline. If the deviation is too large, an abnormality flag is triggered. In some embodiments, the normal fluctuation range is updated periodically to adapt to changes in production rhythm. In practical implementation, if the equipment operating limit is exceeded or the scheduling frequency characteristics are abnormal, the multi-layer optimization analysis mechanism is skipped, and a low optimization level is directly output. The low optimization level is a predefined level identifier indicating that the system is currently unsuitable for complex optimization and basic stability issues should be addressed first. The output method can be sending an alarm to the monitoring system or recording it in the optimization log. Optionally, the specific value of the low optimization level can be subdivided according to the degree of exceeding limits or abnormality to provide a more granular response. In practical implementation, if the equipment operating limit is not exceeded and the scheduling frequency characteristics are normal, the multi-layer optimization analysis mechanism is activated. Activation actions include starting subsequent modules such as calculating structural performance indicators and capturing the dynamic status of material flow to ensure the continuity of the optimization process. Optionally, the activation process may involve resource allocation checks to avoid system resource conflicts. In some embodiments, the entire dynamic priority evaluation process is implemented as an independent software module that runs periodically or is triggered by events, efficiently gatewaying optimization decisions through the above steps and reducing unnecessary computational overhead.

[0026] In practical implementation, the extraction of equipment operating limits and scheduling frequency characteristics involves data preprocessing steps, such as missing value imputation and outlier removal, to ensure data quality. Equipment operating limits may be directly read from equipment sensors or the management platform, while scheduling frequency characteristics require parsing time-series data from task scheduling queues. In some embodiments, this extraction process uses a structured query language to retrieve relevant fields from a relational database and performs preliminary aggregation calculations. It can be understood that the frequency of data extraction is synchronized with the production rhythm to achieve real-time evaluation. In practical implementation, the preset capacity threshold is not static but can be dynamically adjusted according to seasonality or production plans. For example, future load can be predicted using machine learning models, and the threshold can be adaptively updated, enhancing the accuracy of the judgment. When comparing the equipment operating limits with the preset capacity threshold, a numerical comparison function is used, and the result is stored in Boolean form for subsequent logical judgment. In practical implementation, when analyzing whether scheduling frequency characteristics exceed the normal fluctuation range, the calculation of the normal fluctuation range may be based on the moving average method to smooth short-term fluctuations and highlight long-term trends. If the scheduling frequency characteristics show a sudden increase or decrease, it is considered abnormal. Optionally, anomaly detection can also be combined with a rule engine; for example, if multiple consecutive periodic characteristic values ​​exceed the range, the anomaly level is increased. In practical implementation, outputting a low optimization level can be achieved by generating system logs, triggering user interface notifications, or integrating into a decision support system. The specific meaning of a low optimization level is to suggest that the system enter a conservative operating mode to avoid further optimization operations. In some embodiments, a low optimization level may be associated with specific operation instructions, such as reducing production rate or activating backup equipment. In practical implementation, activating the multi-layer optimization analysis mechanism involves initializing relevant computing resources, such as allocating memory and starting analysis threads, to ensure that subsequent steps can be executed seamlessly. It can be understood that the design of the activation conditions aims to balance system burden and optimization benefits, avoiding complex calculations in an unstable state.

[0027] In practical implementation, the decision-making logic for dynamic priority assessment can be implemented using a state machine, where each state corresponds to an assessment step. State transitions are based on comparison and analysis results. For example, when the equipment operating limit is exceeded, the state machine directly jumps to the output low optimization level state; otherwise, it enters the active state. In some embodiments, the assessment process also considers external factors, such as equipment maintenance plans or energy supply conditions, to enrich the decision-making basis. In practical implementation, obtaining the equipment operating limit may involve multi-source data fusion, such as combining real-time sensor readings and equipment archive data to improve the reliability of the limit. The calculation of scheduling frequency characteristics may use time series analysis methods, such as Fast Fourier Transform to identify periodic patterns. Optionally, the feature extraction stage may also include data normalization to eliminate the influence of dimensions. In practical implementation, the setting of preset capacity thresholds relies on domain knowledge, such as determining the safety margin through expert consultation or historical accident analysis. The comparison operation not only checks the numerical magnitude but may also assess trends, such as whether the equipment operating limit is trending upwards, which enhances forward-looking judgment. In practical implementation, the establishment of the normal fluctuation range is based on statistical process control theory. For example, control chart methods are used to set upper and lower limits. When analyzing scheduling frequency characteristics, if the characteristic value falls within the control limits, it is considered normal; otherwise, it is considered abnormal. This method helps to identify potential problems early. In practical implementation, skipping the multi-level optimization analysis mechanism and directly outputting a low optimization level is an optimization measure aimed at reducing computational latency and ensuring system response speed. The output format of the low optimization level can be numerical code or text description for easy subsequent processing. In practical implementation, after activating the multi-level optimization analysis mechanism, the system records the activation timestamp and reason for auditing and optimization tracking. In some embodiments, the activation operation may also trigger resource preloading to accelerate subsequent analysis processes.

