Livestock feed proportioning method based on livestock growth

Abstract

The invention discloses a livestock feed proportioning method based on livestock growth, and relates to the technical field of feed proportioning, and the livestock feed proportioning method comprises the following steps: collecting the body temperature and motion track data and environmental temperature and humidity data of livestock individuals, monitoring the concentration of amino acids in digestive tracts, preprocessing the collected data, and generating a standardized biological data set; analyzing the compensation amount of vitamins and trace elements according to the metabolic fingerprint spectrum, and constructing a multi-objective optimization function in combination with an individual growth curve; solving the multi-objective optimization function through a quantum annealing algorithm, and matching local raw material inventory data of block chain evidence storage to generate an optimization strategy; and executing the optimization strategy and detecting the mixing uniformity, and when the mixing uniformity index exceeds a preset mixing uniformity threshold, starting the compensation device to adjust the optimization strategy. According to the invention, through a cooperation mechanism of dynamic nutritional requirement identification and multi-constraint optimization, the precise regulation and control capability of livestock breeding is significantly improved.

Description

Technical Field

[0001] The invention relates to the technical field of feed ratio, in particular to a livestock feed ratio method based on livestock growth. Background Art

[0002] In recent years, the field of precision livestock farming has gradually integrated IoT sensing, metabolomics, and big data modeling, driving the development of personalized nutritional regulation. Real-time monitoring based on multimodal data has been widely used to assess growth status, dynamically capturing the interaction between livestock and poultry physiological behaviors and the environment. Metabolomics provides a molecular basis for nutritional needs analysis. In terms of feed ratio optimization, traditional models such as linear programming and genetic algorithms, combined with raw material inventory data, have initially achieved a balance between nutrient supply and resource constraints. Blockchain provides trusted evidence support for raw material traceability and inventory management, while quantum computing demonstrates the potential for parallel solutions in complex combinatorial optimization problems.

[0003] Current approaches lack the ability to coordinate dynamic nutrient demand analysis with feed resource optimization in real time. This inability to effectively integrate individual physiological signals, metabolic pathway data, and inventory constraints limits the accuracy and timeliness of nutrient compensation strategies. Metabolomics results are poorly correlated with growth models, and raw material inventory data is not embedded in the optimization process, leading to mismatches between recipes and resources. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a livestock feed ratio method based on livestock growth to solve the problem of coordinated matching between the dynamic nutritional needs of individual livestock and the real-time optimization of feed formula.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a livestock feed ratio method based on livestock growth, which includes collecting body temperature and movement trajectory data of individual livestock and environmental temperature and humidity data, monitoring amino acid concentrations in the digestive tract, preprocessing the collected data, and generating a standardized biological data set;

[0008] The standardized biological data set is input into the pre-trained Gompertz prediction model to calculate the predicted value of the individual growth curve. When the deviation rate exceeds the preset deviation threshold, the metabolomics detection instruction is triggered;

[0009] According to the metabolomics detection instructions, the mass spectrometer is started to perform non-targeted detection on feces and blood samples to generate metabolic fingerprints;

[0010] Analyze the compensation amount of vitamins and trace elements based on metabolic fingerprints, and construct a multi-objective optimization function based on individual growth curves;

[0011] The multi-objective optimization function is solved by the quantum annealing algorithm, matching the local raw material inventory data stored in the blockchain to generate an optimization strategy.

[0012] The optimization strategy is executed and the mixing uniformity is detected. When the mixing uniformity index exceeds a preset mixing uniformity threshold, the compensation device is started to adjust the optimization strategy.

[0013] As a preferred embodiment of the livestock feed ratio method based on livestock growth of the present invention, the specific steps of generating a standardized biological data set are as follows:

[0014] Collect body temperature and movement trajectory data of individual poultry and livestock, and calculate activity intensity index;

[0015] Collect ambient temperature and humidity data and use Grashof number to dynamically calibrate temperature drift error;

[0016] By monitoring the lysine concentration in the digestive tract, the wavelength offset of the Bragg grating is used to generate real-time amino acid concentration data;

[0017] The dynamic time warping algorithm is used to align the time series of body temperature, activity intensity index, amino acid concentration, and ambient temperature and humidity data to generate a millisecond-level synchronized time series dataset.

[0018] The time series dataset is noise filtered, outliers are removed through Mahalanobis distance detection, and a standardized biological dataset is generated.

[0019] As a preferred embodiment of the livestock feed ratio method based on livestock growth of the present invention, the triggering of the metabolomics detection instruction comprises the following specific steps:

[0020] The standardized biological data set was normalized and input into the pre-trained Gompertz prediction model to calculate the individual growth curve prediction value;

[0021] Based on the predicted value of the individual growth curve, the real-time monitoring value of the poultry and livestock weight and movement trajectory data is obtained synchronously, and the deviation rate between the predicted value of the individual growth curve and the real-time monitoring value is output;

[0022] When the deviation rate exceeds a preset deviation threshold, a detection instruction of the priority parameters is sent to the metabolomics detection.

[0023] As a preferred embodiment of the livestock feed ratio method based on livestock growth described in the present invention, the priority parameter refers to the sample detection priority level dynamically assigned according to the real-time monitored amino acid concentration data in the metabolomics detection instruction, which determines the detection order and resource allocation weight of different livestock individuals or samples.

[0024] As a preferred embodiment of the livestock feed ratio method based on livestock growth of the present invention, the steps of generating a metabolic fingerprint are as follows:

[0025] Based on the priority parameters in the metabolomics detection instructions, the mass spectrometer's mass-to-charge ratio scanning range and resolution parameters are dynamically configured to generate a mass spectrometer control file suitable for the sample type;

[0026] According to the mass spectrometer control document, fecal samples were enriched for short-chain fatty acids using solid-phase microextraction, and blood samples were enriched for amino acids using confined electrodialysis to obtain pretreated metabolite-enriched samples;

[0027] The metabolite-enriched sample was introduced into the mass spectrometer, and ionization and fragmentation were performed by dividing the variable window in the data-independent acquisition mode to collect the raw mass spectrometry data;

[0028] A dynamic window peak alignment algorithm was performed on the raw mass spectrometry data, and the metabolite feature vectors were calculated in combination with the KEGG metabolic pathway topology weights to generate a metabolic fingerprint.

