A method and system for optimizing feed formulation based on pet differentiation analysis

Abstract

This invention provides a method and system for optimizing pet feed formulations based on pet differentiation analysis, relating to the field of pet feed optimization technology. It matches standard pet feed formulations based on basic pet information, adjusts them according to environmental and physiological characteristics, and generates adjusted pet feed formulation results. It constructs formulation optimization value constraints based on feed component types and proportion ranges, and builds a formulation optimization vector space and a feed optimization fitness function. Using these as a benchmark, it optimizes the initial pet feed formulation, generating optimized pet feed formulation results. Based on these results, pet feed is prepared. This invention solves the technical problems of insufficient intelligence and limited applicability in existing pet feed formulation optimization methods, as well as insufficient rigor in the analysis and processing of related data, leading to inaccurate final formulation optimization results. It achieves high-quality, adaptable optimization of pet feed formulations.

Description

Technical Field

[0001] This invention relates to the field of pet food optimization technology, specifically to a method and system for optimizing feed formulations based on pet differentiation analysis. Background Technology

[0002] Pet food is an essential part of a pet's survival. It contains a large amount of protein, amino acids, and various trace elements, providing pets with the most basic life support and the nutrients needed for healthy growth and development. However, the formulas of commonly used pet foods are quite general and suitable for many different breeds, resulting in some differences in the adaptability between pet food and pets. To improve the compatibility between pet food and pets, pet food can be optimized to a certain extent, making it more targeted and better suited to pets. Currently, the cross-fertilization method is often used to optimize pet food formulas, but this method has certain limitations.

[0003] In existing technologies, commonly used pet formula optimization methods suffer from insufficient intelligence in the optimization process and limited applicability. Furthermore, the analysis and processing of related data are not rigorous enough, resulting in insufficient accuracy of the final formula optimization results. Summary of the Invention

[0004] This application provides a method and system for optimizing feed formulation based on pet differentiation analysis, which addresses the technical problem that existing pet formulation optimization methods suffer from insufficient intelligence in the optimization process, limited applicability, and insufficient rigor in the analysis and processing of related data, resulting in insufficient accuracy of the final formulation optimization results.

[0005] In view of the above problems, this application provides a method and system for optimizing feed formulation based on pet differentiation analysis.

[0006] Firstly, this application provides a method for optimizing pet feed formulation based on pet differentiation analysis. The method includes: inputting basic pet information into a pet feed standard table and matching it with a pet feed standard formulation; adjusting the pet feed standard formulation according to the characteristics of the breeding environment and the physiological characteristics of the pet to generate a pet feed formulation adjustment result, wherein the pet feed formulation adjustment result includes feed component types and feed component ratio ranges; constructing formulation optimization value constraints based on the feed component types and feed component ratio ranges; traversing the feed component types and the formulation optimization value constraints to construct a formulation optimization vector space, wherein the dimension of the formulation optimization vector space is the same as the dimension of the feed component types; constructing a feed optimization fitness function; optimizing the initial pet feed formulation according to the feed optimization fitness function and the formulation optimization vector space to generate a pet feed formulation optimization result; and preparing pet feed according to the pet feed formulation optimization result.

[0007] Secondly, this application provides a feed formulation optimization system based on pet differentiation analysis. The system includes: a formulation matching module, used to input basic pet information into a pet feed standard table and match pet feed standard formulations; a formulation adjustment module, used to adjust the pet feed standard formulations according to the characteristics of the feeding environment and the physiological characteristics of the pets, generating pet feed formulation adjustment results, wherein the pet feed formulation adjustment results include feed component types and feed component ratio ranges; and a constraint condition construction module, used to construct a formulation optimization system based on the feed component types and the feed component ratio ranges. The system comprises the following modules: a value constraint condition; a space construction module, which iterates through the feed ingredient types and the formula optimization value constraint conditions to construct a formula optimization vector space, wherein the dimension of the formula optimization vector space is the same as the dimension of the feed ingredient types; a function construction module, which constructs a feed optimization fitness function; a formula optimization module, which optimizes the initial pet food formula based on the feed optimization fitness function and the formula optimization vector space to generate a pet food formula optimization result; and a feed preparation module, which prepares pet food based on the pet food formula optimization result.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] This application provides a method for optimizing pet feed formulation based on pet differentiation analysis. The method matches a standard pet feed formulation with basic pet information, adjusts the standard formulation according to environmental and physiological characteristics, and generates an adjusted pet feed formulation result. This result includes feed component types and proportion ranges. Optimization constraints are constructed based on these constraints. An optimization vector space is built by iterating through these constraints, with the dimension of the vector space matching that of the feed component types. A fitness function is then constructed and combined with the optimization vector space to optimize the initial pet feed formulation, generating an optimized pet feed formulation result. Pet feed is then prepared based on this optimized result. This method addresses the technical problems of existing pet feed optimization methods, such as insufficient intelligence in the optimization process, limited applicability, and inadequate analysis of related data, leading to insufficient accuracy in the final optimization result. This method achieves high-quality, adaptable optimization of pet feed formulations. Attached Figure Description

