Cigarette auxiliary material parameter determination method and device and storage medium

Abstract

The invention provides a cigarette auxiliary material parameter determination method and device and a storage medium. The method comprises the following steps: acquiring a cigarette demand input by a user; determining a target auxiliary material parameter corresponding to the cigarette demand based on a pre-constructed constraint space; and displaying the target auxiliary material parameters. In the embodiment of the invention, the constraint space is pre-constructed, so that global search and local search can be balanced in the optimization process through the simulated annealing algorithm, the target auxiliary material parameters, namely the optimal auxiliary material parameters, are determined, and various types of constraint conditions can be adapted, so that the auxiliary material consumption design efficiency is improved, and the auxiliary material consumption design efficiency is improved. And complex and changeable production requirements can be met.

Description

Technical Field

[0001] The present invention relates to the technical field of cigarette manufacturing, and in particular to a method, a device and a storage medium for determining auxiliary material parameters of cigarettes. Background Art

[0002] Cigarettes are tobacco products rolled into strips and are widely used consumer goods. The production process involves complex material selection and process control. The quality of cigarettes is mainly evaluated by their sensory properties, chemical composition and smoke emission. Among them, smoke emission is an important indicator for measuring cigarette quality, which directly affects consumers' smoking experience and health risks.

[0003] Smoke emission includes tar, nicotine, carbon monoxide and other components. The control and optimization of auxiliary materials corresponding to these components determine the quality of cigarettes. At present, the traditional auxiliary material dosage design method mainly relies on manual experience and trial and error. Due to the complex and changeable production requirements, the efficiency of auxiliary material dosage design through the above method is low and it is difficult to adapt to the complex and changeable production requirements. Summary of the invention

[0004] In view of this, the embodiments of the present invention provide a method, device and storage medium for determining auxiliary material parameters of cigarettes to solve the problems in the prior art of low efficiency in auxiliary material dosage design and difficulty in adapting to complex and changing production needs.

[0005] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0006] The first aspect of the embodiment of the present invention shows a method for determining auxiliary material parameters of a cigarette, the method comprising:

[0007] Obtaining cigarette demand input by the user;

[0008] Determining target auxiliary material parameters corresponding to the cigarette demand based on a pre-constructed constraint space;

[0009] The target auxiliary material parameters are displayed.

[0010] Optional, pre-built constraint space processes include:

[0011] Collecting historical auxiliary material parameter data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand, wherein the historical cigarette demand includes smoke index data indicated by each historical cigarette demand;

[0012] Preprocessing the historical auxiliary material parameter data to obtain auxiliary material parameter characteristic data;

[0013] Constructing a smoke index prediction model based on the auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand;

[0014] A constraint space is constructed based on the smoke index prediction model.

[0015] Optionally, the historical auxiliary material parameter data is preprocessed to obtain auxiliary material parameter characteristic data, including:

[0016] Determine whether there is a non-numeric feature in the historical auxiliary material parameter data;

[0017] If it exists, convert the non-numeric feature into corresponding coded data;

[0018] Replacing the corresponding non-numerical features in the historical auxiliary material parameter data based on the coded data to update the historical auxiliary material parameter data;

[0019] Auxiliary material parameter characteristic data is obtained based on the updated historical auxiliary material parameter data.

[0020] Optionally, the auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand are used to construct a smoke index prediction model, including:

[0021] The auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand are used as a sample set;

[0022] Dividing the sample set into a training set and a test set;

[0023] Using the training set to train the initial model to obtain a trained initial model;

[0024] The initial model is tested using the test set, and training is stopped when it is determined that the loss function value of the initial model is less than or equal to a preset target loss value, and the initial model is used as a smoke index prediction model.

[0025] Optionally, constructing a constraint space based on the smoke index prediction model includes:

[0026] Collect M groups of auxiliary material parameter data;

[0027] For each set of auxiliary material parameter data, the auxiliary material parameter data is processed using the smoke index prediction model, and a smoke index prediction result is output;

[0028] For each set of auxiliary material parameter data, determining whether the smoke index prediction result meets the cigarette demand corresponding to the auxiliary material parameter data;

[0029] If it is in compliance, the auxiliary material parameter data is recorded as valid auxiliary material parameter data;

[0030] A constraint space is constructed based on the effective auxiliary material parameter data and preset constraints.

[0031] Optionally, determining a target auxiliary material parameter corresponding to the cigarette demand based on the constraint space includes:

[0032] Based on the constraint space, determining a feasible solution for each feature;

[0033] Constructing a plurality of combined solutions based on the feasible solutions for each of the features;

[0034] Determine the target solution in the combined solution using the simulated annealing optimization algorithm;

[0035] The target solution is used as the target auxiliary material parameter corresponding to the cigarette demand.

[0036] Optionally, a simulated annealing optimization algorithm is used to determine a target solution in the combined solution, including:

[0037] For each combined solution, the combined solution is processed using a simulated annealing optimization algorithm to determine a new internal energy corresponding to the combined solution;

[0038] If the new internal energy is less than the previous internal energy, determining whether the combined solution is a preset solution;

[0039] If the combined solution is a preset solution, taking the combined solution as the target solution;

[0040] If the combined solution is not a preset solution, updating the combined solution to a new solution, and updating the current internal energy based on the new internal energy of the combined solution;

[0041] If the new internal energy is greater than or equal to the current internal energy, determining that the combined solution can be received based on the reception probability calculated based on the current internal energy and the new internal energy, updating the combined solution to a new solution, and updating the current internal energy based on the new internal energy of the combined solution;

[0042] If it is determined that all the combined solutions have not been processed, and the currently updated initial temperature is greater than the preset temperature threshold, the next combined solution is processed using a simulated annealing optimization algorithm;

[0043] If it is determined that all of the combined solutions have been processed and the currently updated initial temperature is less than a preset temperature threshold, the combined solution in the new solution is used as the target solution.

