Intelligent feed optimal proportioning method based on pig herd nutrition analysis

Abstract

The invention relates to the technical field of livestock breeding, in particular to an intelligent feed optimal proportioning method based on pig herd nutrition analysis, which comprises the following steps: S1, collecting stress index data of a pig herd; s2, analyzing the stress level and the stress source of the swinery; s3, dynamically adjusting nutritional requirement parameters according to the stress level and the stress source, and determining nutrients needing to be adjusted under different stress conditions; s4, making corresponding stress mitigation measures based on the stress source and the stress intensity; s5, generating an optimal feed formula meeting the nutritional requirements of the swinery under the stress condition; and S6, converting the optimal feed formula into a production instruction, and sending the production instruction to feed production equipment. By dynamically adjusting the nutritional requirements of the swinery under different stress conditions and combining intelligent feed production equipment, precise feed formula optimization and efficient production are achieved, the feed utilization rate and the swinery health level are improved, and meanwhile the breeding cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of animal husbandry, and in particular to an intelligent feed optimization ratio method based on pig herd nutrition analysis. Background Art

[0002] With the rapid development of modern pig farming, the nutritional needs of pigs have become increasingly complex. Traditional feed ratio methods mainly rely on experience and fixed formulas, and lack accurate assessment of the dynamic nutritional needs of pigs under different stress conditions. Stressors (such as environmental changes, transportation processes and disease infections) have a direct impact on the metabolism, immunity and health of pigs, and these impacts often lead to changes in the pigs' demand for specific nutrients. Traditional methods make it difficult to adjust feed formulas according to real-time stress states, resulting in low feed utilization efficiency, increased breeding costs, and increased health risks for pigs. At the same time, existing feed production equipment lacks intelligent and automated support, making it difficult to provide scientific and timely nutritional ratios for pigs under stress conditions, further restricting the refined management of the pig industry.

[0003] Existing technologies have many deficiencies in solving the problem of optimizing pig feed under stress conditions. First, the traditional formula method fails to integrate dynamic factors such as pig stress status, breed characteristics and growth stage, and cannot meet the precise demand for nutrients under different stress conditions; second, there is a lot of manual participation in the feed production process, and there are operational errors, which cannot guarantee the execution accuracy and consistency of the feed formula. Based on this, there is an urgent need for an intelligent method that can dynamically evaluate the stress status of pigs and automatically optimize the feed ratio according to real-time nutritional needs, so as to solve the problems of inaccurate nutrient supply, low production efficiency and feed waste in the existing technology. Summary of the invention

[0004] Based on the above objectives, the present invention provides an intelligent feed optimization ratio method based on pig herd nutritional analysis.

[0005] The intelligent feed optimization ratio method based on pig herd nutrition analysis includes the following steps:

[0006] S1: Using physiological sensors and behavioral sensors installed in the pig house, real-time data on stress indicators of the pig herd is collected. The stress indicators include heart rate, respiratory rate, body surface temperature, activity level and sound signals;

[0007] S2: inputting the stress index data into a stress assessment model to analyze the stress level and stressors of the pig herd, wherein the stress assessment model is constructed based on a machine learning algorithm and is capable of distinguishing different types of stressors, including environmental stress, transportation stress and disease stress, and assessing stress intensity;

[0008] S3: According to the stress level and stressor, combined with the breed, physiological state and growth stage of the pig herd, dynamically adjust the nutritional requirement parameters, and determine the nutrients that need to be adjusted under different stress conditions, including vitamin C, vitamin E, selenium, zinc, antioxidants and immune enhancers;

[0009] S4: Based on the stressors and stress intensity, formulate corresponding stress relief measures, including adjusting environmental parameters, optimizing feeding management and adding behavioral enrichment facilities to reduce the impact of stress on pigs;

[0010] S5: Based on the dynamically adjusted nutritional requirement parameters and stress relief measures, the feed raw material database and multi-objective optimization algorithm are used to generate the optimal feed formula that meets the nutritional requirements of the pig herd under stress conditions;

[0011] S6: Convert the optimal feed formula into production instructions and send them to feed production equipment to control the addition ratio, processing parameters and addition sequence of each raw material to produce feed that meets the needs of the pig herd under stress conditions.

[0012] Optionally, the S1 specifically includes:

[0013] S11: installing physiological sensors and behavioral sensors in the pig house, wherein the physiological sensors include a heart rate sensor, a respiratory rate sensor, and a body surface temperature sensor, and the behavioral sensors include an activity detector and a sound collection device;

[0014] S12: using the physiological sensor to collect heart rate, respiratory rate and body surface temperature data of the pigs in real time;

[0015] S13: using the behavior sensor to monitor the activity and sound signals of the pig herd in real time;

[0016] S14: synchronously processing the data collected by the physiological sensor and the behavioral sensor, specifically using a data fusion algorithm to integrate the multi-source data into comprehensive stress index data S.

[0017] Optionally, the S2 specifically includes:

[0018] S21: preprocessing the stress index data S, using signal denoising and data smoothing to eliminate noise and interference in the sensor acquisition process;

[0019] S22: normalizing the preprocessed data, using the minimum-maximum normalization method to map each parameter data to the [0, 1] interval, eliminating the dimensional differences between different indicators, and forming a standardized feature data set;

[0020] S23: extracting characteristic parameters from the standardized characteristic data set, wherein the characteristic parameters include heart rate variability, respiratory rate fluctuation amplitude, body surface temperature gradient, activity change rate and sound spectrum characteristics, and constructing a characteristic vector;

[0021] S24: Use machine learning algorithms to establish a stress assessment model, select support vector machines as classifiers, use labeled historical data for training, learn the characteristic patterns corresponding to different stressors, and form a classification model;

[0022] S25: The feature vector obtained in real time is input into the trained stress assessment model, and the model outputs the corresponding stressor category and stress intensity level.

