A computer-based pig feed formula data optimization processing method and system

Abstract

The application relates to the technical field of feed formula data optimization, in particular to a pig feed formula data optimization processing method and system based on a computer, which comprises the following steps: acquiring parameters corresponding to a target breeding scene and inputting the parameters into a preset collaborative optimization model; wherein the collaborative optimization model is configured to: determine a reference threshold of digestible protein; dynamically compensate the reference threshold to generate an upper limit of dynamic demand of the digestible protein; determine a soluble fiber proportion target coordinated with the upper limit of dynamic demand; take the received upper limit of dynamic demand of the digestible protein and the soluble fiber proportion target as a collaborative optimization target, perform proportioning calculation on raw material available nutrient component parameters, and generate a final feed formula meeting the collaborative optimization target. The application can construct a collaborative optimization model of pig breed genetic characteristics, environmental temperature and raw material available nutrients, and realize dynamic and accurate adaptation of a feed formula.

Description

Technical Field

[0001] This application relates to the technical field of feed formulation data optimization, and in particular to a computer-based method and system for optimizing pig feed formulation data. Background Technology

[0002] With the continuous improvement of the scale and intensification of my country's pig farming industry, the precision of pig feed formulation has become crucial in determining farming efficiency. This need is particularly urgent in typical southern farming regions such as Hunan. These regions are characterized by persistently hot and humid summers and cold, damp winters. They also possess abundant local pig breeds with digestive and metabolic characteristics significantly different from introduced breeds. Furthermore, due to limited storage conditions, the actual availability of key nutrients dynamically declines when feed ingredients are stored in high-temperature and high-humidity environments. These regional factors intertwine, creating a complex application scenario with highly coupled multivariates. Traditional static nutritional standards and empirical formulation models struggle to accurately match nutrient supply with the actual needs of animals in such dynamic scenarios, leading to unstable farming performance, low feed utilization efficiency, and even animal health problems.

[0003] In current technologies, pig feed formulation design has evolved from relying solely on empirical manuals to computer-aided approaches. Mainstream solutions often employ linear programming or goal programming algorithms to mathematically optimize ingredient ratios with the aim of minimizing costs or meeting fixed nutritional standards. Some advanced research has begun to introduce single dynamic factors such as environmental temperature, using simple compensation coefficients to globally adjust energy or protein requirements, or recommending fixed nutritional levels for specific breeds. These technologies have improved the scientific rigor and efficiency of formulation design to some extent. However, when faced with specific scenarios like those in Hunan province, where multiple factors are deeply coupled, existing technologies reveal their inherent limitations. They typically rely on nominal values ​​from nutrient composition tables, failing to effectively incorporate the actual attenuation rates of key active ingredients under localized storage conditions. Furthermore, their nutritional requirement models are often static or single-factor modified, failing to establish a deep correlation with the digestive enzyme activity and nutrient tolerance limits of different pig breeds due to genetic differences. More importantly, the introduction of environmental factors is often isolated, failing to form a systematic and synergistic regulatory logic with the metabolic characteristics of pig breeds and the actual supply capacity of raw materials, resulting in one-sided and lagging formulation adjustments.

[0004] In-depth analysis revealed that existing technologies cannot achieve dynamic collaborative modeling and real-time optimization of the three factors—pig breed genetic characteristics, real-time environmental temperature, and actual usable raw materials—in specific regional scenarios. This deficiency stems from the underlying design logic that treats pig breed, environment, and raw materials as isolated or simply superimposed variables. In the actual operation of large-scale piglet farms in Hunan, this deficiency was amplified dramatically, triggering a chain reaction of problems. Taking a typical summer scenario as an example: the local piglets raised on the farm have significantly lower trypsin activity than introduced breeds due to genetic reasons, which determines a lower tolerance limit for digestible protein. At the same time, the sustained high temperature of 30-35℃ in summer increases the basal metabolic rate of piglets, theoretically increasing the demand for digestible protein. If the soybean meal raw material used at this time has been stored for an extended period in a normal temperature and high humidity warehouse, its actual digestible protein content has significantly decreased. Existing technical solutions lack a synergistic analysis model for all three factors, resulting in fragmented operational logic. They may simply adopt a high-protein formula based on standard ingredient lists or merely increase protein requirements due to high temperatures, completely ignoring the possibility that the specific pig breed's protein tolerance threshold has been reached at high temperatures, and also neglecting the fact that the actual protein provided by the raw materials may be insufficient. This fragmented decision-making leads to feed formulations where digestible protein levels may exceed the pigs' physiological tolerance, while the proportion of fiber used to balance cost or volume is not specifically adjusted based on the impact of high temperatures on intestinal motility and changes in the fiber composition of the raw materials. The direct consequences manifest as a complex coexistence of problems such as abnormally high rates of diarrhea in piglets, failure to meet daily weight gain targets, and intestinal flora imbalance, which cannot be fundamentally cured by adjusting a single nutrient indicator in the formula. The essence of the problem is that existing technologies lack a core algorithm that can quantify the dynamic interaction between "genetically determined tolerance limits", "temperature-driven demand increments", and "supply capacity after raw material decay". Therefore, it is impossible to generate a feed formula that is precisely adapted to the current moment, the current pig herd, and the current raw materials, especially maintaining the optimal dynamic ratio of key nutrients such as digestible protein and soluble fiber. Summary of the Invention

[0005] In order to construct a synergistic optimization model of pig breed genetic characteristics, environmental temperature and available nutrients in raw materials, and to achieve dynamic and precise adaptation of feed formulation, this application provides a computer-based method and system for optimizing pig feed formulation data.

