Heterogeneous graph-based herbivorous animal feed formula optimization system and method

Abstract

The invention provides a herbivorous animal feed formula optimization system and method based on an isomerism graph, and the method comprises the steps: carrying out the multiple shielding processing of the isomerism graph during the construction of herbivorous animal feed ingredients, and determining an incidence matrix of the synergistic effect between the remaining feed raw material nodes after each shielding; further determining a integrating degree vector of each feed raw material during raw material compatibility; extracting reference nutritional demand characteristics from the heterogeneous graph, determining nutritional demand satisfaction rates in various growth stages based on initial growth stage information of the target plant-eating animals, and further generating nutritional demand parameter vectors of the target plant-eating animals; determining the nutrition adaptation degrees of different feed raw materials to the target herbivorous animals through the integrating degree vectors and the nutrition requirement parameter vectors, and determining the optimal feed proportioning scheme of the feed corresponding to the target herbivorous animal variety based on all the nutrition adaptation degrees. By adopting the scheme of the application, the feed for the herbivorous animals can be prepared by combining the synergistic relationship of the feed raw materials on the nutritional requirements of the herbivorous animals.

Description

Technical Field

[0001] This application relates to the field of feed formulation technology, and more specifically, to a system and method for optimizing herbivore feed formulation based on heterogeneous graphs. Background Technology

[0002] Feed formulation refers to the process of scientifically screening, proportionally adjusting, and optimizing the combination of various raw materials based on the core needs of the target herbivorous animal, such as its breed, growth stage, and physiological state, combined with the nutritional components, physicochemical properties, and compatibility rules of various feed ingredients.

[0003] In the field of traditional feed formulation, due to the diversity of target herbivorous animal species, dynamic changes in growth stages, and complex synergistic or antagonistic relationships among different feed ingredients, existing methods often struggle to accurately balance the species-specific nutritional needs of herbivores with the dynamic changes in growth stages. Furthermore, the potential correlations between ingredients are easily overlooked in the optimization of ingredient compatibility, resulting in formulation schemes that either fail to fully meet the precise nutritional needs of herbivores at different growth stages or cause nutrient waste or low absorption efficiency due to unreasonable ingredient combinations. Therefore, how to combine the synergistic relationships between various feed ingredients and the nutritional needs of herbivores to formulate feed for herbivores has become a problem facing the industry. Summary of the Invention

[0004] This application provides a system and method for optimizing herbivore feed formulation based on heterogeneous graphs, which can combine the synergistic relationship between various feed ingredients and the nutritional requirements of herbivores to formulate herbivore feed.

[0005] In a first aspect, this application provides a method for optimizing feed formulation for herbivores based on heterogeneous graphs, wherein multiple feed ingredients are used to formulate feed for target herbivores, characterized in that the method includes the following steps: A heterogeneous graph of herbivore feed formulation is constructed using the target herbivore species, herbivore growth stage, feed ingredients and nutrient composition as nodes. The feed ingredient nodes in the heterogeneous graph are masked multiple times to determine the correlation matrix of the synergistic effect between the remaining feed ingredient nodes after each masking. Based on all the correlation matrices, the compatibility vector of each feed ingredient with other feed ingredients in the raw material matching is calculated by a loss function that maximizes mutual information. The baseline nutrient requirement features of the nutrient component nodes when transferring nutrients to the growth stage nodes of the herbivores are extracted from the heterogeneous graph. Based on the baseline nutrient requirement features and the initial growth stage information of the target herbivores, the nutrient requirement satisfaction rate of the target herbivores at each growth stage is determined. All nutrient requirement satisfaction rates are fused and analyzed with the specific basic data corresponding to the target herbivores species to generate nutrient requirement parameter vectors of the target herbivores at different growth stages. By combining the fit vector and the nutrient requirement parameter vector with the nutrient composition characteristics of all feed ingredients, the nutritional suitability of different feed ingredients for the target herbivores is determined. Based on all the nutritional suitability, all feed ingredients in the feed formula are combined and optimized to obtain the optimal feed formulation scheme for the target herbivorous species.

[0006] In some embodiments, performing multiple masking processes on the feed ingredient nodes in the heterogeneous graph to determine the correlation matrix of the synergistic effects among the remaining feed ingredient nodes after each masking specifically includes: Determine the occlusion ratio of the feed ingredient nodes in the heterogeneous graph; The feed ingredient nodes in the heterogeneous graph are masked multiple times according to the masking ratio, and the remaining heterogeneous graph after each masking is recorded. Select a remaining heterogeneous graph as the selected remaining heterogeneous graph, and determine the association information between each feed ingredient node and other feed ingredient nodes in the selected remaining heterogeneous graph; Based on all the association information, determine the association matrix of the synergistic effects between feed ingredient nodes in the selected remaining heterogeneous graph; Continue to determine the correlation matrix of the synergistic effects among the feed ingredient nodes in the remaining heterogeneous graph.

[0007] In some embodiments, calculating the compatibility vector between each feed ingredient and other feed ingredients in ingredient formulation, based on all correlation matrices and using a loss function that maximizes mutual information, specifically includes: Standardize all the correlation matrices to obtain the standardized correlation matrices. Select one feed ingredient as the selected feed ingredient, and extract the correlation numerical sequence between the selected feed ingredient and other feed ingredients from all standardized correlation matrices; Construct a loss function that maximizes mutual information; Input the numerical sequence of the correlation between the selected feed ingredient and other feed ingredients into the loss function that maximizes mutual information, and output the compatibility vector of the selected feed ingredient and other feed ingredients when the ingredients are matched. Continue to determine the compatibility vector between the remaining feed ingredients and other feed ingredients when formulating them.

[0008] In some embodiments, extracting the baseline nutrient requirement characteristics from the heterogeneous graph when nutrient component nodes transfer nutrients to herbivore growth stage nodes specifically includes: The direct association paths between nutrient nodes and herbivore growth stage nodes are extracted from the heterogeneous graph. The required amounts of each nutrient at each stage of the growth process of herbivores are determined based on the heterogeneous diagram. Based on all the demand corresponding to the herbivore growth stage nodes and the directly associated paths, determine the baseline nutrient demand characteristics when nutrient component nodes transfer nutrients to herbivore growth stage nodes.

[0009] In some embodiments, determining the nutritional requirement satisfaction rate of the target herbivore at each growth stage based on the baseline nutritional requirement characteristics combined with the initial growth stage information of the target herbivore specifically includes: Obtain information on the initial growth stage of the target herbivore; The nutrient component feature vectors of the growth stage nodes are determined based on the aforementioned baseline nutrient requirement characteristics. The initial growth stage information and the nutrient component feature vector are fused to determine the nutrient requirement satisfaction rate of the target herbivore at each growth stage.