[0023] Optionally, the S24 specifically includes:

[0024] S241: constructing the labeled historical data into a training set, wherein the training set includes a plurality of samples, each sample consisting of a feature vector and a corresponding stressor label;

[0025] S242: Initialize the support vector machine model and set the loss function to the Hinge loss function to evaluate the cost of classification errors;

[0026] S243: Use the training set to train the support vector machine model. The specific training process aims to find the optimal w and b by minimizing the following objective function;

[0027] S244: Optimize the objective function by gradient descent method, and iteratively update the weight vector w and the bias b;

[0028] S245: After the training is completed, the optimal classification hyperplane, that is, the optimal w and b values, is obtained, and a stress assessment model is formed.

[0029] Optionally, the S25 specifically includes:

[0030] S251: The feature vector X obtained in real time real Input the trained stress assessment model;

[0031] S252: Calculate input feature vector X real The distance between the classification hyperplane of the support vector machine model and the trained weight vector w and bias b are used to calculate the decision value f(X real );

[0032] S253: According to the decision value f(X real ) using the sign function sign(f(X real )) Output the corresponding stressor category y class , when y class =1, indicating environmental stress; when yclass =2, indicating transportation stress; when y class =3, indicating disease stress;

[0033] S254: Calculate the stress intensity level based on the classification decision value f(X real ) absolute value, stress intensity L stress The level of classification includes three levels: low, medium and high.

[0034] Optionally, the S3 specifically includes:

[0035] S31: Obtain basic information of the pig herd, including breed, physiological status and growth stage, and integrate them to form a dynamic data set D pig ;

[0036] S32: Determine the stress level L of the pig herd based on the output of the stress assessment model stress and stressor type class , and combined with the stress level L stress and stressors, to preliminarily determine the effects of stress on the metabolism and immune system of the pig herd;

[0037] S33: Using the nutrient requirement model, according to the dynamic data set D of the pig herd pig and stress conditions stress and class Make dynamic adjustments to determine the amount of adjustment required for nutrients;

[0038] Specifically, when the stress intensity level is high and the stressor is disease stress, priority is given to increasing immune-enhancing nutrients by 10% to 20%, including vitamin C, vitamin E, selenium and zinc;

[0039] When the stress intensity level is moderate and the stressor is transportation stress, increase the supply of antioxidant nutrients and energy by 10% to 20%, including vitamin A, energy supplements and electrolytes;

[0040] When the stress intensity level is low and the stressor is environmental stress, maintain basic nutritional supply and slightly increase stress-resistant nutrients by 5% to 10%;

[0041] S34: According to the nutrient adjustment amount, dynamically adjust the specific proportion of the required nutrients, and specifically calculate the required nutrient supply amount N for each stressor and stress level adjusted , the formula is: N adjusted =N base +ΔN, where N base is the basic nutritional requirement of the pig herd under non-stress conditions, and ΔN is the adjustment amount of nutritional requirements under stress conditions.

[0042] Optionally, the S4 specifically includes:

[0043] S41: For environmental stress, when the stress intensity is high, adjust the temperature and humidity to the comfortable range of the pigs; specifically for fattening pigs, the comfortable temperature range is set at 20℃ to 24℃; for piglets, the comfortable temperature range is set at 25℃ to 28℃; humidity is controlled in the range of 60% to 75%; when the stress intensity is medium, increase the air intake by 20% to 30% to improve the air quality; when the stress intensity is low, make regular environmental parameter adjustments and increase the ventilation volume by 15% to 20% to ensure air circulation;

[0044] S42: For transport stress, when the stress intensity is high, adjust the activity space of the pigs to at least 0.6 square meters per pig; provide sufficient water sources, set the water supplement amount to 300 ml / hour per pig, and add electrolyte solution during transportation, 10-15 ml per pig per day; when the stress intensity is medium, shorten the transportation time, the maximum transportation time limit is 6 hours, and arrange appropriate rest in the middle, at least 30 minutes; when the stress intensity is low, routinely supplement antioxidants and electrolytes, 5 grams of antioxidants and 20 ml of electrolyte solution per pig;

[0045] S43: For disease stress, when the stress intensity is high, immediately isolate the pigs showing stress response and conduct drug treatment. The vitamin C supplement is 500 mg per pig; the vitamin E supplement is 100 mg per pig; the selenium supplement is 0.2 mg per pig; when the stress intensity is medium, increase the supply of immune enhancers, supplement 50 mg of vitamin C and 25 mg of vitamin E per pig per day; adjust the stocking density, reduce the number of pigs in each pen to less than 3 per square meter, and reduce the risk of cross infection; when the stress intensity is low, take preventive measures and add ingredients that enhance immunity. The supplement amount is 100 mg of vitamin E and 100 mg of vitamin C per pig per month.

[0046] Optionally, the S5 specifically includes:

[0047] S51: extracting corresponding feed raw material information from a feed raw material database according to the dynamically adjusted nutritional requirement parameters and stress relief measures, wherein the feed raw material database includes nutritional components, energy density, digestibility, price, and inventory parameters of each raw material to form a raw material selection set R;

[0048] S52: establishing a multi-objective optimization model based on the extracted raw material selection set R and the nutritional requirement parameters, wherein the optimization objectives include nutritional objectives, cost objectives and availability objectives;

[0049] S53: Apply genetic algorithm to perform multi-objective optimization, gradually approach the optimal solution by searching and iterating in the multi-dimensional objective space, and then obtain the optimal feed formula, which includes the specific amount of each raw material.