[0006] In a first aspect, this application provides a computer-based method and system for optimizing pig feed formulation data, which adopts the following technical solution: A computer-based method for optimizing pig feed formulation data includes the following steps:

[0007] Obtain the genetic characteristic parameters of the pig breed, real-time environmental temperature parameters, and raw material available nutrient parameters corrected for the current storage conditions corresponding to the target breeding scenario;

[0008] The genetic characteristics parameters of the pig breed, the real-time ambient temperature parameters, and the available nutrient parameters of the raw materials are input into a pre-set collaborative optimization model. The collaborative optimization model is configured to: determine a baseline threshold for digestible protein based on the genetic characteristics parameters of the pig breed; dynamically compensate the baseline threshold based on the deviation of the real-time ambient temperature parameters from a preset temperature range to generate a dynamic upper limit for digestible protein requirements; and determine a target proportion of soluble fiber that aligns with the dynamic upper limit for requirements based on the temperature compensation range corresponding to the dynamic compensation.

[0009] Receive the upper limit of dynamic demand for digestible protein and the target proportion of soluble fiber, which are output by the collaborative optimization model and matched with the current target aquaculture scenario;

[0010] Using the received upper limit of dynamic demand for digestible protein and the target proportion of soluble fiber as co-optimization objectives, the available nutrient parameters of the raw materials are calculated to generate a final feed formula that meets the co-optimization objectives.

[0011] Optionally, the step of dynamic compensation based on real-time ambient temperature parameters in the collaborative optimization model includes: when the real-time ambient temperature parameters exceed the upper limit threshold of a preset temperature range, applying a first compensation intensity to compensate the benchmark threshold; when the real-time ambient temperature parameters are lower than the lower limit threshold of a preset temperature range, applying a second compensation intensity less than the first compensation intensity to compensate the benchmark threshold.

[0012] Optionally, the step of determining the target proportion of soluble fibers in the collaborative optimization model further includes:

[0013] When the real-time ambient temperature parameter is higher than the upper limit threshold of the preset temperature range, the first fiber correlation ratio is matched with the upper limit of the dynamic demand for digestible protein.

[0014] When the real-time ambient temperature parameter is lower than the lower limit threshold of the preset temperature range, the upper limit of the dynamic demand for digestible protein is matched with a second fiber association ratio that is higher than the first fiber association ratio.

[0015] When the real-time ambient temperature parameter is between the upper and lower thresholds, the upper limit of the dynamic demand for digestible protein is matched with a third fiber association ratio that is between the first fiber association ratio and the second fiber association ratio.

[0016] Optionally, the steps are as follows: using the received upper limit of the dynamic demand for digestible protein and the target proportion of soluble fiber as co-optimization objectives, the available nutrient parameters of the raw materials are calculated for proportioning, and the following steps are performed:

[0017] Based on the upper limit of the dynamic demand for digestible protein, combined with a preset safety margin coefficient, the optimized target content of digestible protein is determined.

[0018] Based on the optimized target content of the digestible protein and the fiber association ratio matched by the collaborative optimization model, the collaborative target content of soluble fiber is determined.

[0019] Perform a formulation calculation to ensure that the total digestible protein content of the formula, calculated based on the available nutrient parameters of the raw materials, is not higher than the optimized target content of the digestible protein, and that the total soluble fiber content of the formula meets the constraints defined based on the synergistic target content of the soluble fiber.

[0020] Optional, also includes:

[0021] The actual aquaculture effect data is judged based on multiple preset effect thresholds;

[0022] If the diarrhea rate is determined to exceed the preset diarrhea threshold, the optimized target content of the digestible protein and the synergistic target content of the soluble fiber used in subsequent calculations will be adjusted downwards simultaneously.

[0023] If the daily weight gain is determined to be lower than the preset weight gain threshold, the compensation intensity used for dynamic compensation in the collaborative optimization model is adjusted upwards.

[0024] If the gut microbiota balance index is determined to be lower than the preset microbiota balance threshold, the synergistic target content of the soluble fiber used in subsequent calculations is adjusted upwards.

[0025] Optionally, before generating the final feed formulation, the following steps are also included:

[0026] When the total soluble fiber content of the formula calculated based on the available nutrient parameters of the raw materials fails to meet the constraint conditions defined by the synergistic target content of the soluble fiber, the total soluble fiber content of the formula is quantitatively corrected by applying a preset enzymatic hydrolysis control method until it meets the constraint conditions.

[0027] Optionally, the genetic characteristic parameters of the pig breed are gene expression data related to the activity of protein digestive enzymes; the available nutrient parameters of the raw materials are the actual content of digestible protein and the actual content of soluble fiber obtained after correction based on specific temperature and humidity storage conditions.

[0028] Optionally, the collaborative optimization model is further configured as follows:

[0029] Outside the preset temperature range, an enhanced compensation range is defined that is associated with the periodic climate characteristics of the target aquaculture scenario;

[0030] When the real-time ambient temperature parameter falls within the enhanced compensation range, the first compensation intensity or the second compensation intensity is adaptively modulated based on the heat tolerance metabolic regulation data associated with the genetic characteristic parameters of the pig breed, and the upper limit of the dynamic demand for digestible protein and the target proportion of soluble fiber are generated according to the modulated compensation intensity.

[0031] Secondly, this application provides a computer-based pig feed formulation data optimization processing system, which adopts the following technical solution: A computer-based pig feed formulation data optimization processing system, comprising:

[0032] The data acquisition module is configured to acquire the genetic characteristic parameters of the pig breed, the real-time ambient temperature parameters, and the raw material available nutrient parameters corrected by the current storage conditions for the target breeding scenario.

[0033] The collaborative optimization module, which has a built-in collaborative optimization model, is configured to: determine a baseline threshold for digestible protein based on the genetic characteristic parameters of the pig breed; dynamically compensate the baseline threshold according to the deviation of the real-time ambient temperature parameter from the preset temperature range to generate a dynamic upper limit for the demand of digestible protein; and determine a target for the proportion of soluble fiber that is coordinated with the dynamic upper limit for the demand based on the temperature compensation range corresponding to the dynamic compensation.

