Intelligent production comprehensive management method and system for realizing timed feeding

Abstract

The application relates to the technical field of intelligent production management, and discloses an intelligent production comprehensive management method and system for realizing timed feeding, which comprises the following steps: acquiring biological feature data of a feeding object, modeling to generate feeding demand model data, and confirming a timed and quantitative or dynamically adjusted feeding mode; based on different modes, an initial plan is set and optimized to generate time period adjustment data and adaptive range data; dynamic feeding parameter adjustment data is generated through prediction and analysis to construct a control strategy; feedback data is collected to evaluate efficiency and optimize the strategy. The system comprises a feeding mode identification module, a timed feeding plan module, a dynamic feeding adjustment module and a strategy optimization module. Through multi-dimensional data integration and dynamic adjustment, the application realizes precise feeding management, improves feeding efficiency and feed utilization rate, is suitable for the field of intelligent production management, and solves the problems of insufficient flexibility and precision of a traditional feeding mode.

Description

Technical Field

[0001] This invention relates to the field of intelligent production management technology, specifically to an intelligent production comprehensive management method and system for realizing timed feeding. Background Technology

[0002] In traditional feeding production management, feeding plans often rely on manual experience, making it difficult to accurately match the actual needs of the animals being fed, resulting in low feeding efficiency and serious feed waste. With the large-scale development of the breeding and animal husbandry industries, higher demands are being placed on the intelligent and precise management of feeding.

[0003] Traditional timed feeding methods have significant limitations. Firstly, fixed feeding times and amounts cannot adapt to the varying nutritional needs of animals at different growth stages, in different health states, and under changing environmental conditions. For example, young animals grow rapidly and have higher nutritional requirements, but traditional timed and quantitative methods may not be able to adjust feeding amounts in a timely manner, affecting their growth and development. When animals are sick, their feeding behavior and nutritional needs change, and traditional methods struggle to respond in real time. Secondly, manual monitoring and adjustment of feeding plans consume significant manpower and resources and are prone to errors due to human factors, failing to achieve dynamic and precise management.

[0004] Existing dynamic feeding adjustment programs also have shortcomings. While some systems attempt to adjust based on the biological characteristics of the animals being fed, data collection is not comprehensive enough, simply considering basic information such as weight and age, failing to analyze key factors such as health indicators and historical feeding behavior data in depth. This results in insufficient accuracy and effectiveness of dynamic adjustments. Furthermore, the impact of environmental factors on the feeding behavior of animals is often overlooked. For example, changes in environmental parameters such as temperature, humidity, and light can significantly affect an animal's appetite and feeding rate, but existing systems lack in-depth analysis of the correlation between environmental parameters and feeding behavior, making it impossible to adjust feeding strategies promptly according to environmental changes.

[0005] In the process of formulating and optimizing feeding plans, traditional methods lack scientific modeling and data analysis tools. Feeding demand modeling is not precise enough to accurately reflect the actual needs of the feeding subjects; time slot priority allocation and conflict detection rely on manual judgment, lacking systematicity and standardization, which can easily lead to unreasonable feeding time slot arrangements and affect feeding effectiveness. At the same time, the feeding efficiency evaluation system is imperfect, unable to provide timely feedback on problems during feeding execution, making it difficult to achieve adaptive optimization of feeding strategies.

[0006] With the development of information technology, intelligent and automated technologies are increasingly widely used in production management. However, in the field of feeding management, how to integrate and analyze multi-dimensional information such as biometric data, environmental parameters, and feeding behavior to construct a precise feeding demand model, achieve an organic combination of timed and quantitative feeding modes and dynamic adjustment modes, and improve feeding efficiency through real-time feedback and adaptive optimization remain pressing technical challenges. This invention aims to solve the above-mentioned problems in traditional feeding management by introducing advanced data processing, modeling, and optimization technologies, thereby improving the level of intelligence and overall management efficiency of feeding production. Summary of the Invention

[0007] The purpose of this invention is to provide an intelligent production management method and system for realizing timed feeding, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an intelligent production integrated management method for timed feeding, the method comprising:

[0009] Q1: Obtain the biometric data of the feeding object; model the feeding demand based on the biometric data of the feeding object to generate feeding demand model data; confirm the feeding mode of the feeding equipment based on the feeding demand model data to obtain feeding mode data, which includes timed and quantitative mode and dynamic adjustment mode.