[0010] Figure 1 This application provides a schematic flowchart of a feed formulation optimization method based on pet differentiation analysis;

[0011] Figure 2 This application provides a schematic diagram illustrating the process of generating pet food formula adjustment results in a feed formula optimization method based on pet differentiation analysis;

[0012] Figure 3 This application provides a schematic diagram of the pet food formulation optimization result generation process in a pet differentiation analysis-based feed formulation optimization method.

[0013] Figure 4 This application provides a schematic diagram of a feed formulation optimization system based on pet differentiation analysis.

[0014] Explanation of reference numerals in the attached diagram: Formula matching module a, Formula adjustment module b, Constraint construction module c, Space construction module d, Function construction module e, Formula optimization module f, Feed preparation module g. Detailed Implementation

[0015] This application provides a method and system for optimizing pet feed formulations based on pet differentiation analysis. It matches standard pet feed formulations with basic pet information, adjusts them according to environmental and physiological characteristics, and generates adjusted pet feed formulation results. It constructs optimization value constraints based on feed component types and proportion ranges, and builds an optimization vector space and a fitness function. Using these as a benchmark, it optimizes the initial pet feed formulation, generating optimized pet feed formulation results. Based on these results, pet feed is prepared. This addresses the technical problem in existing pet feed formulation optimization methods where insufficient intelligence in the optimization process, limited applicability, and inadequate analysis of related data lead to insufficient accuracy in the final optimization results.

[0016] Example 1

[0017] like Figure 1 As shown, this application provides a method for optimizing feed formulation based on pet differentiation analysis, the method comprising:

[0018] Step S100: Input the pet's basic information into the pet food standard table and match the pet food standard formula;

[0019] Specifically, this application provides a feed formulation optimization method based on pet differentiation analysis. This method adapts industry-standard pet feed formulations to pets based on their environmental and physiological characteristics. The optimization generates a feed formulation that matches various pet indicators. Basic pet information is obtained and input into the industry-standard pet feed formulation table. The table matches the pet's basic information with the table to obtain a suitable pet feed formulation. Different types of pets require different nutritional components and intake amounts. Based on the pet's basic information, a corresponding pet feed formulation is determined. This pet feed formulation serves as the initial formulation for subsequent optimization and adjustment.

[0020] Step S200: Adjust the standard formula of the pet food according to the characteristics of the breeding environment and the physiological characteristics of the pet to generate the pet food formula adjustment result, wherein the pet food formula adjustment result includes the type of feed ingredients and the range of feed ingredient proportions;

[0021] Specifically, based on the pet's basic information, the characteristics of the pet's living environment and physiological characteristics are determined. For example, physiological indices such as the pet's heart rate, respiratory rate, blood pressure, and pH are used as the pet's physiological characteristics, and the temperature and humidity of the pet's living environment are used as the living environment temperature. Different types and breeds of pets have different living environments and physiological characteristics. Based on this, the feed component types and feed component ratio ranges of the standard pet food formula are adjusted accordingly. By prioritizing each characteristic, the standard pet food formula is adaptively adjusted to obtain multiple initial formula adjustment results. Furthermore, the intersection of multiple initial formula adjustment results with multiple groups of feed component types and multiple groups of initial pet food component ratio ranges is obtained. The intersection result generates the feed component types and feed component ratio ranges to determine the pet food formula adjustment result. Obtaining the pet food formula adjustment result lays a solid foundation for further feed optimization.

