Tobacco leaf module formula intelligent design method and system and storage medium

Abstract

The invention relates to the technical field of cigarette digital formula design, in particular to a tobacco leaf module formula intelligent design method and system and a storage medium, and the method comprises the following steps: obtaining raw tobacco samples of a current period and the number of modules needing to be planned; obtaining an infrared spectrum of the crude tobacco sample; predicting the infrared spectrum of the crude tobacco sample by adopting a near-infrared prediction model to obtain a crude tobacco quality index prediction value; obtaining a planning constraint condition of the ith module; adopting a module scale maximum planning strategy to obtain a current module tobacco leaf formula and a module scale maximum target; adopting a module cost minimization planning strategy to obtain a current module tobacco leaf formula and a module cost minimization target; adopting a dual-objective planning strategy to obtain a final tobacco leaf formula of the current module; updating inventory data; planning the next module until the planning of n-1 modules is completed; and taking the residual tailings as the formula of the last module. According to the invention, large-scale and low-cost double-target simultaneous realization of each module formula is ensured.

Description

Technical Field

[0001] This invention relates to the field of digital cigarette formulation design technology, specifically to a method, system, and storage medium for intelligent design of tobacco leaf module formulations. Background Technology

[0002] With the continuous development of cigarette brands, the bottlenecks in the supply and demand structure of tobacco raw materials and cost pressures have gradually become prominent. To address the insufficient supply of high-quality tobacco raw materials and reduce tobacco usage costs, cigarette companies have developed new technologies to improve tobacco compatibility, such as pre-processing of formulations, modular formulation pounding and re-drying, and group processing, based on the characteristics of tobacco quality and brand positioning. Among these, modular formulation pounding and re-drying of sheet tobacco involves scientifically blending tobacco raw materials from different production areas, grades, and qualities before processing them in a tobacco re-drying plant to achieve the standardized quality required for cigarette product formulations. Because modular formulation of sheet tobacco can effectively enhance its use value through the design of raw material tobacco blends, it has become a major technical measure and key link commonly adopted by tobacco industrial enterprises. Currently, the design of modular formulations of sheet tobacco mainly relies on the experience of professional formulators combined with blending principles over the years, chemical quality, and sensory quality factors, while also considering the module scale and tobacco usage costs to a certain extent. This approach suffers from problems such as high workload, strong subjectivity, low efficiency, and difficulty in finding the optimal blend ratio.

[0003] The design of tobacco flake module formulations primarily revolves around the overall optimization of module quality, module size, and module cost. First, module quality is the foundation of formulation design and the primary constraint. Second, tobacco flake module design should prioritize resource allocation for higher-grade modules. This is because higher-grade modules can serve as individual components not only in the formulation of their own grade of cigarettes but also in formulations of lower-grade cigarettes, exhibiting broader adaptability and improving the overall quality of lower-grade cigarettes. Finally, for higher-grade module formulations, the overall optimization goal is to maximize module size while minimizing cost, which is a bi-objective programming optimization solution. The maximum size optimization is a linear objective function, while the minimum cost optimization is a non-linear objective function. For this complex multi-module, multi-objective, dynamic inventory-based resource allocation optimization problem, relying on traditional formulation experience is unlikely to yield the optimal resource allocation solution, often resulting in significant resource waste due to the inability to fully utilize resources.

[0004] Therefore, it is urgent to improve the digital design capabilities of tobacco leaf module formulations to ensure stable product quality, achieve rational utilization of tobacco raw materials, and maintain the long-term safety of enterprise operation. Summary of the Invention

[0005] The purpose of this invention is to provide a method, system, and storage medium for intelligent design of tobacco module formulations, in order to solve the technical problems of high subjectivity, low efficiency, difficulty in achieving optimal ratios, and waste of resources in the prior art.

[0006] To achieve the above objectives, the present invention provides a method for designing tobacco leaf module formulations, comprising: Obtain the raw tobacco sample for the current period and the number of modules to be planned; The infrared spectrum of the raw tobacco sample in the current period is obtained based on the raw tobacco sample in the current period; The infrared spectrum of the raw tobacco sample in the current period is predicted using the near-infrared prediction model of the previous period in order to obtain the predicted value of the raw tobacco quality index. Obtain the planning constraints for the i-th module; Based on the predicted values ​​of the raw tobacco quality indicators and the planning constraints of the current module, the current module tobacco formula and the maximum module size target are obtained by adopting the module size maximization planning strategy. Based on the predicted values ​​of the raw tobacco quality indicators and the planning constraints of the current module, a module cost minimization planning strategy is adopted to obtain the current module tobacco leaf formula and the module cost minimization target. Based on the objectives of maximizing module size and minimizing module cost for the current module, a dual-objective programming strategy is adopted to obtain the final tobacco leaf formula for the current module. Update current raw tobacco inventory data; Determine whether the planning of n-1 modules has been completed; if the planning of n-1 modules has been completed, use the remaining waste material as the formula for the nth module. If the planning of n-1 modules is not completed, increment i by 1 and return to the step of obtaining the planning constraints of the i-th module.