[0010] In some embodiments, the fusion analysis of all nutritional requirement satisfaction rates with specific basic data corresponding to the target herbivore species to generate a vector of nutritional requirement parameters for the target herbivore at different growth stages specifically includes: Obtain specific basic data on the target herbivorous animal species; By correlating and aggregating all nutritional requirement satisfaction rates with the specific basic data of the target herbivorous animal species, we can obtain aggregated information on the relationship between the species characteristics and nutritional suitability of the target herbivorous animal. Based on the aggregated information, a vector of nutritional requirements parameters for the target herbivore at different growth stages is generated.

[0011] In some embodiments, determining the nutritional suitability of different feed ingredients for the target herbivore by combining the fit vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients specifically includes: Determine the nutritional characteristics of all feed ingredients; The matching degree between the nutritional components of each feed ingredient and the specific needs of herbivorous animal breeds is determined based on the nutritional requirement parameter vector and the nutritional component characteristics. The nutritional suitability of each feed ingredient for the target herbivore is determined based on the matching degree between the nutritional composition of each feed ingredient and the specific needs of the herbivore species, and the fit vector.

[0012] In some embodiments, optimizing the combination of all feed ingredients in the feed formulation based on all nutritional compatibility to obtain the optimal feed formulation for the target herbivore species specifically includes: Obtain the initial feed formulation plan for all feed ingredients in the feed formula; The initial feed formulation scheme is optimized and adjusted based on all nutritional compatibility factors to obtain the optimal feed formulation scheme for the target herbivore species.

[0013] In some embodiments, initial growth stage information of the target herbivorous animal is obtained from the breeding management system database.

[0014] Secondly, this application provides a heterogeneous graph-based system for optimizing herbivore feed formulations, comprising: The module is used to construct a heterogeneous graph for herbivore feed formulation, with the target herbivore species, herbivore growth stage, feed ingredients and nutrients as nodes. The processing module is used to perform multiple masking processes on the feed ingredient nodes in the heterogeneous graph, thereby determining the correlation matrix of the synergistic effect between the remaining feed ingredient nodes after each masking, and calculating the compatibility vector of each feed ingredient with other feed ingredients in the raw material matching based on all the correlation matrices and through the loss function that maximizes mutual information. The processing module is further configured to extract the baseline nutritional requirement features from the heterogeneous graph when transferring nutrients from the nutrient component nodes to the growth stage nodes of the herbivores, determine the nutritional requirement satisfaction rate of the target herbivores at each growth stage based on the baseline nutritional requirement features and the initial growth stage information of the target herbivores, and fuse all the nutritional requirement satisfaction rates with the specific basic data corresponding to the target herbivores species to generate a nutritional requirement parameter vector of the target herbivores at different growth stages. The execution module is used to determine the nutritional suitability of different feed ingredients for the target herbivorous animal by combining the fit degree vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients, and to optimize the combination of all feed ingredients in the feed formula based on all nutritional suitability to obtain the optimal feed formulation scheme for the target herbivorous animal species.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The heterogeneous graph-based herbivore feed formulation optimization system and method provided in this application first constructs a heterogeneous graph for herbivore feed formulation, using the target herbivore species, herbivore growth stage, feed ingredients, and nutrient components as nodes. The feed ingredient nodes in the heterogeneous graph are then subjected to multiple masking processes to determine the correlation matrix of the synergistic effects between the remaining feed ingredient nodes after each masking. Based on all correlation matrices, the compatibility vector between each feed ingredient and other feed ingredients in feed formulation is calculated using a loss function that maximizes mutual information. Finally, the baseline nutrient requirement characteristics for nutrient component nodes when transferring nutrients to herbivore growth stage nodes are extracted from the heterogeneous graph, based on... The baseline nutritional requirements characteristics, combined with the initial growth stage information of the target herbivores, determine the nutritional requirement satisfaction rate of the target herbivores at each growth stage. All nutritional requirement satisfaction rates are then fused and analyzed with the specific basic data corresponding to the target herbivore breed to generate a nutritional requirement parameter vector for the target herbivores at different growth stages. The fit vector and the nutritional requirement parameter vector are then combined with the nutritional component characteristics of all feed ingredients to determine the nutritional suitability of different feed ingredients for the target herbivores. Based on all nutritional suitability, all feed ingredients in the feed formulation are combined and optimized to obtain the optimal feed formulation scheme corresponding to the target herbivore breed.

[0016] Therefore, this application demonstrates that, in the feed formulation process, firstly, a heterogeneous graph is constructed using the target herbivore species, growth stage, feed ingredients, and nutrient components as nodes. This enables the establishment of a visual framework for multi-factor relationships, breaking the limitations of traditional feed formulation where elements are scattered and isolated. It clarifies the intrinsic connections between species characteristics, growth requirements, and raw material nutrition, laying a comprehensive data foundation for subsequent synergistic analysis. By repeatedly masking the feed ingredient nodes in the heterogeneous graph and determining the fit vector using the correlation matrix and mutual information loss function, the synergistic or antagonistic patterns between different ingredients can be accurately identified. By quantifying the compatibility of ingredient combinations, nutrient waste or conflicts caused by blind combinations are avoided, thus improving the synergistic utilization of ingredients. The scientific basis of this approach lies in its extraction of baseline nutritional requirements from heterogeneous graphs at different growth stages. This is combined with breed-specific basic data to generate a nutritional requirement parameter vector, enabling precise positioning of nutritional needs. This approach considers both the dynamic changes in different growth stages of herbivores and the unique physiological needs of each breed, solving the problem of insufficient adaptability of general nutritional programs. By determining nutritional compatibility through fit vectors, nutritional requirement parameter vectors, and the nutritional composition characteristics of raw materials, and optimizing raw material combinations, a dual optimization of nutritional satisfaction and raw material synergy can be achieved. The resulting feed formulation ensures optimal synergy among raw materials while precisely meeting the specific nutritional needs of the target herbivores, significantly improving the breeding effect and economic value of the feed. Using this approach, feed formulation for herbivores can be based on the synergistic relationship between various feed ingredients and the nutritional needs of herbivores. Attached Figure Description

[0017] Figure 1 This is an exemplary flowchart of a method for optimizing herbivore feed formulation based on heterogeneous graphs, according to some embodiments of this application. Figure 2 This is an exemplary flowchart illustrating the determination of an association matrix according to some embodiments of this application; Figure 3 This is an exemplary flowchart illustrating the determination of a vector of nutritional requirements parameters according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a heterogeneous graph-based herbivore feed formulation optimization system according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device for implementing a method for optimizing feed formulations for herbivores based on heterogeneous graphs, according to some embodiments of this application. Detailed Implementation

[0018] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] refer to Figure 1 The figure is an exemplary flowchart of a heterogeneous graph-based herbivore feed formulation optimization method according to some embodiments of this application. The heterogeneous graph-based herbivore feed formulation optimization method mainly includes the following steps: In step 101, a heterogeneous graph of herbivore feed formulation is constructed using the target herbivore species, herbivore growth stage, feed ingredients and nutrients as nodes.

