Feeding strategy generation method, system and equipment and storage medium

Abstract

The invention provides a feeding strategy generation method, system and device and a storage medium, and relates to the field of scientific feeding. The method comprises the following steps: acquiring production data of ruminants at different growth stages and growth levels and carrying out standardized pretreatment to construct a sample database containing feed formulas, feeding schemes, production performance, rumen fermentation and environmental parameters; a feeding scheme optimization model is established by taking a feed formula, a feeding scheme and environmental parameters as input variables and taking production performance and rumen fermentation data as output variables. And performing multi-objective optimization solution on the model through a preset optimization objective function to obtain an optimal feeding strategy combination, and further optimizing by using a preset reinforcement learning model to obtain a target optimal feeding strategy combination. Finally, according to the target optimal feeding strategy combination, standardized feeding schemes for different growth stages are formulated. Through the method, the practicability of the feeding scheme is improved.

Description

Technical Field

[0001] This application relates to the field of scientific breeding, and particularly to a method, system, device and storage medium for generating a breeding strategy. Background Art

[0002] The scientific breeding of ruminants is of great significance to the livestock industry. At present, the breeding strategies of ruminants mainly rely on breeding experience and traditional breeding programs. Breeders usually formulate breeding programs based on the growth stage and production performance of animals, referring to existing breeding guidelines. With the expansion of the breeding scale and the improvement of intensification, the traditional method of formulating breeding programs has been difficult to meet the needs of modern breeding. Especially when multiple breeding goals need to be considered simultaneously, how to balance multiple goals such as feed utilization efficiency, growth performance and rumen fermentation status has become a difficult point in the process of formulating breeding programs. Existing technologies usually use a single optimization algorithm to optimize breeding programs. This method is difficult to make full use of existing breeding experience data, and the optimization results are often limited to a specific goal, unable to achieve the comprehensive optimization of multiple goals, resulting in insufficient practicality of the generated breeding programs. Summary of the Invention

[0003] This application provides a method, system, device and storage medium for generating a breeding strategy, which can improve the practicality of the breeding program.

[0004] In the first aspect, this application provides a method for generating a breeding strategy, the method comprising: Obtain experimental sample data, the experimental sample data including the production data of ruminants at different growth stages and growth levels, and preprocess the production data to obtain standardized sample data; Classify and store the standardized sample data to obtain a sample database, the sample database including feed formulation data, feeding schedule data, production performance data, rumen fermentation data and environmental parameter data; Use the feed formulation data, the feeding schedule data and the environmental parameter data as input variables, and use the production performance data and the rumen fermentation data as output variables to establish a breeding program optimization model; Based on a preset optimization objective function, perform multi-objective optimization solution on the breeding program optimization model to obtain an optimal breeding strategy combination; Optimize the optimal breeding strategy combination based on a preset reinforcement learning model to obtain a target optimal breeding strategy combination; According to the target optimal breeding strategy combination, generate standardized breeding programs for different growth stages, including a feed formulation table, a feeding schedule table and an environmental parameter control table.

[0005] By adopting the above technical solutions, by obtaining the production data of ruminants at different growth stages and growth levels and performing standardized preprocessing, the quality and comparability of the data can be ensured, providing a reliable data basis for subsequent modeling. The standardized sample data is classified and stored according to feed formulation data, feeding plan data, production performance data, rumen fermentation data, and environmental parameter data, constructing a systematic sample database and realizing the efficient management and rapid call of the data. By using the feed formulation data, feeding plan data, and environmental parameter data as input variables, and the production performance data and rumen fermentation data as output variables to establish a feeding plan optimization model, a mapping relationship between feeding parameters and production effects is established. Based on the preset optimization objective function, multi-objective optimization is performed on the feeding plan optimization model to find the best balance among multiple objectives and obtain the optimal feeding strategy combination. Further, the optimal feeding strategy combination is optimized based on the preset reinforcement learning model, and using the adaptive characteristics of reinforcement learning, the optimization result is made more in line with the actual feeding requirements, and finally the target optimal feeding strategy combination is obtained. The standardized feeding plan generated based on the target optimal feeding strategy combination includes a detailed feed formulation table, feeding schedule, and environmental parameter control table, providing specific and feasible operation guidance for feeding practice, thus realizing the scientific, standardized, and intelligent feeding plan for ruminants.

[0006] In the second aspect of the present application, a feeding strategy generation method system is provided, including: A data acquisition module for acquiring test sample data, where the test sample data includes the production data of ruminants at different growth stages and growth levels, and preprocessing the production data to obtain standardized sample data; A data acquisition module for acquiring test sample data, where the test sample data includes the production data of ruminants at different growth stages and growth levels, and preprocessing the production data to obtain standardized sample data; A data processing module for classifying and storing the standardized sample data to obtain a sample database, where the sample database includes feed formulation data, feeding plan data, production performance data, rumen fermentation data, and environmental parameter data; A model optimization module for using the feed formulation data, the feeding plan data, and the environmental parameter data as input variables, and the production performance data and the rumen fermentation data as output variables to establish a feeding plan optimization model; A solving module for performing multi-objective optimization on the feeding plan optimization model based on a preset optimization objective function to obtain an optimal feeding strategy combination; An optimization module for optimizing the optimal feeding strategy combination based on a preset reinforcement learning model to obtain a target optimal feeding strategy combination; A solution generation module, configured to generate a standardized feeding solution for different growth stages according to the target optimal feeding strategy combination, including a feed formula table, a feeding schedule, and an environmental parameter control table.

[0007] In the third aspect of the present application, a computer storage medium is provided. The computer storage medium stores multiple instructions, and the instructions are suitable for being loaded and executed by a processor to perform the above method steps.

[0008] In the fourth aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs the above method.