[0007] Optionally, the predicted values ​​for raw tobacco quality indicators include: Chemical quality indicators include total sugar, reducing sugar, nicotine, chlorine, total nitrogen, potassium, pH, chlorogenic acid, starch, sugar-to-alkali ratio, and sugar-to-nitrogen ratio; Sensory quality indicators include light aroma, medium aroma, strong aroma, aroma type, poor aroma type, part of aroma, elegance 1 and elegance 2.

[0008] Optionally, based on the predicted values ​​of the raw tobacco quality indicators and the planning constraints of the current module, the module size maximization planning strategy is used to obtain the current module tobacco leaf formula and the module size maximization target, including: The usage amount of each type of raw tobacco and the maximum target module size are calculated based on formulas (1) and (2). (1) (2) in, This is the negative of the maximum size of the current module. The usage amount of each type of raw tobacco, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. This represents the lower limit for the amount of each type of tobacco leaf used. This is the upper limit for the amount of each type of tobacco leaf used.

[0009] Optionally, based on the predicted values ​​of the raw tobacco quality indicators and the planning constraints of the current module, the module cost minimization planning strategy is used to obtain the current module tobacco leaf formula and the module cost minimization objective, including: With the goal of minimizing module cost, a general nonlinear optimization algorithm is used to solve for the module cost, which is then used as the first minimum cost solution. With the goal of minimizing module cost, an intelligent optimization algorithm is used to solve for the module cost, which serves as the second minimum cost solution. The better solution between the first and second cost-minimizing solutions is selected as the module cost-minimizing objective.

[0010] Optionally, with the objective of minimizing module cost, a general nonlinear optimization algorithm is used to solve for the module cost. The first solution for minimizing cost includes: The minimum cost of the current module is calculated using formulas (3) and (4). (3) (4) in, To achieve the minimum cost for the current module, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. This represents the lower limit for the amount of each type of tobacco leaf used. This is the upper limit for the amount of each type of tobacco leaf used.

[0011] Optionally, with the objective of minimizing module cost, an intelligent optimization algorithm is used to solve for the module cost. This second method for minimizing cost includes: The minimum cost of the current module is calculated using formulas (5) and (6). (5) (6) in, To achieve the minimum cost for the current module, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The coefficients of the penalty function are... and Let be the inequality penalty function. The equation penalty function is... The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints.

[0012] Optionally, based on the objectives of maximizing module size and minimizing module cost, a bi-objective programming strategy is adopted to obtain the final tobacco leaf formulation for the current module, including: The final tobacco leaf formula for the current module is obtained according to formulas (7) and (8). (7) (8) in, This is the comprehensive deviation function. For module size, To maximize module size, To balance the magnitudes, For weighting coefficients, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, To minimize module cost, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. This represents the lower limit for the amount of each type of tobacco leaf used. This is the upper limit for the amount of each type of tobacco leaf used.

[0013] Optionally, the method further includes: obtaining a near-infrared prediction model for the current period based on the raw smoke sample of the current period, for prediction of the next period.

[0014] On the other hand, the present invention also provides an intelligent design system for tobacco leaf module formulations, the system comprising: The raw tobacco quality index prediction module is used to obtain predicted values ​​of raw tobacco quality indexes using a near-infrared prediction model. The scale maximization module is used to obtain the module tobacco leaf formula and the module scale maximization target based on the predicted values ​​of the raw tobacco quality indicators and the module planning constraints. The cost minimization module is used to obtain the module tobacco leaf formula and the module cost minimization target based on the predicted values ​​of the raw tobacco quality indicators and the module planning constraints. The dual-objective programming module is used to obtain the final tobacco leaf formula for the current module based on the objectives of maximizing module size and minimizing module cost. The processor is connected to the raw tobacco quality index prediction module, the scale maximization module, the cost minimization module, and the bi-objective programming module, and the processor is configured to execute any of the methods described above.

[0015] In another aspect, the present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement any of the methods described above.

[0016] Beneficial effects of this invention: This invention, based on the quality requirements of modular formulations, prioritizes the resource needs of high-grade modular formulations. It solves for the scale objective through linear programming to maximize module size and for the cost objective through nonlinear programming to minimize module cost. Based on these two methods, a dual-objective loss function of scale and cost is constructed, resulting in a three-step progressive final module configuration scheme. This enables precise design of multiple modular formulations, significantly improving module design efficiency, scientifically controlling the rational allocation of raw tobacco resources, and ensuring the simultaneous achievement of the dual objectives of large-scale production and low cost for each modular formulation. This method can flexibly design tobacco leaf module formulations based on different objectives, meeting the needs of modular formulation design work in various formulation scenarios. It enhances the digitalization level of tobacco leaf module formulation design, ensuring stable product quality, achieving rational utilization of tobacco raw materials, and maintaining the long-term safety of enterprise operations.