[0029] As a preferred embodiment of the livestock feed ratio method based on livestock growth of the present invention, the specific steps of constructing a multi-objective optimization function are as follows:

[0030] Based on the metabolic fingerprint, a dynamic metabolic flux probability distribution model is constructed to output the flux evolution parameters of vitamins and trace elements;

[0031] Using the flux evolution parameters, the time convolution integral of the individual growth curve is calculated to generate the compensation requirement vector of vitamins and trace elements;

[0032] A multi-objective optimization function is constructed based on the compensation demand vector.

[0033] As a preferred embodiment of the livestock feed ratio method based on livestock growth of the present invention, the generation optimization strategy comprises the following specific steps:

[0034] The multi-objective optimization function is encoded into an optimization matrix adapted to the quantum annealer. The nutrient matching, growth deviation suppression, and cost minimization are comprehensively weighted through dynamic weight parameters to generate a weighted matrix.

[0035] Based on the local raw material inventory data stored on the blockchain, the inventory scarcity of each raw material is calculated and the annealing rate parameters of the quantum annealer are dynamically adjusted;

[0036] The optimization matrix is ​​mapped to the quantum bit topology of the quantum annealer, and the quantum bit state is decoded after the annealing operation to generate the optimization strategy.

[0037] As a preferred embodiment of the livestock feed ratio method based on livestock growth of the present invention, the startup compensation device adjustment optimization strategy comprises the following specific steps:

[0038] Control the motor speed and feed valve opening of the feed mixing device according to the optimization strategy to generate the initial mixed feed;

[0039] The near-infrared spectrometer is used to scan multiple sampling points of the initial mixed feed in real time to obtain the spectral distribution data of each nutrient;

[0040] Calculating a mixing uniformity index based on the spectral distribution data, and triggering a compensation device to add nutrients to the mixed feed when the mixing uniformity index exceeds a preset mixing uniformity threshold;

[0041] According to the nutrient content of the compensated mixed feed, the raw material ratio weight in the optimization strategy is dynamically adjusted to generate an updated optimization strategy;

[0042] The updated optimization strategy is re-input into the feed mixing device until the mixing uniformity index is lower than the preset mixing uniformity threshold.

[0043] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the livestock feed ratio method based on livestock growth as described in the first aspect of the present invention is implemented.

[0044] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the livestock feed ratio method based on livestock growth as described in the first aspect of the present invention is implemented.

[0045] The beneficial effects of the present invention are: through the synergistic mechanism of dynamic nutritional demand identification and multi-constraint optimization, the precise control capability of poultry and livestock farming is significantly improved. The adaptive triggering strategy based on growth deviation can accurately activate the detection process at the early stage of metabolic imbalance, and realize the targeted analysis of vitamin and trace element compensation needs; the quantum optimization engine integrating blockchain inventory data breaks through the real-time generation of the global optimal formula under complex constraints, greatly improving the efficiency of raw material utilization; the closed-loop compensation control mechanism ensures the uniformity of feed nutrient distribution and biological effectiveness by dynamically adjusting the ratio weights, and guarantees the dynamic adaptation of individual growth and resource supply. It realizes the collaborative optimization of the entire link from physiological monitoring to formula execution, providing efficient and reliable technical support for precision farming. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 The figure is a flow chart of the livestock feed ratio method based on livestock growth.

[0048] Figure 2 This is a flow chart of data collection and standardization processing.

[0049] Figure 3 Construct a flow chart for metabolic detection and optimization model.

[0050] Figure 4 Adjust the flow chart for strategy execution and feedback. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0054] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides a livestock feed ratio method based on livestock growth, comprising the following steps:

[0055] S1. Collect body temperature and movement trajectory data of individual poultry and livestock, as well as environmental temperature and humidity data, monitor amino acid concentrations in the digestive tract, pre-process the collected data, and generate a standardized biological data set.

[0056] Furthermore, the body temperature and movement trajectory data of individual poultry and livestock are collected to calculate the activity intensity index;

[0057] It should be noted that the body temperature data of individual poultry and livestock are collected through smart ear tags that integrate the MAX30205 chip, and the body temperature data is recorded at fixed time intervals. The body temperature data is transmitted wirelessly to the gateway; the motion trajectory data is collected using UWB positioning tags, and the three-dimensional coordinates are recorded at fixed time intervals.

[0058] Deploy UWB positioning tags to record movement trajectories every 5 seconds, and calculate the activity intensity index based on the trajectory coordinates. The expression is:

[0059]

[0060] Where A is the activity intensity index, x t Expressed as the horizontal coordinate of the current time t, y t It is represented as the vertical coordinate of the current time t, t represents the current time, x t-1 Expressed as the horizontal coordinate of the moment before the current moment t, y t-1 It represents the vertical coordinate of the moment before the current moment t, and n represents the coordinate position index. It represents the cumulative value of all coordinate displacements from t=1 to t=n, and Δt represents the time interval.

[0061] Collect ambient temperature and humidity data and use Grashof number to dynamically calibrate temperature drift error;

[0062] Specifically, the sensor collects ambient temperature and humidity data, recording it at regular intervals. The current Grashof number is calculated using the difference between the sensor surface temperature and the ambient air temperature, the sensor's characteristic length, the acceleration of gravity, the air's volume expansion coefficient, and the air's kinematic viscosity. The resulting Grashof number is compared with a pre-stored reference Grashof number, and the original temperature reading is corrected using a linear compensation formula to eliminate sensor drift errors caused by convection.