[0022] Step S300: Based on the feed component type and the feed component ratio range, construct the formula optimization value constraint conditions;

[0023] Step S400: Traverse the feed component types and the formula optimization value constraints to construct a formula optimization vector space, wherein the dimension of the formula optimization vector space is the same as the dimension of the feed component types;

[0024] Specifically, the pet food formula adjustment results are obtained by adjusting the standard pet food formula. Then, the feed ingredient types and feed ingredient ratio ranges are identified and extracted. Using the feed ingredient ratio ranges as limiting standards, the formula optimization value constraints are constructed for multiple corresponding feed ingredient types. For example, if the protein component ratio range of the feed ingredient type is 3-6%, this range can be used as the protein optimization value constraint. The feed ingredient types and optimization value constraints are then matched one-to-one, and then integrated to generate the formula optimization value constraints.

[0025] Furthermore, the formula optimization vector space is constructed by traversing the feed ingredient types and the formula optimization value constraints. The formula optimization vector space is the space for optimizing the initial pet food formula. The formula optimization vector space is divided equally to obtain multiple spatial dimensions that are the same as the feed ingredient type dimension. The feed ingredient type corresponds one-to-one with the multiple spatial dimensions. The corresponding feed ingredient type is modified and adjusted in the spatial dimensions. The construction of the formula optimization value constraints and the formula optimization vector space provides a theoretical basis and optimization space for subsequent optimization and adjustment of the initial pet food formula.

[0026] Step S500: Construct the feed optimization fitness function;

[0027] Step S600: Optimize the initial pet food formula according to the feed optimization fitness function and the formula optimization vector space to generate the pet food formula optimization result;

[0028] Step S700: Prepare pet food according to the pet food formula optimization results.

[0029] Specifically, the cost screening degree of feed is used as the first fitness, and the selection screening degree is used as the second fitness. The feed optimization fitness function is constructed based on the first fitness and the second fitness. The first fitness is inversely proportional to the feed optimization fitness function, and the second fitness is directly proportional to the feed optimization fitness function. The fitness of various pet food formulas constructed based on the feed optimization fitness function is calculated, providing reference information for subsequent judgment of the merits of various pet food formulas.

[0030] Furthermore, a set of multi-dimensional data is randomly extracted from the formula optimization vector space and combined to form the t-th feed formula. The fitness of the feed formula is then calculated based on the feed optimization fitness function to obtain the fitness calculation result. When the fitness calculation result is satisfactory, data extraction and combination are performed again to determine the (t-1)-th feed formula. Fitness is calculated again based on the feed optimization fitness function. Furthermore, the fitness of the t-th feed formula and the (t-1)-th feed formula is compared, and the feed formula with the larger fitness is determined as the current optimal formula. The feed formula extraction and combination are performed again, and the above operations are repeated to iteratively update the feed formula until a preset number of iterations is met. The current optimal feed formula is taken as the pet food formula optimization result. Furthermore, based on the pet food formula optimization result, pet food is prepared according to the feed component type and component ratio to improve the matching degree between pet food and pets.

[0031] Furthermore, such as Figure 2 As shown, the step S200 of this application further includes adjusting the standard pet food formula based on the characteristics of the breeding environment and the physiological characteristics of the pet to generate a pet food formula adjustment result:

[0032] Step S210: Iterate through the pet's physiological characteristics and the breeding environment characteristics, prioritize them, and generate a feature ranking result;

[0033] Step S220: Iterate through the pet's physiological characteristics and the breeding environment characteristics to adjust the standard pet food formula and generate multiple sets of initial adjustment results for the pet food formula;

[0034] Step S230: Based on the initial adjustment results of the multiple sets of pet food formulas, obtain the multiple initial pet food ingredient types and the multiple initial pet food ingredient ratio ranges;

[0035] Step S240: Based on the feature sorting results, perform intersection calculation on the multiple initial pet food ingredient types to generate the feed ingredient type;

[0036] Step S250: Based on the feed ingredient type and the feature sorting result, the intersection of the multiple initial pet food ingredient ratio ranges is calculated to generate the feed ingredient ratio range.

[0037] Specifically, the process involves iterating through the pet's physiological characteristics and the feeding environment characteristics to determine their respective weights in influencing the pet. Based on this weighting, the characteristics are prioritized and ranked to generate a ranking result. Further iterations of the pet's physiological characteristics and the feeding environment characteristics are then performed to make targeted adjustments to the standard pet food formula, generating multiple sets of initial adjustment results for pet food formulas. Each set of pet's physiological characteristics and the feeding environment characteristics corresponds one-to-one with the multiple sets of initial adjustment results for pet food formulas.