[0028] In implementation, the overall process of dynamic priority assessment is designed to be configurable. For example, the sensitivity of preset capacity thresholds or normal fluctuation ranges can be adjusted through configuration files to adapt to different production environments. Assessment results are fed back to the scheduling center in real time for immediate decision support. In implementation, the extraction of equipment operating limits and scheduling frequency characteristics utilizes efficient data capture techniques, such as application programming interface calls or database polling, to ensure low latency. Comparison and analysis steps employ optimized algorithms, such as binary search or hash matching, to improve processing efficiency. In implementation, when a low optimization level is output, the system may simultaneously generate a detailed report explaining the specific values ​​and context of exceeding limits or anomalies, helping operators quickly diagnose problems. In implementation, the condition checks for activating the multi-layer optimization analysis mechanism are atomic operations, avoiding race conditions and ensuring system stability in concurrent environments. This design, therefore, guarantees the robustness and practicality of the production scheduling optimization method.

[0029] Example 2: See Figure 2 In practical implementation, the first step is to normalize the upper limit of equipment operation and scheduling frequency characteristics. The normalization process uses a min-max scaling method to linearly transform the original values ​​of these characteristics to a dimensionless range between zero and one. A weighted fusion algorithm is then applied to generate a structural performance index. This algorithm takes the normalized upper limit of equipment operation and scheduling frequency characteristics as input, assigning a dynamic weight coefficient to each characteristic. These weight coefficients are dynamically adjusted based on the importance of the characteristic within the current production cycle. The algorithm calculates a comprehensive value as the structural performance index using a linear weighted sum formula. This index quantitatively reflects the stability and efficiency potential of the overall production line structure. In some embodiments, the weight coefficients of the weighted fusion algorithm can be determined through historical data regression analysis or pre-set by domain experts based on the production line configuration. It can be understood that the normalization process eliminates the influence of different dimensions, enabling the upper limit of equipment operation and scheduling frequency characteristics to be fused and calculated on the same scale.

[0030] In practical implementation, flow sensing devices are deployed along the production line's transmission path to measure material flow velocity and flow fluctuations in real time. These devices primarily include ultrasonic flow meters and photoelectric flow sensors, which are installed at intervals on key nodes of the material conveying pipes or conveyor belts. Real-time measurement of material flow velocity means that the sensors acquire the linear or volumetric velocity of the material flow at a fixed sampling frequency (e.g., once per second). Flow fluctuations are quantified by calculating the standard deviation or coefficient of variation of the flow velocities at multiple consecutive sampling points. A fluid continuity model is used to calculate the material throughput per unit time as the dynamic state of the material flow. Based on the law of conservation of mass, the fluid continuity model treats the production line as a flow system. By integrating the material flow velocity flowing through the pipe cross-section within a specific time period, the material throughput for that time period is calculated. The material throughput value characterizes the overall flow efficiency and stability of the material. It is understood that the deployment density and location of the flow sensing devices directly affect the accuracy of capturing the dynamic state of the material flow, and usually require optimized design based on the production line layout. In some embodiments, the calculation of the dynamic state of the material flow also considers the physical properties of the material, such as density and viscosity, to modify the fluid continuity model and improve the accuracy of the throughput calculation.

[0031] In practical implementation, the production line is intelligently partitioned by integrating structural performance indicators and material flow dynamics. These indicators and dynamics are input into a clustering analysis model to generate partition reference values. The model typically employs the K-means clustering algorithm, treating the indicators and data as two-dimensional data points. Iterative clustering is performed based on the Euclidean distance between these data points, and the coordinates of the final cluster centers serve as the partition reference values. The production line is then divided into logical segments based on these reference values, with each segment acting as an evaluation unit. This division process is based on the projection of the partition reference values ​​onto physical space, grouping geographically adjacent and similarly characterized equipment into a single logical segment. Optionally, hierarchical clustering or the DBSCAN algorithm can also be used to adapt to different data distribution characteristics. In practice, the material demand intensity (MDI) and historical failure frequency are calculated for each assessment unit. The MDI is derived by analyzing the unit's equipment load and material flow model. Specifically, the average load rate of all equipment within the assessment unit under the current task is calculated, and then the required material input intensity to meet the load is derived by combining this with the material flow model (usually a linear or nonlinear model describing the material input-output relationship). The historical failure frequency is obtained by statistically analyzing the ratio of the number of failures to the runtime of the assessment unit within a historical period. The historical period is typically set to the past six months or one year, and the ratio is the total number of failures divided by the total operating hours, resulting in a frequency index representing the probability of failure occurring per unit of time. Optionally, real-time order information may be incorporated as a correction factor in the calculation of MDI, making the derived results closer to actual production needs.