[0044] Optionally, also include:

[0045] Count all target solutions;

[0046] For each target solution, determine whether the constraint state flag of the target solution is -1;

[0047] If so, output the auxiliary material parameters corresponding to the target solution.

[0048] The second aspect of the embodiment of the present invention shows a device for determining parameters of auxiliary materials of cigarettes, the device comprising:

[0049] An acquisition unit, used for acquiring cigarette demand input by a user;

[0050] A processing unit, used for determining target auxiliary material parameters corresponding to the cigarette demand based on a pre-constructed constraint space;

[0051] The display unit is used to display the target auxiliary material parameters.

[0052] A third aspect of an embodiment of the present invention shows a storage medium, wherein the storage medium includes a storage program, wherein when the program is running, the device where the storage medium is located is controlled to execute a method for determining auxiliary material parameters of a cigarette as shown in an embodiment of the present invention.

[0053] Based on the above-mentioned embodiment of the present invention, a method, device and storage medium for determining auxiliary material parameters of cigarettes are provided, the method comprising: obtaining cigarette requirements input by a user; determining target auxiliary material parameters corresponding to the cigarette requirements based on a pre-constructed constraint space; and displaying the target auxiliary material parameters. In the embodiment of the present invention, a constraint space is pre-constructed so that the global search and the local search can be balanced during the optimization process through a simulated annealing algorithm, thereby determining the target auxiliary material parameters, i.e., the optimal auxiliary material parameters, and being able to adapt to various types of constraint conditions, thereby improving the efficiency of auxiliary material dosage design and being able to adapt to complex and changeable production requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0055] Figure 1 A schematic diagram of a constrained space construction process according to an embodiment of the present invention;

[0056] Figure 2 A schematic flow chart of a method for determining auxiliary material parameters of a cigarette according to an embodiment of the present invention;

[0057] Figure 3 This is a schematic structural diagram of a device for determining auxiliary material parameters of cigarettes according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] It should be noted that the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0061] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0062] See also Figure 1 , is a schematic diagram of a constrained space construction process according to an embodiment of the present invention, wherein the method comprises:

[0063] Step S101: collecting historical auxiliary material parameter data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demands, wherein the cigarette demands include smoke index data indicated by each historical cigarette demand;

[0064] Among them, the historical auxiliary material parameter data consists of three types of indicators, namely, cigarette paper indicators, tipping paper indicators and filter rod indicators, specifically including cigarette paper indicator data, tipping paper indicator data and filter rod indicator data.

[0065] Auxiliary material parameter data refers to the data of auxiliary material parameters.

[0066] Cigarette paper index data include characteristics of cigarette paper such as air permeability, width, basis weight, citric acid content, potassium content and sodium content;

[0067] Tipping paper index data include the width, air permeability, hole band distance from edge, perforation type and number of hole rows of tipping paper;

[0068] The filter rod index data include the length, circumference, pressure drop, wrapping paper air permeability, tow denier, tow denier and filter rod type.

[0069] Flue gas index data include tar content, nicotine content, CO content and other indicators.

[0070] Step S102: pre-processing the historical auxiliary material parameter data to obtain auxiliary material parameter characteristic data;

[0071] The specific implementation process of step S102 includes the following steps:

[0072] Step S11: Determine whether there is a non-numeric feature in the historical auxiliary material parameter data. If so, execute step S12; if not, directly execute step S13.

[0073] In the specific implementation of step S11, each feature in the historical auxiliary material parameter data is traversed to determine whether there is a feature that is not a numerical feature. If so, step S12 is executed; if not, step S13 is directly executed.

[0074] For example, features such as punch type and filter rod type are not numerical and need to be converted into data values;

[0075] Step S12: converting the non-numeric feature into corresponding coded data;

[0076] The coded data at least includes first coded data and second coded data.

[0077] In the specific implementation of step S12, if it is determined that the non-numeric feature in the historical auxiliary material parameter data is a perforation type, the first coding matrix is ​​searched to determine the first coding data corresponding to the perforation type in the tipping paper index data. If the non-numeric feature is a filter rod type in the filter rod index data, the second coding matrix is ​​searched to determine the second coding data corresponding to the perforation type in the tipping paper index data.

[0078] Specifically, the first coding matrix is ​​used to store the correspondence between the puncturing type and the coded data, and the specific implementation process includes:

[0079] Different puncturing types are converted into codes to obtain a first coding matrix.

[0080] It should be noted that the punching type refers to the types such as no punching, laser punching, electrostatic punching, and natural hole formation;

[0081] Specifically, for each puncture type, the encoder is called to encode it, and a coding vector of length A is obtained, in which the i-th element is 1 and the rest are 0; here i is the index position corresponding to the category of the classification feature in the sample;

[0082] Construct a coding matrix of length A from the coding vectors of all puncturing types and assign unique indexes to all puncturing types;

[0083] Different filter rod types are coded and converted to obtain a second coding matrix.

[0084] Specifically, for each filter rod type, the encoder is called to encode it, and the encoded filter rod types are combined to obtain a coding vector with a length of B;

[0085] The encoding vectors of all filter rod types are used to construct the corresponding encoding matrix, and unique indexes are assigned to all punch types.

[0086] It should be noted that the coding conversion method of the filter rod type is the same as the coding conversion method of the punching type, and they can refer to each other.

[0087] In the embodiment of the present invention, other data that is not numerical data also needs to be encoded in a vector manner first, and the corresponding encoded data will be determined.