[0010] Q2: Based on the timed and quantity-based mode, set the initial feeding plan for the feeding demand model data to generate the timed and quantity-based initial plan data; divide the feeding time period priority into the timed and quantity-based initial plan data to generate the timed and quantity-based time period adjustment data; based on the dynamic adjustment mode, set the dynamic feeding plan for the feeding demand model data to generate the dynamic adjustment initial plan data; analyze the feeding parameter adaptation range of the dynamic adjustment initial plan data to generate the dynamic adjustment adaptation range data.

[0011] Q3: Predict feeding behavior patterns from dynamic adjustment adaptation range data to generate dynamic feeding behavior prediction data; adjust dynamic feeding parameters from the initial dynamic adjustment plan data based on the dynamic feeding behavior prediction data to generate dynamic feeding parameter adjustment data; construct feeding control strategies from the feeding pattern data using dynamic feeding parameter adjustment data and time-based quantitative adjustment data to generate feeding pattern control strategies.

[0012] Q4: Collect feeding execution feedback data for the feeding mode control strategy to obtain a feeding execution feedback dataset; evaluate the feeding efficiency of the feeding execution feedback dataset to generate feeding efficiency evaluation data; and adaptively optimize the feeding mode control strategy using the feeding efficiency evaluation data to generate a dynamic feeding optimization strategy.

[0013] Preferably, Q1 includes the following steps:

[0014] Q11: Obtain biometric data of the animals being fed;

[0015] Q12: Perform data preprocessing on the biological characteristic data of the feeding objects to generate standard biological characteristic data. The data preprocessing includes data normalization, outlier removal, feature dimension alignment, and data noise reduction.

[0016] Q13: Perform biometric identification on standard biometric data to obtain growth stage data of the feeding object; based on the growth stage data, perform feeding demand modeling to generate feeding demand model data;

[0017] Q14: Confirm the feeding mode of the feeding equipment based on the feeding demand model data, and generate feeding mode data, which includes timed and quantitative mode and dynamic adjustment mode.

[0018] Preferably, Q14 includes the following steps:

[0019] Q141: Analyze the health indicators of the feeding subjects using the feeding demand model data to generate health status data of the feeding subjects; analyze the feeding behavior using the feeding demand model data to generate historical feeding behavior data.

[0020] Q142: Calculate the feeding pattern matching degree based on health status data and historical feeding behavior data, and generate feeding pattern matching degree data;

[0021] Q143: Compare the feeding pattern matching data with the preset standard pattern matching threshold. When the matching data is greater than the standard threshold, confirm the dynamic adjustment mode.

[0022] Q144: When the matching degree data is less than or equal to the standard threshold, confirm the timed and quantitative mode; integrate the dynamic adjustment mode and the timed and quantitative mode into the feeding mode data.

[0023] Preferably, Q2 includes the following steps:

[0024] Q21: Divide the feeding demand model data into initial time periods based on the timed and quantitative mode to generate initial timed and quantitative plan data; synchronize the feeding demand model data with time period priority based on the initial plan data to generate time period priority data.

[0025] Q22: Use time period priority data to detect time period conflicts in the feeding demand model data and generate time period conflict detection data; use the time period conflict detection data to adjust the time period priority of the initial time and quantity plan data and generate time and quantity time period adjustment data.

[0026] Q23: Based on the dynamic adjustment mode, dynamically allocate nutrients to the feeding demand model data to generate dynamic adjustment initial plan data; simulate feeding based on the dynamic adjustment initial plan data to generate dynamic feeding simulation data.

[0027] Q24: Perform nutritional requirement fluctuation analysis on dynamic feeding simulation data to generate nutritional requirement fluctuation data; based on the nutritional requirement fluctuation data, perform feeding parameter adaptation range analysis on dynamic feeding simulation data to generate dynamic adjustment adaptation range data.

[0028] Preferably, Q22 includes the following steps:

[0029] Q221: Use time period priority data to dynamically filter feeding time periods from feeding demand model data to obtain high-priority time period data;

[0030] Q222: Perform time period overlap analysis on high-priority time period data and feeding demand model data to generate time period overlap detection data;

[0031] Q223: When the overlap detection data of the time period exceeds the preset threshold, the time period offset adjustment is performed on the initial time-quantitative plan data to generate the first time period adjustment data;

[0032] Q224: When the overlap detection data of the time period does not exceed the threshold, the time period priority data is optimized for time period density to generate the second time period adjustment data;

[0033] Q225: Integrate the adjustment data for the first time period and the adjustment data for the second time period into timed and quantitative time period adjustment data.

