Intelligent scheduling method for optimizing time distribution of different grouping batches in tobacco shred production

Abstract

The invention discloses an intelligent scheduling method for optimizing different grouping batch time distribution in tobacco shred production, and belongs to the technical field of tobacco production, and the method comprises the following steps: S1, data collection and processing: carrying out data collection through a sensor arranged on a tobacco shred production line, arranging the collected data into a time sequence data set, and carrying out data processing; s2, model training: training a prediction model by using the processed data, and adjusting model parameters to optimize the prediction model; and S3, optimizing a production plan: analyzing a result of the prediction model, and determining an optimal scheduling scheme. Prediction model training is carried out by utilizing historical data and real-time data, so that the prediction model can adapt to continuously changing production conditions, then prediction data is output in real time by utilizing the trained model, a prediction result is analyzed to optimize a production plan, batch scheduling is dynamically adjusted, the resource utilization rate is optimized, the downtime is reduced to the maximum extent, and the production efficiency is improved. The overall production efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of tobacco production, and in particular to an intelligent scheduling method for optimizing time allocation of different group batches in tobacco production. Background Art

[0002] Cut tobacco production plays a key role in determining product quality, and this stage involves the careful cutting, drying and conditioning of tobacco leaves to produce uniform cuts that ensure consistency in size, moisture content and texture. The uniformity of these cuts is critical, as any deviation can lead to significant changes in the final product, affecting consumer satisfaction and brand loyalty.

[0003] Traditional scheduling methods for tobacco production management have long involved fixed schedules based on historical data and average production conditions, but this approach is insufficient in a dynamic environment where unforeseen events (such as machine failures and sudden changes in material properties) can disrupt production, leading to delays and inefficiencies.

[0004] Based on the above problems, an intelligent scheduling method is proposed to optimize the time allocation of different group batches in tobacco production, and to actively adjust the production plan according to the prediction. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent scheduling method for optimizing the time allocation of different grouping batches in tobacco production, so as to solve the problems in the background technology.

[0006] To achieve the above object, the present invention provides an intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production, comprising the following steps: S1. Data collection and processing: Data is collected through sensors installed on the tobacco production line, and the collected data is organized into time series data sets and processed; S2, model training: use the processed data to train the prediction model and adjust the model parameters to optimize the prediction model; S3. Optimize production plan: Analyze the results of the prediction model and determine the optimal scheduling plan.

[0007] Preferably, the data collected by the sensors in S1 include machine performance, environmental conditions and material properties in tobacco production, wherein machine performance includes speed, throughput and operating status; environmental conditions include temperature, humidity and airflow; and material properties include moisture content, fiber strength and tobacco leaf texture.

[0008] Preferably, in S1, the time series data set is represented as: ; in, Indicates time t The complete dataset, Indicates the machine performance. Indicates environmental conditions, Represents material properties.

[0009] Preferably, the data processing process of S1 includes the following steps: 1) Noise reduction: Apply filtering algorithms to eliminate irrelevant fluctuations and outliers in the data collected during tobacco production; 2) Normalization: Scaling the data to a standard range ensures consistency and suitability of analysis; 3) Data fusion: combining the processed data into a unified data set; Processed dataset It is expressed as: ; in, Indicates denoising of the original data set; is a standardized function, and its mathematical expression is: ; in, represents the mean of the data set, Represents the standard deviation of the data set.

[0010] Preferably, the specific steps of S2 include: 1) Select a machine learning model based on the characteristics of production data and build a prediction model as follows: ; in, For time The predicted value of is the constant term coefficient, For the Input features, is the coefficient of the first-order term; is the quadratic coefficient, corresponding to the input feature and The interaction of is the total number of input features; 2) Historical data collected on the tobacco production line and real-time datasets Perform data fusion, input the fused data into the prediction model for training, and adjust the model parameters.

[0011] Preferably, during the model training process, the loss function is defined as: ; in, are model parameters, including , and ; For the The actual value of the samples, is the model's predicted value for the i-th sample; N is the total number of input features; Then the gradient descent method is used to update the model parameters: ; in, represents the learning rate, is the number of iterations of gradient descent, is 1, 2, 3, ..., is the loss function L with parameters The gradient at The prediction model is continuously trained through multiple sets of data, the model parameters are optimized, and the accuracy of the prediction model is improved.