[0063] Calculate the current Grashof number using the expression:

[0064]

[0065] Where G is the current Grashof number, g is the acceleration due to gravity, β is the volume expansion coefficient of air, ΔR is the difference between the sensor surface temperature and the ambient air temperature, L is the characteristic length of the sensor, and v is the kinematic viscosity of air.

[0066] It should be noted that the volume expansion coefficient of air represents the relative rate of change of gas volume when the temperature changes, and its source is based on the ideal gas state equation; for example, at room temperature of 25°C, β≈0.00336.

[0067] By monitoring the lysine concentration in the digestive tract, the wavelength offset of the Bragg grating is used to generate real-time amino acid concentration data;

[0068] Specifically, the lysine concentration in the digestive tract is monitored by an implantable photonic crystal fiber sensor. The reflection wavelength of the fiber Bragg grating integrated in the photonic crystal fiber sensor shifts as the lysine concentration changes. The wavelength offset of the fiber Bragg grating is detected by a high-resolution spectrometer, and the wavelength offset is input into a calibration equation to calculate the real-time lysine concentration data.

[0069] The lysine concentration in the terminal ileum was monitored using an implantable photonic crystal fiber sensor. The amino acid concentration data was calculated using the fiber Bragg grating wavelength offset. The expression is:

[0070]

[0071] Where [Lys] represents the lysine concentration, K represents the sensor sensitivity coefficient, δλ represents the fiber Bragg grating wavelength offset, λ represents the fiber Bragg grating wavelength, δ represents the offset, λ0 represents the initial Bragg wavelength, and b represents the sensor baseline offset calibration constant. Expressed as a normalized ratio of wavelength shifts.

[0072] It should be noted that the sensor sensitivity coefficient is obtained by calibrating the photonic crystal fiber sensor in a lysine gradient solution, for example, a value of 8.37. The high sensitivity is due to the light field localization enhancement effect of the photonic crystal structure.

[0073] The dynamic time warping algorithm is used to align the time series of body temperature, activity intensity index, amino acid concentration, and ambient temperature and humidity data to generate a millisecond-level synchronized time series dataset.

[0074] Specifically, taking body temperature as the reference sequence, the minimum cumulative distance paths between the activity intensity index, amino acid concentration, ambient temperature, ambient humidity and the body temperature sequence are calculated respectively to generate a dynamic time warping path matrix; based on the backtracking alignment points of the path matrix, the timestamps of the activity intensity index, amino acid concentration, ambient temperature, and ambient humidity are mapped to the millisecond time points corresponding to the body temperature sequence; for the unaligned local time intervals, the linear interpolation method is used to fill the missing values, and the millisecond timestamps of all sequences are merged to generate a millisecond-level synchronized time series dataset containing body temperature, activity intensity index, amino acid concentration, ambient temperature and ambient humidity.

[0075] The time series dataset is noise filtered, outliers are removed through Mahalanobis distance detection, and a standardized biological dataset is generated.

[0076] It should be noted that the sliding window mean method is used to smooth high-frequency fluctuations. For example, the window size is set to 5 seconds, and random noise is reduced by averaging the data of adjacent time points. The covariance matrix of each time point in the time series data set and the overall mean is calculated, and then the Mahalanobis distance value of each time point is calculated based on the covariance matrix. For example, the time points corresponding to the Mahalanobis distance values ​​exceeding 3 times the standard deviation of the mean are marked as outliers and eliminated. The time series data set after noise filtering and outlier elimination is Z-score standardized. The mean of the corresponding column of each biological indicator data is subtracted and divided by the standard deviation to ensure that the mean of each biological indicator is 0 and the standard deviation is 1.

[0077] Specifically, outliers are removed through Mahalanobis distance detection, and the expression is:

[0078]

[0079] Where D is the Mahalanobis distance, x is the multidimensional data vector to be detected, D(x) is the outlier removed by the Mahalanobis distance, μ is the mean vector of the data set, Z is the vector transpose operation, ∑ -1 It is represented as the inverse matrix of the covariance matrix, and (x-μ) is represented as the deviation vector between the multidimensional data vector to be detected and the mean vector of the data set.

[0080] S2. Input the standardized biological data set into the pre-trained Gompertz prediction model to calculate the predicted value of the individual growth curve. When the deviation rate exceeds the preset deviation threshold, the metabolomics detection instruction is triggered.

[0081] The standardized biological data set was normalized and input into the pre-trained Gompertz prediction model to calculate the individual growth curve prediction value;

[0082] It should be noted that the observed values ​​of each sample are extracted and linearly mapped to the interval [0, 1] using a range transformation. For example, the minimum sample value is set to 0, the maximum sample value is set to 1, and intermediate values ​​are scaled proportionally. The processed data retains the original time series structure and is arranged in timestamp order before being input into the Gompertz prediction model for pre-training. A nonlinear optimization algorithm is used to minimize the mean squared error between the predicted values ​​and the normalized observed values. Parameters are iteratively adjusted until convergence. The optimal parameters are saved and the Gompertz prediction model is validated to achieve accurate growth curve prediction and abnormal deviation detection. The prediction results can be restored to a standardized biological dataset through inverse normalization.

[0083] Specifically, the growth curve prediction value is calculated as follows:

[0084]

[0085] Where S(t) represents the predicted value of the growth curve, t0 is the initial time, represents the cumulative sum from t0 to t, exp represents the exponential function, β represents the exponential decay coefficient, τ represents the historical moment, t-τ represents the time difference between the current moment t and the historical moment τ, W(τ) represents the predicted weight at the historical moment τ, W′(τ) represents the actual observed weight at the historical moment τ, α represents the lysine sensitivity coefficient, Δ[Lys](τ) represents the deviation of the lysine concentration at the historical moment τ from the baseline value, dτ represents the historical moment differential, and γ represents the activity intensity index weight. Expressed as the second derivative of the activity intensity index.