[0017] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 A flowchart of a tobacco leaf module formulation design method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the maximum solution result for module 1 according to one embodiment of the present invention; Figure 3 This is a schematic diagram of the minimum cost solution result for module 1 according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the biobjective programming solution result of module 1 according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the maximum solution result for module 2 according to one embodiment of the present invention; Figure 6 This is a schematic diagram of the minimum cost solution result for module 2 according to one embodiment of the present invention; Figure 7 This is a schematic diagram of the biobjective programming solution result of module 2 according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the maximum solution result for module 3 according to one embodiment of the present invention; Figure 9 This is a schematic diagram of the minimum cost solution result for module 3 according to one embodiment of the present invention; Figure 10 This is a schematic diagram of the biobjective programming solution result of module 3 according to an embodiment of the present invention; Figure 11 This is a schematic diagram of the calculation results of module 4 according to one embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0020] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0021] like Figure 1 The diagram shows a flowchart of a tobacco module formulation design method according to an embodiment of the present invention. Figure 1 In this design method, the steps may include: In step S10, the raw tobacco sample for the current period and the number of modules to be planned are obtained; In step S11, the infrared spectrum of the raw tobacco sample for the current period is obtained based on the raw tobacco sample for the current period. In step S12, the near-infrared prediction model of the previous period is used to predict the infrared spectrum of the raw tobacco sample in the current period in order to obtain the predicted value of the raw tobacco quality index. In step S13, the planning constraints of the i-th module are obtained; In step S14, based on the predicted values ​​of raw tobacco quality indicators and the planning constraints of the current module, the current module tobacco formula and the maximum module size target are obtained by adopting the module size maximization planning strategy. In step S15, based on the predicted values ​​of raw tobacco quality indicators and the planning constraints of the current module, a module cost minimization planning strategy is adopted to obtain the current module tobacco leaf formula and the module cost minimization target. In step S16, a bi-objective programming strategy is adopted to obtain the final tobacco leaf formula for the current module based on the objectives of maximizing module size and minimizing module cost. In step S17, update the current raw tobacco inventory data; In step S18, it is determined whether the planning of n-1 modules has been completed; In step S19, if it is determined that the planning of n-1 modules has been completed, the remaining tailings are used as the formula for the nth module. If the planning of n-1 modules is not completed, increment i by 1 and return to the step of obtaining the planning constraints of the i-th module.

[0022] In such Figure 1 In the tobacco leaf module formulation design method shown, step S10 is used to obtain the raw tobacco sample for the current cycle and the required number of modules. In this embodiment, the specific method for obtaining the raw tobacco sample for the current cycle can be to collect raw tobacco samples allocated from various county-level production areas from tobacco re-drying plants, with each sample randomly collected 5 times from different positions on the tobacco bale, and mixed to form a tobacco sample. Step S11 is used to obtain the infrared spectrum of the raw tobacco sample for the current cycle based on the raw tobacco sample for the current cycle. A suitable amount of tobacco sample is taken and the near-infrared spectrum is collected using the same near-infrared spectrometer. Specifically, in this example, the raw tobacco sample can be dried at below 50 ℃ for about 4 hours until it can be crushed by fingers. It is then ground into powder using a cyclone mill, and the sample powder is passed through a 60-mesh (250 μm pore size) sieve. A suitable amount of powder was placed in a sample cup, and a fixed-mass weight was placed on top of the sample to allow it to compact naturally. The response values ​​(1609 points per near-infrared curve) at wavenumbers of 3800–10000 were then collected using a Fourier transform near-infrared spectrometer to obtain the near-infrared spectral data of the raw tobacco. In this example, the tobacco sample was refrigerated at a temperature below 4°C.

[0023] Step S12 is used to predict the quality indicators of this year's raw tobacco using the previous year's near-infrared model and the near-infrared spectra collected in step S11. Specifically, in this example, these may include 11 chemical quality indicators: total sugar, reducing sugar, nicotine, chlorine, total nitrogen, potassium, pH, chlorogenic acid, starch, sugar-to-alkali ratio, and sugar-to-nitrogen ratio; and 8 sensory quality indicators: light aroma, medium aroma, strong aroma, aroma type, aroma type difference, part of the tobacco, elegance 1, and elegance 2.