[0088] For example: Taking color as a feature, the categories range: red, blue, and green;

[0089] Call the encoder to encode red, blue, and green, and build the color encoding matrix:

[0090] red blue green 1 0 0 0 1 0 0 0 1

[0091] Next, assign indices to red, blue, and green, for example, red = 0; blue = 1; green = 2;

[0092] If the sample feature is green 2, the corresponding encoding vector is [0,0,1] according to the 2-color encoding matrix;

[0093] Step S13: replacing the corresponding non-numerical features in the historical auxiliary material parameter data based on the coded data to update the historical auxiliary material parameter data.

[0094] In the process of specifically implementing step S13, if it is determined that the non-numeric feature existing in the historical auxiliary material parameter data is a punching type, the punching type data in the tipping paper index data is replaced by the first coding data to update the tipping paper index data in the historical auxiliary material parameter data; when the non-numeric feature is the filter rod type in the filter rod index data, the filter rod type data in the filter rod index data is replaced by the second coding data to update the filter rod index data in the historical auxiliary material parameter data.

[0095] Optionally, after executing step S13, the method includes determining the index of the feature that is not a numerical feature, and if it is a tipping paper index, performing a null value search on the tipping paper index data after the feature replacement, and replacing the searched null value with a value of 0 to update the tipping paper index data;

[0096] If it is a filter rod index, a null value search is performed on the filter rod index data after feature replacement, and the searched null value is replaced with a value of 0 to update the filter rod index data;

[0097] If it is a cigarette paper index, a null value search is performed on the cigarette paper index data after feature replacement, and the searched null value is replaced with a value of 0 to update the cigarette paper index data.

[0098] Step S14: Acquire auxiliary material parameter characteristic data based on the updated historical auxiliary material parameter data.

[0099] It should be noted that the extreme value includes the maximum value and the minimum value.

[0100] It should be noted that the specific implementation of step S14 includes the following steps:

[0101] Step S21: construct an empty N-dimensional feature array, and write the updated historical auxiliary material parameter data into the N-dimensional feature array to obtain an initial array;

[0102] In the process of implementing step S21, an empty N-dimensional feature array is constructed, and the values ​​corresponding to the features in the cigarette paper index data, the tipping paper index data and the filter rod index data are written into the N-dimensional feature array to obtain an initial array, that is, the cigarette paper index data, the tipping paper index data and the filter rod index data are arrayed.

[0103] It should be noted that N is set according to actual conditions or multiple experiments.

[0104] Step S22: performing normalization processing based on the initial array to obtain a target value for each feature;

[0105] The specific implementation of step S22 includes the following steps:

[0106] Step S31: extracting the maximum values ​​of all features in the initial array and the corresponding feature values;

[0107] In the specific implementation of step S31, the maximum and minimum values ​​of all features in the cigarette paper index data, tipping paper index data and filter rod index data in the updated historical auxiliary material parameter data, as well as the feature value of each feature are extracted.

[0108] Step S32: for each feature in the initial array, normalize the maximum value and the characteristic value corresponding to the feature to obtain a target value;

[0109] In the specific implementation of step S32, for each feature in the initial array, the maximum value and the characteristic value corresponding to the feature are substituted into formula (1) for normalization to obtain the target value.

[0110] Formula (1):

[0111]

[0112] Among them, X nor is the data corresponding to the normalized feature, that is, the target value, X min is the minimum value under this feature; X max is the maximum value under this feature; X is the eigenvalue corresponding to the feature to be normalized.

[0113] Step S23: updating the initial array based on the target value corresponding to each feature to obtain an auxiliary material parameter feature array.

[0114] In the specific implementation of step S23, the target value corresponding to each feature is replaced with the feature value corresponding to the feature in the initial array to obtain the auxiliary material parameter feature array.

[0115] Step S24: obtaining auxiliary material parameter feature data in the auxiliary material parameter feature array.

[0116] In the specific process of implementing step S24, each target value in the auxiliary material parameter feature array, ie, auxiliary material parameter feature data, is obtained.

[0117] Step S103: constructing a smoke index prediction model by combining the auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand;

[0118] The specific implementation process of step S103 includes the following steps:

[0119] Step S41: taking the auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand as a sample set;

[0120] Step S42: Divide the sample set into a training set and a test set;

[0121] In the specific process of implementing step S42, the sample set is divided into a training set and a test set.

[0122] Step S43: using the training set to train the initial model to obtain a trained initial model;

[0123] Step S44: testing the initial model using the test set, stopping training when it is determined that the loss function value of the initial model is less than or equal to a preset target loss value, and using the initial model as a smoke index prediction model.

[0124] In the process of implementing step S43 and step S44, the initial model is trained using the training set to obtain a trained initial model, and the initial model is tested using the test set. If the loss function value output by the initial model is less than or equal to the preset target loss value, the training is stopped and the initial model is used as the flue gas index prediction model. If the loss function value output by the initial model is greater than the preset target loss value, the initial model is continuously trained based on the training set.

[0125] In the embodiment of the present invention, the root mean square error function is used as the loss function of the initial model, and minimizing the loss function value of the initial model is used as the training goal, that is, the preset target loss value is to minimize the loss function value of the initial model;

[0126] The loss function can be shown as formula (2).

[0127] Formula (2):

[0128]

[0129] Among them, RMSE is the root mean square error, that is, the loss function; R is the number of training sets in the sample set, and Fd is the dth measured flue gas index data, F d ' is the dth predicted smoke index.