[0086] It should be noted that the exponential decay coefficient is determined by half-life analysis of growth deviation data, so that the weight of historical data decays exponentially over time, for example, a value of 0.23; the lysine sensitivity coefficient is obtained by optimizing the metabolic inhibition effect through the quantum annealing algorithm, for example, a value of 0.17, which is used to adjust the inhibitory effect of lysine concentration deviation on the integral term; the activity intensity index weight is calibrated by the statistical correlation between motion mutation events and growth abnormalities, quantifying the contribution intensity of acceleration to the deviation value, for example, a value of 1.5.

[0087] Based on the predicted value of the individual growth curve, the real-time monitoring value of the poultry and livestock weight and movement trajectory data is obtained synchronously, and the deviation rate between the predicted value of the individual growth curve and the real-time monitoring value is output;

[0088] It should be noted that based on the predicted values ​​of the individual growth curve, poultry and livestock weight data is collected at preset time intervals using an electronic weighing device, for example, real-time weight is recorded every 2 hours. Movement trajectory data is continuously acquired using a high-precision GPS positioning device and a three-axis acceleration sensor, for example, the position coordinates are updated every 30 seconds and the displacement per unit time is calculated. The difference between the real-time weight data and the predicted value of the individual growth curve is calculated. For example, if the predicted weight is 50kg and the real-time weight is 48kg, the absolute deviation is 2kg. The movement trajectory displacement is also normalized with the predicted activity intensity of the individual growth curve. For example, if the predicted displacement is 10km per day and the real-time displacement is 8km, the relative deviation rate is 20%. The comprehensive deviation rate between the predicted value of the individual growth curve and the real-time monitoring value is output.

[0089] When the deviation rate exceeds a preset deviation threshold, a detection instruction of the priority parameters is sent to the metabolomics detection.

[0090] It should be noted that the preset deviation threshold refers to a pre-set critical value used to determine the degree of deviation between the predicted weight of an individual's growth curve and the real-time monitored weight, specifically expressed as a numerical ratio or absolute difference. For example, if the predicted weight of an individual's growth curve is 50 kg, the deviation threshold is set at 5% (i.e., 2.5 kg). A real-time monitored weight value below 47.5 kg or above 52.5 kg is considered to have exceeded the deviation rate limit.

[0091] When the deviation rate does not exceed the preset deviation threshold, the normal detection process continues.

[0092] Specifically, when the deviation rate between the predicted value of the individual growth curve and the real-time monitoring value exceeds a preset deviation threshold, for example, the deviation rate exceeds 5%, a priority parameter is generated, the priority parameter is associated with the metabolomics detection instruction, and the metabolomics detection instruction carrying the priority parameter is sent to the mass spectrometer through the communication interface, triggering the mass spectrometer to adjust the detection queue order according to the priority parameter.

[0093] It should be noted that the priority parameter refers to the sample detection priority level dynamically assigned according to the real-time monitored amino acid concentration data in the metabolomics detection instructions, which determines the detection order and resource allocation weight of different poultry and livestock individuals or samples.

[0094] S3. According to the metabolomics detection instructions, the mass spectrometer is started to perform non-targeted detection on feces and blood samples to generate metabolic fingerprint maps.

[0095] Based on the priority parameters in the metabolomics detection instructions, the mass spectrometer's mass-to-charge ratio scanning range and resolution parameters are dynamically configured to generate a mass spectrometer control file suitable for the sample type;

[0096] Specifically, based on the priority parameters in the metabolomics detection instructions, the priority parameters are mapped to a preset mass-to-charge ratio scanning range and resolution parameter mapping table. For example, when the priority parameter is 0.6, the corresponding mass-to-charge ratio scanning range is 100-1500 m / z and the resolution is 70000; the corresponding scanning mode configuration template is called according to blood or feces, for example, the polarity switching mode is enabled for blood samples and the single ion mode is enabled for feces samples; the priority parameters are combined with blood or feces to generate a mass spectrometer control file containing the mass-to-charge ratio range, resolution and scanning mode.

[0097] It should be noted that the preset mass-to-charge ratio scanning range mapping table refers to a pre-established correspondence table between priority parameters and mass / charge ratio detection intervals of the mass spectrometer, which is used to limit the molecular weight range of target metabolites for non-targeted detection. For example, when the priority parameter is 0.3, the corresponding scanning range is 50-800m / z (covering small molecule metabolites such as amino acids and short-chain fatty acids), and when the priority parameter is 0.8, the corresponding scanning range is 200-2000m / z (covering macromolecular metabolites such as lipids and polysaccharides). The mapping table is dynamically adjusted according to the statistical results of the historical metabolite molecular weight distribution. For example, 80% of the metabolites in blood samples are distributed in the 200-1500m / z interval, and 60% of the metabolites in fecal samples are distributed in the 50-800m / z interval.

[0098] The preset resolution parameter mapping table refers to the correspondence table between the pre-set priority parameters and the mass resolution capability of the mass spectrometer, which is used to control the accuracy of distinguishing ions with adjacent mass-to-charge ratios. For example, a priority parameter of 0.3 corresponds to a resolution of 30,000 (rapid screening mode, which can distinguish isotope peak spacing of 0.01 m / z), and a priority parameter of 0.8 corresponds to a resolution of 120,000 (high-precision mode, which can distinguish trace metabolites with a spacing of 0.001 m / z). The mapping table is dynamically configured based on the abundance and interference level of the target metabolites. For example, high-priority detection requires improved resolution to reduce matrix effect interference, while low-priority detection requires reduced resolution to shorten the scanning cycle.