[0024] In this example, the modules that need to be planned are sorted according to quality priority, and the formulation planning is performed sequentially. In this example, the number of modules can be 3 to 5. Step S13 is used to obtain the planning constraints of the modules that need to be planned. First, starting with the module with the highest quality, the planning constraints of the module with the highest quality are obtained. Then, in step S14, the current module tobacco formulation and the maximum module size target are obtained by using the module size maximization planning strategy. In this example, the specific method for using the module size maximization planning strategy to solve the usage of raw tobacco in the module can be, for example, using formula (1) and formula (2) to solve the usage of each type of raw tobacco and the maximum module size target. (1) (2) in, This is the negative of the maximum size of the current module. The usage amount of each type of raw tobacco, and These are the constraints for the quality index inequality. The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module's quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. It is a 1-vector, meaning that the amount of certain tobacco leaves used is directly specified. These selected tobacco leaves are style-characteristic tobacco leaves for the module, and are based on total inventory or empirical rules; This represents the lower limit for the usage of each type of tobacco leaf, the specified minimum usage, or 0. This specifies the upper limit for the usage of each type of tobacco leaf, either the maximum usage amount or the tobacco leaf inventory. In this example, the linear programming solution can be obtained using the built-in Matlab linear programming function `linprog`. The result uses the optimal solution from three algorithms: dual-simplex, interior-point-legacy, and interior-point. The endpoint of the solution is... The change is less than 10 -6 .

[0025] Step S15 is used to obtain the current module tobacco leaf formula and the module cost minimization objective using a module cost minimization planning strategy. Based on the characteristics of a certain tobacco leaf module this year and the cost (purchase unit price) of the raw tobacco sample this year, the amount of raw tobacco to be used in the module is solved through nonlinear programming with the objective of minimizing the module cost. In this embodiment, in order to obtain a better solution, step S15 can use two models: a general nonlinear optimization algorithm and an intelligent optimization algorithm are used to calculate separately, and the better solution is selected as the final solution.

[0026] In this example, the specific method for solving the module cost using a general nonlinear optimization algorithm can be to call the built-in nonlinear programming function fmincon in Matlab and use the optimal solutions of three algorithms: sequential quadratic programming (SQP), interior-point method, and active-set method, with the endpoint being... The change is less than 10 -6 Specifically, the minimum cost of the current module can be solved, for example, using formulas (3) and (4): (3) (4) in, To achieve the minimum cost for the current module, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. It is a 1-vector, meaning that the amount of certain tobacco leaves used is directly specified. These selected tobacco leaves are style-characteristic tobacco leaves for the module, and are based on total inventory or empirical rules; This represents the lower limit for the usage of each type of tobacco leaf, the specified minimum usage, or 0. The upper limit for the usage of each type of tobacco leaf, which is the specified maximum usage or the inventory of tobacco leaves.

[0027] In this example, the specific method for solving the module cost using the intelligent optimization algorithm can be, for example, using formulas (5) and (6) to solve for the minimum cost of the current module, with the endpoint being... The change is less than 10 -6 : (5) (6) in, To achieve the minimum cost for the current module, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The coefficients of the penalty function are... and Let be the inequality penalty function. The equation penalty function is... It is a 1-vector, meaning that the amount of certain tobacco leaves used is directly specified. These selected tobacco leaves are style-characteristic tobacco leaves for the module, based on total inventory or empirical rules. The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Usage quantity The upper and lower limit inequality conditions. Specify the minimum usage or 0. This refers to the specified maximum usage or tobacco leaf inventory. In this example, the penalty function coefficient... The value can be 10 6 .

[0028] In this example, the algorithms could be: Sparrow Search Algorithm (SSA), Ion Swarm Optimization Algorithm (PSO), Simulated Annealing Algorithm (SA), Genetic Algorithm Optimization Algorithm (GA), Gray Wolf Algorithm (GWO), Improved Gray Wolf Algorithm (IGWO), Whale Algorithm (WOA), Improved Whale Algorithm (EWOA), Ant Colony Algorithm (ALO), Butterfly Algorithm (MFO), Dragonfly Algorithm (DA), Locust Algorithm (GOA), Improved Locust Algorithm (IGOA), Multi-Version Optimizer (MVO), Sine & Cosine Algorithm (SCA), Jellyfish Algorithm (SS), Balance Optimizer (EO), Vulture Algorithm (AVOA), Artificial Hummingbird Algorithm (AHA), Artificial Gorilla Army Algorithm (GTO), Flow Direction Algorithm (FDA), Arithmetic Optimization Algorithm (AO), Pelican Algorithm (POA), Chameleon Algorithm (CSA), Snake Algorithm (SO), Northern Goshawk Algorithm (NGO), Meerkat Algorithm (DMOA), Shark Algorithm (WSO), and Alibaba and the Forty Thieves Algorithm (AFT). Choose any number of calculation results from the 29 possible solutions and select the optimal solution.