[0020] In practical implementation, when constructing the heterogeneity diagram of herbivore feed formulation, it is first necessary to clarify the specific division of four core nodes: the target herbivore breed node covers the specific breed type of the herbivore to be formulated, with each breed as an independent node and labeled with its biological characteristics; the herbivore growth stage node is divided into multiple key stages according to the growth cycle of the target herbivore, with each stage node associated with the growth indicators and basic data of physiological requirements for that stage; the feed ingredient node contains all the single ingredients to be selected, with each ingredient node recording its name, origin, and basic component content range; the nutrient component node covers all kinds of nutrients required for the growth of herbivore, with each nutrient component node specifying its chemical composition, physiological function, and recommended intake information; next, the relationship between nodes is established: between the herbivore breed node and the growth stage node, a corresponding connection is established according to the breed-specific growth cycle, and the growth parameters of different breeds at each stage are labeled. Among them, the growth parameters at each stage represent the core indicators used to quantitatively describe the growth and development status, physiological function level, and production-related characteristics of the herbivore breed within its growth cycle stage. The indicators specifically include average daily weight gain, end-of-stage body weight, and feed conversion ratio, reflecting growth rate and feed conversion efficiency; body size indicators such as height, length, and chest circumference, describing body condition and organ development status; digestive organ development parameters such as rumen volume and rumen microbial count; production performance-related parameters that connect growth with target production needs, such as daily milk yield and milk fat percentage during lactation in dairy cows, and carcass percentage and backfat thickness during fattening in beef cattle; and health and metabolic parameters that ensure the physiological basis of growth, such as blood calcium, blood phosphorus, and blood glucose, as well as the incidence of nutrition-related diseases; feed ingredient nodes and nutrition. Between component nodes, weighted connections are established based on the actual measured nutrient content of the raw materials, with the weight value corresponding to the content ratio of a specific nutrient in the raw material; between herbivore growth stage nodes and nutrient component nodes, associations are established based on the nutritional requirement standards of that stage, and the required threshold of each nutrient is marked; at the same time, feed ingredient nodes can establish associations based on actual compatibility experience or known synergistic / antagonistic effects, thereby forming a heterogeneous graph containing multiple types of nodes and rich association information; other methods can also be used in other embodiments, which are not limited here.

[0021] It should be noted that the heterogeneous graph in this application represents a structured data model of multiple different types of nodes and specific relationships between nodes corresponding to the target herbivores. It takes the target herbivore species, herbivore growth stage, feed ingredients and nutrients as four core nodes. Each node carries detailed attribute information of the corresponding field. At the same time, different nodes are linked together through targeted connection relationships, intuitively presenting the complex relationship between herbivore species, growth stage, feed ingredients and nutrients, and providing comprehensive structured data support for feed formulation analysis.

[0022] In step 102, the feed ingredient nodes in the heterogeneous graph are masked multiple times to determine the correlation matrix of the synergistic effect between the remaining feed ingredient nodes after each masking. Based on all the correlation matrices, the compatibility vector of each feed ingredient with other feed ingredients in the raw material matching is calculated by using a loss function that maximizes mutual information.

[0023] In some embodiments, reference Figure 2 As shown, this diagram is an exemplary flowchart for determining the association matrix in some embodiments of this application. In this embodiment, the feed ingredient nodes in the heterogeneous graph are subjected to multiple masking processes, and the association matrix of the synergistic effect among the remaining feed ingredient nodes after each masking can be determined by the following steps: First, in step 1021, the shading ratio of the feed ingredient nodes in the heterogeneous graph is determined; Secondly, in step 1022, the feed raw material nodes in the heterogeneous graph are masked multiple times according to the masking ratio, and the remaining heterogeneous graph after each masking is recorded. Then, in step 1023, a remaining heterogeneous graph is selected as the selected remaining heterogeneous graph, and the association information between each feed ingredient node and other feed ingredient nodes in the selected remaining heterogeneous graph is determined. Therefore, in step 1024, the association matrix of the synergistic effect between feed raw material nodes in the selected remaining heterogeneous graph is determined based on all the association information; Finally, in step 1025, the correlation matrix of the synergistic effects between feed ingredient nodes in the remaining heterogeneous graph is determined.

[0024] In specific implementation, when determining the occlusion ratio of feed ingredient nodes in the heterogeneous graph, it is necessary to consider the total number and diversity of feed ingredient nodes, typically setting an occlusion ratio range of 10%-30%. For example, based on experimental verification or industry experience, when the number of ingredient nodes is 50-100, a 20% occlusion ratio is selected. This ensures that the remaining nodes after each occlusion still reflect the core relationships between ingredients, while generating sufficiently diverse node combination scenarios through multiple occlusions. Simultaneously, it avoids situations where the occlusion ratio is too high, resulting in incomplete remaining node association information, or too low, leading to insufficient differences between different occlusion scenarios. Secondly, the feed ingredient nodes in the heterogeneous graph are then processed according to the occlusion ratio. During multiple masking processes, nodes of a corresponding proportion are selected from all feed ingredient nodes as masking objects. For example, a random sampling method is used. Before each masking, the random seed is reset to ensure sampling independence. The masking operation is achieved by marking the node status as "invalid" (the node is not deleted, but its ability to participate in association calculation is masked). At the same time, the ID, original attributes, and connection relationships with other nodes of the remaining valid raw material nodes are recorded after each masking, forming a residual heterogeneous graph data file containing node sets, node attributes, and association structures. The masking operation is repeated 20-50 times to obtain multiple sets of different residual heterogeneous graphs. Other methods can also be used in other embodiments, which are not limited here.

[0025] In specific implementation, when determining the association information between each feed ingredient node and other feed ingredient nodes in the selected remaining heterogeneous graph, all feed ingredient nodes in the selected remaining heterogeneous graph are first traversed to extract the direct association information between feed ingredient nodes, including known synergistic effects, antagonistic effects, and effect intensity levels between ingredients. At the same time, indirect association information such as the overlap degree of commonly associated nutrient component nodes (which can be quantified by Jaccard coefficient) and the shortest path length in the heterogeneous graph are calculated. This information is classified and organized into structured association information, wherein the association information represents the association situation between feed ingredient nodes, and the association type and quantification value of each pair of ingredient nodes are clearly defined. Other methods can also be used to determine this in other embodiments, which are not limited here.

[0026] In addition, in specific implementation, when determining the association matrix of synergistic effects between feed ingredient nodes in the selected remaining heterogeneous graph based on all the associated information, a square matrix is ​​constructed with the remaining feed ingredient nodes in the selected remaining heterogeneous graph as rows and columns. The matrix dimension is consistent with the number of remaining nodes. For each element in the matrix, the direct association strength (synergy is positive, antagonism is negative, and values ​​are assigned from 1 to 5 according to the strength level) and indirect association degree (such as the overlap of nutritional components multiplied by 0.3) between ingredients are integrated by weighted summation, and the values ​​are mapped to the interval [-1,1]. For example, the maximum-minimum standardization can be used to make the matrix elements intuitively reflect the direction and strength of the association between ingredients, and finally form the association matrix of synergistic effects between feed ingredient nodes in the selected remaining heterogeneous graph. Other methods can also be used to determine this in other embodiments, which are not limited here.