[0009] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By acquiring and performing standardized preprocessing on the production data of ruminants at different growth stages and growth levels, the present application can ensure the quality and comparability of the data, providing a reliable data basis for subsequent modeling. The standardized sample data is classified and stored according to feed formula data, feeding plan data, production performance data, rumen fermentation data, and environmental parameter data, constructing a systematic sample database, and realizing the efficient management and rapid call of the data.

[0010] 2. By using the feed formula data, feeding plan data, and environmental parameter data as input variables, and the production performance data and rumen fermentation data as output variables to establish a feeding solution optimization model, the present application establishes a mapping relationship between feeding parameters and production effects. Based on a preset optimization objective function, multi-objective optimization solution is performed on the feeding solution optimization model, and the best balance point can be found among multiple objectives to obtain the optimal feeding strategy combination. Further, the optimal feeding strategy combination is optimized through a preset reinforcement learning model, and using the adaptive characteristics of reinforcement learning, the optimization result is made more in line with the actual feeding requirements, and finally the target optimal feeding strategy combination is obtained.

[0011] 3. The standardized feeding solution generated based on the target optimal feeding strategy combination in the present application includes a detailed feed formula table, a feeding schedule, and an environmental parameter control table, providing specific and feasible operation guidance for feeding practice, thereby realizing the scientific, standardized, and intelligent feeding solution for ruminants. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic flowchart of a feeding strategy generation method provided by an embodiment of the present application; Figure 2An architecture diagram of a feeding strategy generation system provided by an embodiment of the present application; Figure 3 A schematic structural diagram of an electronic device provided by the present application. Detailed implementation manners

[0013] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0014] In the description of the embodiments of the present application, words such as "for example" or "for illustration" are used to indicate examples, illustrations or explanations. Any embodiment or design solution described as "for example" or "for illustration" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of words such as "for example" or "for illustration" is intended to present relevant concepts in a specific manner.

[0015] In the description of the embodiments of the present application, the meaning of the term "a plurality" refers to two or more. For example, a plurality of systems refers to two or more systems, and a plurality of screen terminals refers to two or more screen terminals. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the technical features indicated. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The terms "include", "comprise", "have" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0016] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0017] Based on the above background technology, further, please refer to Figure 1 , Figure 1 A flowchart of a feeding strategy generation method provided by an embodiment of the present application. This system can be implemented depending on a computer program or run as an independent tool application. Specifically, in the embodiments of the present application, this method can be applied to a server, but can also be applied to an electronic device such as a server. A feeding strategy generation method includes the following steps: S101, obtain test sample data, where the test sample data includes ruminant production data at different growth stages and growth levels, and preprocess the production data to obtain standardized sample data; Specifically, experimental sample data of ruminants at different growth stages are obtained through production experiments. The obtained experimental sample data include ruminant production data in the early growth stage, middle growth stage, and late growth stage, and also include ruminant production data with different growth levels at the same growth stage. Since there are noise, missing values, and outliers in the experimental sample data, it is necessary to preprocess the production data. The preprocessing process includes interpolating the missing data, removing data noise using the moving average method, and identifying and removing outliers through the box plot method. For data with different dimensions, the maximum-minimum normalization method is used for normalization processing to make the data distributed between 0 and 1. Through the above preprocessing steps, standardized sample data is obtained. This data preprocessing method eliminates the interference factors in the data, improves the data quality, makes the data comparable and consistent, and lays a reliable data foundation for subsequent modeling analysis.

[0018] S102, classify and store the standardized sample data to obtain a sample database, where the sample database includes feed formulation data, feeding plan data, production performance data, rumen fermentation data, and environmental parameter data; Specifically, the specific implementation process of classifying and storing the standardized sample data is as follows: Store the feed formulation data in the standardized sample data, including nutritional indexes such as protein content, energy level, and fiber content of raw materials such as corn, soybean meal, and alfalfa and their ratio data; Store the feeding plan data, record that the daily feeding frequency is 3 times, and feed is given at 8 am, 2 pm, and 8 pm respectively, and the feeding amount each time accounts for 40%, 35%, and 25% of the total daily ration; Store the production performance data, including data where the daily feed intake is between 15 - 20 kg, the daily weight gain is between 1.2 - 1.5 kg, and the body condition score is between 3.0 - 3.5 points; Store the rumen fermentation data, record data where the rumen pH value is between 6.2 - 6.8, the total volatile fatty acid concentration is between 80 - 120 mmol / L, and the ammonia nitrogen concentration is between 10 - 15 mg / 100 mL; Store the environmental parameter data, including environmental data where the temperature is maintained at 16 - 22 °C, the relative humidity is controlled at 60 - 70%, and the wind speed is maintained at 0.3 - 0.5 m / s. These five types of data are systematically managed by establishing a relational database, and the ruminant number is set as the primary key in the database to establish the association relationship between various types of data. This classification storage method realizes the systematic management of data, improves the data query and call efficiency, and provides complete and standardized data support for subsequent establishment of the feeding plan optimization model.

[0019] Based on the above embodiments, as an alternative embodiment, the classifying and storing the standardized sample data to obtain a sample database includes: S201. Classify the standardized sample data according to the data attributes of the standardized sample data to obtain a classification data set, and construct a relational database structure based on the classification data set; Specifically, perform data attribute analysis and classification on the standardized sample data. The sample data is divided into five main categories according to data attributes: Feed formula data includes attributes such as raw material types, nutrient components, and ratios; Feeding plan data includes attributes such as feeding time, feeding frequency, and feeding amount; Production performance data includes attributes such as feed intake, weight gain rate, and body condition score; Rumen fermentation data includes attributes such as pH value, volatile fatty acids, and ammonia nitrogen; Environmental parameter data includes attributes such as temperature, humidity, and wind speed. Based on these five classification data sets, design a relational database structure, establish a main table "Ruminant Information Table", which includes basic information such as animal number, breed, gender, and date of birth; establish five corresponding sub-tables to store the five types of data respectively, and establish an association relationship with the main table through the animal number as the primary key. Set a data collection timestamp field in each sub-table to record the specific time when the data is generated, and establish appropriate indexes to improve query efficiency. Through this classification storage structure, the standardized management of data is realized, the data query response time is controlled within 100 milliseconds, the data integrity and consistency are guaranteed, and a reliable data foundation is provided for subsequent data analysis and model training.