[0032] In implementation, key evaluation units are selected based on a comprehensive score derived from material demand intensity and historical failure frequency. The comprehensive score is calculated using a weighted summation method, where weights are assigned to both material demand intensity and historical failure frequency, and the weighted values ​​are summed to obtain the comprehensive score for each evaluation unit. Evaluation units with high comprehensive scores are selected as key evaluation units, which are the focus of subsequent performance monitoring and optimization. It is understood that setting the weights for material demand intensity and historical failure frequency requires a trade-off between production efficiency and system reliability, typically determined through multi-objective optimization methods. In implementation, the cluster analysis model requires pre-specifying the number of clusters K, which can be determined using the elbow rule or silhouette coefficient method to ensure the rationality of the partitioning. The construction of the material flow model relies on a deep understanding of the production process and may require refinement through multiple on-site measurements and parameter identification. The statistical analysis of historical failure frequency relies on the maintenance record database of the equipment management system, requiring accurate and complete data recording. Optionally, the results of intelligent partitioning can be visualized on the production line monitoring interface to help operators intuitively understand the partitioning situation. In some embodiments, the screening of the comprehensive score will set a threshold, and only the evaluation units that exceed the threshold will be identified as key evaluation units. The threshold can be dynamically adjusted based on historical data or operational objectives.

[0033] In practice, the entire process, from calculating structural performance indicators to identifying key evaluation units, is executed continuously and automatically. The calculation of structural performance indicators provides a basis for macro-level evaluation, while the capture of dynamic material flow status reflects micro-level operational conditions. The integration of these two aspects allows intelligent partitioning to have both static structural and dynamic performance perspectives. Clustering analysis models transform complex data into actionable partitioning reference values, and the division of logical segments decomposes the massive production line into manageable evaluation units. The calculation of material demand intensity and failure history frequency characterizes the characteristics of the evaluation units from both demand and reliability dimensions. A comprehensive score based on material demand intensity and failure history frequency enables multi-indicator decision-making, ultimately accurately identifying key evaluation units. In essence, this implementation method systematically identifies key links in the production line through a data-driven approach, providing clear targets for subsequent in-depth optimization. Optionally, the results of intelligent partitioning and key evaluation unit identification are stored in a dedicated database and a partitioning report is generated for management review and confirmation.

[0034] Example 3: See Figure 3In practical implementation, a high-precision sensor array is used to collect the production sequence and operating parameters of the key evaluation unit in real time. The high-precision sensor array consists of a laser rangefinder, an infrared thermal imager, and a vibration accelerometer. These sensors are distributed and installed on the input / output ports and core moving parts of the key evaluation unit. The production sequence refers to the number of products or semi-finished products produced by the key evaluation unit at consecutive time points. The operating parameters include a series of physical quantities such as motor current, spindle speed, operating temperature, and vibration amplitude of the key evaluation unit. The sampling frequency of the high-precision sensor array is set at the kilohertz level to ensure that rapid transient changes during the production process can be captured. The collected raw data is transmitted to the central processing unit for caching and preprocessing via a fieldbus network. In some embodiments, the deployment scheme of the high-precision sensor array is customized according to the equipment type and process characteristics of the key evaluation unit. For example, vibration sensors are deployed in particular for rotating equipment, while temperature sensors are deployed in priority for heating equipment. It is understood that the data quality of the high-precision sensor array directly determines the accuracy of subsequent analysis, therefore, regular calibration and maintenance are required.

[0035] In practical implementation, a sliding window variance calculation is performed on the output sequence to obtain the operational oscillation index. The sliding window variance calculation calculates the variance of the output sequence data within a fixed-length time window. The window length is typically set as an integer multiple of a complete production cycle; for example, if the basic production cycle time is 10 minutes, the window length can be set to 30 minutes. The magnitude of the variance reflects the degree of fluctuation in output output within that time period; a larger variance indicates greater operational instability. The operational oscillation index is the calculated variance value sequence, which quantitatively describes the dynamic changes in the output stability of key assessment units. The operational oscillation index and output rate are input into the impact prediction model, which outputs the impact coefficient of each key assessment unit on scheduling disturbances. The impact prediction model is an artificial neural network model trained on historical data. Its input layer nodes correspond to the feature vectors of the operational oscillation index and output rate, and its output layer nodes correspond to the impact coefficient value. The impact coefficient is a value between 0 and 1; a larger value indicates that the key assessment unit is more sensitive to scheduling disturbances and has a greater impact on the overall production system. It is understandable that the choice of window length in sliding window variance calculation needs to strike a balance between sensitivity and stability. A window that is too short will cause the index to be overly sensitive, while a window that is too long will mask short-term fluctuations. In some embodiments, sliding window variance calculation uses an overlapping window approach, where the window slides by a fixed step size (such as 1 / 4 of the window length) each time to generate a smoother sequence of running oscillations and improve the time-series resolution of the analysis.