[0027] It should be noted that the association matrix in this application represents the association relationship between feed raw material nodes in the process of feed formulation. The positive and negative values ​​and magnitudes of the elements in the association matrix intuitively reflect the direction (positive for synergy and negative for antagonism) and strength of the association, thus presenting the association pattern between the remaining raw material nodes in the system.

[0028] In some embodiments, the following steps can be taken to calculate the compatibility vector between each feed ingredient and other feed ingredients in ingredient formulation, based on all correlation matrices and through a loss function that maximizes mutual information: Standardize all the correlation matrices to obtain the standardized correlation matrices. Select one feed ingredient as the selected feed ingredient, and extract the correlation numerical sequence between the selected feed ingredient and other feed ingredients from all standardized correlation matrices; Construct a loss function that maximizes mutual information; Input the numerical sequence of the correlation between the selected feed ingredient and other feed ingredients into the loss function that maximizes mutual information, and output the compatibility vector of the selected feed ingredient and other feed ingredients when the ingredients are matched. Continue to determine the compatibility vector between the remaining feed ingredients and other feed ingredients when formulating them.

[0029] In practice, when standardizing all correlation matrices, the maximum-minimum standardization method is used to map the values ​​to the range [-1, 1] for the distribution range of element values ​​in each correlation matrix (the formula is: standardized value = (original value - minimum value) / (maximum value - minimum value) × 2 - 1). For example, in a matrix, the correlation value between raw material A and B is 3. The minimum value of the matrix element is -5 and the maximum value is 5. After standardization, it is (3+5) / (5+5) × 2 - 1 = 0.6. This process yields the standardized correlation matrices and ensures data comparability by eliminating the dimensional differences between different matrices.

[0030] In addition, in specific implementation, the sequence of correlation values ​​between the selected feed ingredient and other feed ingredients can be extracted from all standardized correlation matrices in the following way: traverse all standardized correlation matrices, locate the row (or column) where the selected feed ingredient is located in each matrix, extract the element values ​​corresponding to other feed ingredients in that row, and arrange them according to the matrix number to form a sequence. For example, the correlation values ​​between the selected feed ingredient and other feed ingredients extracted from 20 matrices are 0.3, 0.25, ..., 0.4, forming the sequence of correlation values ​​between the selected feed ingredient and other feed ingredients. Other extraction methods can also be used in other embodiments, which are not limited here.

[0031] In constructing the loss function to maximize mutual information, the associated numerical sequences of the selected raw material and other raw materials are used as random variables. The loss function is defined as 1 minus the sum of mutual information between the sequences (i.e., Loss = 1 - ΣI(X,Y), where I(X,Y) is the mutual information between sequences X and Y). The goal is to maximize mutual information by minimizing the loss function. For example, when the mutual information between the selected raw material and one other raw material is 0.8 and with another raw material is 0.6, the initial loss is 1 - (0.8 + 0.6) = -0.4. Further optimization is needed to improve the mutual information. A differentiable loss function is constructed using the tensor flow framework. The calculation of mutual information between sequences can be achieved by preprocessing the associated numerical sequences, including removing outliers and standardizing or discretizing them according to the data type. The data is processed to ensure it meets the basic requirements for mutual information calculation. Data types include continuous and discrete types. Based on the data type, a mutual information calculation method is selected to calculate the mutual information between sequences. For example, if the sequence is discrete, the joint probability distribution and the marginal probability distribution of the two sequences are constructed by statistically analyzing the frequency of occurrence of each value in the sequence, and then substituted into the mutual information definition formula for calculation. If the sequence is continuous, the probability density function of the sequence is estimated by kernel density estimation, and then the mutual information is solved by integral operation. Alternatively, the mutual information estimation tool in the tensor flow framework that supports differentiable calculation can be directly used to model and learn the preprocessed sequence, and finally output the mutual information value I(X,Y) between the selected raw material and the associated sequences of other raw materials. Other methods can also be used in other embodiments, which are not limited here.

[0032] In addition, in specific implementation, when the associated numerical sequence of the selected feed ingredients is input into the loss function, the numerical values ​​in the sequence are used as input features of the model parameters. The gradient descent optimization algorithm is used to iteratively adjust the parameters so that the loss function converges to the minimum value (i.e., mutual information is maximized). For example, for the associated sequence of the selected feed ingredients, after 500 iterations, the loss stabilizes at -1.2. At this time, in the output parameter vector, each element corresponds to the compatibility degree between the selected feed ingredient and other ingredients, forming a compatibility degree vector between the selected feed ingredient and other feed ingredients when the ingredients are combined. Other methods can also be used in other embodiments, which are not limited here.

[0033] It should be noted that the fit vector in this application represents the degree of compatibility of a feed ingredient with all other feed ingredients when they are combined. It can be used to reflect the association patterns between ingredients under different shading scenarios, and provide a quantitative basis for subsequent screening of reasonable feed ingredient combinations and formulation of scientific formulation schemes.

[0034] In step 103, the baseline nutrient requirement features when nutrient component nodes transfer nutrients to herbivore growth stage nodes are extracted from the heterogeneous graph. Based on the baseline nutrient requirement features and the initial growth stage information of the target herbivore, the nutrient requirement satisfaction rate of the target herbivore at each growth stage is determined. All nutrient requirement satisfaction rates are fused and analyzed with the specific basic data corresponding to the target herbivore species to generate a nutrient requirement parameter vector of the target herbivore at different growth stages.

[0035] In some embodiments, extracting the baseline nutrient requirement characteristics from the heterogeneous graph when nutrient component nodes transfer nutrients to herbivore growth stage nodes can be achieved through the following steps: The direct association paths between nutrient nodes and herbivore growth stage nodes are extracted from the heterogeneous graph. The required amounts of each nutrient at each stage of the growth process of herbivores are determined based on the heterogeneous diagram. Based on all the demand corresponding to the herbivore growth stage nodes and the directly associated paths, determine the baseline nutrient demand characteristics when nutrient component nodes transfer nutrients to herbivore growth stage nodes.

[0036] In specific implementation, when extracting the direct association paths between nutrient nodes and herbivore growth stage nodes from a heterogeneous graph, it is necessary to traverse the edge set of the heterogeneous graph, filter out edges where the source node type is "nutrient" and the target node type is "herbivore growth stage", record the connection relationship of each edge (e.g., "vitamin A - chick stage" "protein - fattening herbivore stage") and the attributes of the edge (e.g., association strength, function type), and exclude paths that are indirectly connected through other nodes (e.g., feed ingredients), to obtain the direct association paths between each nutrient node and each herbivore growth stage node. This step can be achieved by querying a graph database. Here, a direct association path means a connection path between a nutrient node and a herbivore growth stage node that does not pass through other types of nodes. Other extraction methods can be used in other embodiments, which are not limited here.