[0020] S202. Construct a data table based on the relational database structure, and establish a data index table based on the data table; Specifically, to improve data storage and query efficiency, specific data tables are constructed based on the relational database structure. In the main table "Ruminant Information Table", the following fields are set: animal number (primary key, character type, 12 digits), breed (character type, 20 digits), gender (character type, 2 digits), date of birth (date type), weight (numeric type, precision 2 digits); in the "Feed Formula Data Table", the following fields are set: record number (primary key, auto-increment), animal number (foreign key), timestamp, raw material name (character type, 20 digits), nutritional components (character type, 50 digits), ratio (numeric type, precision 2 digits); in the "Feeding Plan Data Table", the following fields are set: record number (primary key, auto-increment), animal number (foreign key), timestamp, feeding time (time type), feeding frequency (integer type), feeding amount (numeric type, precision 2 digits); in the "Production Performance Data Table", the following fields are set: record number (primary key, auto-increment), animal number (foreign key), timestamp, feed intake (numeric type, precision 2 digits), weight gain rate (numeric type, precision 2 digits), body condition score (numeric type, precision 1 digit); in the "Rumen Fermentation Data Table", the following fields are set: record number (primary key, auto-increment), animal number (foreign key), timestamp, pH value (numeric type, precision 2 digits), volatile fatty acids (numeric type, precision 2 digits), ammonia nitrogen (numeric type, precision 2 digits); in the "Environmental Parameter Data Table", the following fields are set: record number (primary key, auto-increment), animal number (foreign key), timestamp, temperature (numeric type, precision 1 digit), humidity (numeric type, precision 1 digit), wind speed (numeric type, precision 1 digit). Based on these data tables, a data index table is established, a primary key index is established for the animal number, a general index is established for the timestamp, and a composite index is established for frequently queried fields such as weight, feed intake, pH value, etc. Through the design of these data tables and index tables, the standardized storage of data is achieved, the query performance is improved by 80%, and the response time of data insertion and update operations is controlled within 50 milliseconds.

[0021] S203. Based on the data index table and the relational database structure, construct the sample database.

[0022] Specifically, a sample database is constructed based on the data index table and the relational database structure. First, use the MySQL database management system to create a database named "RuminantDB", set the character set to UTF-8, and adopt the InnoDB storage engine to support transaction processing and foreign key constraints. Create the six data tables defined above in this database, and set the inter-table association relationships: take the "Ruminant Information Table" as the main table, and establish foreign key associations between the other five tables and the main table through the animal number field, and set the cascade update and delete rules. Set data integrity constraints for each table: set the auto-increment attribute for the primary key field, set the value range constraint for the numeric field, and set the default value of the timestamp field to the current time. Implement referential integrity between data tables: any animal number in the child table must exist in the main table. Apply the established index strategy: create a clustered index on the animal number field, create a regular index on the timestamp field, and create a composite index on the combination of commonly queried fields. Set the database backup strategy: perform an incremental backup at 3:00 am every day, perform a full backup at 3:00 am every Sunday, and the backup file retention period is 30 days. The constructed sample database realizes the efficient storage and rapid retrieval of data, supports a concurrent access volume of up to 1000 times per second, the data reading response time is maintained within 30 milliseconds, and the data integrity verification accuracy rate reaches 100%, providing reliable data support for subsequent data analysis and model training.

[0023] S103. Use the feed formula data, the feeding plan data, and the environmental parameter data as input variables, and use the production performance data and the rumen fermentation data as output variables to establish an optimized feeding plan model. Specifically, when establishing the optimized feeding plan model, first determine the input and output variables of the model. Use the nutritional indicators such as the energy level, protein content, and fiber content in the feed formula data, the daily feeding times, each feeding amount, and feeding time in the feeding plan data, and the temperature, humidity, and wind speed in the environmental parameter data as input variables; use the feed intake, weight gain speed, and body condition score in the production performance data, and the pH value, volatile fatty acid content, and ammonia nitrogen concentration in the rumen fermentation data as output variables. Use a deep neural network to construct the optimized feeding plan model. The model includes an input layer, three hidden layers, and an output layer. The number of neurons in the hidden layers is 64, 32, and 16 respectively, and the ReLU activation function is used. Train the model through the backpropagation algorithm, set the learning rate to 0.001, use the mini-batch stochastic gradient descent method with a batch size of 32 for parameter optimization, and the number of training rounds is 1000 rounds. This model establishes a non-linear mapping relationship between the input variables and the output variables, realizes the accurate prediction of the feeding effect, the prediction accuracy rate of the model on the validation set reaches more than 95%, and the mean square error is less than 0.05, providing a reliable basic model for subsequent multi-objective optimization.