[0036] In practice, the output rate is determined by the slope of a linear regression of the change in output per unit time. The unit time is typically set as a standard production time unit (e.g., 1 hour or 1 shift). The change in output refers to the cumulative change in output within that time unit. The linear regression slope is the slope of the output-time curve fitted using the least squares method. This slope directly reflects the production efficiency trend of the key evaluation unit; a positive slope and a larger value indicate faster output growth and higher production efficiency. The training data for the impact prediction model comes from scheduling disturbance events and their subsequent impact records in historical production records. Each training sample includes the operational oscillation indicators and output rate data of the key evaluation unit before and after the disturbance as input features, and the degree of system performance impact as assessed by experts or actually observed as the output label. Model training uses an error backpropagation algorithm to optimize network weights until the error between the predicted output and the true label reaches a preset threshold. It is understandable that the calculation of the linear regression slope is sensitive to outliers; therefore, a smoothing filter is usually applied to the output sequence before fitting to eliminate the influence of random interference. In some embodiments, the calculation of output rate also takes into account the differences in product type. By introducing a standard product equivalence coefficient, the output of different products is converted into comparable standard output, and then linear regression analysis is performed to ensure the comparability of output rate in mixed production scenarios of different products.

[0037] In practice, the time interval from the start to the end of a task in a key evaluation unit is recorded as the task execution duration. The task start time is defined as the timestamp when raw materials enter the key evaluation unit or when the equipment start signal is triggered. The task end time is defined as the timestamp when finished products are produced or when the equipment stop signal is triggered. The time interval is calculated by the difference between the two timestamps, accurate to the second. The task execution duration directly reflects the efficiency of the key evaluation unit in completing a specific production task. The influence coefficient and task execution duration are standardized using the Z-score normalization method. This involves subtracting the mean of historical data from the original values ​​of the influence coefficient and task execution duration, and then dividing by the standard deviation of historical data. The transformed data conforms to a standard normal distribution with a mean of 0 and a standard deviation of 1. Standardization eliminates the differences in units and numerical range between the influence coefficient and task execution duration, allowing them to be integrated under the same standard. A nonlinear mapping function is applied to combine the standardized values ​​into a performance quantification score. The nonlinear mapping function uses the Sigmoid function, and its specific expression is: Where: symbol Represents the final calculated performance quantification score, with the symbol... Represents the standardized influence coefficient, with the sign... Represents the standardized task execution time, symbol and These are the weighting coefficients assigned to the influence coefficient and the task execution time, respectively. and weighting coefficients The value is determined using the analytic hierarchy process (AHP) to reflect the relative importance of the two factors in performance evaluation; the performance quantification score is... It is a value between 0 and 1, and the higher the score, the better the overall performance of the key evaluation unit.

[0038] In practice, the entire process from monitoring to performance quantification is executed automatically on a periodic basis, for example, an analysis cycle is initiated every 0.5 hours or every hour. Real-time data from the high-precision sensor array provides the data foundation for calculating operational oscillation indicators and output rates. Sliding window variance and linear regression slope are effective mathematical tools for characterizing the dynamic behavior of key evaluation units. The impact prediction model links behavioral characteristics with disturbance resistance capabilities. Records of task execution time reflect basic efficiency, standardized transformation ensures data comparability, and nonlinear mapping functions achieve the scientific fusion of multiple indicators. It can be understood that the performance quantification score provides a unified metric for horizontal comparison between different key evaluation units and prepares the ground for subsequent weight factor fusion and comprehensive optimization index calculation. Optionally, the performance quantification score, along with the equipment identifier and timestamp of the key evaluation unit, is stored in the performance database and a performance trend report is generated to support management decisions on equipment maintenance and production scheduling. The impact prediction model is periodically incrementally trained using the latest production data to adapt to changes in the production system over time and maintain prediction accuracy.