[0037] In addition, in specific implementation, when determining the required amount of each nutrient for each growth stage node of a herbivore based on the heterogeneous graph, the attribute fields of the target growth stage node can be queried first. For example, the raw data stored in key-value pairs can be obtained through a graph database access tool (e.g., {"protein requirement":"20%","calcium requirement":"0.8%"}). Then, these key-value pairs are traversed, and for each key (e.g., "protein requirement"), the nutrient type is extracted by string truncation (taking the core name "protein" before "requirement"). For the corresponding value (e.g., "20%), the required amount value is extracted by string replacement (removing "%") and type conversion. Thus, the required amount of each nutrient for each growth stage node of the herbivore is obtained, where the required amount represents the parameter of the degree of requirement of each nutrient for each growth stage node of the herbivore. Other methods can also be used to determine this in other embodiments, which are not limited here.

[0038] Furthermore, in practical implementation, when determining the baseline nutrient requirement characteristics based on the demand of nodes at each herbivore growth stage and the direct association paths, the demand should be used as the core parameter for propagation weighting, and the direct association paths should be weighted accordingly. For each herbivore growth stage node, if the demand of a herbivore growth stage node for nutrient node A is 20% (weight 0.2), and the association strength between A and that stage is 0.8, then the propagation contribution value of the direct association path between that herbivore growth stage node and the nutrient node is 0.2 × 0.8 = 0.16. The propagation contribution values ​​of all nutrient nodes to that growth stage through direct association paths are summarized to form a feature vector containing the propagation strength from each nutrient node to the herbivore growth stage node. This process can be achieved using the message passing mechanism in graph neural networks, such as using a deep graph library (DeepGraph). The nutrient requirement features are weighted according to demand by the Library (DGL) or attention weights are assigned to different paths through a graph attention network. Finally, all feature vectors are aggregated to obtain the baseline nutrient requirement features when nutrient nodes transfer nutrients to herbivore growth stage nodes. Other methods can also be used to determine the nutrient requirement in other embodiments, which are not limited here.

[0039] It should be noted that the baseline nutritional requirement characteristics in this application represent the comprehensive characteristics of the ability of nutrient component nodes to transmit nutritional requirements to herbivore growth stage nodes. They reflect the actual needs, adaptability, and correlation of different nutrients for specific growth stages and serve as a key bridge connecting nutrients and the growth needs of herbivores.

[0040] In some embodiments, determining the nutritional requirement satisfaction rate of the target herbivore at each growth stage based on the baseline nutritional requirement characteristics and the initial growth stage information of the target herbivore can be achieved through the following steps: Obtain information on the initial growth stage of the target herbivore; The nutrient component feature vectors of the growth stage nodes are determined based on the aforementioned baseline nutrient requirement characteristics. The initial growth stage information and the nutrient component feature vector are fused to determine the nutrient requirement satisfaction rate of the target herbivore at each growth stage.

[0041] When obtaining initial growth stage information of target herbivores, it is necessary to extract the initial growth stage information of target herbivores from the breeding management system database, sensor records, or manually entered data. The initial growth stage information includes herbivore breed, current growth stage name, starting age, initial weight, and health status label. The initial growth stage information represents the core information set of basic characteristics and status of target herbivores at the beginning of a specific growth cycle, and is mainly used to define the growth background and basic needs baseline of herbivores.

[0042] In specific implementation, when determining the nutrient component feature vector of the growth stage node based on the baseline nutrient requirement characteristics, it is necessary to use the nutrient component type (such as protein, calcium, lysine, vitamin) as the dimension, and quantify and integrate the propagation characteristics of the propagation intensity, demand matching degree, and adaptability direction corresponding to each nutrient component into vector elements; for example, sorted according to a preset nutrient component list, the first dimension corresponds to the propagation intensity of protein (e.g., 0.18), the second dimension corresponds to the demand matching degree of calcium (e.g., 15%), and the third dimension corresponds to the adaptability direction of lysine (positive adaptability is recorded as 1, negative adaptability as -1). A fixed-length nutrient component feature vector is constructed through a numerical Python array, wherein the nutrient component feature vector represents the structured and numerical vector of the propagation characteristics of various nutrients associated with the growth stage node, and the vector length is consistent with the number of nutrient types to ensure that the feature vector structure of different growth stages is consistent; other methods can also be used to determine this in other embodiments, which are not limited here.

[0043] In addition, in specific implementation, when fusing the initial growth stage information and the nutrient component feature vector, the categorical features (such as breed and health status) in the initial information are first converted into numerical vectors, such as by one-hot encoding (e.g., "white-feathered broiler" is encoded as [1,0,0]) or embedding encoding. The numerical initial information (such as age and weight) is standardized to the [0,1] interval, for example by standardization using the maximum-minimum value. The processed initial information vector (e.g., length 5) is combined with the nutrient component feature vector (e.g., length 10) to form a fused feature vector of length 15, such as by feature concatenation or by weighted summation. Finally, the fused feature vector is used as input and imported into a pre-trained prediction model (e.g., based on random forest). The model (using regression models, graph neural networks, or rule-based reasoning engines) learns and integrates the correlation patterns contained in the feature vectors (e.g., "broiler chicken + 30 days old + protein transmission intensity 0.18" corresponds to "protein supply 22%)", outputs the recommended supply amount, adaptation level, and constraints (e.g., calcium-phosphorus ratio range) of each nutrient component for each growth stage, and calculates the nutritional requirement satisfaction rate of the target herbivore at each growth stage based on all recommended supply amounts. The above calculation can be implemented using regression analysis or analysis of variance in mathematical statistics. For example, the matching ratio of each recommended supply amount to the baseline nutritional requirement can be used as the nutritional requirement satisfaction rate for the corresponding growth stage. Other methods can also be used to determine this in other embodiments, which are not limited here.

[0044] It should be noted that the nutritional requirement satisfaction rate in this application represents the parameter value of the degree to which the individual nutritional requirements of the target herbivores are met during their growth stages. In essence, it is a precise definition and scientific recommendation of the nutrients required by herbivores at specific growth stages, used to guide feed formulation optimization or feeding management, and to ensure an efficient match between the nutritional requirements of herbivores and the feed supply capacity.

[0045] In some embodiments, reference Figure 3 As shown, this figure is an exemplary flowchart of determining the nutritional requirement parameter vector in some embodiments of this application. In this embodiment, the nutritional requirement satisfaction rate of all nutritional requirements is fused and analyzed with the specific basic data corresponding to the target herbivore species to generate the nutritional requirement parameter vector of the target herbivore at different growth stages. This can be achieved by the following steps: First, in step 1031, specific basic data of the target herbivorous animal species are obtained; Secondly, in step 1032, all nutritional requirement satisfaction rates are correlated and aggregated with the specific basic data of the target herbivorous animal species to obtain aggregated information between the species characteristics and nutritional suitability of the target herbivorous animal. Finally, in step 1033, a vector of nutritional requirements parameters for the target herbivore at different growth stages is generated based on the aggregated information.