[0099] According to the mass spectrometer control document, fecal samples were enriched for short-chain fatty acids using solid-phase microextraction, and blood samples were enriched for amino acids using confined electrodialysis to obtain pretreated metabolite-enriched samples;

[0100] It should be noted that the sample type refers to the type of biological sample to be analyzed in metabolomics testing, such as stool samples or blood samples, and the scan mode parameter refers to the ion scanning configuration used by the mass spectrometer when performing non-targeted detection (such as positive and negative ion alternating scanning mode or single ion scanning mode).

[0101] Specifically, based on the sample type and scan mode parameters defined in the mass spectrometer control file, a solid-phase microextraction fiber was used to adsorb short-chain fatty acids from fecal samples. For blood samples, a confined electrodialysis device applied an electric field to direct the migration of amino acids into the enrichment chamber, removing large proteins. After pretreatment, the metabolite-enriched samples showed increased short-chain fatty acid concentrations in fecal samples and improved amino acid recovery in blood samples, resulting in enriched samples that matched the detection sensitivity of the mass spectrometer control file.

[0102] The metabolite-enriched sample was introduced into the mass spectrometer, and ionization and fragmentation were performed by dividing the variable window in the data-independent acquisition mode to collect the raw mass spectrometry data;

[0103] Specifically, the metabolite-enriched sample is introduced into the ion source of the mass spectrometer through an automatic sample injector. According to the data-independent acquisition mode defined in the mass spectrometer control file, the variable window is divided according to the mass-to-charge ratio range. Electrospray ionization is used to generate parent ions in each window, which are then fragmented by collision-induced dissociation after quadrupole screening to generate daughter ions. Full-scan mass spectrometry data are collected at a speed of 10 spectra per second, for example, the parent ion scan range is 100-2000 m / z, the resolution is 70000, and the daughter ion scan range is 50-2000 m / z, the resolution is 35000, to generate raw mass spectrometry data containing ion abundance, mass-to-charge ratio and fragmentation information.

[0104] A dynamic window peak alignment algorithm was performed on the raw mass spectrometry data, and the metabolite feature vectors were calculated in combination with the KEGG metabolic pathway topology weights to generate a metabolic fingerprint.

[0105] Specifically, a dynamic window peak alignment algorithm is performed on the original mass spectrometry data, and the window width is dynamically adjusted based on the mass-to-charge ratio of the parent ion and the retention time drift; according to the KEGG metabolic pathway topology weight, the pathway nodes related to vitamin and trace element metabolism are extracted, and the pathway weight weighted calculation is performed on the characteristic peak abundance to generate a metabolite feature vector containing mass-to-charge ratio-retention time-weighted abundance; the feature vectors are arranged in ascending order according to mass-to-charge ratio to generate a metabolic fingerprint map with mass-to-charge ratio as the horizontal axis and weighted abundance as the vertical axis.

[0106] S4. Analyze the compensation amount of vitamins and trace elements based on the metabolic fingerprint map, combine it with the individual growth curve, and construct a multi-objective optimization function.

[0107] Furthermore, based on the metabolic fingerprint, a dynamic metabolic flux probability distribution model is constructed to output the flux evolution parameters of vitamins and trace elements;

[0108] Specifically, based on the mass-to-charge ratio-retention time-weighted abundance data in the metabolic fingerprint, the weighted abundance time series of vitamins and trace elements were extracted, and the KEGG metabolic pathway topology weight was combined as the prior probability distribution. The posterior probability density function of the metabolic flux was calculated by Markov chain Monte Carlo sampling; the time series data was input into the dynamic Bayesian network, and the flux conditional probability table was updated in hours. The flux evolution parameters of vitamins and trace elements were output, including the flux mean change rate, fluctuation amplitude and covariance coefficient with the growth curve.

[0109] Using the flux evolution parameters, the time convolution integral of the individual growth curve is calculated to generate the compensation requirement vector of vitamins and trace elements;

[0110] Specifically, based on the flux mean change rate, fluctuation amplitude and covariance coefficient with the growth curve in the flux evolution parameters, combined with the dynamic parameters of the individual growth curve, the flux evolution parameters and the individual growth curve are convolved and integrated in the time dimension. For example, the convolution kernel function of the vitamin flux mean change rate and the growth rate is an exponential decay function, and the integration result is converted into the hourly vitamin compensation dose; the covariance integral value of the trace element flux and the growth curve is calculated simultaneously. For example, the integral weight corresponding to the zinc ion covariance coefficient of -0.3 is 0.7, and a compensation demand vector containing the compensation dose, time series distribution and compensation form of vitamins and trace elements is generated. For example, the vitamin compensation demand vector is [dose 0.5 mg / h, time series 0-24h, sustained-release form], and the zinc ion compensation demand vector is [dose 2 mg / h, time series 8-18h, liquid form].

[0111] A multi-objective optimization function is constructed based on the compensation demand vector.

[0112] It should be noted that the multi-objective optimization function includes nutritional matching, growth deviation suppression and cost minimization; nutritional matching refers to the degree of consistency between the compensatory dose, temporal distribution and morphological parameters of vitamins and trace elements in feed and the dynamic metabolic flux evolution parameters analyzed from the metabolic fingerprint map. For example, the compensatory dose of vitamins needs to cover the physiological demand increment corresponding to the mean change rate of metabolic flux, and the compensatory temporal distribution needs to match the regulation window of flux fluctuation amplitude to ensure the dynamic balance between nutritional supply and metabolic demand.

[0113] Growth deviation inhibition refers to reducing the comprehensive deviation rate between the predicted value and the real-time monitoring value of the individual growth curve through compensation strategies. For example, the negative correlation between the vitamin compensation dose and the growth rate covariance coefficient can inhibit the lag in weight gain, and the integral weight of the trace element flux evolution parameter and the motion trajectory displacement can reduce the fluctuation of the activity intensity index, thereby maintaining the steady-state convergence of the growth curve.