[0039] Example 4: In specific implementation, a standard test load is applied to the key evaluation unit and its energy consumption to output ratio is measured to obtain the energy efficiency. The standard test load is a standardized workload that simulates a typical production task. Its parameter settings, such as processing speed and material input rate, are determined according to industry specifications or internal enterprise standards. The application process is automatically executed by a preset program in the control system to ensure the consistency and repeatability of the load conditions. Energy consumption is measured using a high-precision energy meter installed on the main power supply circuit of the key evaluation unit. The energy meter records active energy consumption at fixed time intervals. Output is measured by counting the number of qualified products using a counter built into the key evaluation unit or a specially installed photoelectric sensor. The energy efficiency is obtained by calculating the ratio of total energy consumption to total output within a specific time period. This ratio reflects the energy consumed by the key evaluation unit to produce a unit of product. The lower the value, the higher the energy efficiency. In practice, the load is gradually increased until the system reaches its stability limit, and the maximum output value is recorded as the maximum throughput capacity. The gradual increase in load employs a step-by-step strategy, increasing the load by a preset percentage (e.g., 5%) at fixed time intervals (e.g., ten minutes). After each load increase, the system is allowed to stabilize. Stability is determined by the output volatility of the key evaluation unit being below a threshold and the equipment operating parameters being within normal ranges. The system stability limit is the critical point at which further load increases would lead to equipment alarms, product quality exceeding standards, or system instability. The average hourly output value recorded during a stable operating cycle before this critical point is the maximum throughput capacity. Maximum throughput capacity characterizes the peak capacity of the key evaluation unit under sustainable operating conditions. It is understandable that the design of the standard test load needs to cover the typical operating modes of the key evaluation unit to ensure that the measured energy utilization rate is representative.

[0040] In practical implementation, the entropy method is used to weight energy consumption utilization rate and maximum throughput capacity to generate unit weight factors. The entropy method is an objective weighting method based on information entropy. Its calculation process first requires constructing an evaluation matrix, where rows correspond to different key evaluation units, and columns correspond to the two indicators: energy consumption utilization rate and maximum throughput capacity. The evaluation matrix is ​​normalized to eliminate the influence of dimensions. Then, the information entropy of each indicator is calculated. Information entropy reflects the dispersion of the indicator data; the greater the dispersion of the indicator, the smaller the information entropy, and the greater the weight of that indicator in the evaluation. Finally, the weight coefficient of each indicator is calculated based on the information entropy. The unit weight factor is the weighted calculation result of the weight of the energy consumption utilization rate indicator and the weight of the maximum throughput capacity indicator, reflecting the relative contribution of the two indicators in determining the importance of the key evaluation unit. It can be understood that the advantage of the entropy method is that its weights are entirely determined by the distribution characteristics of the data itself, avoiding the bias of subjective judgment and making the allocation of unit weight factors more objective and fair. Optionally, when constructing the evaluation matrix, if there are outliers in energy consumption utilization or maximum throughput, data cleaning should be performed first, such as using the three-standard-deviation rule to remove outlier data points, in order to ensure the robustness of the entropy method calculation.

[0041] In practice, the performance quantification results and unit weighting factors are integrated to output the comprehensive optimization index of the coarse-line four-line production system under the current scheduling scenario. The performance quantification score of each key evaluation unit is multiplied by its corresponding unit weighting factor to obtain a weighted score. The performance quantification score is a value between zero and one obtained from the previous step of performance quantification of the key evaluation units. The unit weighting factor is a coefficient greater than zero calculated using the entropy method. The multiplication operation scales the performance quantification score according to the importance of its corresponding key evaluation unit; the higher the importance of the key evaluation unit, the greater its contribution to the final result. The weighted scores of all key evaluation units are summed, aggregating the contributions of each key evaluation unit into a total scalar value. This value reflects the overall performance level of the entire production system after considering the different importance of each unit. The cumulative sum is converted into a comprehensive optimization index between zero and one hundred using normalization. Normalization employs a linear transformation method, dividing the cumulative sum by the sum of the unit weight factors of all key evaluation units and then multiplying by one hundred. This ensures the comprehensive optimization index falls within an intuitive score range of zero to one hundred, facilitating horizontal comparisons and trend analysis for managers. The comprehensive optimization index is a relative value; its level depends not only on the actual performance of each key evaluation unit but also on the allocation of unit weight factors. It comprehensively reflects the optimization status of the production system across multiple dimensions, including energy efficiency, capacity, and performance. Optionally, the calculation results of the comprehensive optimization index, along with metadata such as timestamps and production batch numbers, are stored in the system database and can be displayed in real-time on the monitoring dashboard, providing immediate reference for production scheduling decisions. Refer to Table 1, which illustrates the process of calculating the weighted score and comprehensive optimization index from the performance quantification scores of key evaluation units and unit weight factors. Table 1: Calculation Table of Weights and Scores for Key Assessment Units In practice, the sum of the weighted scores is... The sum of the unit weight factors for all key evaluation units is Normalization is applied. Therefore, the comprehensive optimization index is 79. It is understood that this example is only for illustrating the calculation logic; in actual systems, the number of key evaluation units may be much greater, and the calculation is entirely automated. In some embodiments, the calculation cycle of the unit weight factor may differ from the update cycle of the performance quantification score. For example, the unit weight factor may be recalculated monthly based on the previous month's operating data, while the performance quantification score is updated hourly to ensure the stability of weight allocation while maintaining the real-time performance evaluation. In specific implementations, the process of applying standard test loads needs to ensure production safety. Dedicated protection logic is set up to immediately stop load increases and fall back to a safe load level once abnormal equipment parameters are detected. Maximum throughput capacity determination is usually scheduled during planned equipment maintenance periods or low-load production periods to minimize the impact on normal production. Energy utilization rate measurements consider corrections for environmental factors, such as compensating for electricity meter readings based on measured ambient temperature to improve data accuracy. When calculating unit weight factors using the entropy method, if the data variation of a certain indicator is extremely small, a minimum weight lower limit is set for that indicator to prevent its effect from being excessively weakened. Optionally, in addition to numerical form, the comprehensive optimization index can also be supplemented with color coding or level divisions to enhance the intuitiveness of the results presentation. The calculation module that integrates performance quantification results and unit weight factors is designed as a highly cohesive and loosely coupled software service, which can be called by other modules in the production management system through an application programming interface.