[0046] In practice, firstly, specific basic data of the target herbivorous animal breed is extracted from the breed database corresponding to the target herbivorous animal. This specific basic data represents a set of fundamental information accumulated through long-term breeding practices and scientific research, including the unique biological characteristics, nutritional requirements, physiological and metabolic features, and breeding suitability of the target herbivorous animal breed. This includes breed genetic characteristics (such as the high lean meat percentage gene in "Durocrandra vulgaris"), standard nutritional requirement thresholds for the breed (such as a crude protein standard of 16% during the fattening period), and metabolic characteristic parameters (such as an energy conversion ratio of 0.85). Secondly, when aggregating all nutritional requirement fulfillment rates with the target herbivorous animal breed, all nutritional requirement fulfillment rates are correlated with the specific basic data for calculation. For example, the deviation rate between the nutritional requirement fulfillment rate and the breed's standard nutritional supply threshold (such as a recommended nutritional requirement fulfillment rate of 16%) is calculated. The data includes 0.5%, a deviation rate of 3.125% from the standard of 16%, and matching coefficients between the fit level and breed requirements (e.g., a coefficient of 1.0 for "high fit" and 0.7 for "medium fit"). It also integrates breed-specific nutritionally sensitive components (e.g., the breed's high requirement weight for lysine). Finally, this data is integrated into a structured aggregated information table containing "breed name - nutritional component deviation rate - fit matching coefficient - sensitive component weight". This structured aggregated information table serves as the aggregated information between the breed characteristics and nutritional fit of the target herbivore. For example, the above steps can be implemented by using a panda merging function in a programmable language to associate the nutritional requirement satisfaction rate with the breed characteristic data table. The aggregated information represents the result of aggregating information between the breed characteristics and nutritional fit of the target herbivore. Other methods can also be used in other embodiments, which are not limited here.

[0047] In specific implementation, when generating the nutrient requirement parameter vector based on the aggregated information, the vector dimensions must first be determined: covering the variety-specific dimension (such as the genetic characteristic coding value of the variety, the quantitative value of metabolic characteristics), the dynamic requirement dimension of the growth stage (such as the deviation rate of each nutrient component, the adaptation matching coefficient), and the sensitive component dimension (such as the weight value of sensitive nutrients); then, the various types of data in the aggregated information are quantified and standardized, for example, the genetic characteristics of the variety are converted into unique heat codes (such as "Duchesnea grandiflora herbivore" is coded as [1,0,0]), and the numerical data of deviation rate and matching coefficient are standardized to the [0,1] interval, such as through... The maximum-minimum standardization method is used. Finally, the standardized quantified values ​​are combined into a vector according to the preset dimension order. This vector is used as a nutritional requirement parameter vector that includes the breed specificity of herbivores and the dynamic requirements of growth stages. For example, the vector of a Duchesnea grandis herbivore during the fattening period may be [1, 0.85, 0.03125, 1.0, 0.9], where the first position is the breed code, the second position is the metabolic characteristic value, the third position is the crude protein deviation rate, the fourth position is the adaptation matching coefficient, and the fifth position is the lysine sensitive weight. The vector can be constructed by numerical array to ensure that the dimensions are fixed and the units are uniform.

[0048] It should be noted that the nutrient requirement parameter vector in this application represents the parameter vector of nutrient requirements of the target herbivores at different growth stages. It reflects the dynamic adjustment of requirements as the growth stages progress, and provides structured input data for subsequent feed formulation optimization, dynamic decision-making on nutrient supply, and intelligent control of the breeding system, thereby realizing the transformation from "experience-based feeding" to "data-driven precision nutrition".

[0049] In step 104, the nutritional suitability of different feed ingredients for the target herbivorous animal is determined by combining the fit vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients. Based on all nutritional suitability, all feed ingredients in the feed formula are combined and optimized to obtain the optimal feed formulation scheme for the target herbivorous animal species.

[0050] In some embodiments, determining the nutritional suitability of different feed ingredients for a target herbivore by combining the fit vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients can be achieved through the following steps: Determine the nutritional characteristics of all feed ingredients; The matching degree between the nutritional components of each feed ingredient and the specific needs of herbivorous animal breeds is determined based on the nutritional requirement parameter vector and the nutritional component characteristics. The nutritional suitability of each feed ingredient for the target herbivore is determined based on the matching degree between the nutritional composition of each feed ingredient and the specific needs of the herbivore species, and the fit vector.

[0051] In practice, when determining the nutritional characteristics of all feed ingredients, it is necessary to extract the core nutritional data of each ingredient from the feed ingredient database, ingredient testing report, or industry standards. All core nutritional data are used as the nutritional characteristics of all feed ingredients. The core nutritional data includes the content (units such as %, g / kg) of crude protein, amino acids (such as lysine and methionine), minerals (such as calcium and phosphorus), vitamins, etc. In other embodiments, other methods can be used to determine the nutritional characteristics, which are not limited here.

[0052] In addition, in specific implementation, when determining the matching degree based on the nutrient requirement parameter vector and nutrient component characteristics, the nutrient component characteristics of the feed raw materials are first converted into feature vectors of the same dimension as the nutrient requirement parameter vector (e.g., filling in the standardized content of each component in the raw material according to the order of nutrient components in the vector); then, the matching degree of the nutrient components of each feed raw material with the specific needs of herbivorous animals is determined by calculating the similarity (e.g., cosine similarity, Pearson correlation coefficient) or weighted deviation value (e.g., assigning higher weights to components sensitive to breed-specific needs, calculating the weighted sum of the absolute deviations between the raw material components and the needs). For example, if the protein dimension weight in the nutrient perception vector is 0.3, and the protein content of a certain raw material deviates from the needs by 2%, then the component contributes a deviation of 0.006. After summing all component deviations and normalizing, a matching degree in the 0-1 interval is obtained (1 is a perfect match). The matching degree is a parameter of the degree of matching between the nutrient components of the feed raw materials and the specific needs of herbivorous animals. In other embodiments, other methods can also be used to determine this, which are not limited here.

[0053] In addition, in specific implementation, when determining nutritional suitability based on the matching degree and compatibility vector, the two need to be weighted and fused: For each feed ingredient, based on the matching degree between the ingredient and the needs of herbivores (e.g., 60%), combined with the average compatibility value of the ingredient with other potential compatible ingredients in the compatibility vector (e.g., 40%, to avoid overall nutritional imbalance due to a single ingredient being compatible but conflicting with other ingredients), a comprehensive score is calculated using the formula "Nutritional Suitability = Matching Degree × 0.6 + Average Compatibility Value × 0.4"; the score is then standardized (e.g., mapped to 0-100 points), and suitability levels are divided according to the score (e.g., above 80 points is "high suitability"), finally obtaining the nutritional suitability of each feed ingredient for the target herbivores. The above scheme can be batch calculated using panda application functions, or multi-ingredient parallel processing can be achieved using numerical matrix operations; other methods can also be used in other embodiments, which are not limited here.

[0054] It should be noted that the nutritional suitability in this application refers to the parameter value of the degree of nutritional suitability of feed ingredients to target herbivores. It can intuitively judge the rationality of a certain ingredient or formula in terms of nutritional supply, provide data support for precision feed formula design and nutritional regulation in the breeding process, reduce nutritional waste or deficiency, and improve breeding efficiency.