[0012] Preferably, in S3, the objective function of optimizing production includes minimizing downtime and maximizing throughput, and the expression of the objective function is: ; in, Represents the scheduling plan, is the objective function, represents the minimum scheduling solution of the objective function; For downtime, For quality, is the throughput, , , is the weight factor; The weight assigned to the downtime goal depends on the cost of downtime, including equipment depreciation, labor costs, and potential losses from order delays. The weights assigned to quality objectives depend on the costs of quality problems, including rework, scrap, reduced customer satisfaction, and loss of brand reputation. The weight assigned to the throughput target depends on the benefits of increased throughput, including market share expansion, increased revenue, and the ability to meet customer demand; In the objective function expression, the scheduling vector is , , represents the set of feasible schedules given the current constraints, where the constraints are machine performance, material properties, and environmental conditions; Solve the objective function to obtain the optimal schedule for: .

[0013] Preferably, in the objective function, the downtime is affected by machine performance, environmental conditions and material properties, and is expressed as: ; Throughput is affected by machine performance and material properties and is expressed as: ; The quality depends on the deviation of environmental conditions, expressed as: ; According to the sensitivity analysis, the sensitivity of mass to environmental conditions is expressed as: .

[0014] Therefore, the present invention provides an intelligent scheduling method for optimizing the time allocation of different grouping batches in tobacco production, which uses historical data and real-time data to train a prediction model so that it can adapt to changing production conditions, and then uses the trained model to output prediction data in real time, analyzes the prediction results to optimize the production plan, dynamically adjusts batch scheduling, optimizes resource utilization, minimizes downtime, and improves overall production efficiency.

[0015] The present invention improves efficiency, stabilizes production results, and maintains consistency and quality of tobacco crushed material production by dynamically adjusting production plans based on real-time data and predictive analysis, and making proactive adjustments based on potential interruptions in the forecast results. This can help manufacturers reduce operating costs while maximizing throughput. The flexibility and scalability of the system make it suitable for various industries, providing practical solutions to the challenges posed by fluctuating production conditions and different material properties.

[0016] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic diagram of the system architecture of this embodiment; Figure 2 This is a downtime distribution diagram of an embodiment of the present invention; Figure 3 FIG. 4 is a throughput distribution diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0019] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0020] Example The present invention provides an intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production. The overall system architecture is as follows: Figure 1 As shown, the specific method includes the following steps: S1. Data collection and processing: Sensors installed on the tobacco production line are used to collect data related to machine performance, environmental conditions, and material properties. Machine performance includes speed, throughput, and operating status; environmental conditions include temperature, humidity, and airflow; and material properties include moisture content, fiber strength, and tobacco leaf texture.

[0021] The collected data are organized into a time series dataset, represented as, ; in, Indicates time t The complete dataset, Indicates the machine performance. Indicates environmental conditions, Represents material properties.

[0022] And the data processing of the above data set includes the following steps: 1) Noise reduction: Apply filtering algorithms to eliminate irrelevant fluctuations and outliers in the data; 2) Normalization: Scaling the data to a standard range ensures consistency and suitability of analysis; 3) Data fusion: combining the processed data into a unified data set; Processed dataset It is expressed as: ; in, Indicates denoising of the original data set; is a standardized function, and its mathematical expression is: ; in, represents the mean of the data set, Represents the standard deviation of the data set.

[0023] S2. Model training: Use the processed data to train the prediction model and adjust the model parameters to optimize the prediction model. Specifically: 1) Select a machine learning model based on the characteristics of production data and build a prediction model as follows: ; in, For time The predicted value of is the constant term coefficient, For the input features, is the coefficient of the first-order term; is the quadratic coefficient, corresponding to the input feature and The interaction of is the total number of input features; 2) Historical data collected on the tobacco production line and real-time datasets Perform data fusion and input the fused data into the prediction model for training. Data fusion is to fuse the data representing the same features in two data sets to form a new data set. For example, the machine performance in the two data sets is combined into a unified data set. Adjust the model parameters and first define the loss function as: ; in, are model parameters, including , and ; For the i The actual value of the samples, is the model's predicted value for the i-th sample; N is the total number of input features; Then the gradient descent method is used to update the model parameters: ; in, Represents the learning rate, which is a hyperparameter in the gradient descent method. It controls the speed at which the model parameters are updated. is the number of iterations of gradient descent, It is 1, 2, 3, ..., the number of iterations refers to the number of times the model parameters are updated during the training process, and each iteration is an update of the model parameters; is the loss function L with parameters The gradient at The prediction model is continuously trained through multiple sets of data, the model parameters are optimized, and the accuracy of the prediction model is improved.