[0114] Cost minimization refers to selecting the most cost-effective raw material combination and compensation form through local raw material inventory data and market price mapping tables stored in blockchain, while satisfying the constraints of nutritional matching and growth deviation suppression.

[0115] Specifically, a multi-objective optimization function is constructed based on the compensation demand vector, and the expression is:

[0116]

[0117] Where, It is represented as the value of the multi-objective optimization function, a is represented as the decision variable vector, erf is represented as the error function, P is represented as the nutritional matching degree, r N It is expressed as the mean of the nutritional dimension benchmark, N is expressed as the nutritional dimension benchmark, r is expressed as the mean, σ N is represented by the standard deviation of the nutritional dimension, σ is represented by the standard deviation, G(a) is represented by the growth deviation inhibition term, κ is represented by the growth inhibition attenuation coefficient, λ is represented by the cost weight adjustment factor, C(a) is represented by the cost function, and C max is the cost upper limit, ε is the cost nonlinear coefficient, sinc(·) is the normalization function, π is the mathematical constant, T′ is the time period, Expressed as the deviation of normalized time phase.

[0118] It should be noted that the growth deviation inhibition term calculates the cumulative deviation between actual and predicted body weight through time convolution integral; the growth inhibition attenuation coefficient is determined by exponential fitting of historical growth deviation data to control the attenuation rate of the deviation impact, for example, the value is 2.3; the cost weight adjustment factor is dynamically adjusted based on the supply chain volatility and is updated by the real-time raw material price variance calculation of the blockchain, for example, the value is 1.2; the cost nonlinearity coefficient is calibrated through marginal effect experiments in the high-cost range to strengthen the penalty for excess costs, for example, the value is 1.5.

[0119] S5. Solve the multi-objective optimization function through the quantum annealing algorithm, match the local raw material inventory data stored in the blockchain, and generate an optimization strategy.

[0120] Furthermore, the multi-objective optimization function is encoded into an optimization matrix adapted to the quantum annealer, and the nutritional matching, growth deviation suppression, and cost minimization are comprehensively weighted through dynamic weight parameters to generate a weighted matrix.

[0121] Specifically, the nutritional matching, growth deviation suppression and cost minimization in the multi-objective optimization function are respectively encoded into a quadratic unconstrained binary optimization matrix adapted to the quantum annealing machine. The difference between the vitamin compensation dose and the required dose in the nutritional matching is converted into a linear coefficient of a binary variable; the weighted sum of the absolute value of the vitamin time series distribution covariance and the absolute value of the zinc ion covariance in the growth deviation suppression is converted into a quadratic coupling term; the cost minimization is based on the mapping of the raw material inventory data stored on the blockchain and the compensation form cost into a quadratic term; the safety stock and the nutritional safety upper limit in the constraints are respectively converted into penalty terms to generate a QUBO optimization matrix.

[0122] Based on the local raw material inventory data stored on the blockchain, the inventory scarcity of each raw material is calculated and the annealing rate parameters of the quantum annealer are dynamically adjusted;

[0123] It should be noted that local raw material inventory data stored on the blockchain refers to real-time inventory information of local raw materials required for livestock feed production, recorded through blockchain technology. This includes immutable distributed ledger data such as raw material type, quantity, production batch, and storage location. For example, soybean meal and fish meal inventory levels are stored in blocks as hash values. Each inventory change is verified through a consensus mechanism and a new block is generated to ensure data traceability and authenticity.

[0124] The quantum annealing machine's annealing rate parameter is a core parameter that controls the decay rate of quantum fluctuations during the execution of the quantum annealing algorithm. It is used to adjust the balance between the quantum tunneling effect and thermal energy perturbations during the annealing process, directly affecting the convergence speed of the optimized solution and the probability of global optimality. For example, when the annealing rate parameter is set to 20 milliseconds, the algorithm completes the annealing path from the initial high-temperature state to the ground state within 20 milliseconds. If the rate is too fast, it may lead to local minima, while too slow a rate will increase the computational time but increase the probability of global convergence to over 95%.

[0125] Based on the local raw material inventory data stored in the blockchain, the inventory scarcity of each raw material is calculated as follows:

[0126]

[0127] Where, It is represented by the inventory scarcity of raw material i, i is the raw material index, The real-time inventory of raw material i is represented by blockchain evidence. is the safety stock of raw material i, η is the arrival time sensitivity coefficient, It represents the expected arrival time of raw material i, and T2 represents the maximum allowable arrival delay time.

[0128] The optimization matrix is ​​mapped to the quantum bit topology of the quantum annealer, and the quantum bit state is decoded after the annealing operation to generate the optimization strategy.

[0129] S6. Execute the optimization strategy and detect the mixing uniformity. When the mixing uniformity index exceeds a preset mixing uniformity threshold, start the compensation device to adjust the optimization strategy.

[0130] Furthermore, the motor speed of the feed mixing device and the opening of the feed valve are controlled according to the optimization strategy to generate the initial mixed feed;

[0131] Specifically, according to the raw material ratio weights in the optimization strategy, such as soybean meal weight 0.6 and fish meal weight 0.4, and compensation form parameters, such as vitamin slow-release particles and zinc ion liquid, the raw material ratio weights are converted into motor speed and feed valve opening control instructions of the feed mixing device. For example, the soybean meal feed valve opening is set to 60% opening according to the weight 0.6, and the fish meal feed valve opening is set to 40% opening according to the weight 0.4; the motor speed is adjusted according to the mixing uniformity requirements of the compensation form, for example, the slow-release particles require a speed of 1200rpm and the liquid compensation requires a speed of 800rpm, and the weighted average is taken; the control instruction drives the feed mixing device to operate, and the feed valve is fed synchronously according to the set opening to generate an initial mixed feed. For example, the output feed contains 60% soybean meal, 40% fish meal, 0.5mg / kg vitamin content, and 2mg / kg zinc ion content.