[0029] Step S16 is used to obtain the final tobacco formula for the current module by employing a bi-objective programming strategy based on the objectives of maximizing module size and minimizing module cost. In this example, the objectives could be the maximum size obtained from the maximum module size programming solution in step S14 and the minimum cost obtained from the minimum module cost programming solution in step S15, with a magnitude balancing parameter and weight allocation introduced to construct the loss function. This transforms the bi-objective programming problem into a single-objective problem, with the endpoint being... The change is less than 10 -6 Specifically, in this example, the final tobacco formulation for the current module can be obtained, for example, using formulas (7) and (8): (7) (8) in, This is the comprehensive deviation function. For module size, To maximize the module size objective, the solution is found in step S14. To balance the magnitudes, For weighting coefficients, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The objective of minimizing module cost is solved in step S15. The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. It is a 1-vector, meaning that the amount of certain tobacco leaves used is directly specified. These selected tobacco leaves are style-characteristic tobacco leaves for the module, based on total inventory or empirical rules. This represents the lower limit for the usage of each type of tobacco leaf, the specified minimum usage, or 0. This represents the upper limit for the usage of each type of tobacco leaf, specifying either the maximum usage or the inventory level of tobacco leaves. In this example, to ensure a balance between the scale and cost objectives in the dual-objective planning, a certain limit can be set. When users' demand for module size outweighs their demand for module cost, then ,on the contrary When there is no bias towards dual objectives .

[0030] After completing the planning for the current module, update the current raw tobacco inventory data in step S17. Steps S18 and S19 are used to continue planning the next module following steps S13 to S17. The planning solution starts with the module with the highest quality. The last module does not require planning; all raw tobacco tailings inventory are directly treated as one module, and the module's quality, weight, and cost are calculated. In this example, the number of modules in a re-drying plant is generally 3 to 5, determined by experience.

[0031] The embodiments of the present invention also include obtaining a near-infrared prediction model for the current period based on the raw tobacco sample of the current period, for use in predicting the next period. In this example, obtaining the near-infrared prediction model for the current period includes: In step S20, the chemical quality indicators, sensory evaluation indicators, and near-infrared spectra of the raw tobacco for this year are obtained; In step S21, a near-infrared prediction model for this year is constructed.

[0032] In this example, the specific method for obtaining the chemical quality indicators, sensory evaluation indicators and near-infrared spectra of the raw tobacco in step S20 can be to take an appropriate amount of samples from the refrigerated raw tobacco of the current year to test the chemical quality indicators and sensory evaluation indicators, and to collect near-infrared spectra. All samples are tested within 6 months.

[0033] The determination of 11 chemical quality indicators: After preparing the samples according to "YC / T31-1996 Preparation and Moisture Determination of Tobacco and Tobacco Products Samples", the following standards were applied: "YC / T159-2019 Determination of Water-Soluble Sugars in Tobacco and Tobacco Products", "YC / T468-2021 Determination of Total Alkaloids in Tobacco and Tobacco Products", "YC / T162-2011 Determination of Chlorine in Tobacco and Tobacco Products", and "YC / T161-2002 Determination of Total Nitrogen in Tobacco and Tobacco Products". The contents of six conventional chemical components of tobacco leaves—total sugar, reducing sugar, nicotine, chlorine, total nitrogen, and potassium—were determined according to the standards in YC / T217-2007 "Determination of Potassium in Tobacco and Tobacco Products." The pH, chlorogenic acid, and starch content of tobacco leaves were determined according to YC / T222-2007 "Determination of pH in Tobacco and Tobacco Products," YC / T202-2006 "Determination of Polyphenolic Compounds Chlorogenic Acid, Hyoscyamine, and Rutin in Tobacco and Tobacco Products," and YC / T216-2013 "Determination of Starch in Tobacco and Tobacco Products—Continuous Flow Method." The dry basis percentage content of the chemical indicators was calculated, and the ratio of reducing sugar to total alkaloids was calculated to obtain the sugar-alkaloid ratio, and the ratio of reducing sugar to total nitrogen was calculated to obtain the sugar-nitrogen ratio.

[0034] The determination of 8 sensory evaluation indicators: The sensory quality indicators were optimized and designed based on the usage characteristics of tobacco companies and the accuracy of near-infrared spectroscopy modeling. These indicators include 6 items: light aroma, medium aroma, strong aroma, aroma type, aroma type difference, part, elegance 1 and elegance 2. The three indicators—light aroma, medium aroma, and strong aroma—are binary indicators determined through commonly accepted tobacco leaf samples. That is, if a raw tobacco is judged to be light aroma, the light aroma value is 1, and the medium and strong aroma values ​​are both 0; if a raw tobacco is judged to be medium aroma, the medium aroma value is 1, and the light and strong aroma values ​​are both 0; if a raw tobacco is judged to be strong aroma, the strong aroma value is 1, and the light and medium aroma values ​​are both 0. The part of the tobacco is a ternary indicator determined through commonly accepted tobacco leaf samples, that is, 1, 2, and 3 represent the upper, middle, and lower parts, respectively. Among the light, medium, and strong aromas, if the light aroma value is the largest, the aroma type value is 1; if the medium aroma value is the largest, the aroma type value is 2; if the strong aroma value is the largest, the aroma type value is 3. The aroma type difference is the light aroma value minus the strong aroma value. Elegance 1 and Elegance 2 are continuous quantitative indicators, each with a score of 5 to 10. To ensure the reliability of the results, a professional group of 7 people scored them, and referring to the scoring rules of competitive sports, the two highest scores and two lowest scores were removed, and the average of the remaining three scores was taken.