[0024] S3. Optimize production plan: Analyze the results of the prediction model and determine the optimal scheduling plan.

[0025] The objective function for optimizing production includes minimizing downtime and maximizing throughput. The expression of the objective function is: ; in, Represents the scheduling plan, is the objective function, represents the minimum scheduling solution of the objective function; For downtime, For quality, is the throughput, , , is the weight factor, The weight assigned to the downtime goal depends on the cost of downtime, including equipment depreciation, labor costs, and potential losses from order delays. The weights assigned to quality objectives depend on the costs of quality problems, including rework, scrap, reduced customer satisfaction, and loss of brand reputation. The weight assigned to the throughput target depends on the benefits of increased throughput, including market share expansion, increased revenue, and the ability to meet customer demand; in, , , The specific settings usually follow these steps: Goal analysis: First, analyze the company's strategic goals and current production status to determine which goals are more important; Cost-benefit analysis: Evaluate the impact of each objective on costs and benefits, including direct costs (such as downtime costs, quality defect costs) and indirect costs (such as loss of brand reputation, reduced customer satisfaction); Weight allocation: Based on the analysis results, weight factors are allocated; these weights can be determined based on expert opinions, historical data analysis, or through experimental methods (such as A / B testing); Adjustment and optimization: During the implementation process, the weight factors are continuously adjusted according to the actual results and changes in the production environment to achieve the best production plan.

[0026] The setting of weight factors is an iterative and dynamic process that needs to be adjusted according to actual conditions. Usually, the sum of these weight factors is 1 to ensure that the relative importance of each item in the objective function is comparable. However, this is not a hard and fast rule, and sometimes the sum of weight factors can be different from 1 according to specific circumstances.

[0027] Downtime is affected by machine performance, environmental conditions and material properties and is expressed as: ; Throughput is affected by machine performance and material properties and is expressed as: ; The quality depends on the deviation of environmental conditions, expressed as: ; According to sensitivity analysis, the impact of environmental fluctuations on product quality is quantified, and the sensitivity of quality relative to environmental conditions is expressed as: .

[0028] In the objective function expression, the scheduling vector is , , represents the set of feasible schedules given the current constraints, where the constraints are machine performance, material properties, and environmental conditions; Solve the objective function to obtain the optimal schedule for: .

[0029] In order to verify the rationality of the above method, a specific example is used for experimental verification. The specific steps are as follows: 1) Develop a comprehensive simulation environment that simulates real-world production conditions, including dynamic changes in machine performance, environmental conditions, and material properties collected during the tobacco shred production process; 2) Generate the data required for 1,000 tobacco shreds production processes to replicate typical behavior patterns in tobacco shreds production. The generated data reflects key production parameters such as machine performance, environmental conditions, and material properties; machine performance data simulates operational factors such as speed, throughput, and stability; environmental conditions focus on variables such as temperature and humidity; and material properties capture characteristics such as moisture content, texture, and fiber strength of tobacco leaves.

[0030] 3) Using the prediction model in this method, we train the model using 80% of the generated data as the training set and 20% of the generated data as the test set. We use hyperparameter optimization techniques to fine-tune the model to ensure that it accurately captures the nonlinear relationship between the input features. 4) Two schemes are used for comparison to verify the feasibility of the method of the present invention.

[0031] Solution 1 (Baseline): Implement the baseline scheduling strategy, set the production parameters to the average value, and predict the downtime and throughput as a reference; Option 2 (Optimization): Based on the output of the prediction model, the optimization strategy in this method is used to dynamically adjust production parameters to reduce downtime and maximize throughput.

[0032] Compare the results of the above two schemes by drawing a map, such as Figure 2 As shown, the optimized downtime (red dashed line) is significantly lower than the baseline (blue dashed line), showing not only a reduction in average downtime but also a reduction in variability, indicating a more stable and predictable production process.