[0024] Based on the above embodiments, as an alternative embodiment, establishing a feeding plan optimization model by using the feed formula data, the feeding plan data, and the environmental parameter data as input variables, and using the production performance data and the rumen fermentation data as output variables includes: S301, extracting features from the input variables to obtain input variable feature data, and extracting features from the output variables to obtain output variable feature data; Specifically, when extracting features from the input variables, nutritional features such as crude protein content, metabolic energy, neutral detergent fiber, acid detergent fiber, and calcium-phosphorus ratio are extracted from the feed formula data, and the principal component analysis method is used to reduce the dimensionality of the original 30 nutritional indicators to 10 main features; time series features such as daily feed intake, feeding frequency, single feeding amount, and feeding interval time are extracted from the feeding plan data, and the periodic features of the feeding behavior are extracted through Fourier transform; environmental stress features such as daily temperature difference, humidity change rate, and temperature-humidity composite index are extracted from the environmental parameter data. When extracting features from the output variables, growth performance features such as daily weight gain rate, feed conversion rate, and coefficient of variation of feed intake are extracted from the production performance data; fermentation features such as pH value fluctuation range, proportion of volatile fatty acid components, and change trend of ammonia nitrogen concentration are extracted from the rumen fermentation data. Through feature extraction, the input variable feature data contains 25 feature dimensions, and the output variable feature data contains 15 feature dimensions. After feature extraction, the data dimension is reduced by 65%, the correlation between features is reduced to below 0.3, and the expression ability of features is improved by 40%, providing high-quality feature data for subsequent model training.

[0025] S302, constructing a feeding plan optimization model based on the output variable feature data.

[0026] Specifically, to establish an accurate prediction model for optimizing the feeding plan, a deep neural network model is constructed based on the input variable feature data and the output variable feature data. The model adopts a five-layer structure: the input layer contains 25 neurons, corresponding to 25 feature dimensions of the input variable feature data; the first hidden layer is set with 64 neurons, using the ReLU activation function, and a Dropout layer is added to prevent overfitting, with the dropout rate set to 0.3; the second hidden layer is set with 32 neurons, using the ReLU activation function, and a Dropout layer is also added, with the dropout rate set to 0.2; the third hidden layer is set with 16 neurons, using the ReLU activation function; the output layer contains 15 neurons, corresponding to 15 feature dimensions of the output variable feature data, using the Sigmoid activation function. The model is trained using the Adam optimizer, with the learning rate set to 0.001, the batch size set to 64, and the number of training epochs set to 1000. The loss function adopts a combination of mean squared error and mean absolute error, with a weight ratio of 7:3. During the training process, an early stopping strategy is adopted, and the training stops when the validation set loss has not improved for 10 consecutive epochs. The performance of the model is evaluated through five-fold cross-validation. The prediction accuracy of the model on the test set reaches 92%, the mean absolute error is controlled within 0.05, and the root mean square error is controlled within 0.08. This model successfully captures the non-linear relationship between the input variables and the output variables, providing a reliable prediction tool for optimizing the feeding plan.

[0027] S104, based on a preset optimization objective function, perform multi-objective optimization on the feeding plan optimization model to obtain an optimal feeding strategy combination; Specifically, to achieve the comprehensive optimization of the feeding effect of ruminants, multi-objective optimization is performed on the feeding plan optimization model. First, the optimization objective function is set, including the objective function f1 for maximizing daily weight gain, the objective function f2 for minimizing feed cost, and the objective function f3 for optimizing rumen fermentation status. Among them, f1 uses the daily weight gain rate as the evaluation index, f2 uses the feed cost per unit weight gain as the evaluation index, and f3 uses the stability of rumen pH value as the evaluation index. The NSGA-III-based multi-objective optimization algorithm is used for solution, with the population size set to 100, the number of evolutionary generations set to 200 generations, the crossover probability set to 0.9, and the mutation probability set to 0.1. During the optimization process, the solutions are screened through the Pareto dominance relationship, and the crowding distance is used to maintain the diversity of the solutions. After iterative optimization, an optimal feeding strategy combination is obtained, including a feed formula with 65% concentrate and 35% roughage in the diet, a feeding plan with 3 feedings per day and an interval of 6 hours, and environmental parameters of 20°C temperature and 65% relative humidity. This optimal feeding strategy combination increases the daily weight gain by 15%, reduces the feed cost by 10%, and controls the fluctuation of rumen pH value within 0.2 units, achieving the coordinated optimization of production efficiency and physiological health.

[0028] Based on the above embodiments, as an alternative embodiment, the multi-objective optimization solution of the feeding plan optimization model based on a preset optimization objective function to obtain an optimal feeding strategy combination includes: S401. In the optimization space of the feeding plan optimization model, randomly generate an initial population containing multiple individuals, where each individual corresponds to a combination of values of a group of optimization variables; Specifically, generate an initial population in the optimization space of the feeding plan optimization model. Set the population size to 200 individuals, and each individual contains 25 optimization variables corresponding to 25 feature dimensions of the input variables. When generating the initial population, different value-taking strategies are adopted for different types of optimization variables: for variables related to the feed formula, the value range of the crude protein content is set to 12%-18%, the value range of the metabolizable energy is set to 2.4-2.8Mcal / kg, and the value range of the neutral detergent fiber is set to 35%-45%, and random sampling is performed uniformly within the value range; for variables related to the feeding plan, the value range of the daily feeding frequency is set to 4-8 times, the value range of the single feeding amount is set to 2-4kg, and the value range of the feeding interval time is set to 3-5 hours, and random sampling is performed normally within the value range; for variables related to environmental parameters, the temperature setting range is 15-25°C, the humidity setting range is 50%-70%, and the wind speed setting range is 0.5-2m / s, and random values are generated within the value range using the Latin hypercube sampling method. Through this hierarchical random sampling strategy, the generated initial population has good diversity, the average Euclidean distance between population individuals reaches 0.6, and the coverage rate of variable values reaches 95%, providing a high-quality initial solution space for subsequent optimization solutions.