[0038] Furthermore, based on the initial adjustment results of the multiple sets of pet food formulas, the pet food ingredient types and ingredient ratio ranges corresponding to each adjustment result are collected to obtain the multiple sets of initial pet food ingredient types and the multiple sets of initial pet food ingredient ratio ranges. Based on the feature ranking results, the intersection of the multiple sets of initial pet food ingredient types is obtained. In the feature ranking results, based on the order of arrangement, the priority of the multiple sets of initial pet food ingredient types corresponding to each feature decreases sequentially. During the intersection finding process, if there is no intersection between the later and earlier ranked types, the earlier ranked type is taken as the standard. The intersection finding results are integrated to generate the feed ingredient types. Similarly, based on the feed ingredient types and the feature ranking results, the intersection of the multiple sets of initial pet food ingredient ratio ranges is obtained to determine the intersection finding result. The intersection finding result represents the optimal comprehensive adjustment based on the specific characteristics of the pet, generating the feed ingredient ratio range. The acquisition of the feed ingredient types and the feed ingredient ratio ranges provides basic information for subsequent formula adjustment and optimization.

[0039] Furthermore, the step S210 of this application, which involves prioritizing the pet's physiological characteristics and the characteristics of the living environment to generate a feature ranking result, further includes:

[0040] Step S211: When the preset update cycle is met, upload the feature parameter sorting data through the first database, the second database, and up to the Nth database;

[0041] Step S212: Update the priority sorting table according to the sorting data based on the feature parameters;

[0042] Step S213: Input the pet's physiological characteristics and the breeding environment characteristics into the updated priority sorting table to generate the feature sorting result.

[0043] Specifically, a preset update cycle is established. For example, the pet's growth stage can be used as the preset update cycle for adaptive adjustment. When the preset update cycle is met, the feature parameter sorting data is uploaded based on the first database, the second database, and up to the Nth database. These data come from various participating parties, such as pet hospitals, feed research institutes, and feed production factories. When the uploaded sorting data conflict, the sorting data with the highest overlap is taken for any feature. If the overlap is the same, the two features can be in the same sort, i.e., have the same priority. The feature parameter sorting data is adapted to the pet's current state. The priority sorting table is updated based on the feature parameter sorting data. Furthermore, the pet's physiological characteristics and the breeding environment characteristics are input into the updated priority sorting table. The feature sorting result is generated by adjusting the feature sorting. The feature sorting result is obtained by retrieving the feature parameter sorting data and performing feature priority sorting, which can effectively ensure the fit between the feature sorting result and the actual situation.

[0044] Furthermore, the step S220 of this application, which involves adjusting the standard pet food formula based on the pet's physiological characteristics and the characteristics of its living environment to generate multiple sets of initial adjustment results for the pet food formula, further includes:

[0045] Step S221: Obtain the pet physiological characteristic record dataset, the breeding environment characteristic record dataset, and the pet food formula adjustment identifier dataset to generate the model building dataset;

[0046] Step S222: Construct a dataset based on the model and train the Mth decision tree;

[0047] Step S223: Obtain the amount of data in the model construction dataset where the Mth decision tree does not meet the preset accuracy, and determine whether the preset data amount is met;

[0048] Step S224: If satisfied, merge the first decision tree, the second decision tree, and so on up to the Mth decision tree to generate a pet food formula adjustment model.

[0049] Specifically, the process involves collecting the pet physiological characteristic record dataset, the breeding environment characteristic record dataset, and the pet food formula adjustment identifier dataset. Based on these datasets, a model construction dataset is generated. Multiple corresponding datasets are extracted from the model construction dataset for decision tree learning to obtain multiple decision trees. These decision trees correspond to datasets of various pets. Analysis and prediction are performed based on the pet physiological characteristic record dataset and the corresponding breeding environment characteristic record dataset to generate corresponding adjustment data scales. Furthermore, the adjustment data scales are compared with the pet food formula adjustment identifier dataset, and the deviation is calculated. The deviation calculation result is then used to further determine the Mth decision. Whether the deviation calculation result of the tree meets the preset accuracy rate, the preset accuracy rate refers to the scale range for measuring the deviation calculation result. If it does not meet the preset accuracy rate, the amount of data in the model construction dataset of the Mth decision tree that does not meet the preset accuracy rate is obtained, and it is further determined whether the preset data amount is met. The preset data amount is a threshold range for limiting the amount of unqualified data. When the preset data amount is met, it means that the influence scale of the unqualified data is within an acceptable limit. The first decision tree, the second decision tree and up to the Mth decision tree are merged to generate a pet food formula adjustment model. The pet food formula adjustment model is used as an auxiliary tool to adaptively adjust the pet food formula.