[0132] The near-infrared spectrometer is used to scan multiple sampling points of the initial mixed feed in real time to obtain the spectral distribution data of each nutrient;

[0133] Specifically, five sampling points are set above the initial mixed feed conveyor belt using a near-infrared spectrometer, and a fiber optic probe is installed at each sampling point to collect reflectance spectrum data; for example, sampling point 1 is at the starting end of the conveyor belt and sampling point 5 is at the end, and the scanning range of each probe covers a feed area with a diameter of 10 cm; the spectral data are preprocessed by first-order derivative and standard normal variable transformation, for example, the absorbance derivative peak at a wavelength of 1300 nm corresponds to the characteristic peak of vitamins, and the absorbance at a wavelength of 2100 nm corresponds to the characteristic peak of zinc ions; the absorbance mean and variance at the characteristic wavelength are extracted, for example, the vitamin absorbance mean at sampling point 1 is 0.65 and the variance is 0.02, and the zinc ion absorbance mean is 0.43 and the variance is 0.03, and a spectral distribution data set containing five sampling points is generated, and the wavelength, absorbance mean, variance and timestamp are recorded for each data point.

[0134] Calculating a mixing uniformity index based on the spectral distribution data, and triggering a compensation device to add nutrients to the mixed feed when the mixing uniformity index exceeds a preset mixing uniformity threshold;

[0135] Specifically, based on the mean absorbance values ​​of vitamins and zinc ions at each sampling point in the spectral distribution data, the coefficient of variation (the percentage ratio of the standard deviation to the mean) for each nutrient is calculated. The two coefficients of variation are weighted and summed using a preset weight factor to generate a mixing uniformity index. When the mixing uniformity index exceeds a preset mixing uniformity threshold, the compensation amount of vitamins or zinc ions is calculated based on the direction and magnitude of the deviation between the mean absorbance and the target value. The servo motor-driven compensation device is triggered to add the corresponding nutrient powder to the mixed feed at a set flow rate. Injection continues until the compensated mixing uniformity index drops below the preset mixing uniformity threshold.

[0136] It should be noted that the coefficient of variation is calculated based on the mean and standard deviation of the absorbance of vitamins and zinc ions in the spectral distribution data: for vitamins, the mean absorbance of five sampling points (for example, 0.65, 0.63, 0.71, 0.58, 0.67) is collected to calculate the mean (0.648) and the standard deviation (0.049), and the coefficient of variation is the ratio of the standard deviation to the mean; for zinc ions, the mean absorbance (for example, 0.43, 0.45, 0.39, 0.42, 0.47) is calculated to calculate the mean (0.432) and the standard deviation (0.032), and the coefficient of variation is the ratio of the standard deviation to the mean; the two coefficients of variation are weighted and summed according to the weight coefficient, for example, a vitamin weight of 0.6 and a zinc ion weight of 0.4, to generate a mixing uniformity index.

[0137] According to the nutrient content of the compensated mixed feed, the raw material ratio weight in the optimization strategy is dynamically adjusted to generate an updated optimization strategy;

[0138] Specifically, the compensation dosage parameters in the compensation demand vector are updated according to the nutrient content of the compensated mixed feed; the updated compensation demand vector is input into the multi-objective optimization function, the weight coefficients of vitamins and zinc ions in the nutritional matching degree are recalculated, and the raw material ratio weights in cost minimization are adjusted in combination with the dynamic changes of local raw material inventory data stored in the blockchain; the updated multi-objective optimization function is re-solved through the quantum annealing machine to generate an optimization strategy that adapts to the current nutrient content and inventory constraints.

[0139] The updated optimization strategy is re-input into the feed mixing device until the mixing uniformity index is lower than the preset mixing uniformity threshold.

[0140] Specifically, the soybean meal feed valve opening, fish meal feed valve opening and compensation dosage parameters in the updated optimization strategy are converted into motor speed control instructions and feed valve opening control signals of the feed mixing device, driving the feed mixing device to execute the mixing process under the new parameters; the five sampling points of the mixed feed are scanned in real time by a near-infrared spectrometer to calculate the updated mixing uniformity index. If it still exceeds the preset mixing uniformity threshold, the compensation amount is recalculated according to the absorbance mean deviation, and the servo motor-driven compensation device is triggered to inject additional vitamin powder at a set flow rate; the mixing, spectral detection and compensation processes are repeated until the mixing uniformity index drops below the preset mixing uniformity threshold, generating a mixed feed that meets the uniformity requirements.

[0141] It should be noted that the preset mixing uniformity threshold is based on regulatory requirements for feed ingredient uniformity and the stability requirements of specific nutritional additives. It is usually quantified using the coefficient of variation or absorbance standard deviation of near-infrared spectroscopy. For example, the value for common livestock and poultry feed is [5,10].

[0142] This embodiment also provides a computer device, which is suitable for the livestock feed ratio method based on livestock growth, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the livestock feed ratio method based on livestock growth proposed in the above embodiment.

[0143] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0144] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the livestock feed ratio method based on livestock growth as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0145] In summary, the present invention significantly improves the precision control capability of livestock farming through: the synergistic mechanism of dynamic nutritional demand identification and multi-constraint optimization. The adaptive triggering strategy based on growth deviation can accurately activate the detection process at the early stage of metabolic imbalance, and realize the targeted analysis of vitamin and trace element compensation needs; the quantum optimization engine integrating blockchain inventory data breaks through the real-time generation of global optimal formulas under complex constraints, greatly improving the efficiency of raw material utilization; the closed-loop compensation control mechanism ensures the uniformity of feed nutrient distribution and biological effectiveness by dynamically adjusting the ratio weights, and guarantees the dynamic adaptation of individual growth and resource supply. It realizes the full-link collaborative optimization from physiological monitoring to formula execution, providing efficient and reliable technical support for precision farming.