[0033] like Figure 3 As shown in the figure, compared with the baseline throughput (green dashed line), the optimized throughput curve moves to the right, and the throughput (red dashed line) shows a significant increase, reflecting that the optimized strategy has higher production efficiency. The effectiveness of the optimized strategy is further demonstrated by visual comparison of key indicators.

[0034] Therefore, the present invention relates to an intelligent scheduling method for optimizing the time allocation of different grouping batches in tobacco production, using historical data and real-time data to train a prediction model so that it can adapt to changing production conditions, and then using the trained model to output prediction data in real time, analyzing the prediction results to optimize the production plan, dynamically adjust batch scheduling, optimize resource utilization, minimize downtime, and improve overall production efficiency.

[0035] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. An intelligent scheduling method for optimizing the time allocation of different batches in tobacco production, It is characterized in that The following steps are involved: S1. Data collection and processing: Data is collected through sensors installed on the tobacco production line, and the collected data is organized into time series data sets and processed; S2, model training: use the processed data to train the prediction model and adjust the model parameters to optimize the prediction model; S3. Optimize production plan: Analyze the results of the prediction model and determine the optimal scheduling plan.

2. According to claim 1, an intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production, Features: The data collected by the sensors in S1 include machine performance, environmental conditions and material properties in tobacco production, wherein machine performance includes speed, throughput and operating status; environmental conditions include temperature, humidity and airflow; material properties include moisture content, fiber strength and tobacco leaf texture.

3. According to claim 1, an intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production, Features: In S1, the time series dataset is represented as: ； in, Indicates time t The complete dataset, Indicates the machine performance. Indicates environmental conditions, Represents material properties.

4. According to claim 1, an intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production, Features: The data processing process of S1 includes the following steps: 1) Noise reduction: Apply filtering algorithms to eliminate irrelevant fluctuations and outliers in the data collected during tobacco production; 2) Normalization: Scaling the data to a standard range ensures consistency and suitability of analysis; 3) Data fusion: combining the processed data into a unified data set; Processed dataset It is expressed as: ； in, Indicates denoising of the original data set; is a standardized function, and its mathematical expression is: ； in, represents the mean of the data set, Represents the standard deviation of the data set.

5. According to claim 1, an intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production, Features: The specific steps of S2 include: 1) Select a machine learning model based on the characteristics of production data and build a prediction model as follows: ； in, For time The predicted value of is the constant term coefficient, For the Input features, is the coefficient of the first-order term; is the quadratic coefficient, corresponding to the input feature and The interaction of is the total number of input features; 2) Historical data collected on the tobacco production line and real-time datasets Perform data fusion, input the fused data into the prediction model for training, and adjust the model parameters.

6. An intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production according to claim 5, Features: During the model training process, the loss function is defined as: ； in, are model parameters, including , and ; For the i The actual value of the samples, is the model's predicted value for the i-th sample; N is the total number of input features; Then the gradient descent method is used to update the model parameters: ； in, represents the learning rate, is the number of iterations of gradient descent, is 1, 2, 3, ..., is the loss function L with parameters The gradient at The prediction model is continuously trained through multiple sets of data, the model parameters are optimized, and the accuracy of the prediction model is improved.

7. An intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production according to claim 1, Features: In S3, the objective function of optimizing production includes minimizing downtime and maximizing throughput, and the expression of the objective function is: ； in, Represents the scheduling plan, is the objective function, represents the minimum scheduling solution of the objective function; For downtime, For quality, is the throughput, is the weight assigned to the downtime goal, depending on the cost of downtime; The weight assigned to the quality goal depends on the cost of the quality problem; is the weight assigned to the throughput target, depending on the benefit of increasing throughput; In the objective function expression, the scheduling vector is , , represents the set of feasible schedules given the current constraints, where the constraints are machine performance, material properties, and environmental conditions; Solve the objective function to obtain the optimal schedule for: 。 8. An intelligent scheduling method for optimizing the time allocation of different group batches in tobacco production according to claim 7, Features: In the objective function, downtime is affected by machine performance, environmental conditions and material properties, and is expressed as: ； Throughput is affected by machine performance and material properties and is expressed as: ； The quality depends on the deviation of environmental conditions, expressed as: ； According to the sensitivity analysis, the sensitivity of mass to environmental conditions is expressed as: 。

Images