[0050] Furthermore, the determination of whether the preset data volume is met, step S223 of this application also includes:

[0051] Step S2231: If the preset accuracy is not met, construct a dataset based on the model that does not meet the preset accuracy, and construct the (M+1)th decision tree;

[0052] Step S2232: Determine the amount of data in the model construction dataset for which the (M+1)th decision tree does not meet the preset accuracy, and determine whether the preset data amount is met.

[0053] Specifically, the data volume of the model construction dataset for which the Mth decision tree does not meet the preset accuracy is obtained, and it is determined whether the preset data volume is met. If it is not met, it indicates that the data volume is too large and has too great an impact on subsequent formula optimization. Based on the model construction dataset for which the preset accuracy is not met, the M+1th decision tree is constructed. The preset accuracy is further compared and determined on the M+1th decision tree based on the above analysis steps. The data volume of the model construction dataset for which the preset accuracy is not met is obtained, and the preset data volume is determined again. By performing secondary processing and analysis on the data for which the preset data volume is not met, the accuracy of adjusting the pet food formula is improved.

[0054] Furthermore, the construction of the feed optimization fitness function, step S500 of this application further includes:

[0055] Step S510: Obtain the first fitness function:

[0056]

[0057] Among them, C t Let c be the cost screening degree of the t-th feed group. k Cost estimation for the k-th type of feed. Let K be the proportion of the kth type of feed, and K be the total number of feed types in the tth group.

[0058] Step S520: Obtain the second fitness function:

[0059] F t =v t α *f t β

[0060] Among them, F t To select the screening degree, v t f represents the rate of weight gain for pets fed the t-th diet. t Let be the frequency of pet illness in the t-th group of feed, and α and β be weighting indices;

[0061] Step S530: Construct the feed optimization fitness function from the first fitness function and the second fitness function:

[0062]

[0063] Among them, A t γ and δ are weighting indices for optimizing the fitness of the t-th group of feed.

[0064] Specifically, the feed optimization fitness function is constructed to perform the optimization standard calculation of the initial pet food formula, based on the first fitness calculation formula. Calculate the cost screening degree of feed, where C t Let c be the cost screening degree of the t-th feed group. k Cost estimation for the k-th type of feed. Let K be the proportion of the k-th feed type, and K be the total number of feed types in the t-th group. The cost screening degree is determined by superimposing the cost estimates of multiple feeds and the feed proportions. The cost screening degree is inversely proportional to the feed optimization fitness, based on the second fitness calculation formula F. t =v t α *f t β Calculate the selection screening degree, where F t To select the screening degree, v t f represents the rate of weight gain for pets fed the t-th diet. t Let be the frequency of pet illness in the t-th feed group, and α and β be weighting indices that can be set according to the actual scenario. The default values ​​are α = 1 and β = 2. The feed optimization fitness function is constructed based on the first fitness function and the second fitness function, using the formula... Calculate feed optimization fitness, where A t The optimal fitness of the t-th feed group is defined by γ and γ', which are weight indices that can be set according to the actual scenario. The default values ​​are γ = 1 and γ' = 1. The optimization analysis of the pet food formula is performed based on the feed optimization fitness function.

[0065] Furthermore, such as Figure 3 As shown, the step S600 of this application further includes optimizing the initial pet food formula based on the feed optimization fitness function and the formula optimization vector space to generate the pet food formula optimization result;

[0066] Step S610: Randomly select values ​​from the formula optimization vector space to obtain the t-th feed formula;

[0067] Step S620: Calculate the t-th feed formula according to the feed optimization fitness function to generate the t-th fitness;

[0068] Step S630: When the fitness of the tth degree satisfies the preset fitness, obtain the fitness of the (t-1)th degree;

[0069] Step S640: Determine whether the (t-1)th fitness is less than or equal to the tth fitness;

[0070] Step S650: If it is less than or equal to, add the (t-1)th feed formula to the elimination data group, and continue iterating based on the tth feed formula;

[0071] Step S660: If it is greater than 1, add the t-th feed formula to the elimination data group and continue iterating based on the (t-1)-th feed formula;

[0072] Step S670: When the preset number of iterations is met, the initial pet food formula is replaced according to the feed formula being iterated, and the optimized pet food formula is generated.