[0034] The formulation calculation module is configured to use the upper limit of the dynamic demand for digestible protein and the target proportion of soluble fiber output by the co-optimization module as co-optimization targets, and to perform proportion calculations on the available nutrient parameters of the raw materials to generate a feed formulation that meets the co-optimization targets.

[0035] Optionally, a feedback adjustment module is also included, which is configured as follows:

[0036] Receive breeding effect data collected from actual feeding based on the feed formula;

[0037] The breeding effect data is compared with preset thresholds for diarrhea rate, daily weight gain, and gut microbiota balance.

[0038] Based on the comparison results, the system dynamically generates adjustment instructions for the dynamic compensation logic in the collaborative optimization model or correction instructions for the proportion constraints in the formula calculation module, and feeds these instructions back to the corresponding modules for iterative calculation until a final feed formula that meets the preset effect threshold is generated.

[0039] In summary, this application includes the following beneficial technical effects:

[0040] This method, by constructing a collaborative optimization model that integrates pig genetic characteristics, real-time ambient temperature, and available nutrients in raw materials modified by storage conditions, achieves dynamic unified analysis and decision-making for these three highly coupled variables for the first time. It fundamentally overcomes the one-sidedness and lag of traditional formulation technologies that isolate or simply superimpose factors, thus ensuring that feed formulations can match the real physiological needs of animals and the actual supply capacity of raw materials in real time and accurately in complex and ever-changing actual breeding scenarios. This is the most critical innovation in solving the core defects of existing technologies.

[0041] This method dynamically adjusts the upper limit of digestible protein requirements and matches the corresponding soluble fiber ratio target based on the deviation of real-time temperature from the preset suitable range. It establishes a dynamic synergistic relationship between nutrients, enabling the formula to not only respond to changes in individual nutrient requirements caused by temperature changes, but also maintain the physiological balance between key nutrients. This effectively reduces intestinal burden and diarrhea risk in hot seasons and promotes nutrient absorption and utilization in cold seasons, significantly improving the adaptability of feed formula to environmental fluctuations and feeding stability.

[0042] This method introduces a closed-loop feedback adjustment mechanism based on actual breeding effect data. It can dynamically adjust the compensation intensity in the collaborative optimization model or the target content constraint in the formula calculation based on the monitoring results of key indicators such as diarrhea rate, daily weight gain, and gut microbiota. This drives the system to perform iterative optimization, so that the model parameters and formula results can continuously self-correct and improve with the accumulation of breeding practice. This ensures the long-term reliability of the optimization effect and the continuous adaptability to specific farm conditions, and realizes the leap from static optimization to dynamic self-learning. Attached Figure Description

[0043] Figure 1 It is a flowchart of the overall collaborative optimization process. Detailed Implementation

[0044] The following is in conjunction with the appendix Figure 1 This application will be described in further detail.

[0045] This application discloses a computer-based method for optimizing pig feed formulation data. The method first involves pre-implementation preparation, as follows:

[0046] 1. Parameter calibration

[0047] Genetic characteristic parameters of pig breeds: The relative expression level of the TRY1 gene in the target pig breed (which is directly related to the activity of protein digestive enzymes) was obtained by gene sequencing. The values ​​were determined to be 0.8 for local native piglets, 1.2 for Duroc, and 1.0 for Xianghong crossbred. This parameter was verified based on the Hunan pig breeding industry database and multiple batches of digestion experiments.

[0048] Temperature parameters: The preset suitable temperature range is 15℃-25℃ (upper limit threshold 25℃, lower limit threshold 15℃), and the summer enhanced compensation range is defined as 30℃-35℃ and the winter enhanced compensation range as 5℃-10℃, which conforms to the periodic climate characteristics of Hunan.

[0049] Raw material correction standards: For storage at room temperature (20℃-25℃) for more than 30 days, digestible protein (ADCP) is corrected to 90% of the nominal value and soluble fiber (SDF) is corrected to 110%; for storage at high temperature and high humidity (>25℃, >70% humidity) for more than 15 days, ADCP is corrected to 85% and SDF is corrected to 120%; for storage at low temperature (<10℃), both are retained at 100%. This standard has been confirmed by comparative experiments on raw material storage, with an error of ≤5%.

[0050] Effect thresholds: Referring to industry standards and large-scale farming data, a diarrhea rate threshold of 5%, a daily weight gain threshold of 200g / day, and a gut microbiota balance threshold (lactic acid bacteria / Escherichia coli) of 2:1 were set as benchmarks for subsequent effect evaluation.

[0051] 2. Equipment Deployment

[0052] A temperature sensor with an accuracy of ±0.2℃ is deployed 20cm directly above the piglet feed trough, and real-time temperature is collected every 10 minutes to ensure accurate acquisition of environmental parameters; a near-infrared spectrometer (NIR) is equipped to detect the nominal content of ADCP and SDF in the raw materials, providing data support for the correction of raw material parameters; smart feed troughs, high-definition cameras, weighing equipment and intestinal flora sequencers are deployed to collect data on feed intake, diarrhea, daily weight gain and flora balance, respectively, to build a system for collecting data on breeding effects.

[0053] 3. Model Initialization

[0054] The collaborative optimization model incorporates logic for benchmark threshold calculation, temperature compensation, fiber ratio matching, and proportioning. The initial parameters are based on industry data and previous experimental settings, with a preset safety margin coefficient of 90% (to avoid protein overload), a cellulase SDF degradation efficiency of 15%, and a xylanase SDF degradation efficiency of 42%. These parameters were determined through enzymatic hydrolysis experiments.

[0055] After the pre-implementation preparations are completed, the implementation steps will begin, as follows:

[0056] S1: Multi-source data acquisition and preprocessing

[0057] like Figure 1 As shown, this step relies on the equipment deployment and parameter calibration results completed in the pre-implementation preparation stage to obtain the core data of the target aquaculture scenario and carry out preprocessing, providing accurate and non-redundant input information for subsequent collaborative optimization model calculations.