[0130] The specific process of implementing step S43 includes:

[0131] First, set the framework of the initial model, that is, preset the weights and biases of each layer of the neural network of the initial model;

[0132] Next, the feedforward network is used as the network type of the initial model;

[0133] Optionally, the result interval of tar content in the output result of the smoke index prediction model is (0, D1], D1 is the upper limit of tar content, D1>0; the result interval of CO content is (0, D2], D2 is the upper limit of CO content, D2>0; the result interval of nicotine content is (0, D3], D3 is the upper limit of nicotine content, D3>0.

[0134] At this time, the network layer structure of the initial model includes input layer, hidden layer and output layer.

[0135] Among them, the input layer is used to receive the characteristic data of material parameters, finished tobacco data, physical parameters of cigarettes and corresponding historical cigarette demand, etc.; the hidden layer has a two-layer structure, each layer has C neurons, and the RELU activation function is used; the output layer is used to output the predicted smoke indicators.

[0136] For each network layer, weighted sum calculation is performed using formula (3) to obtain the weighted sum of the current network layer.

[0137] z j =W j ×a+b j

[0138] Among them, z j represents the weighted sum of the jth layer, W j represents the weight of the jth layer, b j represents the j-th layer bias, and a represents the sample set;

[0139] The RELU activation function is used to calculate the output of each layer, and the function formula is f(t)=max(0, t);

[0140] Taking the output of each layer as the input of the next layer, the expression of this forward propagation process is:

[0141] f(W1×a+b1) -＞ W2×f(W1×a+b1)+b 2-＞ ...W k ×f(W k-1 ×a+b k-1 )+bk

[0142] Among them, W1 is the weight of the first layer of neural network, b1 is the bias of the first layer of neural network, W2 is the weight of the second layer of neural network, b2 is the bias of the second layer of neural network, and so on. k is the k-th layer neural network weight, b k is the bias of the k-th neural network layer, and k is the total number of neural network layers excluding the input layer.

[0143] Optionally, when training or running a neural network, the activation value of the output layer (i.e., the output of the neuron) becomes very large, exceeding the range of computer numerical representation (usually the representation range of floating-point numbers), resulting in errors or instability in numerical calculations, indicating that the output structure explodes at this time. This situation may cause the performance of the neural network to degrade or even fail completely.

[0144] Step S104: constructing a constraint space based on the flue gas index prediction model.

[0145] The specific process of implementing step S104 includes the following steps:

[0146] Step S51: collecting M groups of auxiliary material parameter data;

[0147] In the specific implementation of step S51, the user may input M groups of auxiliary material parameter data through the system, where M is a positive integer greater than or equal to 2.

[0148] Optionally, it also includes collecting the cigarette demand corresponding to each set of auxiliary material parameter data.

[0149] Step S52: for each set of auxiliary material parameter data, the auxiliary material parameter data is processed using the smoke index prediction model, and a smoke index prediction result is output;

[0150] In the specific implementation of step S52, for each set of auxiliary material parameter data, the smoke index prediction model obtained through the training is used to process the auxiliary material parameter data to output a smoke index prediction result;

[0151] Step S53: for each set of auxiliary material parameter data, determine whether the smoke index prediction result meets the cigarette demand corresponding to the auxiliary material parameter data; if so, execute step S54; if not, discard it.

[0152] It should be noted that the smoke index prediction result includes each smoke index data, such as tar content, nicotine content and CO content. The cigarette demand includes the smoke index data indicated by the user;

[0153] In the process of implementing step S53, for each set of auxiliary material parameter data, it is determined whether the tar content in the smoke index prediction result is less than or equal to the tar content in the cigarette demand. If so, it is compared whether the nicotine content in the smoke index prediction result is less than or equal to the nicotine content in the cigarette demand. If so, the CO content is continued to be compared until it is determined that each smoke index data in the smoke index prediction result meets the corresponding cigarette demand, then step S54 is executed. If any item is no, the corresponding auxiliary material parameter data is discarded.

[0154] Wherein, the cigarette demand includes smoke index data indicated by each historical cigarette demand;

[0155] Flue gas index data include tar content, nicotine content and CO content.

[0156] Step S54: Record the auxiliary material parameter data as effective auxiliary material parameter data.

[0157] In the specific process of implementing step S54, all auxiliary material parameter data outputting smoke index prediction results are recorded, and the auxiliary material parameter data meeting the cigarette requirements are recorded as effective auxiliary material parameter data.

[0158] Step S55: constructing a constraint space based on the effective auxiliary material parameter data and preset constraints.

[0159] The specific implementation of step S54 includes the following steps:

[0160] Step S61: constructing a search space based on the effective auxiliary material parameter data;

[0161] In the specific implementation of step S61, the effective auxiliary material parameter data is feature-classified to obtain data sets corresponding to different features, and different search spaces are constructed based on the data sets corresponding to the feature classifications;

[0162] That is to say, extract the value corresponding to the air permeability of cigarette paper in the cigarette paper index data from the effective auxiliary material parameter data, and make it a data set; then extract the value corresponding to the width of the tipping paper in the tipping paper index data from the effective auxiliary material parameter data, and make it a data set; and so on, until the data set corresponding to each feature is determined; then use the feature as the initial state of the search space, and the data set to which the feature belongs as the target state of the search space to construct the search space.

[0163] It should be noted that the search space refers to the process by which the computer finds a way to transform from the initial state to the target state and solves the correct answer to a problem based on given conditions.

[0164] Step S62: constraining the search space based on the preset constraints to obtain a corresponding constraint space.

[0165] It should be noted that the preset constraints refer to the auxiliary material parameter rules set according to the cigarette standards. For example, the air permeability of cigarette paper in the cigarette paper index data must meet the range a, the width of the tipping paper in the tipping paper index data must be less than b, and the length of the filter rod in the filter rod index data must be less than c, etc.

[0166] Among them, cigarette standards refer to standards set by national standards or standards set by experts.