[0035] This year, the near-infrared spectra of raw tobacco were collected twice. The near-infrared spectral acquisition method is the same as in step S10. Although long-term storage at low temperature has little impact on the quality indicators of raw tobacco, the moisture content will generally change significantly. Since near-infrared spectra are greatly affected by moisture, the spectra must be collected again to ensure the accuracy of subsequent modeling.

[0036] In this example, step S21, constructing the near-infrared prediction model for this year, can be based on the near-infrared spectra and 19 tobacco quality indicators collected in step S20. A relatively stable near-infrared spectrum-tobacco quality indicator prediction model can be established using partial least squares method. The coefficient of determination R² for the fit between the true and predicted values ​​of the validation and test sets is required to be greater than 0.95.

[0037] Due to the instability of the near-infrared model caused by the difference in tobacco leaf year, in order to ensure the accuracy of the model, the near-infrared prediction model established this year can only be used to predict the quality indicators of raw tobacco samples for the following year.

[0038] The present invention also provides the following embodiments.

[0039] Taking 39 raw tobacco resources commissioned by China Tobacco Corporation to a tobacco re-drying plant in 2024 as an example, modules were designed according to the above method. According to the 2024 plan, this batch of resources was planned to be divided into four modules, ranked from highest to lowest quality: ultra-high-end module (Module 1), high-end module (Module 2), mid-range module (Module 3), and low-end module (Module 4). Module design proceeded sequentially from highest to lowest quality. Except for the low-end module, which consisted of leftover materials, the planning objectives for the other three modules were to maximize scale and minimize cost. Basic information on the production area, grade, weight, and cost of the 39 raw tobacco samples is shown in Table 1.

[0040] Table 1. Information on 139 raw tobacco samples and values ​​of 19 quality indicators.

[0041] Thirty-nine raw tobacco samples were collected from a tobacco re-drying plant. Each sample was randomly collected five times from different locations on the tobacco bale, with 500 g collected each time, and the samples were mixed to form a 2.5 kg raw tobacco sample. Near-infrared spectra of the 39 samples were collected according to step S10 of the technical plan, and the remaining tobacco samples were refrigerated in a constant temperature and humidity room at less than 4 ℃.

[0042] Using the 2023 near-infrared model and the near-infrared spectra of 39 tobacco varieties collected above, the following 11 chemical quality indicators for raw tobacco in 2024 were predicted and calculated: total sugar, reducing sugar, nicotine, chlorine, total nitrogen, potassium, pH, chlorogenic acid, starch, sugar-to-alkali ratio, and sugar-to-nitrogen ratio; and 8 sensory quality indicators: light aroma, medium aroma, strong aroma, aroma type, aroma type difference, part of tobacco plant, elegance 1, and elegance 2. The predicted values ​​for the 19 quality indicators are shown in Table 1.

[0043] Based on the quality requirements of China Tobacco Corporation for tobacco leaf modules, the quality indicators for the first three modules are shown in Table 2. The fourth low-end module uses all remaining tailings and calculates the quality indicators.

[0044] Table 2 Quality Indicators for the First Three Modules

[0045] Based on the style characteristics and long-term usage experience of the tobacco leaf modules, the required origin types and usage scales for each module are specified in Table 3. Table 3 shows that Module 1 includes all inventory from the 5 required origins numbered 1-5; Module 2 includes the remaining inventory from the 6 required origins numbered 6-11; Module 3 includes the remaining inventory from the 8 required origins numbered 12-19; ​​and Module 4 uses the remaining inventory from all origins of the tobacco waste. The lower limit (lb) for each tobacco leaf usage is set to 0, and the upper limit (ub) is the current remaining inventory.

[0046] Table 3 Style characteristics of each module

[0047] To improve computational efficiency, this patent uses Matlab Appdesigner to develop a visual computational script - the Intelligent Planning System for Tobacco Leaf Module Formula (IOAs).

[0048] Plan the solution for module 1 first, in descending order of module quality: solve for the maximum size of module 1 according to step S14, such as... Figure 2 As shown: "Input Data Table" is Table 1; the three tables in "Constraints" are the quality indicators in Table 2, the module style characteristics in Table 3, and the upper and lower limits of raw tobacco scale, respectively; "Algorithm Decision" shows the results of tobacco leaf usage and quality indicators calculated by various selected models; "Planning Algorithm" is the algorithm selection window for three different types of planning; "Result Formula Sheet" shows the module tobacco leaf usage and quality indicator results corresponding to the algorithm finally selected in "Algorithm Decision"; "Planning Solution" is the function button area.