[0050] Optionally, the S53 specifically includes:

[0051] S531: Initialize the parameters of the genetic algorithm, set the population size, maximum number of iterations, crossover probability and mutation probability, and randomly generate an initial population, where each individual represents a feed formula, including the amount of each raw material;

[0052] S532: define a fitness function for evaluating the quality of each feed formula;

[0053] S533: Execute the selection operation, using the roulette wheel selection method to select the individual with the highest fitness from the current population for reproduction;

[0054] S534: performing a crossover operation, using a single-point crossover method, exchanging genes of two selected feed formulas at any position to generate new offspring individuals;

[0055] S535: performing a mutation operation to randomly modify genes of some feed formulas with a predetermined mutation probability, thereby increasing the diversity of the population and avoiding local optimal solutions;

[0056] S536: Evaluate the fitness of the new individuals after crossover and mutation, calculate the nutritional satisfaction, cost control and raw material availability of the newly generated feed formula, and select the optimal solution through the fitness function;

[0057] S537: Repeat the above steps of selection, crossover, mutation and fitness evaluation until the maximum number of iterations is reached or the fitness is no longer significantly improved, and generate the final optimal feed formula.

[0058] Optionally, the S6 specifically includes:

[0059] S61: exporting the raw material composition and corresponding dosage data of the optimal feed formula from the feed optimization module, wherein the formula data includes the specific dosage, mixing ratio and nutritional information of each raw material, and forms a basic data set for feed production;

[0060] S62: Inputting the basic data set into the feed production control device, the control device is connected to the automatic batching device, and automatically converts the raw material data into executable batching instructions, the instructions including the order, amount and mixing time of each raw material;

[0061] S63: The batching instructions are sent to the automatic batching equipment, and the equipment weighs, transports, mixes and processes the raw materials according to the instructions to ensure that the operation is carried out in accordance with the requirements of the optimal feed formula.

[0062] Beneficial effects of the present invention:

[0063] The present invention monitors the stress state of the pig herd in real time, combines the pig herd's breed, physiological state and growth stage, dynamically adjusts the nutritional demand parameters, and uses a multi-objective optimization algorithm to generate the optimal feed formula. This method can accurately respond to the pig herd's demand for specific nutrients under different stress conditions, ensure that the feed formula is scientific and reasonable, reduce breeding costs, and ensure the healthy growth of the pig herd under stress conditions.

[0064] The present invention ensures the stability and consistency of feed production and reduces quality deviations caused by human factors through the full monitoring and automatic adjustment of the production process by an intelligent system. At the same time, the combination of formula optimization and production automation improves feed production efficiency, enhances the pig herd's anti-stress ability, and promotes the sustainable development of the pig industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 This is a schematic diagram of an intelligent feed optimization ratio method according to an embodiment of the present invention;

[0067] Figure 2 The figure is a schematic diagram of the stress assessment model construction process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0069] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0070] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0071] like Figure 1-Figure 2 As shown, the intelligent feed optimization ratio method based on pig herd nutrition analysis includes the following steps:

[0072] S1: Using physiological sensors and behavioral sensors installed in the pig house, real-time data on stress indicators of the pig herd is collected. Stress indicators include heart rate, respiratory rate, body surface temperature, activity level and sound signals;

[0073] S2: Input stress index data into the stress assessment model to analyze the stress level and stressors of the pig herd. The stress assessment model is built based on a machine learning algorithm and can distinguish different types of stressors, including environmental stress, transportation stress, and disease stress, and assess the stress intensity.

[0074] S3: Dynamically adjust the nutritional requirement parameters according to the stress level and stressor, combined with the breed, physiological state and growth stage of the pig herd, and determine the nutrients that need to be adjusted under different stress conditions, including vitamin C, vitamin E, selenium, zinc, antioxidants and immune enhancers;

[0075] S4: Based on the stressor and stress intensity, formulate corresponding stress relief measures, including adjusting environmental parameters (temperature, humidity, light), optimizing feeding management (reducing density, improving ventilation) and adding behavioral enrichment facilities to reduce the impact of stress on the pig herd;

[0076] S5: Based on the dynamically adjusted nutritional requirement parameters and stress relief measures, the optimal feed formula that meets the nutritional requirements of the pig herd under stress conditions is generated using the feed raw material database and a multi-objective optimization algorithm that comprehensively considers nutritional balance, feed palatability, cost control, and raw material availability;

[0077] S6: Convert the optimal feed formula into production instructions and send them to the feed production equipment to control the addition ratio, processing parameters and addition sequence of each raw material to produce feed that meets the needs of the pig herd under stress conditions.

[0078] S1 specifically includes:

[0079] S11: Install physiological sensors and behavioral sensors in the pig house. The physiological sensors include a heart rate sensor, a respiratory rate sensor, and a body surface temperature sensor. The behavioral sensors include an activity detector and a sound collection device.

[0080] S12: Physiological sensors are used to collect heart rate, respiratory rate and body temperature data of pigs in real time. The heart rate sensor uses non-contact microwave radar technology to detect the tiny displacement of the pig's chest cavity, obtain heart beat signals, calculate heart rate data, and avoid interference with the pigs; the respiratory rate sensor uses infrared thermal imaging technology to capture the thermal radiation changes in the pig's nostrils, analyze the periodic changes of temperature over time, calculate the respiratory rate, and ensure the accuracy of the data; the body temperature sensor uses high-precision infrared temperature measurement technology to measure the body temperature distribution of pigs, eliminate the influence of ambient temperature through multi-point measurement, and obtain accurate body temperature data;

[0081] S13: Use behavioral sensors to monitor the activity and sound signals of the pigs in real time. The activity detector uses video image analysis technology and high-definition cameras installed in the pig house to capture the movement images of the pigs, and uses a target tracking algorithm to analyze the movement trajectory and speed of the pigs and calculate the activity index; the sound collection device collects the sound signals of the pigs, and uses digital signal processing technology to perform spectrum analysis and pattern recognition on the sound signals, extract the voiceprint features of the pigs, and judge their emotional state and stress level;