[0073] Specifically, values ​​are randomly selected from multiple dimensions of the formulation optimization vector space, integrated, and processed to generate the t-th feed formulation. The fitness of the t-th feed formulation is calculated based on the feed optimization fitness function to generate the t-th fitness. It is then determined whether the t-th fitness meets a preset fitness threshold, which is a fitness threshold used to determine whether the combined feed formulation is qualified. When the t-th fitness meets the preset fitness, a set of data is randomly extracted from the formulation optimization vector space to determine the (t-1)-th feed formulation. The (t-1)-th fitness is calculated based on the feed optimization fitness function, and a comparison is further performed between the t-th fitness and the (t-1)-th fitness. When the (t-1)-th fitness is less than or equal to the t-th fitness, it indicates that the (t-1)-th fitness is qualified. If the fitness of a given feed formula is unqualified, the (t-1)th feed formula is added to the elimination data group, and the t-th feed formula is used as the current optimal formula for continued iteration. When the fitness of the (t-1)th feed formula is greater than that of the t-th feed formula, it indicates that the feed formula used for the t-th feed formula is unqualified, and the t-th feed formula is added to the elimination data group. The (t-1)th feed formula is used as the current optimal formula for continued iteration. A preset number of iterations is set, which is the maximum number of times the fitness of the feed formula is compared and updated. When the preset number of iterations is met, the feed formula currently being iterated is used as the optimal formula to replace the initial pet food formula. The optimal formula is used as the pet food formula optimization result. By iteratively updating multiple pet food formulas, the final optimized pet food formula becomes more adaptable.

[0074] Example 2

[0075] Based on the same inventive concept as the feed formulation optimization method based on pet differentiation analysis in the foregoing embodiments, such as Figure 4 As shown, this application provides a feed formulation optimization system based on pet differentiation analysis, the system comprising:

[0076] Formula matching module a, which is used to input basic pet information into the pet food standard table and match the pet food standard formula;

[0077] Formula adjustment module b, which is used to adjust the standard formula of pet food according to the characteristics of the breeding environment and the physiological characteristics of pets, and generate pet food formula adjustment results, wherein the pet food formula adjustment results include feed component types and feed component ratio ranges;

[0078] Constraint construction module c, which is used to construct formula optimization value constraints based on the feed component type and the feed component ratio range;

[0079] Space construction module d, which is used to traverse the feed component type and the formula optimization value constraint to construct the formula optimization vector space, wherein the dimension of the formula optimization vector space is the same as the dimension of the feed component type;

[0080] Function construction module e, which is used to construct the feed optimization fitness function;

[0081] Formula optimization module f, which is used to optimize the initial pet food formula according to the feed optimization fitness function and the formula optimization vector space, and generate pet food formula optimization results;

[0082] Feed preparation module g, which is used to prepare pet food based on the pet food formula optimization results.

[0083] Furthermore, the system also includes:

[0084] The feature sorting module is used to traverse the pet's physiological characteristics and the breeding environment characteristics to sort them by priority and generate feature sorting results.

[0085] The initial formula adjustment module is used to iterate through the pet's physiological characteristics and the characteristics of the breeding environment to adjust the standard formula of the pet food and generate multiple sets of initial adjustment results for the pet food formula;

[0086] The component information acquisition module is used to acquire the component types and component ratio ranges of multiple initial pet food based on the initial adjustment results of the multiple sets of pet food formulas.

[0087] The ingredient type generation module is used to obtain the feed ingredient type by intersecting the multiple initial pet food ingredient types according to the feature sorting result.

[0088] The component ratio range generation module is used to obtain the feed component ratio range by intersecting the multiple initial pet food component ratio ranges according to the feed component type and the feature sorting result.

[0089] Furthermore, the system also includes:

[0090] The data upload module is used to upload feature parameter sorting data through the first database, the second database, up to the Nth database when the preset update cycle is met.

[0091] A sorting table update module is used to update the priority sorting table according to the sorting data based on the feature parameters.

[0092] The sorting result generation module is used to input the pet's physiological characteristics and the breeding environment characteristics into the updated priority sorting table to generate the feature sorting result.

[0093] Furthermore, the system also includes:

[0094] The dataset generation module is used to acquire a pet physiological characteristic record dataset, a breeding environment characteristic record dataset, and a pet food formula adjustment identifier dataset to generate a model building dataset.