[0049] Solve for the minimum cost of module 1 in step S15, such as... Figure 3 As shown.

[0050] Depend on Figure 1 and Figure 2 It can be seen that the dual-simplex algorithm yields the largest solution, resulting in goal1 = 34909, while the SQP algorithm yields the smallest solution, resulting in goal2 = 73.72. And k=1 is set. Solve the biobjective programming problem according to step S16, and choose as follows: Figure 4Results for method2. To improve the practical feasibility of the module design, tobacco leaves with a usage of less than 20 in the solved module are directly ignored (e.g., ...). Figure 4 The result within the red box is the corrected result. Figure 4 It can be seen that the final design size of Module 1 is 34156, and the cost is 74.75.

[0051] After completing the solution for Module 1, update the 2024 raw tobacco inventory. The solution results for Module 2 are as follows: Figures 5 to 7 As shown. After completing the solution for Module 2, update the 2024 raw tobacco inventory. The solution results for Module 3 are as follows. Figures 8 to 10 As shown. After solving Module 3, update the 2024 raw tobacco inventory. Module 4 uses all waste materials and calculates the scale and cost (e.g. Figure 10 (As shown).

[0052] The results of solving the four modules are summarized in Tables 4 to 6. Thus, the formulation design for the four tobacco leaf modules of a tobacco re-drying plant is complete.

[0053] Table 44 summarizes the solution results for all 44 modules (scale and cost).

[0054] Table 54 summarizes the solution results (usage) for all 54 modules.

[0055] Table 64 summarizes the solution results for all 64 modules (quality indicators).

[0056] While designing the modular formulation, China Tobacco's chemical analysis and sensory evaluation personnel must update the 2024 near-infrared model within a few months. Appropriate samples will be taken from refrigerated 2024 raw tobacco to test chemical quality indicators and evaluate sensory indicators, and near-infrared spectra will be collected. All samples should be tested within six months. Once the model is completed, it will be used to predict the quality indicators of 2025 raw tobacco samples.

[0057] On the other hand, the present invention also provides an intelligent design system for tobacco modular formulations. The system includes a raw tobacco quality index prediction module, a scale maximization module, a cost minimization module, a bi-objective programming module, and a processor. The raw tobacco quality index prediction module is used to obtain predicted values ​​of raw tobacco quality indicators using a near-infrared prediction model. The scale maximization module is used to obtain the modular tobacco formulation and the module scale maximization objective based on the predicted raw tobacco quality index values ​​and module planning constraints. The cost minimization module is used to obtain the modular tobacco formulation and the module cost minimization objective based on the predicted raw tobacco quality index values ​​and module planning constraints. The bi-objective programming module is used to obtain the final tobacco formulation of the current module based on the module scale maximization objective and the module cost minimization objective. The processor is connected to the raw tobacco quality index prediction module, the scale maximization module, the cost minimization module, and the bi-objective programming module, and the processor is configured to execute any of the methods described above.

[0058] In another aspect, the present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement any of the methods described above.

[0059] Beneficial effects of this invention: This invention, based on the quality requirements of modular formulations, prioritizes the resource needs of high-grade modular formulations. It solves for the scale objective through linear programming to maximize module size and for the cost objective through nonlinear programming to minimize module cost. Based on these two methods, a dual-objective loss function of scale and cost is constructed, resulting in a three-step progressive final module configuration scheme. This enables precise design of multiple modular formulations, significantly improving module design efficiency, scientifically controlling the rational allocation of raw tobacco resources, and ensuring the simultaneous achievement of the dual objectives of large-scale production and low cost for each modular formulation. This method can flexibly design tobacco leaf module formulations based on different objectives, meeting the needs of modular formulation design work in various formulation scenarios. It enhances the digitalization level of tobacco leaf module formulation design, ensuring stable product quality, achieving rational utilization of tobacco raw materials, and maintaining the long-term safety of enterprise operations.

[0060] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0061] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0062] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0063] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0064] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0065] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0066] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0067] It should also be noted that 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. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0068] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for designing a tobacco leaf module formulation, characterized in that, The design method includes: Obtain the raw tobacco sample for the current period and the number of modules to be planned; The infrared spectrum of the raw tobacco sample in the current period is obtained based on the raw tobacco sample in the current period; The infrared spectrum of the raw tobacco sample in the current period is predicted using the near-infrared prediction model of the previous period in order to obtain the predicted value of the raw tobacco quality index. Obtain the planning constraints for the i-th module; Based on the predicted values ​​of the raw tobacco quality indicators and the planning constraints of the current module, the current module tobacco formula and the maximum module size target are obtained by adopting the module size maximization planning strategy. Based on the predicted values ​​of the raw tobacco quality indicators and the planning constraints of the current module, a module cost minimization planning strategy is adopted to obtain the current module tobacco leaf formula and the module cost minimization target. Based on the objectives of maximizing module size and minimizing module cost for the current module, a dual-objective programming strategy is adopted to obtain the final tobacco leaf formula for the current module. Update current raw tobacco inventory data; Determine whether the planning of n-1 modules has been completed; If the planning of n-1 modules is completed, the remaining tailings will be used as the formula for the nth module. If the planning of n-1 modules is not completed, increment i by 1 and return to the step of obtaining the planning constraints of the i-th module.