[0055] In some embodiments, optimizing the combination of all feed ingredients in the feed formulation based on all nutrient compatibility to obtain the optimal feed formulation for the target herbivore species can be achieved through the following steps: Obtain the initial feed formulation plan for all feed ingredients in the feed formula; The initial feed formulation scheme is optimized and adjusted based on all nutritional compatibility factors to obtain the optimal feed formulation scheme for the target herbivore species.

[0056] In practice, the initial feed formulation scheme for all feed ingredients in the feed formula is obtained from the feed database corresponding to the target herbivorous animal. The initial feed formulation scheme is the basic blueprint for feed combination optimization. It mainly includes the core information of the target herbivorous animal (such as breed, growth stage, and nutritional requirements, such as the crude protein requirement of laying hens during the egg-laying period being 16%-18%), a list of available feed ingredients (including ingredient names and categories, such as energy ingredient corn, protein ingredient soybean meal, and mineral additive dicalcium phosphate, etc.), the initial addition ratio of each ingredient (expressed as a percentage or weight percentage, such as corn 55%, soybean meal 20%, and wheat bran 10%), basic attribute data of the ingredients (such as nutrient content, unit cost, procurement channels and supply restrictions), the estimated total nutrient content calculation results (such as whether the total content of crude protein, calcium, and lysine in the scheme is close to the requirement threshold), and preliminary cost estimates (such as the total cost of raw materials per ton of feed), cost upper limit constraints, and preliminary compatibility screening records (such as whether obvious ingredient antagonistic combinations are avoided).

[0057] Furthermore, in practical implementation, when optimizing and adjusting the initial feed formulation based on all nutritional compatibility scores, it is necessary to first convert the nutritional compatibility scores of each raw material into weight coefficients (assigning higher weights to raw materials with high compatibility scores). This can be achieved through linear transformation or piecewise weighted summation. Then, a weighted summation of all nutritional compatibility scores should be performed, and an optimization model should be established with "nutritional requirement compliance rate," "raw material compatibility score," and "weighted sum of compatibility scores" as core indicators. This model can be established, for example, through a multi-objective genetic algorithm or linear weighted summation. The proportion of raw materials should be adjusted iteratively through the optimization model (e.g., increasing the proportion of raw materials with a compatibility score ≥ 0.8 by 5%-10%, decreasing the proportion of raw materials with a nutritional compatibility score ≤ 0.5, or replacing them with highly compatible alternative raw materials). For example, the genetic algorithm in existing technologies can be used as the iterative engine for the optimization model, while also incorporating commonly used technologies in the feed industry. The safety thresholds of feed ingredients and the cost limits involved in the initial feed formulation plan are determined. Simultaneously, the nutritional components of the adjusted plan are verified to meet the target requirements (e.g., whether the measured protein value is within the standard range) and whether there are any incompatibilities in the combination of raw materials (e.g., checking whether the coexistence ratio of antagonistic raw materials exceeds the limit through the compatibility matrix). Based on the adjustment of the raw material ratio, the initial feed formulation plan is optimized in multiple rounds. This can be achieved with the help of heuristic algorithms (such as genetic algorithms) or professional feed formulation software until the plan meets the nutritional requirements with a compliance rate of ≥95%, a compatibility score of ≥90 points, and the weighted sum of the fitness is maximized. Finally, the optimal feed formulation plan is output, which includes the optimized raw material ratio, measured nutritional components, total fitness score, and compatibility assessment. This optimal feed formulation plan is used as the optimal feed formulation plan for the target herbivore species.

[0058] It should be noted that the optimal feed formulation scheme in this application refers to a feed formulation scheme that achieves a balance of multiple dimensions such as nutritional suitability, raw material compatibility, economy and safety by scientifically selecting raw materials, adjusting proportions and optimizing combinations, under the premise of meeting the specific needs of the target herbivores. It can be used to determine the feed formulation for the target herbivores, and can improve the production performance (such as weight gain, egg production and milk production) and health level of herbivores, while maximizing the breeding benefits.

[0059] Furthermore, in another aspect of this application, in some embodiments, this application provides a herbivore feed formulation optimization system based on heterogeneous graphs, referring to... Figure 4 The figure is a schematic diagram of a heterogeneous graph-based herbivore feed formulation optimization system according to some embodiments of this application. The heterogeneous graph-based herbivore feed formulation optimization system 400 includes: a construction module 401, a processing module 402, and an execution module 403, which are described below: Construction module 401, in this application, is mainly used to construct a heterogeneous graph of herbivore feed formulation with the target herbivore species, herbivore growth stage, feed ingredients and nutrient components as nodes. Processing module 402, in this application, is used to perform multiple masking processes on the feed ingredient nodes in the heterogeneous graph, thereby determining the correlation matrix of the synergistic effect between the remaining feed ingredient nodes after each masking, and calculating the compatibility vector of each feed ingredient with other feed ingredients in the material matching process through a loss function that maximizes mutual information based on all the correlation matrices. It should be noted that the processing module 402 in this application is also used to extract the baseline nutritional requirement characteristics when the nutrient component nodes transfer nutrition to the growth stage nodes of the herbivores from the heterogeneous graph, determine the nutritional requirement satisfaction rate of the target herbivores at each growth stage based on the baseline nutritional requirement characteristics and the initial growth stage information of the target herbivores, and fuse all the nutritional requirement satisfaction rates with the specific basic data corresponding to the target herbivores species to generate a nutritional requirement parameter vector of the target herbivores at different growth stages. The execution module 403 in this application is mainly used to determine the nutritional suitability of different feed ingredients for the target herbivorous animal by combining the fit degree vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients, and to optimize the combination of all feed ingredients in the feed formula based on all nutritional suitability to obtain the optimal feed formulation scheme for the target herbivorous animal species.

[0060] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for optimizing herbivore feed formulation based on heterogeneous graphs.

[0061] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device for implementing a heterogeneous graph-based herbivore feed formulation optimization method according to some embodiments of this application. The heterogeneous graph-based herbivore feed formulation optimization method in the above embodiments can be implemented through... Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0062] Processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0063] The communication bus 502 can be used to transmit information between the aforementioned components.

[0064] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 503 may exist independently and be connected to processor 501 via communication bus 502. Memory 503 may also be integrated with processor 501.

[0065] The memory 503 stores program code for executing the scheme of this application, and its execution is controlled by the processor 501. The processor 501 executes the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0066] Communication interface 504 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0067] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0068] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0069] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for optimizing herbivore feed formulation based on heterogeneous graphs.