[0082] S14: Synchronously process the data collected by the physiological sensor and the behavioral sensor. Specifically, a data fusion algorithm is used to integrate the multi-source data into the comprehensive stress index data S. The data fusion algorithm adopts a weighted data fusion model, assigns different weights according to the reliability and importance of each sensor data, and combines them to form a comprehensive stress index S. The formula is: S = w 1 ×HR n +w 2 ×RR n +w 3 ×BT n +w 4 ×AL n +w 5 ×SV n , among which, HR n is the normalized value of heart rate data; RR n is the normalized value of respiratory rate data; BTn is the normalized value of the body surface temperature data; AL n is the normalized value of activity data; SV n is the normalized value of the sound signal data; w 1 ,w 2 ,w 3 ,w 4 ,w 5 is the weight coefficient corresponding to each parameter, which needs to satisfy w 1 +w 2 +w 3 +w 4 +w 5 =1; The comprehensive stress index S reflects the overall stress level of the pig herd and provides reliable data support for subsequent stress response evaluation.

[0083] S2 specifically includes:

[0084] S21: Preprocess the stress index data S, use signal denoising and data smoothing to eliminate noise and interference in the sensor acquisition process; improve data quality, where signal denoising uses wavelet denoising technology to decompose the collected stress index signal S(t) into multiple layers of wavelet coefficients c j , where S(t) represents the stress index signal collected at time t, c j is the wavelet coefficient of the jth layer, threshold processing is performed on the high-frequency noise, the signal is reconstructed after filtering out the noise, and the denoised stress index data is obtained; the specific steps are:

[0085] First, the signal S(t) is transformed by wavelet to obtain the wavelet coefficients c of each layer. j ;

[0086] Then, the wavelet coefficients c of the high frequency part are j Apply soft threshold processing to remove noise components and retain low-frequency information;

[0087] Finally, the signal S′(t) is reconstructed by inverse wavelet transform, where S′(t) is the denoised signal;

[0088] The data smoothing adopts the sliding average method to smooth the denoised signal S′(t). Specifically, the sliding window size is set to N, and the N data points in each window are averaged to reduce the impact of short-term fluctuations. The calculation formula of the smoothed data S″(t) is: Among them, S′(ti) is the data at the ti-th time point after denoising, and N is the size of the sliding window. Through the denoising and smoothing processing in the above steps, the random noise and abnormal fluctuations generated in the sensor acquisition process are effectively eliminated, and a smoother and more accurate stress index signal is obtained, which lays the foundation for subsequent feature extraction and evaluation.

[0089] S22: normalizing the preprocessed data, using the minimum-maximum normalization method to map each parameter data to the [0, 1] interval, eliminating the dimensional differences between different indicators, and forming a standardized feature data set;

[0090] S23: extracting characteristic parameters from the standardized characteristic data set, the characteristic parameters including heart rate variability, respiratory rate fluctuation amplitude, body surface temperature gradient, activity change rate and sound spectrum characteristics, and constructing a characteristic vector;

[0091] S24: Use machine learning algorithms to establish a stress assessment model, select support vector machine (SVM) as a classifier, use labeled historical data for training, learn the characteristic patterns corresponding to different stressors (environmental stress, transportation stress, disease stress), and form a classification model;

[0092] S25: The feature vectors acquired in real time are input into the trained stress assessment model, and the model outputs the corresponding stressor category and stress intensity level, so as to accurately distinguish different types of stressors and evaluate the stress level. Through the above steps, the stress state of the pig herd can be accurately identified and classified, which provides a scientific basis for the subsequent dynamic adjustment of nutritional needs and the formulation of stress relief measures, and improves the practicality and intelligence of the system.

[0093] The stress assessment model established in S24 specifically includes:

[0094] S241: Construct the labeled historical data into a training set, which includes multiple samples, each of which is represented by a feature vector X i and the corresponding stressor label y i Composition; eigenvector X i =[x 1 , x 2 , ..., x n ] represents the characteristic parameters extracted from historical data, including heart rate variability, respiratory rate fluctuation amplitude, body surface temperature gradient, activity change rate and sound spectrum characteristics, y i is the stressor label, with values ​​of {1, 2, 3}, corresponding to environmental stress, transportation stress, and disease stress, respectively;

[0095] S242: Initialize the support vector machine (SVM) model and set the loss function to the Hinge loss function to evaluate the cost of classification errors. The goal of SVM is to find a classification hyperplane so that samples of different categories can be correctly distinguished as much as possible. The classification hyperplane is determined by the parameter vector w and the bias b, which is expressed as: w·X+b=0, where w is the weight vector, X is the input feature vector, and b is the bias term.

[0096] S243: Use the training set to train the support vector machine model. The specific training process aims to find the optimal w and b by minimizing the following objective function. The expression of the minimization function is: Among them, ∥w∥ 2 is the bi-norm of the weight vector, which is used to maximize the classification interval; C is the regularization parameter, which is used to balance the classification error and the size of the interval; m is the number of training samples; y i is the stressor label of sample i, X i is the feature vector of the sample;

[0097] S244: Optimize the objective function by gradient descent method, and iteratively update the weight vector w and bias b; the rule for each iterative update is: Among them, η is the learning rate, L is the loss function, and are the gradients of weight w and bias b respectively;

[0098] S245: After the training is completed, the optimal classification hyperplane, that is, the optimal w and b values, is obtained, and a stress assessment model is formed; the model can input the real-time input feature vector X into the hyperplane equation, and use the classification decision function to judge the stress source category y, and its calculation formula is: y=sign(w·X+b), where sign is a symbol function used to output the classification result, y=1 represents environmental stress, y=2 represents transportation stress, and y=3 represents disease stress; by selecting a support vector machine as a classifier for training, an accurate stress assessment model is formed, which effectively utilizes the characteristic patterns of historical data and can accurately distinguish different types of stressors. The use of the SVM algorithm ensures the high classification accuracy of the model, improves the stress assessment performance of the system, and provides a reliable foundation for subsequent nutritional adjustment and feed optimization.