[0034] Preferably, Q23 includes the following steps:

[0035] Q231: Perform environmental parameter correlation analysis on feeding demand model data based on dynamic adjustment mode to generate environmental correlation parameter data;

[0036] Q232: Predict feeding behavior from environmental parameter data to generate dynamic feeding prediction data; optimize nutrient allocation from feeding demand model data based on dynamic feeding prediction data to generate dynamic nutrient allocation data.

[0037] Q233: Simulate and set feeding parameters for feeding equipment using dynamic nutrient distribution data to generate dynamic adjustment initial plan data;

[0038] Q234: Verify the stability of feeding parameters based on dynamically adjusted initial plan data, and generate dynamic feeding simulation data.

[0039] Preferably, Q3 includes the following steps:

[0040] Q31: Predict feeding rate trends from dynamically adjusted adaptation range data to generate dynamic feeding rate prediction data;

[0041] Q32: Analyze the correlation impact of environmental parameters based on dynamic foraging rate prediction data, and generate environmental correlation impact data;

[0042] Q33: Based on environmental impact data, dynamically adjust the feeding time period of the initial plan data to generate dynamic feeding parameter adjustment data;

[0043] Q34: Construct a feeding pattern control strategy by adjusting data through dynamic feeding parameters and timed and quantitative time periods.

[0044] Preferably, Q31 includes the following steps:

[0045] Q311: Perform time-series clustering on dynamically adjusted adaptation range data to generate foraging behavior time-period clustering data;

[0046] Q312: Perform feeding rate gradient analysis based on time series clustering results to generate feeding rate gradient change data;

[0047] Q313: By using the data on the gradient change of feeding rate to perform trend fitting on the data of dynamic adjustment adaptation range, dynamic feeding rate prediction data is generated.

[0048] Preferably, Q33 includes the following steps:

[0049] Q331: Based on environmental impact data, the feeding time priority of the dynamically adjusted initial plan data is rearranged to generate time priority rearranged data;

[0050] Q332: Perform nutrient allocation balance calculations on time-priority rearranged data to generate dynamic nutrient balance data;

[0051] Q333: Optimize feeding time parameters using dynamic nutritional balance data to generate dynamic feeding parameter adjustment data.

[0052] Preferably, the present invention further includes an intelligent production integrated management system for realizing timed feeding, used to execute the intelligent production integrated management method as described above, the system comprising:

[0053] The feeding pattern recognition module is used to acquire the biological characteristic data of the feeding object; to model the feeding requirements of the biological characteristic data and generate feeding requirement model data; and to confirm the feeding pattern based on the model data, generating a timed and quantitative pattern or a dynamic adjustment pattern.

[0054] The timed feeding plan module is used to set the initial feeding plan based on the timed and quantitative mode, generate time period priority data, and generate timed and quantitative time period adjustment data through time period conflict detection and adjustment.

[0055] The dynamic feeding adjustment module is used to set dynamic feeding plans based on dynamic adjustment modes and generate dynamic adjustment adaptation range data; it generates dynamic feeding parameter adjustment data through feeding behavior prediction and environmental correlation analysis.

[0056] The strategy optimization module is used to collect feeding execution feedback data, evaluate feeding efficiency and optimize control strategies, and generate dynamic feeding optimization strategies to execute feeding operations.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] In terms of data processing and modeling, the accuracy and usability of the data are ensured by acquiring the biological characteristic data of the feeding subjects and performing data preprocessing (including data normalization, outlier removal, feature dimension alignment, and data noise reduction). Biometric identification is performed on the standard biological characteristic data to obtain the growth stage data of the feeding subjects. Based on this, feeding requirements modeling is performed to generate accurate feeding requirements model data. This modeling method based on actual biological characteristics can more accurately reflect the nutritional needs of the feeding subjects at different growth stages, providing a reliable basis for the confirmation of subsequent feeding patterns and the formulation of feeding plans. It changes the traditional extensive modeling method that relies on manual experience and improves the scientificity and applicability of the model.