2. The design method according to claim 1, characterized in that, The predicted values ​​for raw tobacco quality indicators include: Chemical quality indicators include total sugar, reducing sugar, nicotine, chlorine, total nitrogen, potassium, pH, chlorogenic acid, starch, sugar-to-alkali ratio, and sugar-to-nitrogen ratio; Sensory quality indicators include light aroma, medium aroma, strong aroma, aroma type, poor aroma type, part of aroma, elegance 1 and elegance 2.

3. The design method according to claim 1, characterized in that, Based on the predicted values ​​of the raw tobacco quality indicators and the current module's planning constraints, the current module's tobacco leaf formula and the maximum module size objective are obtained using a module-scale maximization planning strategy, including: The usage amount of each type of raw tobacco and the maximum target module size are calculated based on formulas (1) and (2). ，（1） ，（2） in, This is the negative of the maximum size of the current module. The usage amount of each type of raw tobacco, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. This represents the lower limit for the amount of each type of tobacco leaf used. This is the upper limit for the amount of each type of tobacco leaf used.

4. The design method according to claim 1, characterized in that, Based on the predicted values ​​of the raw tobacco quality indicators and the current module's planning constraints, the module cost minimization planning strategy is used to obtain the current module's tobacco leaf formula and the module's minimum cost objective, including: With the goal of minimizing module cost, a general nonlinear optimization algorithm is used to solve for the module cost, which is then used as the first minimum cost solution. With the goal of minimizing module cost, an intelligent optimization algorithm is used to solve for the module cost, which serves as the second minimum cost solution. The better solution between the first and second cost-minimizing solutions is selected as the module cost-minimizing objective.

5. The design method according to claim 4, characterized in that, With the objective of minimizing module cost, a general nonlinear optimization algorithm is used to solve for the module cost. The first solution for minimizing cost includes: The minimum cost of the current module is calculated using formulas (3) and (4). ，（3） ，（4） in, The minimum cost for the current module. The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. This represents the lower limit for the amount of each type of tobacco leaf used. This is the upper limit for the amount of each type of tobacco leaf used.

6. The design method according to claim 4, characterized in that, With the objective of minimizing module cost, an intelligent optimization algorithm is used to solve for the module cost. The second minimum cost solution includes: The minimum cost of the current module is calculated using formulas (5) and (6). ，（5） ，（6） in, The minimum cost for the current module. The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, The coefficients of the penalty function are... and Let be the inequality penalty function. The equation penalty function is... The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints.

7. The design method according to claim 1, characterized in that, Based on the objectives of maximizing module size and minimizing module cost for the current module, a dual-objective programming strategy is adopted to obtain the final tobacco leaf formulation for the current module, including: The final tobacco leaf formula for the current module is obtained according to formulas (7) and (8). ，（7） ，（8） in, This is the comprehensive deviation function. For module size, To maximize module size, To balance the magnitudes, For weighting coefficients, The unit cost of each type of raw tobacco. The usage amount of each type of raw tobacco, To minimize module cost, The input values ​​are the quality indicators for each tobacco leaf. This represents the lower limit of the module quality indicator constraints. This represents the upper limit of the module quality indicator constraints. Used to fix the amount of a specific tobacco leaf used. This represents the lower limit for the amount of each type of tobacco leaf used. This is the upper limit for the amount of each type of tobacco leaf used.

8. The design method according to claim 1, characterized in that, The method further includes: obtaining a near-infrared prediction model for the current period based on the raw tobacco sample of the current period, for use in the prediction of the next period.

9. A smart design system for tobacco leaf modular formulation, characterized in that, The system includes: The raw tobacco quality index prediction module is used to obtain predicted values ​​of raw tobacco quality indexes using a near-infrared prediction model. The scale maximization module is used to obtain the module tobacco leaf formula and the module scale maximization target based on the predicted values ​​of the raw tobacco quality indicators and the module planning constraints. The cost minimization module is used to obtain the module tobacco leaf formula and the module cost minimization target based on the predicted values ​​of the raw tobacco quality indicators and the module planning constraints. The dual-objective programming module is used to obtain the final tobacco leaf formula for the current module based on the objectives of maximizing module size and minimizing module cost. The processor is connected to the raw tobacco quality index prediction module, the scale maximization module, the cost minimization module, and the bi-objective programming module, and the processor is configured to perform the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a processor, implement the method as described in any one of claims 1 to 8.

Images