[0058] S11: Data Acquisition

[0059] Collection of genetic characteristic parameters of pig breeds: The data acquisition module uses a gene sequencer to perform gene sequencing on the target pig population and extracts the relative expression level data of the TRY1 gene, which is directly related to the activity of protein digestive enzymes. This set of data was determined with reference to the Hunan Pig Breeding Industry Database and the results of multiple batches of pig digestion experiments. Specifically, the relative expression level of the TRY1 gene is 0.8 in local native piglets, 1.2 in Duroc piglets, and 1.0 in Xianghong crossbred piglets.

[0060] Real-time ambient temperature parameter acquisition: The data acquisition module collects real-time ambient temperature parameters of the piglet shed through a temperature sensor pre-deployed 20cm directly above the piglet feed trough. This temperature sensor has an accuracy of ±0.2℃, and the data acquisition frequency is set to once every 10 minutes. During the acquisition process, the time point and corresponding temperature value of each acquisition are recorded simultaneously to ensure the continuity of temperature data and provide a reliable basis for subsequent judgment of the degree of temperature deviation from the preset range.

[0061] Raw material nutrient parameter collection: The data acquisition module first uses a pre-deployed near-infrared spectrometer to detect the actual content of digestible protein (ADCP) and soluble fiber (SDF) in the feed raw materials. Then, considering the current storage conditions of the raw materials (including storage temperature, humidity, and storage time), the parameters are corrected according to the raw material correction standards established during the pre-implementation preparation phase. The specific correction logic is as follows: For storage at room temperature (20℃-25℃) for more than 30 days, ADCP is corrected to 90% of the raw material's nominal content at the manufacturer's label, and SDF is corrected to 110% of the raw material's nominal content at the manufacturer's label; for storage at high temperature and high humidity (temperature > 25℃, humidity > 70%) for more than 15 days, ADCP is corrected to 85% of the raw material's nominal content at the manufacturer's label, and SDF is corrected to 120% of the raw material's nominal content at the manufacturer's label; for storage at low temperature (temperature < 10℃), both ADCP and SDF are retained at 100% of the raw material's nominal content at the manufacturer's label.

[0062] S12: Data Filtering

[0063] The data acquisition module automatically filters and processes the collected raw data. Based on the core parameter types identified during the pre-implementation preparation phase, the module automatically removes fields irrelevant to feed formulation optimization, including information such as pig breed coat color, outdoor temperature, and raw material procurement channels. Simultaneously, it retains five core fields: relative expression level of the TRY1 gene, real-time ambient temperature, raw material ADCP correction value, raw material SDF correction value, and raw material storage conditions. This filtering operation effectively reduces data redundancy, prevents irrelevant data from interfering with subsequent parameter calculations, and ensures the relevance and accuracy of the input data.

[0064] S2: Collaborative Optimization Model Calculation

[0065] This step, based on the core data preprocessed by S1, determines the core objective of formula optimization through the logical operation of the collaborative optimization model. The model input data is directly taken from the key fields filtered by S1, ensuring that the data transmission is consistent and accurate, and avoiding irrelevant information from interfering with the calculation results.

[0066] S21: Determine the baseline threshold for digestible protein (ADCP)

[0067] The collaborative optimization model receives the porcine genetic characteristic parameters (relative expression level of the TRY1 gene) output by S1 and calculates the ADCP baseline threshold based on these parameters. This calculation logic originates from multiple batches of porcine digestion experiments: researchers conducted 30-day digestive and metabolic monitoring on piglets with different TRY1 expression levels and found a positive correlation between the relative expression level of the TRY1 gene and the activity of protein digestive enzymes in piglets, which in turn formed a linear relationship with the upper limit of ADCP tolerance: when the relative expression level of the TRY1 gene was 0, the baseline ADCP tolerance of piglets was 14%; with each increase of 0.1 in TRY1 expression level, the ADCP tolerance increased by 0.5%. Based on this experimental conclusion, the formula for calculating the ADCP baseline threshold was derived: ADCP baseline threshold (%) = 14% + relative expression level of the TRY1 gene × 5%.

[0068] The model calculates baseline thresholds for the three pig breeds mentioned in S1: the relative expression level of the TRY1 gene in local native piglets is 0.8, and its ADCP baseline threshold is 14% + 0.8 × 5% = 18%; the relative expression level of the TRY1 gene in Duroc piglets is 1.2, and its ADCP baseline threshold is 14% + 1.2 × 5% = 20%; the relative expression level of the TRY1 gene in Xianghong hybrid piglets is 1.0, and its ADCP baseline threshold is 14% + 1.0 × 5% = 19%.

[0069] S22: Upper limit of dynamic requirements for the production of digestible protein (ADCP)

[0070] The collaborative optimization model first calls the real-time ambient temperature parameters output by S1 and compares them with the preset temperature range (15℃-25℃) during the pre-implementation preparation stage. After determining the degree of temperature deviation, it performs dynamic compensation based on the characteristics of the pig breed.

[0071] Compensation intensity determination: The model sets the compensation intensity based on the results of multiple batches of metabolic experiments in summer and winter. When the real-time ambient temperature parameter exceeds the upper limit threshold (25℃) of the preset temperature range, the first compensation intensity of 0.5% / ℃ is applied; when the real-time ambient temperature parameter is lower than the lower limit threshold (15℃) of the preset temperature range, the second compensation intensity of 0.3% / ℃ is applied.

[0072] Enhanced Compensation Range Modulation: If the real-time ambient temperature parameter falls within the enhanced compensation range defined in the pre-implementation preparation phase (30℃-35℃ in summer, 5℃-10℃ in winter), the model will adaptively modulate the compensation intensity by combining heat tolerance metabolic regulation data associated with pig breed genetic characteristics. For example, for local native piglets with weak heat tolerance, within the summer enhanced compensation range (30℃-35℃), to avoid the calculated value exceeding their physiological tolerance limit, the model will adaptively lower the compensation intensity from the baseline of 0.5% / ℃ to 0.45% / ℃; Duroc piglets have strong heat tolerance, and the compensation intensity remains unchanged at 0.5% / ℃ within the same temperature range.