[0095] A decision tree training module, which is used to construct a dataset based on the model and train the Mth decision tree;

[0096] A preset data volume judgment module is used to obtain the data volume of the model construction dataset for which the Mth decision tree does not meet the preset accuracy, and to determine whether the preset data volume is met.

[0097] The adjustment model generation module is used to merge the first decision tree, the second decision tree, and up to the Mth decision tree if the conditions are met, to generate a pet food formula adjustment model.

[0098] Furthermore, the system also includes:

[0099] A decision tree construction module is used to construct the (M+1)th decision tree based on the model that does not meet the preset accuracy if the conditions are not met.

[0100] The data volume judgment module is used to determine the data volume of the model construction dataset for which the (M+1)th decision tree does not meet the preset accuracy, and to determine whether the preset data volume is met.

[0101] Furthermore, the system also includes:

[0102] The first fitness function acquisition module is used to acquire the first fitness function:

[0103]

[0104] Among them, C t Let c be the cost screening degree of the t-th feed group. k Cost estimation for the k-th type of feed. Let K be the proportion of the kth type of feed, and K be the total number of feed types in the tth group.

[0105] The second fitness function acquisition module is used to acquire the second fitness function.

[0106] F t =v t α *f t β

[0107] Among them, F t To select the screening degree, v t f represents the rate of weight gain for pets fed the t-th diet. t Let be the frequency of pet illness in the t-th group of feed, and α and β be weighting indices;

[0108] A feed optimization fitness function construction module is used to construct the feed optimization fitness function from the first fitness function and the second fitness function.

[0109]

[0110] Among them, A t γ and δ are weighting indices for optimizing the fitness of the t-th group of feed.

[0111] Furthermore, the system also includes;

[0112] A feed formulation acquisition module is used to randomly select values ​​from the formulation optimization vector space to obtain the t-th feed formulation.

[0113] A fitness calculation module is used to calculate the fitness of the t-th feed formula based on the feed optimization fitness function, and generate the fitness of the t-th feed formula.

[0114] A fitness acquisition module is used to acquire the (t-1)th fitness when the t-th fitness satisfies a preset fitness.

[0115] A fitness determination module is used to determine whether the (t-1)th fitness is less than or equal to the tth fitness;

[0116] The formula iteration module is used to add the (t-1)th feed formula to the elimination data group if it is less than or equal to the tth feed formula, and continue iterating based on the tth feed formula.

[0117] A feed formulation iteration module is used to add the t-th feed formulation to the elimination data group if the value is greater than the t-1-th feed formulation, and continue iterating based on the t-1-th feed formulation.

[0118] The formula optimization result generation module is used to replace the initial pet food formula according to the feed formula being iterated when a preset number of iterations is met, and generate the pet food formula optimization result.

[0119] Through the foregoing detailed description of a feed formulation optimization method based on pet differentiation analysis, those skilled in the art can clearly understand the feed formulation optimization method and system based on pet differentiation analysis in this embodiment. As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the method section description.

[0120] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing feed formulation based on pet differentiation analysis, characterized in that, The method includes: Enter the pet's basic information into the pet food standard table and match the pet food standard formula; The standard pet food formula is adjusted according to the characteristics of the breeding environment and the physiological characteristics of the pets to generate pet food formula adjustment results, wherein the pet food formula adjustment results include feed component types and feed component ratio ranges; Based on the feed component type and the feed component ratio range, construct the formula optimization value constraint conditions; Traverse the feed component types and the formula optimization value constraints to construct a formula optimization vector space, wherein the dimension of the formula optimization vector space is the same as the dimension of the feed component types; Construct a feed optimization fitness function; The initial pet food formula is optimized based on the feed optimization fitness function and the formula optimization vector space to generate the pet food formula optimization result; Pet food was prepared based on the optimized pet food formula results. The process of adjusting the standard pet food formula based on the characteristics of the breeding environment and the physiological characteristics of the pet to generate a pet food formula adjustment result includes: The pet's physiological characteristics and the living environment characteristics are traversed and prioritized to generate a feature ranking result; The standard pet food formula is adjusted by iterating through the pet's physiological characteristics and the characteristics of the breeding environment, generating multiple sets of initial adjustment results for the pet food formula; Based on the initial adjustment results of the multiple sets of pet food formulas, obtain the types of initial pet food ingredients and the range of proportions of initial pet food ingredients. Based on the feature sorting results, the intersection of the multiple initial pet food ingredient types is calculated to generate the feed ingredient type; Based on the feed ingredient type and the feature sorting result, the intersection of the multiple initial pet food ingredient ratio ranges is calculated to generate the feed ingredient ratio range. The process of traversing the pet's physiological characteristics and the living environment characteristics, prioritizing them, and generating a feature ranking result includes: When the preset update cycle is met, the feature parameter sorting data is uploaded through the first database, the second database, and up to the Nth database. The priority sorting table is updated based on the sorted data according to the aforementioned feature parameters; Input the pet's physiological characteristics and the breeding environment characteristics into the updated priority ranking table to generate the feature ranking result; The process of iterating through the pet's physiological characteristics and the characteristics of its living environment to adjust the standard pet food formula generates multiple sets of initial adjustment results for the pet food formula, including: Acquire a dataset of pet physiological characteristics, a dataset of living environment characteristics, and a dataset of pet food formula adjustment identifiers, and generate a dataset for model building. Build a dataset based on the model and train the Mth decision tree; Obtain the amount of data in the model construction dataset for which the Mth decision tree does not meet the preset accuracy, and determine whether the preset data amount is met. If the conditions are met, the first decision tree, the second decision tree, and so on up to the Mth decision tree are merged to generate a pet food formula adjustment model.