[0042] See Figure 4 This chart presents the comprehensive optimization analysis results of key evaluation units in the production system in terms of energy efficiency and capacity. The relationship between energy utilization and maximum throughput capacity of each key unit is presented in the form of a bubble chart. The size of each bubble represents the weight factor of that unit in the system, and the color intensity reflects its performance quantification level. The overall distribution shows that units with lower energy utilization and higher throughput capacity exhibit better overall performance, while the differences in bubble size highlight the varying importance of different units to the overall system optimization. This chart visually reveals the balance between system energy efficiency and capacity, providing important basis for identifying optimization priorities and resource allocation.

[0043] Example 5: In specific implementation, equipment configuration parameters and historical scheduling records are read from the production management database. The production management database is a relational database management system deployed in the factory information network. Equipment configuration parameters are stored in a data table named "Equipment Basic Information Table". This table contains structured fields such as equipment number, equipment model, rated capacity, maximum allowable operating speed, production line number, and installation location coordinates. Historical scheduling records are obtained from a data table named "Production Scheduling History Table". Historical scheduling records contain detailed entries such as task ID, product code, planned start time, planned end time, actual start time, actual end time, executing equipment number, and task status (completed / interrupted / queued). The read operation is implemented through a structured query language. For example, a SELECT statement is executed to select the currently valid equipment configuration parameters and historical scheduling records within the last three months from the corresponding data table. The read process has a data verification mechanism to mark records with empty values ​​or incorrect formats. In practical implementation, an IoT gateway captures the timestamps and task types of production task triggering events in real time. The IoT gateway is a hardware device connecting the production site control network to the upper-level information system, such as an industrial gateway supporting the OPCUA protocol. The IoT gateway monitors signal changes from programmable logic controllers or distributed control systems in real time. When a new production task is detected, the IoT gateway immediately captures the event and records a timestamp accurate to milliseconds. Simultaneously, the IoT gateway parses the identification field in the task instruction packet to determine the task type, which includes predefined categories such as regular batch production tasks, emergency order insertion tasks, equipment debugging tasks, and sample trial production tasks. It is understandable that the stability and integrity of the production management database are crucial, requiring regular data backups and consistency checks to ensure the accuracy and reliability of the read equipment configuration parameters and historical scheduling records.

[0044] In practical implementation, equipment configuration parameters, historical scheduling records, and real-time task data are integrated into a structured scheduling configuration dataset. This integration process is completed within a dedicated data processing service. This service first assigns a unique dataset identifier to each record, then aligns and correlates the equipment configuration parameters, historical scheduling records, and real-time task data according to the time dimension. For example, it links the three types of data through equipment numbers and foreign key relationships. Data cleaning operations are performed during the integration process, including removing duplicate records, filling in missing values, and correcting obvious outliers. The final structured scheduling configuration dataset is stored in a unified tabular format, with each row representing a complete data record containing all relevant fields of equipment static parameters, historical performance indicators, and real-time task information. In some embodiments, the structured scheduling configuration dataset can be specifically represented as a wide table, whose fields include, but are not limited to: record ID, timestamp, equipment number, equipment model, rated capacity, current task type, task priority, planned output, actual output, task status, equipment real-time operating speed, duration of this task execution, and associated historical average completion rate. This wide table structure facilitates multi-dimensional queries and correlation analysis by subsequent analysis modules. Optionally, the integration process can also incorporate a data quality scoring mechanism to calculate a quality score for each record. Records with low quality scores can be flagged or excluded to improve the overall reliability of the dataset.