[0167] The number of preset constraints is at least one.

[0168] In the specific implementation of step S62, the search space is conditionally constrained according to the preset constraint, and the constrained search space is used as the constraint space.

[0169] In an embodiment of the present invention, a smoke index prediction model is constructed by collecting and preprocessing historical auxiliary material parameter data, as well as finished tobacco data, cigarette physical index data and smoke index data, thereby reducing the number of trial and error times in the design process, greatly improving the efficiency of auxiliary material parameter design, and greatly reducing the labor input cost during design; so that the smoke index prediction model can be used to accurately control the release of tar, nicotine and CO content in the smoke of cigarettes, so as to meet the customized needs of different markets and consumers for smoke indexes of new products; then the constraint space constructed by the smoke index prediction model is used; so that the simulated annealing algorithm can balance the global search and the local search during the optimization process, so as to find the optimal auxiliary material parameter combination from the constraint space, and can adapt to various types of constraints; thereby improving the efficiency of auxiliary material dosage design, and being able to adapt to complex and changeable production needs.

[0170] Based on the constraint space shown in the above embodiment of the present invention, the embodiment of the present invention also discloses a flow chart of a method for determining parameters of auxiliary materials of cigarettes, such as Figure 2 As shown, the method includes:

[0171] Step S201: Obtaining cigarette demand input by a user.

[0172] Among them, the expected smoke index input by the user, that is, the cigarette demand, is obtained;

[0173] Step S202: determining target auxiliary material parameters corresponding to the cigarette demand based on the constraint space;

[0174] The specific implementation process of step S202 includes the following steps:

[0175] Step S71: Determine a feasible solution for each feature based on the constraint space;

[0176] In the specific implementation of step S71, the data set corresponding to each feature in the constraint space is processed according to the preset constraints, and the value that meets the preset constraints is determined, and the value is used as a feasible solution.

[0177] That is, for the preset constraints, the data set corresponding to each feature under the constraint space is traversed, the values ​​of the data set that do not meet the preset constraints are removed, and the remaining values ​​are used as feasible solutions for the features.

[0178] Optionally, if the user sets a fixed value corresponding to a feature when inputting cigarette demand, then during the execution of step S71, the value corresponding to the feature is directly used as a feasible solution, and the value corresponding to the feature does not participate in the processing of step S71.

[0179] Step S72: construct multiple combined solutions based on the feasible solutions of each feature.

[0180] Among them, the combined solution includes each feature of the auxiliary material parameters, such as the air permeability, width, basis weight, citric acid content, potassium content and sodium content of cigarette paper; the width, air permeability, hole band distance from the edge, perforation type and number of hole rows of the tipping paper; and the length, circumference, pressure drop, forming paper permeability, single denier of the tow, total denier of the tow and filter rod type and other features.

[0181] In the specific implementation of step S72, the feasible solutions of each feature are traversed, and each feasible solution of each feature is combined with each feasible solution in each other feature to obtain a combined solution constructed by the feasible solutions of each feature.

[0182] It should be noted that the combined solution refers to the combination of auxiliary material parameters of different circumferences and cigarette lengths obtained from the constraint space.

[0183] Step S73: Determine the target solution in the combined solution using a simulated annealing optimization algorithm;

[0184] The specific implementation of step S73 includes the following steps:

[0185] Step S81: for each combined solution, the combined solution is processed using a simulated annealing optimization algorithm to determine a new internal energy Energy_new corresponding to the combined solution.

[0186] In the specific implementation of step S81, the parameters such as the temperature threshold temp_limits, the cooling rate alpha, the initial temperature temperature, and the number of iterations L in the simulated annealing optimization algorithm are initialized, and then the new internal energy Energy_new, the current internal energy Energy_current, and the optimal internal energy Energy_best, that is, the initial value of the target internal energy, are defined; for each combined solution, the current internal energy Energy_new corresponding to the combined solution is determined by the simulated annealing optimization algorithm.

[0187] Specifically, the process of determining the current internal energy Energy_new corresponding to the combined solution by using the simulated annealing optimization algorithm includes:

[0188] Calling the trained flue gas index prediction model to process the combined solution to obtain a predicted value;

[0189] Among them, the predicted values ​​include predicted smoke index data, namely predicted tar content, predicted nicotine content and predicted CO content.

[0190] The deviation between the preset value and the cigarette requirement is taken as the new internal energy Energy_new.

[0191] It should be noted that data for which the expected smoke index is not set in the cigarette demand will not be included in the calculation.

[0192] In a specific implementation, if the cigarette demand does not include the expected tar content, it is only necessary to calculate the deviation between the predicted nicotine content and the expected nicotine content to obtain a first nicotine deviation; calculate the deviation between the predicted CO content and the expected CO content to obtain a first CO deviation; and combine the first nicotine deviation and the first CO deviation as the new internal energy Energy_new; similarly, if the cigarette demand does not include the expected nicotine content, or / and the expected CO content, the implementation process of the new internal energy Energy_new is the same as the implementation process of the new internal energy Energy_new that does not include the expected tar content, and they can refer to each other.

[0193] If the cigarette demand includes expected tar content, expected nicotine content and expected CO content, the deviation of each predicted index content (predicted tar content, predicted nicotine content, predicted CO content) and its corresponding expected index content is calculated respectively, and they are combined to obtain the new internal energy Energy_new.

[0194] Step S82: Determine whether the new internal energy Energy_new is less than the current internal energy Energy_current, if so, execute step S83, if not, execute step S86.

[0195] It should be noted that if the new internal energy Energy_new is the first combined solution for simulated annealing optimization, the current internal energy Energy_current is set as the initial value, and the new internal energy Energy_new is compared with the current internal energy Energy_current. If the new internal energy Energy_new is less than or equal to the current internal energy Energy_current, step S83 is executed; if the new internal energy Energy_new is greater than the current internal energy Energy_current, step S86 is executed.