[0146] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A livestock feed ratio method based on livestock growth, characterized by: include, Collect body temperature and movement trajectory data of individual poultry and livestock, as well as environmental temperature and humidity data, monitor amino acid concentrations in the digestive tract, pre-process the collected data, and generate standardized biological data sets; The standardized biological data set is input into the pre-trained Gompertz prediction model to calculate the predicted value of the individual growth curve. When the deviation rate exceeds the preset deviation threshold, the metabolomics detection instruction is triggered; According to the metabolomics detection instructions, the mass spectrometer is started to perform non-targeted detection on feces and blood samples to generate metabolic fingerprints; Analyze the compensation amount of vitamins and trace elements based on metabolic fingerprints, and construct a multi-objective optimization function based on individual growth curves; The multi-objective optimization function is solved by the quantum annealing algorithm, matching the local raw material inventory data stored in the blockchain to generate an optimization strategy. The optimization strategy is executed and the mixing uniformity is detected. When the mixing uniformity index exceeds a preset mixing uniformity threshold, the compensation device is started to adjust the optimization strategy.

2. The livestock feed ratio method based on livestock growth according to claim 1, characterized in that: The specific steps of generating a standardized biological data set are as follows: Collect body temperature and movement trajectory data of individual poultry and livestock, and calculate activity intensity index; Collect ambient temperature and humidity data and use Grashof number to dynamically calibrate temperature drift error; By monitoring the lysine concentration in the digestive tract, the wavelength offset of the Bragg grating is used to generate real-time amino acid concentration data; The dynamic time warping algorithm is used to align the time series of body temperature, activity intensity index, amino acid concentration, and ambient temperature and humidity data to generate a millisecond-level synchronized time series dataset. The time series dataset is noise filtered, outliers are removed through Mahalanobis distance detection, and a standardized biological dataset is generated.

3. The livestock feed ratio method based on livestock growth according to claim 2, characterized in that: The specific steps of triggering the metabolomics detection instruction are as follows: The standardized biological data set was normalized and input into the pre-trained Gompertz prediction model to calculate the individual growth curve prediction value; Based on the predicted value of the individual growth curve, the real-time monitoring value of the poultry and livestock weight and movement trajectory data is obtained synchronously, and the deviation rate between the predicted value of the individual growth curve and the real-time monitoring value is output; When the deviation rate exceeds a preset deviation threshold, a detection instruction of the priority parameters is sent to the metabolomics detection.

4. The livestock feed ratio method based on livestock growth according to claim 3, characterized in that: The priority parameter refers to the sample detection priority level dynamically assigned according to the real-time monitored amino acid concentration data in the metabolomics detection instruction, which determines the detection order and resource allocation weight of different poultry and livestock individuals or samples.

5. The livestock feed ratio method based on livestock growth according to claim 4, characterized in that: The specific steps of generating the metabolic fingerprint are as follows: Based on the priority parameters in the metabolomics detection instructions, the mass spectrometer's mass-to-charge ratio scanning range and resolution parameters are dynamically configured to generate a mass spectrometer control file suitable for the sample type; According to the mass spectrometer control document, solid-phase microextraction was used to enrich short-chain fatty acids in fecal samples, and confined electrodialysis was used to enrich amino acids in blood samples to obtain metabolite-enriched samples; The metabolite-enriched sample was introduced into the mass spectrometer, and ionization and fragmentation were performed by dividing the variable window in the data-independent acquisition mode to collect the raw mass spectrometry data; A dynamic window peak alignment algorithm was performed on the raw mass spectrometry data, and the metabolite feature vectors were calculated in combination with the KEGG metabolic pathway topology weights to generate a metabolic fingerprint.

6. The livestock feed ratio method based on livestock growth according to claim 5, characterized in that: The specific steps of constructing the multi-objective optimization function are as follows: Based on the metabolic fingerprint, a dynamic metabolic flux probability distribution model is constructed to output the flux evolution parameters of vitamins and trace elements; Using the flux evolution parameters, the time convolution integral of the individual growth curve is calculated to generate the compensation requirement vector of vitamins and trace elements; A multi-objective optimization function is constructed based on the compensation demand vector.

7. The livestock feed ratio formulating method based on livestock growth according to claim 6, characterized in that: The specific steps of generating the optimization strategy are as follows: The multi-objective optimization function is encoded into an optimization matrix adapted to the quantum annealer. The nutrient matching, growth deviation suppression, and cost minimization are comprehensively weighted through dynamic weight parameters to generate a weighted matrix. Based on the local raw material inventory data stored on the blockchain, the inventory scarcity of each raw material is calculated and the annealing rate parameters of the quantum annealer are dynamically adjusted; The optimization matrix is ​​mapped to the quantum bit topology of the quantum annealer, and the quantum bit state is decoded after the annealing operation to generate the optimization strategy.

8. The livestock feed ratio method based on livestock growth according to claim 7, characterized in that: The specific steps of starting the compensation device to adjust the optimization strategy are as follows: Control the motor speed and feed valve opening of the feed mixing device according to the optimization strategy to generate the initial mixed feed; The near-infrared spectrometer is used to scan multiple sampling points of the initial mixed feed in real time to obtain the spectral distribution data of each nutrient; Calculating a mixing uniformity index based on the spectral distribution data, and triggering a compensation device to add nutrients to the mixed feed when the mixing uniformity index exceeds a preset mixing uniformity threshold; According to the nutrient content of the compensated mixed feed, the raw material ratio weight in the optimization strategy is dynamically adjusted to generate an updated optimization strategy; The updated optimization strategy is re-input into the feed mixing device until the mixing uniformity index is lower than the preset mixing uniformity threshold.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the livestock feed ratio method based on livestock growth according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the livestock feed ratio method based on livestock growth according to any one of claims 1 to 8 are implemented.

Images