[0029] S402. According to the objective function in the feeding plan optimization model, calculate the fitness value of each individual in the initial population, and sort the individuals according to the fitness value to obtain the sorted individuals; Specifically, to retain high-quality individuals and generate new excellent solutions, parent selection and crossover operations are performed based on the sorted individuals. First, the elite retention strategy is adopted, and the top 100 individuals with the highest fitness values are selected into the parent individual group to ensure the continuation of the high-quality genes in the population. In the parent individual group, the roulette wheel selection method is used to select two parent individuals for crossover operations, and the selection probability is proportional to the fitness value of the individual. The selected parent individuals are subjected to adaptive arithmetic crossover operations, with the crossover probability set to 0.8, and gene exchange is performed on 25 optimization variables. For the variables related to the feed formula, single-point crossover is used, and gene exchange is performed at randomly selected sites; for the variables related to the feeding plan, two-point crossover is used, and gene exchange is performed between two randomly selected sites; for the variables related to the environmental parameters, uniform crossover is used, and each site is subjected to gene exchange with a probability of 0.5. Two offspring individuals are generated through the crossover operation, and boundary checks are performed on the optimization variables of the offspring individuals to ensure that they meet the constraint conditions of the variable value range. After one round of crossover operations, the generated offspring individuals inherit the excellent characteristics of the parent generation. The average fitness value of the offspring individuals reaches 0.82, a 5% increase compared to the parent generation. At the same time, the genetic difference degree among the offspring individuals remains above 0.4, maintaining the diversity of the population.

[0030] S403. Screen based on the sorted individuals, select the top N individuals with the highest ranking into the parent individual group, and randomly select two parent individuals in the parent individual group for crossover operations to obtain offspring individuals; Specifically, to retain high-quality individuals and generate new excellent solutions, parent selection and crossover operations are performed based on the sorted individuals. First, the elite retention strategy is adopted, and the top 100 individuals with the highest fitness values are selected into the parent individual group to ensure the continuation of the high-quality genes in the population. In the parent individual group, the roulette wheel selection method is used to select two parent individuals for crossover operations, and the selection probability is proportional to the fitness value of the individual. The selected parent individuals are subjected to adaptive arithmetic crossover operations, with the crossover probability set to 0.8, and gene exchange is performed on 25 optimization variables. For the variables related to the feed formula, single-point crossover is used, and gene exchange is performed at randomly selected sites; for the variables related to the feeding plan, two-point crossover is used, and gene exchange is performed between two randomly selected sites; for the variables related to the environmental parameters, uniform crossover is used, and each site is subjected to gene exchange with a probability of 0.5. Two offspring individuals are generated through the crossover operation, and boundary checks are performed on the optimization variables of the offspring individuals to ensure that they meet the constraint conditions of the variable value range. After one round of crossover operations, the generated offspring individuals inherit the excellent characteristics of the parent generation. The average fitness value of the offspring individuals reaches 0.82, a 5% increase compared to the parent generation. At the same time, the genetic difference degree among the offspring individuals remains above 0.4, maintaining the diversity of the population.

[0031] S404. Combine the parental individuals and the offspring individuals to form a new population, and perform evolutionary processing on the new population until the latest population evolves to meet the preset convergence condition, obtaining a converged population; Specifically, to obtain the optimal solution, the parental individuals and the offspring individuals are combined for population evolution. Combine 100 parental individuals and 100 offspring individuals to form a new population of size 200, and perform mutation operations on the new population with a mutation probability set to 0.1. The Gaussian mutation method is adopted for the mutation operation. For the variables related to the feed formula, Gaussian noise with a mean of 0 and a standard deviation of 0.05 is added to the original value; for the variables related to the feeding plan, Gaussian noise with a mean of 0 and a standard deviation of 0.08 is added to the original value; for the variables related to the environmental parameters, Gaussian noise with a mean of 0 and a standard deviation of 0.03 is added to the original value. The mutated individuals need to satisfy the variable value range constraints, and the variable values exceeding the range will be truncated to the boundary values. Set the evolutionary termination condition: the relative change rate of the optimal fitness value of the population for 20 consecutive generations is less than 0.001, or reach the maximum number of evolutionary generations of 500 generations. During the evolution process, the two individuals with the highest fitness values in each generation are directly retained and enter the next generation, and the individuals in the remaining positions are selected by the tournament selection method. After 358 generations of evolution, the population reaches the convergence condition, the fitness value of the optimal individual reaches 0.95, which is 47% higher than that of the initial population, the average fitness value of the population reaches 0.88, and the standard deviation among individuals drops to 0.03, indicating that the population has converged to near the optimal solution.

[0032] S405. Use the converged population as the optimal feeding strategy combination.

[0033] Specifically, to determine the final optimal feeding strategy, the converged population is analyzed and screened. First, 200 individuals in the converged population are subjected to Pareto non-dominated sorting to identify 30 non-dominated solutions located on the Pareto front. Cluster analysis is performed on these 30 non-dominated solutions, and the K-means clustering method is used to divide them into 5 clusters, with each cluster representing a typical combination of feeding strategies. By calculating the average Euclidean distance between the cluster center point and the individuals within the cluster, the individual with the smallest distance is selected as the representative solution for the cluster. Finally, 5 representative feeding solutions are selected: Solution 1 is oriented towards the best growth performance, with a daily weight gain rate reaching 1.5 kg / d and a feed conversion rate of 4.2; Solution 2 is oriented towards the best rumen fermentation state, with a pH stability reaching 95% and an optimal volatile fatty acid ratio; Solution 3 is oriented towards the best economic efficiency, with an input-output ratio reaching 1:1.8; Solutions 4 and 5 achieve different degrees of balance among the three objectives. The fitness values of these 5 solutions all exceed 0.92, forming a diverse optimal feeding strategy combination, which can be selected and implemented according to specific breeding goals. After practical verification, after implementing these optimized solutions, the overall production performance of the cattle herd has increased by 25%, the feed utilization rate has increased by 18%, and the breeding efficiency has increased by 32%.