[0073] Dynamic Demand Upper Limit Calculation: The model generates the final demand value using the following formula: When the real-time temperature > 25℃: ADCP dynamic demand upper limit (%) = ADCP baseline threshold + (real-time temperature - 25℃) × corresponding compensation intensity; When the real-time temperature < 15℃: ADCP dynamic demand upper limit (%) = ADCP baseline threshold + (15℃ - real-time temperature) × corresponding compensation intensity. Taking local native piglets as an example, if the real-time ambient temperature collected by S1 is 32℃, and the temperature deviation (32℃ - 25℃) is 7℃, which does not fall within the enhanced compensation range, its ADCP dynamic demand upper limit is 18% + 7 × 0.5% = 21.5%. If the real-time temperature is 33℃ (falling into the summer enhanced compensation range), then calculated according to the modulated intensity, the ADCP dynamic demand upper limit is 18% + (33℃ - 25℃) × 0.45% = 21.6%; if no modulation is performed and 0.5% / ℃ is still used, then the demand upper limit is 18% + 8 × 0.5% = 22.0%. The upper limit of demand after modulation (21.6%) was lower than that without modulation (22.0%), which reflects the protective effect on pig breeds with weaker heat resistance.

[0074] S23: Determine the target proportion of soluble fiber (SDF).

[0075] The collaborative optimization model matches an appropriate SDF ratio target to the upper limit of ADCP dynamic demand based on the temperature range corresponding to the dynamic compensation in S22. This matching logic is based on experiments on the effect of temperature on piglet intestinal function.

[0076] High temperature range (>25℃): When the real-time ambient temperature parameter is higher than the upper limit threshold of the preset temperature range (25℃), the model matches the upper limit of ADCP dynamic demand with the first fiber association ratio of 8%. Experiments show that the intestinal peristalsis speed of piglets is accelerated under high temperature environment, and excessive SDF content can easily lead to increased intestinal burden and diarrhea; an 8% SDF ratio can maintain the basic intestinal peristaltic function and reduce the risk of diarrhea.

[0077] Low temperature range (<15℃): When the real-time ambient temperature parameter is lower than the lower limit threshold (15℃) of the preset temperature range, the model matches the second fiber association ratio of 12%, which is higher than the first fiber association ratio. Winter experiments show that low temperature slows down intestinal peristalsis in piglets, and a 12% SDF content can promote intestinal peristalsis, improve nutrient absorption efficiency, and increase ADCP utilization.

[0078] Normal temperature range (15℃-25℃): When the real-time ambient temperature parameter is between the upper and lower thresholds, the model matches the third fiber association ratio of 10%. This ratio is based on industry-standard large-scale farming and has been confirmed by multiple batches of experiments in spring and autumn. It can balance intestinal peristalsis speed and nutrient absorption efficiency, and is suitable for the physiological needs of piglets in most mild climate scenarios.

[0079] For example, the local native piglets obtained an upper limit of ADCP dynamic demand of 21.5% in S22 (corresponding to a real-time temperature of 32℃, a high-temperature range), and the model matched it with an 8% SDF ratio target, forming an "ADCP-SDF" synergistic optimization combination, providing clear constraints for subsequent formula calculations.

[0080] S3: Feed Formulation Calculation

[0081] like Figure 1 As shown, this step is based on the upper limit of dynamic demand for digestible protein (ADCP) and the target of soluble fiber (SDF) ratio output by the S2 co-optimization model. Using these two as co-optimization targets, the available nutrient parameters of the raw materials after S1 pretreatment are calculated to ensure that the generated formula meets the physiological needs of pig breeds and environmental adaptability.

[0082] S31: Determine the target content for optimization

[0083] ADCP Optimization Target Content Determination: The formulation calculation module calls the ADCP dynamic demand upper limit output by S2, and combines it with the 90% safety margin coefficient preset in the pre-implementation preparation stage to determine the ADCP optimization target content. The basis for setting this safety margin coefficient comes from multiple batches of raw material nutrient fluctuation experiments: the experiments found that the actual ADCP content of feed raw materials may fluctuate naturally upward or downward by about ±5% during storage and transportation. To ensure that the ADCP content provided by the formulation will not exceed the upper limit of the pigs' dynamic demand under any fluctuation conditions, a 10% safety margin is reserved, that is, 90% of the upper limit of dynamic demand is taken as the optimization target content. ADCP Optimization Target Content (%) = Upper Limit of ADCP Dynamic Demand (%) × Safety Margin Coefficient (90%).

[0084] SDF synergistic target content determination: The formulation calculation module directly sets the SDF synergistic target content based on the fiber correlation ratio determined in S23. This logic originates from the matching experiment between temperature range and fiber ratio, aiming to ensure the synergy between intestinal peristalsis and nutrient absorption. For example, when the fiber correlation ratio is 8%, the SDF synergistic target content is 8%. In subsequent formulation calculations, the total SDF content of the formulation should be close to this target value. Taking the local native piglet case in S2 as an example: the upper limit of ADCP dynamic demand output by S2 is 21.5%, and the fiber correlation ratio is 8%. Therefore, the ADCP optimization target content = 21.5% × 90% ≈ 19.35%, and the SDF synergistic target content = 8%, forming a clear dual-objective constraint.

[0085] S32: Proportioning Solution

[0086] The formulation calculation module calls a linear programming ratio algorithm to screen readily available feed ingredients for the target farming scenario (such as extruded soybean meal, wheat middlings, rice husk powder, corn starch, etc.). Based on the available nutrient parameters of the ingredients output by S1 (actual ADCP content, actual SDF content), the addition ratio of each ingredient is determined. The core constraints of the algorithm are: the total ADCP content of the formulation should not exceed the optimized target ADCP content, and the error between the total SDF content of the formulation and the synergistic target SDF content should not exceed ±5%.