[0099] S25 specifically includes:

[0100] S251: The feature vector X obtained in real time real Input the trained stress assessment model, the feature vector is represented as: X real =[x 1 , x 2 , ..., x n ], including feature parameters extracted from real-time data, such as heart rate variability, respiratory rate fluctuation amplitude, body surface temperature gradient, activity change rate and sound spectrum characteristics, ensuring that all feature parameters have been preprocessed and normalized;

[0101] S252: Calculate input feature vector X real The distance between the classification hyperplane of the support vector machine model and the trained weight vector w and bias b are used to calculate the decision value f(X real), the calculation formula is: f(X real )=w·X real +b, where w is the optimal weight vector obtained during training, X real is the feature vector of real-time input, and b is the bias term after training;

[0102] S253: According to the decision value f(X real ) using the sign function sign(f(X real )) Output the corresponding stressor category y class , the formula is: class =sign(f(X real )), where y class =1, indicating environmental stress; when y class =2, indicating transportation stress; when y class =3, indicating disease stress;

[0103] S254: Calculate the stress intensity level based on the classification decision value f(X real ) absolute value, stress intensity L stress The classification of the level includes three levels: low, medium and high; the formula for the absolute value is: stress =|f(X real )|, assuming the stress intensity level is L stress , the specific level is based on I stress The value range is divided into:

[0104] When I stress <T 1 When the intensity is low, the stress level is L stress =1;

[0105] When T 1 ≤I stress <T 2 When the intensity is medium, the stress level is L stress =2;

[0106] When I stress ≥T 2 When the intensity is high, the stress level is L stress =3; where T 1 and T 2 is the stress intensity threshold set according to historical data; finally, the stress source category y is output class and stress intensity level L stress, which is used for the subsequent formulation of stress relief measures and feed formula adjustment; through the above steps, combined with the decision function of the SVM classifier and the stress intensity level classification method, the input feature vector is analyzed in real time, and the stress source category and stress intensity level are accurately output. This process effectively utilizes the decision value of the classification model, ensures the accuracy of stress source identification and the rationality of intensity assessment, provides reliable data support for subsequent stress treatment and management, and improves the practicality and intelligence level of the system.

[0107] S3 specifically includes:

[0108] S31: Obtain basic information of the pig herd, including breed, physiological status (such as weight, health status) and growth stage (such as weaning period, growing period or fattening period). The basic information is updated through regular testing and breeding records, and integrated to form a dynamic data set D pig , used for nutritional needs adjustment;

[0109] S32: Based on the output of the stress assessment model, determine the stress level L of the pig herd stress and stressor type class , and combined with the stress level L stress (low, medium, high) and stressors (environment, transportation, disease), to preliminarily determine the impact of stress on the metabolism and immune system of the pig herd;

[0110] S33: Using the nutrient requirement model, according to the dynamic data set D of the pig herd pig and stress conditions stress and class Make dynamic adjustments to determine the amount of adjustment required for nutrients;

[0111] Specifically, when the stress intensity level is high and the stressor is disease stress, priority is given to increasing immune-enhancing nutrients by 10% to 20%, including vitamin C, vitamin E, selenium and zinc;

[0112] When the stress intensity level is moderate and the stressor is transportation stress, increase the supply of antioxidant nutrients and energy by 10% to 20%, including vitamin A, energy supplements and electrolytes;

[0113] When the stress intensity level is low and the stressor is environmental stress, maintain basic nutritional supply and slightly increase stress-resistant nutrients by 5% to 10%;

[0114] S34: According to the nutrient adjustment amount, dynamically adjust the specific proportion of the required nutrients, and specifically calculate the required nutrient supply amount N for each stressor and stress level adjusted , the formula is: N adjusted =N base +ΔN, where N baseis the basic nutritional requirement of the pig herd under non-stress conditions, and ΔN is the nutritional requirement adjustment under stress conditions, which is precisely adjusted based on the breed, physiological state and growth stage of the pig herd; Where ΔN represents the adjustment of nutritional requirements under stress conditions; N base It is based on the basic nutritional needs under normal physiological conditions (for example, the regular requirements of vitamins, minerals, proteins, etc.); stress is the adjustment factor of stress level on basic nutritional requirements; N i,add It is the adjustment increment of specific nutrients (such as vitamin C, vitamin E, zinc, selenium, etc.) under stress conditions; i is the stress adjustment coefficient of each nutrient; through the above steps, based on the stress source and stress level, combined with the physiological state and growth stage of the pig herd, the nutritional demand parameters are dynamically adjusted to ensure the accuracy and scientificity of nutritional supply, which not only enhances the pig herd's ability to resist stress under stress conditions, but also effectively improves its immunity and overall health level, thereby optimizing the feed formula and improving breeding efficiency.

[0115] S4 specifically includes:

[0116] S41: For environmental stress, when the stress intensity is high, adjust the temperature and humidity to the comfortable range of the pigs; specifically for fattening pigs, the comfortable temperature range is set at 20℃ to 24℃; for piglets, the comfortable temperature range is set at 25℃ to 28℃; humidity is controlled in the range of 60% to 75%; when the stress intensity is medium, increase the air intake by 20% to 30% to improve the air quality; when the stress intensity is low, make regular environmental parameter adjustments and increase the ventilation volume by 15% to 20% to ensure air circulation;

[0117] S42: For transport stress, when the stress intensity is high, adjust the activity space of the pigs to at least 0.6 square meters per pig; provide sufficient water sources, set the water supplement amount to 300 ml / hour per pig, and add electrolyte solution during transportation, 10-15 ml per pig per day; when the stress intensity is medium, shorten the transportation time, the maximum transportation time limit is 6 hours, and arrange appropriate rest in the middle, at least 30 minutes; when the stress intensity is low, routinely supplement antioxidants and electrolytes, 5 grams of antioxidants and 20 ml of electrolyte solution per pig;