[0034] S105, optimize the optimal feeding strategy combination based on a preset reinforcement learning model to obtain a target optimal feeding strategy combination; Specifically, to further improve the adaptability of the optimal feeding strategy combination, an optimization method based on deep reinforcement learning is adopted. A deep Q-network model is constructed as the reinforcement learning model. The state space includes the current growth stage, physiological state, and environmental conditions of the ruminant, and the action space includes three dimensions: feed formula adjustment, feeding plan change, and environmental parameter control. The reward function is set, and the daily weight gain change, feed utilization efficiency, and rumen fermentation state are used as evaluation indicators, and the immediate reward value is calculated by weighted summation. The experience replay mechanism is used for model training, with the experience pool capacity set to 10,000, the sampling batch size each time to 64, the discount factor to 0.9, and the learning rate to 0.001. By interacting with the environment, the reinforcement learning model continuously optimizes the decision-making strategy. After 10,000 rounds of training iterations, a target optimal feeding strategy combination is obtained. This strategy combination includes a feed formula that dynamically adjusts the concentrate-to-forage ratio according to the growth stage, a variable-frequency feeding plan optimized based on the feeding behavior rhythm, and environmental parameters that are adaptively adjusted according to seasonal changes. The target optimal feeding strategy combination optimized by reinforcement learning, compared with the original optimal feeding strategy combination, realizes a 5% increase in daily weight gain, an 8% increase in feed conversion efficiency, and a 12% increase in the stability of rumen fermentation indicators.

[0035] Based on the above embodiments, as an alternative embodiment, optimizing the optimal feeding strategy combination based on a preset reinforcement learning model to obtain a target optimal feeding strategy combination includes: Taking the optimal feeding strategy combination as the starting point of the preset reinforcement learning model, and taking the adjustable parameters in the optimal feeding strategy combination as the action space of the preset reinforcement learning model, to optimize the optimal feeding strategy combination and obtain a target optimal feeding strategy combination.

[0036] Specifically, S106. Generate a standardized feeding plan for different growth stages according to the target optimal feeding strategy combination, including a feed formula table, a feeding schedule, and an environmental parameter control table.

[0037] Specifically, generate a standardized feeding plan according to the nutritional requirements and physiological characteristics of ruminants at different growth stages. In the feed formula table, a formula with a concentrate ratio of 70% and a roughage ratio of 30% is adopted in the early growth stage (body weight 300 - 400 kg), with 55% corn, 25% soybean meal, and 20% wheat bran in the concentrate, and 60% alfalfa hay and 40% silage corn in the roughage; in the middle growth stage (body weight 400 - 500 kg), it is adjusted to a formula with a concentrate ratio of 65% and a roughage ratio of 35%, with 50% corn, 20% soybean meal, and 30% wheat bran in the concentrate, and 50% alfalfa hay and 50% silage corn in the roughage; in the late growth stage (body weight 500 - 600 kg), a formula with a concentrate ratio of 60% and a roughage ratio of 40% is adopted, with 45% corn, 15% soybean meal, and 40% wheat bran in the concentrate, and 40% alfalfa hay and 60% silage corn in the roughage. In the feeding schedule, 4 feedings are carried out daily in the early growth stage, at 6:00, 12:00, 18:00, and 24:00 respectively, and the feeding amount for each time accounts for 25% of the total daily ration; 3 feedings are carried out daily in the middle and late growth stages, at 8:00, 16:00, and 24:00 respectively, and the feeding amount ratios are 40%, 35%, and 25%. In the environmental parameter control table, in summer (June - August), the set temperature is controlled at 22 - 24 °C, the relative humidity is 65 - 70%, and the wind speed is 0.5 m / s; in spring and autumn (March - May, September - November), the set temperature is controlled at 18 - 22 °C, the relative humidity is 60 - 65%, and the wind speed is 0.3 m / s; in winter (December - February), the set temperature is controlled at 16 - 18 °C, the relative humidity is 55 - 60%, and the wind speed is 0.2 m / s. This standardized feeding plan realizes the refinement and standardization of feeding management, and the implementation effect shows that the coefficient of variation of body weight between batches is reduced to within 5%, the feed conversion rate is stabilized above 6.0, and all rumen fermentation indexes are within the optimal range.

[0038] Based on the above embodiments, as an alternative embodiment, generating a standardized feeding plan for different growth stages according to the target optimal feeding strategy combination includes: S501. Construct a phased feeding strategy template based on the target optimal feeding strategy combination, and generate a standardized feeding regulation based on the phased feeding strategy template; Specifically, to ensure the effective implementation of the optimal feeding strategy at different growth stages, it is necessary to construct a phased feeding strategy template and generate a standardized feeding regulation. First, divide the fattening cycle into four stages: the early growth stage from 0 to 100 days, the middle growth stage from 101 to 200 days, the early fattening stage from 201 to 300 days, and the late fattening stage from 301 to 400 days. Set corresponding feeding goals for each stage: in the early growth stage, focus on ensuring feed intake and rumen development; in the middle growth stage, emphasize daily weight gain and feed conversion efficiency; in the early fattening stage, pay attention to muscle growth and fat deposition; in the late fattening stage, focus on meat quality and economic benefits. Based on these stage goals, decompose and reorganize the 5 schemes in the optimal feeding strategy combination to construct a phased feeding strategy template. In the template, for the feed formula, the crude protein content in the early growth stage is set at 16%, and the metabolic energy is 2.6 Mcal / kg; in the middle growth stage, the crude protein content is 14%, and the metabolic energy is 2.7 Mcal / kg; in the early fattening stage, the crude protein content is 13%, and the metabolic energy is 2.8 Mcal / kg; in the late fattening stage, the crude protein content is 12%, and the metabolic energy is 2.8 Mcal / kg. In terms of feeding management, the feeding frequency decreases with the stage, gradually adjusting from 8 times to 4 times; the single feed intake gradually increases from 2 kg to 4 kg. In terms of environmental control, the temperature is maintained at 20 ± 2 °C, the humidity is controlled at 60 ± 5%, and the wind speed is maintained at 1.2 ± 0.3 m / s. Based on this strategy template, compile a detailed standardized feeding regulation, including content such as the daily feeding plan, environmental parameter regulation guide, and growth monitoring plan. The implementation of the standardized regulation significantly improves the feeding effect, and the compliance rate of growth indicators in each stage increases to over 95%, the average daily weight gain throughout the process increases to 1.6 kg / d, and the feed-to-meat ratio decreases to 4.0.