[0087] Taking the summer formula for local native piglets as an example: extruded soybean meal (ADCP 40%, SDF 6%), wheat middlings (ADCP 9%, SDF 8%), rice husk powder (ADCP 2%, SDF 12%), and corn starch (ADCP 1.8%, SDF 0.5%) are selected. After preliminary optimization by the algorithm, the proportion of raw materials is set to 38%, 28%, 17%, and 17% (totaling 100%). The formula calculation module calculates the following based on the following proportions: Total ADCP content = 38%×40%+28%×9%+17%×2%+17%×1.8%=15.2%+2.52%+0.34%+0.306%=18.37%, which is not higher than the ADCP optimization target content of 19.35%, thus meeting the constraint; Total SDF content = 38%×6%+28%×8%+17%×12%+17%×0.5%=2.28%+2.24%+2.04%+0.085%=6.65%, which is significantly different from the synergistic target content of 8% SDF and does not yet meet the ±5% error constraint. It is necessary to proceed to step S33 for enzymatic hydrolysis regulation correction or further adjust the raw material ratio.

[0088] S33: Enzymatic hydrolysis regulation correction

[0089] When the total SDF content of the formulation fails to meet the constraints, the formulation calculation module applies the pre-set enzymatic hydrolysis control method from the pre-implementation preparation stage for quantitative correction. The selected enzymes are cellulase and xylanase, which can indirectly regulate the effective content or functional ratio of soluble fiber (SDF) in the formulation by degrading some of the fiber structure in the raw materials and altering the fiber composition. Enzymatic hydrolysis comparison experiments confirmed that xylanase treatment at a 0.1% addition level significantly increased the effective SDF content of the formulation. For the aforementioned summer formulation case, the formulation calculation module first adjusted the raw material ratio to increase the total SDF content. If adjusting the ratio still fails to achieve the target, 0.1% xylanase can be added for correction to increase the effective functional content of SDF in the formulation.

[0090] S4: Feedback Adjustment and Iterative Optimization

[0091] This step is based on the actual feeding effect of the feed formula generated by S3. By collecting breeding data and comparing it with preset thresholds, the model parameters and formula constraints are dynamically corrected. The feedback data is directly related to the actual application effect of the S3 formula. The adjustment instructions are precisely applied to the collaborative optimization model or formula calculation module to ensure that the subsequent formulas continue to adapt to the target breeding scenario.

[0092] S41: Collect aquaculture effect data

[0093] The feedback adjustment module receives data on the breeding effect after feeding with the S3 optimized formula. Data collection is completed using equipment deployed during the pre-implementation preparation phase: When collecting diarrhea rate data, high-definition cameras deployed in the breeding sheds are used for observation, combined with on-site records made by farm inspectors according to clinical diagnostic standards, to calculate the proportion of piglets with diarrhea to the total number of piglets; when collecting daily weight gain data, the weighing equipment deployed in the early stages is used to measure the weight of each piglet once a week, and the ratio of the weight difference between two adjacent measurements to the interval is calculated; when collecting intestinal flora balance data, intestinal flora sequencers are used to collect piglet fecal samples once every two weeks, and the number of lactic acid bacteria and Escherichia coli in the samples is detected, and their ratio is calculated. The above data collection frequency and detection methods have been verified through multiple batches of breeding, which can ensure data accuracy while avoiding excessive testing that would increase breeding costs.

[0094] S42: Data Comparison and Command Generation

[0095] The feedback adjustment module compares the collected aquaculture performance data with the performance thresholds set during the pre-implementation preparation phase. Based on the comparison results, it generates targeted adjustment instructions: The preset threshold for diarrhea rate is 5%. When the diarrhea rate exceeds 5%, the module dynamically generates instructions to simultaneously adjust the ADCP optimization target content and SDF synergistic target content in subsequent calculations downwards, with an adjustment range of 5%-10%. The preset threshold for daily weight gain is 200g / day. When the daily weight gain is below 200g / day, the module generates instructions to adjust the temperature compensation intensity in the synergistic optimization model upwards, with an adjustment range of 10%-15% of the original compensation intensity. The preset threshold for gut microbiota balance is a lactic acid bacteria to Escherichia coli ratio of 2:1. When the ratio is below 2:1, the module generates instructions to adjust the SDF synergistic target content in subsequent calculations upwards, with an adjustment range of 10%-20% of the original target content.

[0096] S43: Iterative Calculation

[0097] The feedback adjustment module feeds back the generated adjustment instructions to the corresponding modules: the temperature compensation intensity adjustment instruction is fed back to the S2 co-optimization module, and the ADCP and SDF target content adjustment instructions are fed back to the S3 formula calculation module. After receiving the instructions, the corresponding modules update their own core parameters. For example, the co-optimization module updates the compensation intensity, and the formula calculation module updates the target content. Then, the S2 co-optimization model calculation and the S3 feed formula ratio calculation steps are re-executed to generate a new feed formula. To ensure that the model adapts to changes in the scenario in the long term, more than 20 batches of breeding scenario data are included every quarter, covering different temperatures, pig breeds, and raw material storage conditions. The core parameters such as the ADCP baseline threshold and temperature compensation intensity corresponding to the relative expression level of the TRY1 gene in S2 are further corrected to continuously improve the matching degree between the model calculation results and actual breeding needs, until a final feed formula that meets all preset effect thresholds is generated, namely, diarrhea rate ≤5%, daily weight gain ≥200g / day, and lactic acid bacteria to Escherichia coli ratio ≥2:1.