[0196] If the new internal energy Energy_new is not the first combined solution to be optimized by simulated annealing, obtain the updated current internal energy Energy_current, and compare the new internal energy Energy_new with the current internal energy Energy_current. If the new internal energy Energy_new is less than or equal to the current internal energy Energy_current, execute step 83; if the new internal energy Energy_new is greater than the current internal energy Energy_current, execute step S86.

[0197] Step S83: Determine whether the combined solution is a preset solution, if so, execute step S84, if not, execute step S85.

[0198] It should be noted that the preset solution refers to the minimum solution, which is the minimum value of the auxiliary material parameter set in advance according to the cigarette standard.

[0199] In the specific implementation of step S83, it is determined whether the parameters in the current solution tobacco_current, that is, the combined solution, are all minimum values. If so, step S84 is executed, and if not, step S85 is executed.

[0200] Step S84: taking the combined solution as the target solution;

[0201] In the specific implementation of step S84, the current solution tobacco_current is updated with the optimal auxiliary material parameter tobacco_best, and the current internal energy Energy_current is updated with the new internal energy Energy_new, that is, the current internal energy Energy_current is replaced with the new internal energy Energy_new of the combined solution, and the next combined solution is processed, that is, returning to execute step S81.

[0202] Furthermore, the new internal energy Energy_new of the combined solution is used to update the optimal internal energy Energy_best.

[0203] Step S85: updating the combined solution to a new solution, and updating the current internal energy Energy_curren based on the new internal energy Energy_new of the combined solution, so as to process the next combined solution using the simulated annealing optimization algorithm;

[0204] In the process of specifically implementing step S85, the combined solution is directly accepted, and the current solution tobacco_current is updated to the new solution tobacco_new. That is, the combined solution is used to replace the combined solution in the new solution tobacco_new with the current solution tobacco_current to obtain the new solution tobacco_new; and its new internal energy Energy_new is used as the current internal energy Energy_current. That is, the current internal energy Energy_current is replaced with the new internal energy Energy_new of the combined solution.

[0205] Step S86: Determine whether all the combined solutions have been processed. If not, return to step S81. If completed, execute step S87.

[0206] In the specific implementation of step S86, all the combination solutions are traversed to determine whether the traversal is completed. If so, step S87 is executed. If not, the next combination solution is processed, that is, the execution returns to step S81.

[0207] Step S87: Determine whether the currently updated initial temperature is less than the preset temperature threshold. If so, stop the iteration and execute step S88. If not, return to continue executing step S81.

[0208] In the process of implementing step S87, first obtain the initial temperature after the last cooling. If the initial temperature after the last cooling does not exist, set the initial temperature to the preset value, and use the product of the initial temperature temperature and the cooling rate alpha as the new initial temperature, that is, the currently updated initial temperature; determine whether the currently updated initial temperature is less than the preset temperature threshold, that is, determine whether the currently updated initial temperature drops below the preset threshold. If so, stop the iteration and execute step S88. If not, return to continue executing step S81.

[0209] It should be noted that the specific implementation process of step S86 and step S87 can be carried out in the above order or simultaneously.

[0210] Optionally, if the number of iterations is greater than a preset number of iterations, step S88 is also executed.

[0211] Step S88: taking the combined solution in the new solution as the target solution.

[0212] Step S89: Determine whether the combined solution can be received based on the reception probability calculated based on the current internal energy Energy_curren and the new internal energy Energy_new. If so, return to step S85. If not, delete the combined solution and traverse the next combined solution, that is, return to step S81.

[0213] In the process of implementing step S89, the reception probability of the combined solution is calculated using the current internal energy Energy_curren and the new internal energy Energy_new; and it is determined whether the reception probability is greater than the preset probability. If it is greater than or equal to it, the process returns to step S85. If it is less than it, the combined solution is deleted and the next combined solution is traversed, that is, the process returns to step S81.

[0214] Among them, the preset probability is set in advance based on multiple experiments.

[0215] It should be noted that the process of calculating the reception probability of the combined solution using the current internal energy Energy_curren and the new internal energy Energy_new includes: inputting the current internal energy Energy_curren and the new internal energy Energy_new into formula (3) for calculation to obtain the reception probability P of the combined solution.

[0216] Formula (3):

[0217]

[0218] Where p represents the probability of accepting the solution, that is, the acceptance probability, E t Indicates the current internal energy Energy_curren, E t+1 represents the new internal energy Energy_new, g t are the model coefficients, which can be preset to control the number of iterations;

[0219] It should be noted that the target solution is to obtain the optimal auxiliary material parameter combination for different circumferences and cigarette lengths from the constraint space.

[0220] Optionally, after executing step S89, the method further includes:

[0221] Step S91: Count all target solutions;

[0222] In the specific implementation of step S91, the new solution tobacco_new as the target solution and the target solution of step S84 are recorded.

[0223] Step S92: For each target solution, determine whether the constraint state flag State of the target solution is -1. If so, execute step S93. If not, fill the value of the constraint state flag State into the constraint conflict information, output the constraint conflict information, and return to execute step S71 based on the constraint conflict information and the corresponding combined solution to re-determine the feasible solution.

[0224] In the process of implementing step S92, for each target solution, the preset constraints are used to further determine whether there are constraints that affect the target solution. If so, the constraint state flag State of the target solution is set to -1. If not, the value of the constraint state flag State is filled in the constraint conflict information, the constraint conflict information is output, and based on the constraint conflict information and the corresponding combined solution, the execution of step S71 is returned to re-determine the feasible solution.