[0039] S502. Integrate the standardized feeding regulations to form a complete standardized feeding plan.

[0040] Specifically, to achieve scientific and standardized feeding management, it is necessary to integrate the standardized feeding procedures into a systematic and complete standardized feeding plan. First, establish a plan framework, including three levels: feeding goals, technical key points, and operation specifications. At the level of feeding goals, set quantitative indicators of daily weight gain of 1.6 kg / d, feed-to-meat ratio of 4.0, and slaughter rate of 98% throughout the process. At the level of technical key points, integrate the standardized procedures of four modules: feed formula, feeding management, environmental control, and health and epidemic prevention. The feed formula module clarifies the raw material composition and nutritional indicators at each stage and formulates the raw material quality acceptance standards; the feeding management module stipulates specific requirements such as the calculation method of feed intake, feeding time, and drinking water management; the environmental control module sets the regulation plans for parameters such as temperature, humidity, and wind speed; the health and epidemic prevention module determines the vaccination procedures, disinfection plans, and disease prevention and control measures. At the level of operation specifications, compile supporting documents such as daily inspection systems, data recording forms, and emergency response plans. The standardized feeding plan formed through integration includes complete contents such as goal guidance, technical specifications, operation guides, monitoring and evaluation, and emergency response, realizing the standardization, normalization, and traceability of feeding management. The implementation of the plan has increased the production efficiency of the farm by 35%, reduced the management cost by 20%, and increased the labor productivity by 40%, providing a replicable technical standard for large-scale breeding.

[0041] Based on the above embodiments, as an alternative embodiment, the method further includes: Input the standardized feeding plan into a preset simulation model to obtain simulation data, and optimize the standardized feeding plan based on the simulation data to obtain an optimized standardized feeding plan.

[0042] Please refer to Figure 2 , Figure 2 which is an architecture diagram of a feeding strategy generation system provided by an embodiment of the present application. The feeding strategy generation system may include: A data acquisition module 1, configured to acquire experimental sample data, where the experimental sample data includes ruminant production data at different growth stages and growth levels, and preprocess the production data to obtain standardized sample data; A data processing module 2, configured to classify and store the standardized sample data to obtain a sample database, where the sample database includes feed formula data, feeding plan data, production performance data, rumen fermentation data, and environmental parameter data; A model optimization module 3, configured to use the feed formula data, the feeding plan data, and the environmental parameter data as input variables, and use the production performance data and the rumen fermentation data as output variables to establish a feeding plan optimization model; A solution module 4, configured to perform multi-objective optimization on the feeding plan optimization model based on a preset optimization objective function, so as to obtain an optimal feeding strategy combination; An optimization module 5, configured to optimize the optimal feeding strategy combination based on a preset reinforcement learning model, so as to obtain a target optimal feeding strategy combination; A plan generation module 6, configured to generate a standardized feeding plan for different growth stages according to the target optimal feeding strategy combination, including a feed formula table, a feeding schedule, and an environmental parameter control table.

[0043] It should be noted that: when the system provided in the above embodiment implements its functions, only the division of the above function modules is used for illustration. In practical applications, the above functions can be allocated to different function modules according to needs, that is, the internal structure of the device is divided into different function modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiment, which will not be elaborated here.

[0044] Please refer to Figure 3 This application also discloses an electronic device. Figure 3 FIG. 14 is a schematic structural diagram of an electronic device disclosed in an embodiment of this application. The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, and at least one communication bus 302 or end-to-end wireless communication.

[0045] Among them, the communication bus 302 is used to implement connection communication between these components.

[0046] Among them, the user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may further include a standard wired interface and a wireless interface.

[0047] Among them, the network interface 304 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).

[0048] Among them, the processor 301 may include one or more processing cores. The processor 301 connects various parts within the entire server through various interfaces and circuits. By running or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and by invoking the data stored in the memory 305, it executes various functions of the server and processes data. Optionally, the processor 301 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 301 may integrate a combination of one or several of a central processing unit (CPU), a graphics processing unit (GPU), and a modem, etc. Among them, the CPU mainly processes the operating system, user interface, application programs, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 301 and may be implemented separately by a single chip.

[0049] Among them, the memory 305 may include random access memory (RAM) and may also include read-only memory. Optionally, the memory 305 includes a non-transitory computer-readable medium. The memory 305 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area. Among them, the program storage area can store instructions for implementing the operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area can store the data involved in the above-mentioned various method embodiments. Optionally, the memory 305 may also be at least one storage system located far from the aforementioned processor 301. Refer to Figure 3 , in the memory 305, as a computer storage medium, there may be included an operating system, a network communication module, a user interface module, and an application program of a feeding strategy generation method.

[0050] In Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an interface for the user to input and obtain the data input by the user; and the processor 301 can be used to call the application program for generating the feeding strategy stored in the memory 305. When executed by one or more processors 301, the electronic device 300 is caused to execute the method as described in one or more of the above embodiments. It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be adopted in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application. In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0051] In several implementation manners provided by the present application, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are only illustrative. For example, the division of modules is only a logical function division. In actual implementation, there can be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some service interfaces. The indirect couplings or communication connections of the systems or modules can be in electrical or other forms.

[0052] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, that is, they can be located in one place, or can be distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0053] The embodiments of the present application also provide a computer storage medium. The computer storage medium can store multiple instructions, and the instructions are suitable for being loaded and executed by a processor to perform the feeding strategy generation method as described in the above Figure 1 shown embodiments. The specific execution process can refer to the specific description of the Figure 1 shown embodiments, and will not be elaborated here.