[0098] The implementation principle of a computer-based pig feed formulation data optimization method according to an embodiment of this application is as follows: This method integrates and acquires genetic characteristic parameters of pig breeds (such as gene expression levels related to protein digestive enzyme activity), real-time environmental temperature parameters, and raw material available nutrient parameters corrected for storage conditions, and inputs them into a collaborative optimization model. First, a baseline threshold for digestible protein is determined based on genetic characteristics. Then, dynamic compensation is performed based on the degree of temperature deviation from a preset range to generate a dynamic upper limit for digestible protein demand. Simultaneously, a synergistic target for the proportion of soluble fiber is matched according to the temperature compensation range, and these two are used as optimization targets. The method calculates the proportions of raw materials to generate a feed formula adapted to the current scenario. By constructing a dynamic synergistic model among pig breed genetic characteristics, environmental temperature, and the actual available nutrients in the raw materials, this method effectively solves the problems of formula bias and lag caused by the inability of traditional static formula technology to cope with complex scenarios with multiple coupled variables (such as regional climate fluctuations, differences in pig breed digestion, and nutrient decay of raw materials). It achieves real-time and accurate matching between feed formula and breeding scenario, thereby improving feed utilization efficiency, significantly reducing the diarrhea rate of piglets, promoting daily weight gain and maintaining intestinal flora balance, and comprehensively improving breeding efficiency and animal health.

[0099] This application discloses a computer-based pig feed formulation data optimization and processing system. The system includes a data acquisition module, a collaborative optimization module, a formulation calculation module, and a feedback adjustment module. After completing preliminary preparations such as parameter calibration, equipment deployment, and model initialization, the system can be started and run. Each module works collaboratively according to predetermined logic to achieve dynamic and precise optimization of the pig feed formulation. The specific implementation process is as follows:

[0100] The data acquisition module first extracts the relative expression level of the TRY1 gene, which is related to the activity of protein digestive enzymes, in the target pig herd using a gene sequencer. Then, it collects the real-time ambient temperature of the pig house using a temperature sensor. At the same time, it uses a near-infrared spectrometer to detect the nominal content of digestible protein (ADCP) and soluble fiber (SDF) in the feed ingredients and corrects the nutritional parameters based on the storage conditions of the ingredients. After that, the module filters out irrelevant data such as pig breed coat color and outside temperature, and only retains five core fields: relative expression level of TRY1 gene, real-time ambient temperature, corrected value of feed ADCP, corrected value of feed SDF, and storage conditions of feed, to provide accurate input for subsequent calculations.

[0101] The co-optimization module then receives this core data and first calculates the ADCP baseline threshold for different pig breeds based on the formula "ADCP baseline threshold (%) = 14% + TRY1 gene relative expression level × 5%". Next, it compares the real-time temperature with the suitable range of 15℃-25℃. If the temperature is higher than 25℃, the baseline threshold is increased by a first compensation intensity of 0.5% / ℃; if it is lower than 15℃, it is increased by a second compensation intensity of 0.3% / ℃. If the temperature falls within the summer intensive compensation range of 30℃-35℃ or the winter intensive compensation range of 5℃-10℃, the compensation intensity is adjusted based on the pig breed's heat tolerance. Then, the module matches the corresponding SDF ratio target to the upper limit of ADCP dynamic demand based on the temperature range, forming a co-optimization combination of "ADCP dynamic demand upper limit - SDF ratio".

[0102] The formulation calculation module calls this synergistic optimization combination, determines the ADCP optimization target content based on a 90% safety margin coefficient, and sets the SDF synergistic target content according to the matching fiber ratio. Then, it calls the linear programming algorithm to screen locally available raw materials such as extruded soybean meal and wheat middlings for ratio calculation. The total ADCP content of the formulation is required to be no higher than the optimization target content, and the error between the total SDF content and the synergistic target content is no more than ±5%. If the SDF content does not meet the constraints, the module will first try to adjust the raw material ratio. If it still does not meet the standard, enzyme preparations will be added for enzymatic hydrolysis to correct the effective SDF content.

[0103] Finally, the feedback adjustment module collects data on diarrhea rate, daily weight gain, and gut microbiota balance after formula feeding using a high-definition camera, weighing equipment, and gut microbiota sequencer. This data is then compared with preset thresholds: diarrhea rate ≤ 5%, daily weight gain ≥ 200g / day, and gut microbiota lactic acid bacteria to E. coli ratio ≥ 2:1. If the diarrhea rate exceeds the standard, the ADCP and SDF target content is simultaneously reduced by 5%-10%. If the daily weight gain does not meet the standard, the temperature compensation intensity is increased by 10%-15%. If the gut microbiota ratio is below the threshold, the SDF synergistic target content is increased by 10%-20%. The adjustment instructions are then fed back to the corresponding modules, driving them to update parameters and recalculate. In addition, more than 20 batches of aquaculture data from different scenarios are incorporated every quarter to correct the core parameters of the model until a final feed formula that meets all effect thresholds is generated.

[0104] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A computer-based method for optimizing pig feed formulation data, characterized in that, Includes the following steps: Obtain the genetic characteristic parameters of the pig breed, real-time environmental temperature parameters, and raw material available nutrient parameters corrected for the current storage conditions corresponding to the target breeding scenario; The genetic characteristics parameters of the pig breed, the real-time ambient temperature parameters, and the available nutrient parameters of the raw materials are input into a pre-set collaborative optimization model. The collaborative optimization model is configured to: determine a baseline threshold for digestible protein based on the genetic characteristics parameters of the pig breed; dynamically compensate the baseline threshold based on the deviation of the real-time ambient temperature parameters from a preset temperature range to generate a dynamic upper limit for digestible protein requirements; and determine a target proportion of soluble fiber that aligns with the dynamic upper limit for requirements based on the temperature compensation range corresponding to the dynamic compensation. Receive the upper limit of dynamic demand for digestible protein and the target proportion of soluble fiber, which are output by the collaborative optimization model and matched with the current target aquaculture scenario; Using the received upper limit of dynamic demand for digestible protein and the target proportion of soluble fiber as synergistic optimization objectives, the available nutrient parameters of the raw materials are calculated to generate a final feed formula that meets the synergistic optimization objectives. The collaborative optimization model includes the following steps for dynamic compensation based on real-time ambient temperature parameters: when the real-time ambient temperature parameter exceeds the upper limit threshold of a preset temperature range, a first compensation intensity is applied to compensate the benchmark threshold; when the real-time ambient temperature parameter is lower than the lower limit threshold of a preset temperature range, a second compensation intensity less than the first compensation intensity is applied to compensate the benchmark threshold.