[0225] Step S93: outputting auxiliary material parameters corresponding to the target solution.

[0226] In the specific implementation of step S93, it is determined again that there is no constraint conflict in the auxiliary material parameters in the target solution, and the auxiliary material parameters corresponding to the target solution, ie, the optimal auxiliary material parameters, are output.

[0227] Step S203: displaying the target auxiliary material parameters.

[0228] In an embodiment of the present invention, a feasible solution is determined from the constraint space, and then, the simulated annealing algorithm is used to balance the global search and the local search during the optimization process to determine the target solution from the combined solution constructed by the feasible solution; further, it is determined again whether the auxiliary material parameters in the target solution have no constraint conflicts, so as to increase the chance of finding the optimal auxiliary material parameter combination and adapt to various types of constraints.

[0229] Based on the above embodiment of the present invention, a device for determining parameters of auxiliary materials of cigarettes is shown, such as Figure 3 As shown, the device comprises:

[0230] An acquisition unit 301 is used to acquire a cigarette demand input by a user;

[0231] The processing unit 302 is used to determine the target auxiliary material parameters corresponding to the cigarette demand based on the pre-constructed constraint space;

[0232] The display unit 303 is used to display the target auxiliary material parameters.

[0233] The specific principles and execution processes of each unit in the device for determining the auxiliary material parameters of cigarettes disclosed in the above-mentioned embodiment of the present invention are the same as the corresponding contents in the method for determining the auxiliary material parameters of cigarettes provided in the above-mentioned embodiment of the present invention. Please refer to the corresponding parts in the method for determining the auxiliary material parameters of cigarettes disclosed in the above-mentioned embodiment of the present invention, and will not be repeated here.

[0234] In an embodiment of the present invention, a feasible solution is determined from the constraint space, and then, the simulated annealing algorithm is used to balance the global search and the local search during the optimization process to determine the target solution from the combined solution constructed by the feasible solution; further, it is determined again whether the auxiliary material parameters in the target solution have no constraint conflicts, so as to increase the chance of finding the optimal auxiliary material parameter combination and adapt to various types of constraints.

[0235] Optionally, based on the device for determining cigarette auxiliary material parameters shown in the above embodiment of the present invention, the processing unit 302 of the process of pre-constructing the constraint space is specifically used for:

[0236] Collecting historical auxiliary material parameter data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand, wherein the historical cigarette demand includes smoke index data indicated by each historical cigarette demand;

[0237] Preprocessing the historical auxiliary material parameter data to obtain auxiliary material parameter characteristic data;

[0238] Constructing a smoke index prediction model based on the auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand;

[0239] A constraint space is constructed based on the smoke index prediction model.

[0240] The historical auxiliary material parameter data is preprocessed to obtain auxiliary material parameter characteristic data, including:

[0241] Determine whether there is a non-numeric feature in the historical auxiliary material parameter data;

[0242] If it exists, convert the non-numeric feature into corresponding coded data;

[0243] Replacing the corresponding non-numerical features in the historical auxiliary material parameter data based on the coded data to update the historical auxiliary material parameter data;

[0244] Auxiliary material parameter characteristic data is obtained based on the updated historical auxiliary material parameter data.

[0245] The auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand are used to construct a smoke index prediction model, including:

[0246] The auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand are used as a sample set;

[0247] Dividing the sample set into a training set and a test set;

[0248] Using the training set to train the initial model to obtain a trained initial model;

[0249] The initial model is tested using the test set, and training is stopped when it is determined that the loss function value of the initial model is less than or equal to a preset target loss value, and the initial model is used as a smoke index prediction model.

[0250] Optionally, based on the device for determining cigarette auxiliary material parameters shown in the above embodiment of the present invention, a processing unit 302 for constructing a constraint space based on the smoke index prediction model is specifically used for:

[0251] Collect M groups of auxiliary material parameter data;

[0252] For each set of auxiliary material parameter data, the auxiliary material parameter data is processed using the smoke index prediction model, and a smoke index prediction result is output;

[0253] For each set of auxiliary material parameter data, determining whether the smoke index prediction result meets the cigarette demand corresponding to the auxiliary material parameter data;

[0254] If it meets the requirements, the auxiliary material parameter data is recorded as valid auxiliary material parameter data.

[0255] A constraint space is constructed based on the effective auxiliary material parameter data and preset constraints.

[0256] Optionally, based on the device for determining parameters of auxiliary materials of cigarettes shown in the above embodiment of the present invention, the processing unit 302 is specifically configured to:

[0257] Based on the constraint space, determining a feasible solution for each feature;

[0258] Constructing a plurality of combined solutions based on the feasible solutions for each of the features;

[0259] Determine the target solution in the combined solution using the simulated annealing optimization algorithm;

[0260] The target solution is used as the target auxiliary material parameter corresponding to the cigarette demand.

[0261] Among them, the target solution in the combined solution is determined by using the simulated annealing optimization algorithm, including:

[0262] For each combined solution, the combined solution is processed using a simulated annealing optimization algorithm to determine a new internal energy corresponding to the combined solution;

[0263] If the new internal energy is less than the previous internal energy, determining whether the combined solution is a preset solution;

[0264] If the combined solution is a preset solution, taking the combined solution as the target solution;

[0265] If the combined solution is not a preset solution, updating the combined solution to a new solution, and updating the current internal energy based on the new internal energy of the combined solution;

[0266] If the new internal energy is greater than or equal to the current internal energy, determining that the combined solution can be received based on the reception probability calculated based on the current internal energy and the new internal energy, updating the combined solution to a new solution, and updating the current internal energy based on the new internal energy of the combined solution;

[0267] If it is determined that all the combined solutions have not been processed, and the currently updated initial temperature is greater than the preset temperature threshold, the next combined solution is processed using a simulated annealing optimization algorithm;

[0268] If it is determined that all of the combined solutions have been processed and the currently updated initial temperature is less than a preset temperature threshold, the combined solution in the new solution is used as the target solution.