2. The method as described in claim 1, characterized in that, The determination of whether the preset data volume is met also includes: If the preset accuracy is not met, a dataset is constructed based on the model that does not meet the preset accuracy, and the (M+1)th decision tree is constructed. Determine the amount of data in the model construction dataset for which the (M+1)th decision tree does not meet the preset accuracy, and determine whether the preset data amount is met.

3. The method as described in claim 1, characterized in that, The construction of the feed optimization fitness function includes: Obtain the first fitness function: Among them, C t Let c be the cost screening degree of the t-th feed group. k Cost estimation for the k-th type of feed. Let K be the proportion of the kth type of feed, and K be the total number of feed types in the tth group. Obtain the second fitness function: F t =v t α *f t β Among them, F t To select the screening degree, v t f represents the rate of weight gain for pets fed the t-th diet. t Let be the frequency of pet illness in the t-th group of feed, and α and β be weighting indices; The feed optimization fitness function is constructed by combining the first fitness function and the second fitness function: Among them, A t γ and δ are weighting indices for optimizing the fitness of the t-th group of feed.

4. The method as described in claim 3, characterized in that, The step of optimizing the initial pet food formula based on the feed optimization fitness function and the formula optimization vector space to generate pet food formula optimization results includes: The feed formula for the t-th generation is obtained by randomly selecting values ​​from the formula optimization vector space. Based on the feed optimization fitness function, the t-th feed formulation is calculated to generate the t-th fitness. When the fitness of the t-th degree satisfies the preset fitness, obtain the fitness of the (t-1)-th degree. Determine whether the (t-1)th fitness is less than or equal to the tth fitness; If it is less than or equal to, add the (t-1)th feed formula to the elimination data group, and continue iterating based on the tth feed formula; If it is greater than 1, the feed formula of the tth generation is added to the elimination data group, and the iteration continues based on the feed formula of the t-1th generation. When the preset number of iterations is met, the initial pet food formula is replaced according to the feed formula being iterated, and the optimized pet food formula is generated.

5. A feed formulation optimization system based on pet differentiation analysis, characterized in that, The system for implementing the feed formulation optimization method based on pet differentiation analysis as described in any one of claims 1-4, the system comprising: The formula matching module is used to input basic pet information into the pet food standard table and match the pet food standard formula. The formula adjustment module is used to adjust the standard formula of pet food according to the characteristics of the breeding environment and the physiological characteristics of pets, and generate pet food formula adjustment results, wherein the pet food formula adjustment results include feed component types and feed component ratio ranges; A constraint construction module is used to construct formula optimization value constraints based on the feed component type and the feed component ratio range. A space construction module is used to traverse the feed component types and the formula optimization value constraints to construct a formula optimization vector space, wherein the dimension of the formula optimization vector space is the same as the dimension of the feed component types. A function construction module, which is used to construct a feed optimization fitness function; The formula optimization module is used to optimize the initial pet food formula according to the feed optimization fitness function and the formula optimization vector space, and generate the pet food formula optimization result. A feed preparation module is used to prepare pet food based on the pet food formula optimization results.

Images