[0070] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0071] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for optimizing feed formulation for herbivores based on heterogeneous graphs, wherein, The method of formulating feed for target herbivores using multiple feed ingredients is characterized by comprising the following steps: A heterogeneous graph of herbivore feed formulation is constructed using the target herbivore species, herbivore growth stage, feed ingredients and nutrient composition as nodes. The feed ingredient nodes in the heterogeneous graph are masked multiple times to determine the correlation matrix of the synergistic effect between the remaining feed ingredient nodes after each masking. Based on all the correlation matrices, the compatibility vector of each feed ingredient with other feed ingredients in the raw material matching is calculated by a loss function that maximizes mutual information. The baseline nutrient requirement features of the nutrient component nodes when transferring nutrients to the growth stage nodes of the herbivores are extracted from the heterogeneous graph. Based on the baseline nutrient requirement features and the initial growth stage information of the target herbivores, the nutrient requirement satisfaction rate of the target herbivores at each growth stage is determined. All nutrient requirement satisfaction rates are fused and analyzed with the specific basic data corresponding to the target herbivores species to generate nutrient requirement parameter vectors of the target herbivores at different growth stages. By combining the fit vector and the nutrient requirement parameter vector with the nutrient composition characteristics of all feed ingredients, the nutritional suitability of different feed ingredients for the target herbivores is determined. Based on all the nutritional suitability, all feed ingredients in the feed formula are combined and optimized to obtain the optimal feed formulation scheme for the target herbivorous species.

2. The method as described in claim 1, characterized in that, The process of performing multiple masking operations on the feed ingredient nodes in the heterogeneous graph, and then determining the correlation matrix of the synergistic effects among the remaining feed ingredient nodes after each masking, specifically includes: Determine the occlusion ratio of the feed ingredient nodes in the heterogeneous graph; The feed ingredient nodes in the heterogeneous graph are masked multiple times according to the masking ratio, and the remaining heterogeneous graph after each masking is recorded. Select a remaining heterogeneous graph as the selected remaining heterogeneous graph, and determine the association information between each feed ingredient node and other feed ingredient nodes in the selected remaining heterogeneous graph; Based on all the association information, determine the association matrix of the synergistic effects between feed ingredient nodes in the selected remaining heterogeneous graph; Continue to determine the correlation matrix of the synergistic effects among the feed ingredient nodes in the remaining heterogeneous graph.

3. The method as described in claim 1, characterized in that, Based on all the correlation matrices, the compatibility vector between each feed ingredient and other feed ingredients in ingredient formulation is calculated using a loss function that maximizes mutual information. Specifically, this includes: Standardize all the correlation matrices to obtain the standardized correlation matrices. Select one feed ingredient as the selected feed ingredient, and extract the correlation numerical sequence between the selected feed ingredient and other feed ingredients from all standardized correlation matrices; Construct a loss function that maximizes mutual information; Input the numerical sequence of the correlation between the selected feed ingredient and other feed ingredients into the loss function that maximizes mutual information, and output the compatibility vector of the selected feed ingredient and other feed ingredients when the ingredients are matched. Continue to determine the compatibility vector between the remaining feed ingredients and other feed ingredients when formulating them.

4. The method as described in claim 1, characterized in that, The baseline nutrient requirement characteristics extracted from the heterogeneous graph for nutrient transfer from nutrient component nodes to herbivore growth stage nodes specifically include: The direct association paths between nutrient nodes and herbivore growth stage nodes are extracted from the heterogeneous graph. The required amounts of each nutrient at each stage of the growth process of herbivores are determined based on the heterogeneous diagram. Based on all the demand corresponding to the herbivore growth stage nodes and the directly associated paths, determine the baseline nutrient demand characteristics when nutrient component nodes transfer nutrients to herbivore growth stage nodes.

5. The method as described in claim 1, characterized in that, Based on the aforementioned baseline nutritional requirements characteristics and combined with the initial growth stage information of the target herbivores, the nutritional requirement satisfaction rate of the target herbivores at each growth stage is determined, specifically including: Obtain information on the initial growth stage of the target herbivore; The nutrient component feature vectors of the growth stage nodes are determined based on the aforementioned baseline nutrient requirement characteristics. The initial growth stage information and the nutrient component feature vector are fused to determine the nutrient requirement satisfaction rate of the target herbivore at each growth stage.

6. The method as described in claim 1, characterized in that, By fusing and analyzing all nutritional requirement fulfillment rates with specific baseline data corresponding to the target herbivore species, a vector of nutritional requirement parameters for the target herbivore at different growth stages is generated, specifically including: Obtain specific basic data on the target herbivorous animal species; By correlating and aggregating all nutritional requirement satisfaction rates with the specific basic data of the target herbivorous animal species, we can obtain aggregated information on the relationship between the species characteristics and nutritional suitability of the target herbivorous animal. Based on the aggregated information, a vector of nutritional requirements parameters for the target herbivore at different growth stages is generated.

7. The method as described in claim 1, characterized in that, Determining the nutritional suitability of different feed ingredients for the target herbivore by combining the fit vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients specifically includes: Determine the nutritional characteristics of all feed ingredients; The matching degree between the nutritional components of each feed ingredient and the specific needs of herbivorous animal breeds is determined based on the nutritional requirement parameter vector and the nutritional component characteristics. The nutritional suitability of each feed ingredient for the target herbivore is determined based on the matching degree between the nutritional composition of each feed ingredient and the specific needs of the herbivore species, and the fit vector.

8. The method as described in claim 1, characterized in that, Based on all nutritional compatibility factors, the combination and optimization of all feed ingredients in the feed formulation are performed to obtain the optimal feed formulation for the target herbivore species. Specifically, this includes: Obtain the initial feed formulation plan for all feed ingredients in the feed formula; The initial feed formulation scheme is optimized and adjusted based on all nutritional compatibility factors to obtain the optimal feed formulation scheme for the target herbivore species.

9. The method as described in claim 1, characterized in that, Obtain initial growth stage information of the target herbivores from the breeding management system database.

10. A system for optimizing feed formulation for herbivores based on heterogeneous graphs, characterized in that, include: The module is used to construct a heterogeneous graph for herbivore feed formulation, with the target herbivore species, herbivore growth stage, feed ingredients and nutrients as nodes. The processing module is used to perform multiple masking processes on the feed ingredient nodes in the heterogeneous graph, thereby determining the correlation matrix of the synergistic effect between the remaining feed ingredient nodes after each masking, and calculating the compatibility vector of each feed ingredient with other feed ingredients in the raw material matching based on all the correlation matrices and through the loss function that maximizes mutual information. The processing module is further configured to extract the baseline nutritional requirement features from the heterogeneous graph when transferring nutrients from the nutrient component nodes to the growth stage nodes of the herbivores, determine the nutritional requirement satisfaction rate of the target herbivores at each growth stage based on the baseline nutritional requirement features and the initial growth stage information of the target herbivores, and fuse all the nutritional requirement satisfaction rates with the specific basic data corresponding to the target herbivores species to generate a nutritional requirement parameter vector of the target herbivores at different growth stages. The execution module is used to determine the nutritional suitability of different feed ingredients for the target herbivorous animal by combining the fit degree vector and the nutritional requirement parameter vector with the nutritional component characteristics of all feed ingredients, and to optimize the combination of all feed ingredients in the feed formula based on all nutritional suitability to obtain the optimal feed formulation scheme for the target herbivorous animal species.

Images