[0118] S43: For disease stress, when the stress intensity is high, immediately isolate the pigs with stress response and carry out drug treatment. The vitamin C supplement is 500 mg per pig; the vitamin E supplement is 100 mg per pig; the selenium supplement is 0.2 mg per pig; when the stress intensity is medium, increase the supply of immune enhancers, supplement 50 mg of vitamin C and 25 mg of vitamin E per pig per day; adjust the stocking density, reduce the number of pigs in each pen to less than 3 per square meter, and reduce the risk of cross infection; when the stress intensity is low, take preventive measures, add immune-enhancing ingredients, and the supplement amount is 100 mg of vitamin E and 100 mg of vitamin C per pig per month; through the above steps, based on the stressor and stress intensity, specific and scientific stress relief measures are formulated, and various relief measures are dynamically adjusted according to the type and intensity of stress to ensure that the pigs receive timely and appropriate intervention under different stress conditions, reduce the negative impact of stress on health and production performance, improve the pigs' stress resistance and survival rate, and optimize breeding management.

[0119] S5 specifically includes:

[0120] S51: extracting corresponding feed raw material information from a feed raw material database according to the dynamically adjusted nutritional requirement parameters and stress relief measures. The feed raw material database includes the nutritional composition, energy density, digestibility, price and inventory parameters of each raw material to form a raw material selection set R for subsequent formula optimization;

[0121] S52: establishing a multi-objective optimization model based on the extracted raw material selection set R and the nutritional requirement parameters, wherein the optimization objectives include nutritional objectives, cost objectives and availability objectives;

[0122] Nutritional targets are used to ensure that the generated feed formula can meet the nutritional needs of the pig herd under stress conditions, specifically meeting the minimum requirements for nutrients such as protein, energy, vitamins and minerals. The optimization formula is as follows:

[0123]

[0124] Among them, r i is the amount of raw material i, p i is the protein content in raw material i, e i is the energy value in raw material i, N protein and N energy are the adjusted protein and energy requirements, respectively;

[0125] The cost target is used to reduce the total cost of feed formula as much as possible while meeting nutritional requirements. The cost optimization formula is: Among them, c i is the unit price of raw material i, ri is the amount of raw material i;

[0126] The availability objective is used to consider the availability of feed raw materials, ensure that the formula is generated within the current inventory, and avoid the use of scarce raw materials; the constraints are: Among them, s i is the inventory of raw material i;

[0127] S53: Apply genetic algorithms for multi-objective optimization. By searching and iterating in the multi-dimensional objective space, we gradually approach the optimal solution and obtain the optimal feed formula. The formula includes the specific amount of each raw material, which can meet the nutritional needs of the pig herd under stress conditions, while having low cost and efficient raw material utilization.

[0128] S53 specifically includes:

[0129] S531: Initialize the parameters of the genetic algorithm, set the population size, maximum number of iterations, crossover probability and mutation probability, and randomly generate an initial population, where each individual represents a feed formula, including the amount of each raw material, and the initial population is used for the subsequent optimization process;

[0130] S532: Define a fitness function to evaluate the quality of each feed formula. The fitness function is calculated based on multiple objectives, including a nutritional objective (ensuring that the feed meets the nutritional needs of the pig herd), a cost objective (minimizing the cost of the formula), and a raw material availability objective (ensuring that the raw materials used are within the current inventory range);

[0131] S533: Execute a selection operation, using a roulette wheel selection method to select individuals with the highest fitness from the current population for reproduction, and better individuals have a higher probability of being selected to generate the next generation of feed formula;

[0132] S534: Perform a crossover operation, using a single-point crossover method to exchange genes of two selected feed formulas at any position to generate new offspring individuals. This process creates a new feed formula combination by combining the amounts of different raw materials;

[0133] S535: performing a mutation operation to randomly modify genes of some feed formulas (i.e., raw material dosage) with a predetermined mutation probability, thereby increasing the diversity of the population and avoiding local optimal solutions;

[0134] S536: Evaluate the fitness of the new individuals after crossover and mutation, calculate the nutritional satisfaction, cost control and raw material availability of the newly generated feed formula, and select the optimal solution through the fitness function;

[0135] S537: Repeat the above steps of selection, crossover, mutation and fitness evaluation until the maximum number of iterations is reached or the fitness is no longer significantly improved, and generate the final optimal feed formula; by adopting the selection, crossover, mutation and iterative optimization process of the genetic algorithm, the optimal feed formula that meets the nutritional needs of the pig herd under stress conditions is effectively generated. This method comprehensively considers multiple objectives such as nutritional needs, cost and raw material availability, ensures the scientificity, economy and operability of the formula, and provides an optimization solution for pig health management and cost control.

[0136] The optimal feed formula iteration process is as follows:

[0137] Initialize the parameters of the genetic algorithm, set the population size P, the maximum number of iterations G max , crossover probability p c , mutation probability p m And fitness function; randomly generate P initial solutions as the initial population R i , each solution R i Represents a feed formula, including the amount of each raw material r i , the dosage meets the aforementioned nutritional requirements and constraints;

[0138] Define the fitness function F(R i ), used to evaluate each feed formula R i The fitness function is composed of nutrition target, cost target and raw material availability target; the specific calculation formula is as follows:

[0139] Assume that nutritional fitness is F nutrition , the formula is: Among them, r ij For feed formula R i The amount of raw material j in p j is the protein content of raw material j, e j is the energy value of raw material j, N protein and N energy are the adjusted protein and energy requirements, respectively, and n is the number of feed formulations;

[0140] Let the cost fitness be F cost , the formula is: Among them, c j is the unit price of raw material j, r ij For feed formula R i The amount of raw material j;

[0141] Assume the availability fitness is F availability , the formula is: Among them, s j is the inventory of raw material j, if rij If the inventory is exceeded, there will be a penalty point;

[0142] The total fitness function F(R i ) is: F(R i )=w 1 ·F nutrition (R i )+w 2 ·F cost (R i )+w 3 ·F availability (R i ), where w 1 ,w 2 ,w 3 is the fitness weight, satisfying w 1 +w 2 +w 3 =1;

[0143] Select the operation according to the fitness function value F(R i ), the roulette wheel selection method is used to select the better solution from the current population for reproduction; the selection probability P select (R i ) is calculated as: According to the selection probability P select (R i ) Select the feed formula with the highest fitness for subsequent crossover and mutation operations.