[0045] In practical implementation, taking a specific four-line coarse wire production system as an example, this system includes four parallel wire production lines. The data acquisition process is as follows: Equipment configuration parameters are read from the production management database. For example, for the core equipment of production line 1: the double twisting machine, the read equipment configuration parameters include equipment number DT-001, equipment model DT-450, rated capacity of 1800 meters per hour, maximum allowable operating speed of 450 rpm, production line number Line-1, and installation location coordinates A-02. Simultaneously, historical data is read from the production management database. The scheduling records, for example, can be read from the historical scheduling records of the double twisting machine within the past week. One record shows that the task ID is T20231026001, the product code is WIRE-5.0mm, the planned start time is 2023-10-26 08:00:00, the planned end time is 2023-10-26 16:00:00, the actual start time is 2023-10-26 08:02:30, the actual end time is 2023-10-26 15:55:10, the executing equipment number is DT-001, and the task status is completed. In practical implementation, the IoT gateway captures production task trigger signals in real time. For example, at 09:00:00 on 2023-10-27, the IoT gateway detects a digital input signal from the Line-1 programmable logic controller changing from low to high. This signal corresponds to a new production task start command. The IoT gateway immediately records the timestamp 2023-10-27 09:00:00.125 and reads the data frame accompanying the signal, parsing the task type as "regular batch production task" and the product specification as WIRE-4.0mm. Optionally, the IoT gateway can also capture analog signals, such as a set line speed value, as a supplement to the task parameters. In some embodiments, the granularity of real-time capture can be very fine. For example, for a continuously producing extrusion line, the IoT gateway can capture real-time process parameters such as die head temperature and screw speed once per second, and these parameters are also part of the real-time task data.

[0046] In practical implementation, the equipment configuration parameters, historical scheduling records, and real-time task data from the above example are integrated. The integration process first establishes relationships between the data. For example, the static configuration, historical performance, and current new task of the double twisting machine are linked through the equipment number DT-001. The data cleaning operation checks whether the linear speed setpoint in the real-time task data is within the maximum operating speed range allowed by the equipment configuration parameters. If it exceeds the range, the record is marked as pending review. Finally, a complete structured record is generated for this new task trigger event and stored in the structured scheduling configuration dataset. This record contains the equipment's basic capability information, historical completion status, and real-time characteristics of the current task. It can be understood that the structured scheduling configuration dataset actually constructs a data panorama of the past, present, and planned states of the production system. Optionally, the integration process can be periodic batch processing, such as executing every five minutes, or event-driven stream processing, that is, it is performed immediately after each new task trigger event is captured, to achieve near real-time data updates. In practical implementation, the physical storage format of the structured scheduling configuration dataset can be columnar to optimize query performance, while a composite index based on timestamps and device numbers can be established to accelerate data retrieval for specific time periods or specific devices. As the data foundation for all subsequent advanced functions such as dynamic priority evaluation and multi-level optimization analysis, the quality and timeliness of the structured scheduling configuration dataset directly affect the effectiveness of the entire optimization method. Optionally, version management of the dataset can be configured to retain historical versions of the structured scheduling configuration dataset for backtracking analysis and model training.

[0047] See Figure 5 This visualization presents a comprehensive analysis of production task scheduling and execution. The graph uses horizontal task bars to compare the planned and actual execution times of each task, with different colors distinguishing task completion statuses such as completed normally, interrupted, or queued. The length of the task bars reflects the deviation between actual execution and the plan, and the overall layout clearly shows the timing of task execution and resource usage. This visualization method allows for the rapid identification of scheduling conflicts, equipment utilization issues, and task execution efficiency, providing an intuitive reference for optimizing production planning and improving scheduling accuracy.

[0048] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A production scheduling optimization method based on coarse-line four-line production, characterized in that, The method is implemented through the following steps: Collect the initial scheduling parameters and real-time task trigger signals of the coarse-line four-line production system to generate a scheduling configuration dataset; Dynamically prioritize the scheduling configuration dataset to determine whether to enable a multi-level optimization analysis mechanism; When the multi-layer optimization analysis mechanism is enabled, the structural efficiency index of the production line is calculated, and the dynamic status of material flow is captured simultaneously. By integrating structural performance indicators and material flow dynamics, the production line is intelligently partitioned to identify key evaluation units; Continuously monitor the operational oscillation indicators and output rates of key evaluation units, and analyze the impact coefficients of each unit on scheduling disturbances; The performance of key evaluation units is quantified based on the impact coefficient and task execution time. Extract the energy consumption efficiency and maximum throughput of key evaluation units, and derive the unit weighting factors; By integrating the performance quantification results and unit weight factors, a comprehensive optimization index for the coarse-line four-line production system under the current scheduling scenario is output.