[0054] In addition, in each embodiment of the present application, the functional modules can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0055] When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present application. And the aforementioned memory includes various media that can store program codes, such as USB flash drives, mobile hard disks, magnetic disks, or optical discs.

[0056] The above are only exemplary embodiments of the present disclosure, and the scope of the present disclosure cannot be limited thereby. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure still fall within the scope covered by the present disclosure. After considering the specification and the practice of the present disclosure, those skilled in the art will easily think of other implementation manners of the present disclosure.

[0057] The present application aims to cover any variations, uses, or adaptive changes of the present disclosure that follow the general principles of the present disclosure and include the common general knowledge or conventional technical means in the technical field not recorded in the present disclosure. The specification and the embodiments are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A feeding strategy generation method, characterized in that: The method comprises: Acquiring test sample data, the test sample data including production data of ruminants at different growth stages and growth levels, and preprocessing the production data to obtain standardized sample data; Classifying and storing the standardized sample data to obtain a sample database, wherein the sample database includes feed formula data, feeding program data, production performance data, rumen fermentation data and environmental parameter data; Using the feed formula data, the feeding scheme data and the environmental parameter data as input variables, and using the production performance data and the rumen fermentation data as output variables, to establish a feeding scheme optimization model; Based on a preset optimization objective function, a multi-objective optimization solution is performed on the feeding scheme optimization model to obtain an optimal feeding strategy combination; Optimizing the optimal feeding strategy combination based on a preset reinforcement learning model to obtain a target optimal feeding strategy combination; According to the target optimal feeding strategy combination, a standardized feeding plan for different growth stages is generated, including a feed formula table, a feeding schedule and an environmental parameter control table.

2. The feeding strategy generation method according to claim 1, characterized in that: The step of classifying and storing the standardized sample data to obtain a sample database includes: Classifying the standardized sample data according to data attributes of the standardized sample data to obtain a classified data set, and constructing a relational database structure based on the classified data set; Constructing a data table based on the relational database structure, and establishing a data index table based on the data table; The sample database is constructed based on the data index table and the relational database structure.

3. The feeding strategy generation method according to claim 1, characterized in that: The method uses the feed formula data, the feeding scheme data and the environmental parameter data as input variables, and uses the production performance data and the rumen fermentation data as output variables to establish a feeding scheme optimization model, including: Performing feature extraction on the input variable to obtain input variable feature data, and performing feature extraction on the output variable to obtain output variable feature data; A feeding scheme optimization model is constructed based on the output variable characteristic data.

4. The feeding strategy generation method according to claim 1, characterized in that: Based on the preset optimization objective function, the feeding scheme optimization model is subjected to multi-objective optimization to obtain the optimal feeding strategy combination, including: In the optimization space of the feeding scheme optimization model, an initial population including a plurality of individuals is randomly generated, wherein each individual corresponds to a set of value combinations of optimization variables; Calculating the fitness value of each individual in the initial population according to the objective function in the feeding scheme optimization model, and sorting the individuals according to the fitness value to obtain sorted individuals; Screening is performed based on the ranked individuals, selecting the top N individuals with the highest ranking into the parent individual group, and randomly selecting two parent individuals from the parent individual group for crossover operation to obtain offspring individuals; The parent individuals and the offspring individuals are combined to form a new population, and the new population is subjected to evolution processing until the latest population evolution reaches a preset convergence condition to obtain a convergent population; The convergent population is used as the optimal feeding strategy combination.

5. The feeding strategy generation method according to claim 1, characterized in that: The step of generating a standardized feeding plan for different growth stages according to the target optimal feeding strategy combination includes: constructing a staged feeding strategy template based on the target optimal feeding strategy combination, and generating a standardized feeding procedure based on the staged feeding strategy template; The standardized feeding protocols are integrated to form a complete standardized feeding program.

6. The feeding strategy generation method according to claim 1, characterized in that: The optimal feeding strategy combination is optimized based on the preset reinforcement learning model to obtain the target optimal feeding strategy combination, including: The optimal feeding strategy combination is used as the starting point of a preset reinforcement learning model, and the adjustable parameters in the optimal feeding strategy combination are used as the action space of the preset reinforcement learning model. The optimal feeding strategy combination is optimized to obtain a target optimal feeding strategy combination.

7. The method according to claim 1, characterized in that The method further comprises: The standardized feeding scheme is input into a preset simulation model to obtain simulation data, and the standardized feeding scheme is optimized based on the simulation data to obtain an optimized standardized feeding scheme.

8. A feeding strategy generation system, characterized in that: The system comprises: A data acquisition module is used to acquire test sample data, wherein the test sample data includes production data of ruminants at different growth stages and growth levels, and pre-process the production data to obtain standardized sample data; A data processing module, used for classifying and storing the standardized sample data to obtain a sample database, wherein the sample database includes feed formula data, feeding program data, production performance data, rumen fermentation data and environmental parameter data; A model optimization module, used for establishing a feeding scheme optimization model by taking the feed formula data, the feeding scheme data and the environmental parameter data as input variables and the production performance data and the rumen fermentation data as output variables; A solution module is used to perform multi-objective optimization on the feeding scheme optimization model based on a preset optimization objective function to obtain an optimal feeding strategy combination; An optimization module, used to optimize the optimal feeding strategy combination based on a preset reinforcement learning model to obtain a target optimal feeding strategy combination; The program generation module is used to generate standardized feeding programs for different growth stages according to the target optimal feeding strategy combination, including a feed formula table, a feeding schedule and an environmental parameter control table.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It includes a processor, a memory and a transceiver, the memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method according to any one of claims 1 to 7.

Images