[0144] S6 specifically includes:

[0145] S61: The raw material composition and corresponding dosage data of the optimal feed formula are exported from the feed optimization module, the formula data including the specific dosage, mixing ratio and nutritional information of each raw material, and form a basic data set for feed production;

[0146] S62: inputting the basic data set into the feed production control device, the control device is connected to the automatic batching device, and the raw material data is automatically converted into executable batching instructions, the instructions including the order, amount and mixing time of each raw material;

[0147] S63: Send the batching instructions to the automatic batching equipment, and the equipment weighs, transports, mixes and processes the raw materials according to the instructions, ensuring that the operation is carried out in accordance with the requirements of the optimal feed formula to avoid raw material waste and formula deviation; through the above steps, the optimal feed formula can be automatically converted into production instructions, combined with the automatic batching equipment to achieve an efficient and accurate feed production process. This method ensures that each link of feed production is executed according to the optimized formula, avoids human operation errors, improves production efficiency and consistency of feed quality, and provides strong support for the health of the pig herd and the economy of the feed.

[0148] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0149] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent feed optimization ratio method based on pig nutrition analysis, characterized in that: The following steps are involved: S1: Using physiological sensors and behavioral sensors installed in the pig house, real-time data on stress indicators of the pig herd is collected. The stress indicators include heart rate, respiratory rate, body surface temperature, activity level and sound signals; S2: inputting the stress index data into a stress assessment model to analyze the stress level and stressors of the pig herd, wherein the stress assessment model is constructed based on a machine learning algorithm and is capable of distinguishing different types of stressors, including environmental stress, transportation stress and disease stress, and assessing stress intensity; S3: According to the stress level and stressor, combined with the breed, physiological state and growth stage of the pig herd, dynamically adjust the nutritional requirement parameters, and determine the nutrients that need to be adjusted under different stress conditions, including vitamin C, vitamin E, selenium, zinc, antioxidants and immune enhancers; S4: Based on the stressors and stress intensity, formulate corresponding stress relief measures, including adjusting environmental parameters, optimizing feeding management and adding behavioral enrichment facilities to reduce the impact of stress on pigs; S5: Based on the dynamically adjusted nutritional requirement parameters and stress relief measures, the feed raw material database and multi-objective optimization algorithm are used to generate the optimal feed formula that meets the nutritional requirements of the pig herd under stress conditions; S6: Convert the optimal feed formula into production instructions and send them to feed production equipment to control the addition ratio, processing parameters and addition sequence of each raw material to produce feed that meets the needs of the pig herd under stress conditions.

2. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S1 specifically includes: S11: installing physiological sensors and behavioral sensors in the pig house, wherein the physiological sensors include a heart rate sensor, a respiratory rate sensor, and a body surface temperature sensor, and the behavioral sensors include an activity detector and a sound collection device; S12: using the physiological sensor to collect heart rate, respiratory rate and body surface temperature data of the pigs in real time; S13: using the behavior sensor to monitor the activity and sound signals of the pig herd in real time; S14: synchronously processing the data collected by the physiological sensor and the behavioral sensor, specifically using a data fusion algorithm to integrate the multi-source data into comprehensive stress index data S.

3. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S2 specifically includes: S21: preprocessing the stress index data s, using signal denoising and data smoothing to eliminate noise and interference in the sensor acquisition process; S22: normalizing the preprocessed data, using the minimum-maximum normalization method to map each parameter data to the [0, 1] interval, eliminating the dimensional differences between different indicators, and forming a standardized feature data set; S23: extracting characteristic parameters from the standardized characteristic data set, wherein the characteristic parameters include heart rate variability, respiratory rate fluctuation amplitude, body surface temperature gradient, activity change rate and sound spectrum characteristics, and constructing a characteristic vector; S24: Use machine learning algorithms to establish a stress assessment model, select support vector machines as classifiers, use labeled historical data for training, learn the characteristic patterns corresponding to different stressors, and form a classification model; S25: The feature vector obtained in real time is input into the trained stress assessment model, and the model outputs the corresponding stressor category and stress intensity level.

4. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S24 specifically includes: S241: constructing the labeled historical data into a training set, wherein the training set includes a plurality of samples, each sample consisting of a feature vector and a corresponding stressor label; S242: Initialize the support vector machine model and set the loss function to the Hinge loss function to evaluate the cost of classification errors; S243: Use the training set to train the support vector machine model. The specific training process aims to find the optimal w and b by minimizing the following objective function; S244: Optimize the objective function by gradient descent method, and iteratively update the weight vector w and the bias b; S245: After the training is completed, the optimal classification hyperplane, that is, the optimal w and b values, is obtained, and a stress assessment model is formed.

5. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 4 is characterized in that: The S25 specifically includes: S251: The feature vector X obtained in real time real Input the trained stress assessment model; S252: Calculate input feature vector X real The distance between the classification hyperplane of the support vector machine model and the trained weight vector w and bias b are used to calculate the decision value f(X real ); S253: According to the decision value f(X real ) using the sign function sign(f(X real )) Output the corresponding stressor category y class , when y class =1, indicating environmental stress; when y class =2, indicating transportation stress; when y class =3, indicating disease stress; S254: Calculate the stress intensity level based on the classification decision value f(X real ) absolute value, stress intensity L stress The level of classification includes three levels: low, medium and high.

6. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 5 is characterized in that: The S3 specifically includes: S31: Obtain basic information of the pig herd, including breed, physiological status and growth stage, and integrate them to form a dynamic data set D pig S32: Determine the stress level L of the pig herd based on the output of the stress assessment model stress and stressor type class , and combined with the stress level L stress and stressors, to preliminarily determine the effects of stress on the metabolism and immunity of pigs; S33: Using the nutrient requirement model, according to the dynamic data set D of the pig herd pig and stress conditions stress and class Make dynamic adjustments to determine the amount of adjustment required for nutrients; Specifically, when the stress intensity level is high and the stressor is disease stress, priority is given to increasing immune-enhancing nutrients by 10% to 20%, including vitamin C, vitamin E, selenium and zinc; When the stress intensity level is moderate and the stressor is transportation stress, increase the supply of antioxidant nutrients and energy by 10% to 20%, including vitamin A, energy supplements and electrolytes; When the stress intensity level is low and the stressor is environmental stress, maintain basic nutritional supply and slightly increase stress-resistant nutrients by 5% to 10%; S34: According to the nutrient adjustment amount, dynamically adjust the specific proportion of the required nutrients, and specifically calculate the required nutrient supply amount N for each stressor and stress level adjusted , the formula is: N adjusted =N base +ΔN, where N base is the basic nutritional requirement of the pig herd under non-stress conditions, and ΔN is the adjustment amount of nutritional requirements under stress conditions.

7. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S4 specifically includes: S41: For environmental stress, when the stress intensity is high, adjust the temperature and humidity to the comfortable range of the pigs; specifically for fattening pigs, the comfortable temperature range is set at 20℃ to 24℃; for piglets, the comfortable temperature range is set at 25℃ to 28℃; humidity is controlled in the range of 60% to 75%; when the stress intensity is medium, increase the air intake by 20% to 30% to improve the air quality; when the stress intensity is low, make regular environmental parameter adjustments and increase the ventilation volume by 15% to 20% to ensure air circulation; S42: For transport stress, when the stress intensity is high, adjust the activity space of the pigs to at least 0.6 square meters per pig; provide sufficient water sources, set the water supplement amount to 300 ml / hour per pig, and add electrolyte solution during transportation, 10-15 ml per pig per day; when the stress intensity is medium, shorten the transportation time, the maximum transportation time limit is 6 hours, and arrange appropriate rest in the middle, at least 30 minutes; when the stress intensity is low, routinely supplement antioxidants and electrolytes, 5 grams of antioxidants and 20 ml of electrolyte solution per pig; S43: For disease stress, when the stress intensity is high, immediately isolate the pigs showing stress response and conduct drug treatment. The vitamin C supplement is 500 mg per pig; the vitamin E supplement is 100 mg per pig; the selenium supplement is 0.2 mg per pig; when the stress intensity is medium, increase the supply of immune enhancers, supplement 50 mg of vitamin C and 25 mg of vitamin E per pig per day; adjust the stocking density, reduce the number of pigs in each pen to less than 3 per square meter, and reduce the risk of cross infection; when the stress intensity is low, take preventive measures and add ingredients that enhance immunity. The supplement amount is 100 mg of vitamin E and 100 mg of vitamin C per pig per month.

8. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S5 specifically includes: S51: extracting corresponding feed raw material information from a feed raw material database according to the dynamically adjusted nutritional requirement parameters and stress relief measures, wherein the feed raw material database includes nutritional components, energy density, digestibility, price, and inventory parameters of each raw material to form a raw material selection set R; S52: establishing a multi-objective optimization model based on the extracted raw material selection set R and the nutritional requirement parameters, wherein the optimization objectives include nutritional objectives, cost objectives and availability objectives; S53: Apply genetic algorithm to perform multi-objective optimization, gradually approach the optimal solution by searching and iterating in the multi-dimensional objective space, and then obtain the optimal feed formula, which includes the specific amount of each raw material.

9. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S53 specifically includes: S531: Initialize the parameters of the genetic algorithm, set the population size, maximum number of iterations, crossover probability and mutation probability, and randomly generate an initial population, where each individual represents a feed formula, including the amount of each raw material; S532: define a fitness function for evaluating the quality of each feed formula; S533: Execute the selection operation, using the roulette wheel selection method to select the individual with the highest fitness from the current population for reproduction; S534: performing a crossover operation, using a single-point crossover method, exchanging genes of two selected feed formulas at any position to generate new offspring individuals; S535: performing a mutation operation to randomly modify genes of some feed formulas with a predetermined mutation probability, thereby increasing the diversity of the population and avoiding local optimal solutions; S536: Evaluate the fitness of the new individuals after crossover and mutation, calculate the nutritional satisfaction, cost control and raw material availability of the newly generated feed formula, and select the optimal solution through the fitness function; S537: Repeat the above steps of selection, crossover, mutation and fitness evaluation until the maximum number of iterations is reached or the fitness is no longer significantly improved, and generate the final optimal feed formula.

10. The intelligent feed optimization ratio method based on pig nutrition analysis according to claim 1 is characterized in that: The S6 specifically includes: S61: exporting the raw material composition and corresponding dosage data of the optimal feed formula from the feed optimization module, wherein the formula data includes the specific dosage, mixing ratio and nutritional information of each raw material, and forms a basic data set for feed production; S62: Inputting the basic data set into the feed production control device, the control device is connected to the automatic batching device, and automatically converts the raw material data into executable batching instructions, the instructions including the order, amount and mixing time of each raw material; S63: The batching instructions are sent to the automatic batching equipment, and the equipment weighs, transports, mixes and processes the raw materials according to the instructions to ensure that the operation is carried out in accordance with the requirements of the optimal feed formula.

Images