[0269] Optionally, based on the device for determining cigarette auxiliary material parameters shown in the above embodiment of the present invention, the processing unit 302 is further configured to:

[0270] Count all target solutions;

[0271] For each target solution, determine whether the constraint state flag of the target solution is -1;

[0272] If so, output the auxiliary material parameters corresponding to the target solution.

[0273] An embodiment of the present application provides an electronic device, which includes a processor and a memory, wherein the memory is used for determining program codes and data for auxiliary material parameters of cigarettes, and the processor is used for calling program instructions in the memory to execute the steps shown in the method for determining auxiliary material parameters of cigarettes in the above embodiment.

[0274] An embodiment of the present application provides a storage medium, which includes a storage program, wherein when the program is running, the device where the storage medium is located is controlled to execute the method for determining the auxiliary material parameters of cigarettes shown in the above embodiment.

[0275] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0276] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0277] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for determining parameters of auxiliary materials of cigarettes, characterized in that: The method comprises: Obtain cigarette demand input by the user; Determining target auxiliary material parameters corresponding to the cigarette demand based on a pre-constructed constraint space; The target auxiliary material parameters are displayed.

2. The method according to claim 1, characterized in that The process of pre-building constraint spaces includes: Collecting historical auxiliary material parameter data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand, wherein the historical cigarette demand includes smoke index data indicated by each historical cigarette demand; Preprocessing the historical auxiliary material parameter data to obtain auxiliary material parameter characteristic data; Constructing a smoke index prediction model based on the auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand; A constraint space is constructed based on the smoke index prediction model.

3. The method according to claim 2, characterized in that Preprocessing the historical auxiliary material parameter data to obtain auxiliary material parameter characteristic data includes: Determine whether there is a non-numeric feature in the historical auxiliary material parameter data; If it exists, convert the non-numeric feature into corresponding coded data; Replacing the corresponding non-numerical features in the historical auxiliary material parameter data based on the coded data to update the historical auxiliary material parameter data; Auxiliary material parameter characteristic data is obtained based on the updated historical auxiliary material parameter data.

4. The method according to claim 2, characterized in that: The auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand are used to construct a smoke index prediction model, including: The auxiliary material parameter characteristic data, finished tobacco data, cigarette physical parameters and corresponding historical cigarette demand are used as a sample set; Dividing the sample set into a training set and a test set; Using the training set to train the initial model to obtain a trained initial model; The initial model is tested using the test set, and training is stopped when it is determined that the loss function value of the initial model is less than or equal to a preset target loss value, and the initial model is used as a smoke index prediction model.

5. The method according to claim 4, characterized in that Constructing a constraint space based on the flue gas index prediction model includes: Collect M groups of auxiliary material parameter data; For each set of auxiliary material parameter data, the auxiliary material parameter data is processed using the smoke index prediction model, and a smoke index prediction result is output; For each set of auxiliary material parameter data, determining whether the smoke index prediction result meets the cigarette demand corresponding to the auxiliary material parameter data; If it is in compliance, the auxiliary material parameter data is recorded as valid auxiliary material parameter data; A constraint space is constructed based on the effective auxiliary material parameter data and preset constraints.

6. The method according to claim 5, characterized in that Determining target auxiliary material parameters corresponding to the cigarette demand based on the constraint space includes: Based on the constraint space, determining a feasible solution for each feature; Constructing a plurality of combined solutions based on the feasible solutions for each of the features; Determine the target solution in the combined solution using the simulated annealing optimization algorithm; The target solution is used as the target auxiliary material parameter corresponding to the cigarette demand.

7. The method according to claim 6, characterized in that The simulated annealing optimization algorithm is used to determine the target solution in the combined solution, including: For each combined solution, the combined solution is processed using a simulated annealing optimization algorithm to determine a new internal energy corresponding to the combined solution; If the new internal energy is less than the previous internal energy, determining whether the combined solution is a preset solution; If the combined solution is a preset solution, taking the combined solution as the target solution; If the combined solution is not a preset solution, updating the combined solution to a new solution, and updating the current internal energy based on the new internal energy of the combined solution; If the new internal energy is greater than or equal to the current internal energy, determining that the combined solution can be received based on the reception probability calculated based on the current internal energy and the new internal energy, updating the combined solution to a new solution, and updating the current internal energy based on the new internal energy of the combined solution; If it is determined that all the combined solutions have not been processed, and the currently updated initial temperature is greater than the preset temperature threshold, the next combined solution is processed using a simulated annealing optimization algorithm; If it is determined that all of the combined solutions have been processed and the currently updated initial temperature is less than a preset temperature threshold, the combined solution in the new solution is used as the target solution.

8. The method according to claim 6, characterized in that Also includes: Count all target solutions; For each target solution, determine whether the constraint state flag of the target solution is -1; If yes, output the auxiliary material parameters corresponding to the target solution.

9. A device for determining parameters of auxiliary materials of cigarettes, characterized in that: The device comprises: An acquisition unit, used for acquiring cigarette demand input by a user; A processing unit, used for determining target auxiliary material parameters corresponding to the cigarette demand based on a pre-constructed constraint space; The display unit is used to display the target auxiliary material parameters.

10. A storage medium, characterized in that: The storage medium includes a storage program, wherein when the program is running, the device where the storage medium is located is controlled to execute the method for determining the auxiliary material parameters of a cigarette according to any one of claims 1 to 8.

Images