2. The method according to claim 1, characterized in that, The step of determining the target proportion of soluble fiber in the collaborative optimization model further includes: When the real-time ambient temperature parameter is higher than the upper limit threshold of the preset temperature range, the first fiber correlation ratio is matched with the upper limit of the dynamic demand for digestible protein. When the real-time ambient temperature parameter is lower than the lower limit threshold of the preset temperature range, the upper limit of the dynamic demand for digestible protein is matched with a second fiber association ratio that is higher than the first fiber association ratio. When the real-time ambient temperature parameter is between the upper and lower thresholds, the upper limit of the dynamic demand for digestible protein is matched with a third fiber association ratio that is between the first fiber association ratio and the second fiber association ratio.

3. The method according to claim 2, characterized in that, The method uses the received upper limit of the dynamic demand for digestible protein and the target proportion of soluble fiber as co-optimization objectives to calculate the proportion of available nutrient parameters of the raw materials, and performs the following steps: Based on the upper limit of the dynamic demand for digestible protein, combined with a preset safety margin coefficient, the optimized target content of digestible protein is determined. Based on the optimized target content of the digestible protein and the fiber association ratio matched by the collaborative optimization model, the collaborative target content of soluble fiber is determined. Perform a formulation calculation to ensure that the total digestible protein content of the formula, calculated based on the available nutrient parameters of the raw materials, is not higher than the optimized target content of the digestible protein, and that the total soluble fiber content of the formula meets the constraints defined based on the synergistic target content of the soluble fiber.

4. The method according to claim 3, characterized in that, Also includes: The actual aquaculture effect data is judged based on multiple preset effect thresholds; If the diarrhea rate is determined to exceed the preset diarrhea threshold, the optimized target content of the digestible protein and the synergistic target content of the soluble fiber used in subsequent calculations will be adjusted downwards simultaneously. If the daily weight gain is determined to be lower than the preset weight gain threshold, the compensation intensity used for dynamic compensation in the collaborative optimization model is adjusted upwards. If the gut microbiota balance index is determined to be lower than the preset microbiota balance threshold, the synergistic target content of the soluble fiber used in subsequent calculations is adjusted upwards.

5. The method according to claim 3, characterized in that, Before generating the final feed formulation, the following steps are also included: When the total soluble fiber content of the formula calculated based on the available nutrient parameters of the raw materials fails to meet the constraint conditions defined by the synergistic target content of the soluble fiber, the total soluble fiber content of the formula is quantitatively corrected by applying a preset enzymatic hydrolysis control method until it meets the constraint conditions.

6. The method according to claim 1, characterized in that, The genetic characteristic parameters of the pig breed are gene expression data related to the activity of protein digestive enzymes; the available nutrient parameters of the raw materials are the actual content of digestible protein and the actual content of soluble fiber obtained after correction based on specific temperature and humidity storage conditions.

7. The method according to claim 1, characterized in that, The collaborative optimization model is further configured as follows: Outside the preset temperature range, an enhanced compensation range is defined that is associated with the periodic climate characteristics of the target aquaculture scenario; When the real-time ambient temperature parameter falls within the enhanced compensation range, the first compensation intensity or the second compensation intensity is adaptively modulated based on the heat tolerance metabolic regulation data associated with the genetic characteristic parameters of the pig breed, and the upper limit of the dynamic demand for digestible protein and the target proportion of soluble fiber are generated according to the modulated compensation intensity.

8. A computer-based pig feed formulation data optimization and processing system, characterized in that, include: The data acquisition module is configured to acquire the genetic characteristic parameters of the pig breed, the real-time ambient temperature parameters, and the raw material available nutrient parameters corrected by the current storage conditions for the target breeding scenario. The collaborative optimization module, which has a built-in collaborative optimization model, is configured to: determine a baseline threshold for digestible protein based on the genetic characteristic parameters of the pig breed; and dynamically compensate the baseline threshold according to the deviation of the real-time ambient temperature parameter from the preset temperature range to generate a dynamic upper limit for the demand of digestible protein. Based on the temperature compensation range corresponding to the dynamic compensation, a target soluble fiber ratio that is coordinated with the upper limit of the dynamic demand is determined; when the real-time ambient temperature parameter exceeds the upper limit threshold of the preset temperature range, a first compensation intensity is applied to compensate the benchmark threshold; when the real-time ambient temperature parameter is lower than the lower limit threshold of the preset temperature range, a second compensation intensity less than the first compensation intensity is applied to compensate the benchmark threshold. The formulation calculation module is configured to use the upper limit of the dynamic demand for digestible protein and the target proportion of soluble fiber output by the co-optimization module as co-optimization targets, and to perform proportion calculations on the available nutrient parameters of the raw materials to generate a feed formulation that meets the co-optimization targets.

9. The system according to claim 8, characterized in that, It also includes a feedback adjustment module, which is configured as follows: Receive breeding effect data collected from actual feeding based on the feed formula; The breeding effect data is compared with preset thresholds for diarrhea rate, daily weight gain, and gut microbiota balance. Based on the comparison results, the system dynamically generates adjustment instructions for the dynamic compensation logic in the collaborative optimization model or correction instructions for the proportion constraints in the formula calculation module, and feeds these instructions back to the corresponding modules for iterative calculation until a final feed formula that meets the preset effect threshold is generated.

Images