2. The production scheduling optimization method based on coarse-line four-line production according to claim 1, characterized in that, Dynamic priority evaluation is performed on the scheduling configuration dataset to determine whether to enable a multi-level optimization analysis mechanism, specifically including: Extract device operating limits and scheduling frequency characteristics from the scheduling configuration dataset; By comparing the upper limit of equipment operation with the preset capacity threshold, it can be determined whether the production line load exceeds the limit; Analyze whether the scheduling frequency characteristics exceed the normal fluctuation range; If the upper limit of equipment operation is exceeded or the scheduling frequency characteristics are abnormal, the multi-level optimization analysis mechanism is skipped and a low optimization level is directly output; otherwise, the multi-level optimization analysis mechanism is activated.

3. The production scheduling optimization method based on coarse-line four-line production according to claim 2, characterized in that, Calculate the structural efficiency indicators of the production line and simultaneously capture the dynamic status of material flow, specifically including: The upper limit of equipment operation and scheduling frequency characteristics are normalized, and a weighted fusion algorithm is applied to generate structural performance indicators. Deploy flow sensing devices along the production line transmission path to measure material flow rate and flow fluctuations in real time; The fluid continuity model is used to calculate the material throughput per unit time, which is used as the dynamic state of the material flow.

4. The production scheduling optimization method based on coarse-line four-line production according to claim 3, characterized in that, By integrating structural performance indicators and material flow dynamics, the production line is intelligently partitioned to identify key evaluation units, specifically including: Input structural performance indicators and material flow dynamics into the cluster analysis model to generate partition reference values; The production line is divided into logical segments based on the partition reference values, with each segment serving as an evaluation unit. For each assessment unit, calculate its material demand intensity and failure history frequency; Material demand intensity is derived by analyzing unit equipment load and material flow models. The historical frequency of failures is obtained by statistically analyzing the ratio of the number of failures to the runtime of the unit within a historical period. Key evaluation units were selected based on a comprehensive score of material demand intensity and failure history frequency.

5. The production scheduling optimization method based on coarse-line four-line production according to claim 4, characterized in that, Continuously monitor the operational oscillation indicators and output rates of key evaluation units, and analyze the impact coefficients of each unit on scheduling disturbances, specifically including: High-precision sensor arrays are used to collect production sequences and operating parameters of key evaluation units in real time; A sliding window variance calculation is performed on the production series to obtain the operating oscillation index; The output rate is determined by the slope of a linear regression of the change in output per unit time. The model will run the oscillation index and output rate input impact prediction model and output the impact coefficient of each key assessment unit on scheduling disturbances.

6. The production scheduling optimization method based on coarse-line four-line production according to claim 5, characterized in that, Based on the impact coefficient and task execution duration, the performance of key evaluation units is quantified, specifically including: Record the time interval between the start and end of the key evaluation units as the task execution duration; Standardize and convert the impact coefficient and task execution time. A nonlinear mapping function is applied to combine the standardized values ​​into a performance quantization score.

7. The production scheduling optimization method based on coarse-line four-line production according to claim 6, characterized in that, Extract the energy consumption efficiency and maximum throughput capacity of key evaluation units, and derive unit weighting factors, specifically including: Apply standard test loads to key evaluation units and measure their energy consumption to output ratio to obtain energy utilization rate; By gradually increasing the load until the system reaches its stable limit, the maximum output value is recorded as the maximum throughput capacity. The entropy method is used to calculate the energy consumption efficiency and maximum throughput capacity by weighting, and a unit weight factor is generated.

8. The production scheduling optimization method based on coarse-line four-line production according to claim 7, characterized in that, By integrating performance quantification results and unit weighting factors, a comprehensive optimization index for the coarse-line four-line production system under the current scheduling scenario is output, specifically including: The performance quantification score of each key evaluation unit is multiplied by the corresponding unit weight factor to obtain the weighted score; The weighted scores of all key assessment units are summed. The summation is converted into a comprehensive optimization index between zero and one hundred by applying normalization.

9. The production scheduling optimization method based on coarse-line four-line production according to claim 1, characterized in that, Collect initial scheduling parameters and real-time task trigger signals from the coarse-line four-line production system to generate a scheduling configuration dataset, specifically including: Read equipment configuration parameters and historical scheduling records from the production management database; The IoT gateway captures the timestamps and task types of production task trigger events in real time. The device configuration parameters, historical scheduling records, and real-time task data are integrated into a structured scheduling configuration dataset.

10. A production scheduling optimization system based on a coarse-coarse four-line production line, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the coarse-line four-line production scheduling optimization method according to any one of claims 1 to 9.

Images