[0059] In the feeding pattern confirmation stage, health indicator analysis and feeding behavior analysis are performed on the feeding demand model data to generate health status data and historical feeding behavior data. This data is then used to calculate the feeding pattern matching degree and compare it with preset standard thresholds to scientifically confirm whether to adopt a timed and quantitative feeding mode or a dynamic adjustment mode. This differentiated pattern confirmation mechanism allows for flexible selection of the feeding mode based on the actual condition of the feeding subject. When the matching degree is high, a dynamic adjustment mode is used to promptly respond to the specific needs of the feeding subject; when the matching degree is low, a timed and quantitative feeding mode is used to ensure feeding stability, avoiding the limitations of traditional single modes and improving the flexibility and targeting of feeding management.

[0060] During the setup of a timed feeding plan, initial time slots are divided and time slot priorities are synchronized based on a timed and quantitative feeding model. Through time slot conflict detection and adjustment, reasonable timed and quantitative time slot adjustment data is generated. Specifically, feeding time slots are dynamically filtered using time slot priority data, time slot overlap is analyzed, and time slot offset adjustments or time slot density optimizations are made according to different situations to ensure reasonable feeding time slot arrangements and avoid conflicts and unreasonable overlaps. This systematic time slot management approach can optimize feeding time arrangements, improve the rationality and execution efficiency of feeding plans, and reduce feeding chaos and inefficiency caused by improper time slot arrangements.

[0061] For dynamic feeding adjustments, environmental parameter correlation analysis and feeding behavior prediction are performed based on the dynamic adjustment model to generate dynamic feeding prediction data and optimize nutrient allocation. Through simulated feeding and nutrient demand fluctuation analysis, the adaptability range of feeding parameters is determined. Simultaneously, feeding rate trend prediction is performed on the dynamic adjustment adaptability range data, and combined with environmental correlation impact analysis, feeding time and parameters are dynamically adjusted. This dynamic adjustment mechanism, which combines environmental factors with feeding behavior, can respond in real time to environmental changes and the feeding trends of the fed organisms, adjusting feeding parameters and time in a timely manner to ensure that nutrient supply and demand are synchronized, improving feed utilization, reducing waste, and better meeting the nutritional needs of the fed organisms in different environments, thus promoting their healthy growth.

[0062] In terms of strategy optimization, feeding efficiency is evaluated by collecting feeding execution feedback data, and the feeding mode control strategy is adaptively optimized based on the evaluation results to generate a dynamic feeding optimization strategy. This closed-loop feedback mechanism can promptly identify problems in the feeding process, adjust the strategy according to the actual results, and achieve continuous improvement and optimization of feeding management. This enables the system to continuously adapt to changing feeding needs and environmental conditions, maintaining a highly efficient and stable operating state.

[0063] Overall, this invention achieves full automation and intelligence across the entire process from data collection, modeling, pattern confirmation, plan setting to strategy optimization through intelligent processing and refined management in multiple stages. This not only reduces labor costs and improves management efficiency, but also enhances feeding quality and reduces feed waste through precise feeding control and dynamic adjustment. It provides strong technical support for the large-scale and intelligent development of the breeding and animal husbandry industries, and has significant economic and social benefits. Attached Figure Description

[0064] Figure 1 This is a schematic diagram illustrating the working principle of the intelligent production management method for timed feeding as described in this invention.

[0065] Figure 2 A flowchart for processing biometric data of feeding subjects and modeling feeding requirements;

[0066] Figure 3 A flowchart for prioritizing time-based and quantity-based time periods;

[0067] Figure 4 This is a flowchart illustrating the correlation between environmental parameters and simulated feeding under dynamic adjustment mode. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] Please see Figures 1-4 The present invention relates to an intelligent production management method for realizing timed feeding, the specific implementation steps of which are as follows:

[0070] Q1: Obtain biometric data of the feeding subjects and generate feeding pattern data.

[0071] Acquire the biological characteristics data of the feeding subjects, such as weight, growth stage, health indicators, and feeding behavior data. Model feeding requirements using this data to generate a feeding requirement model that reflects the nutritional needs and feeding patterns of the feeding subjects. Based on this model data, determine the feeding modes of the feeding equipment, including a timed and quantitative mode and a dynamic adjustment mode. The timed and quantitative mode is suitable for scenarios with stable demand, while the dynamic adjustment mode is used for scenarios requiring flexible adjustments based on real-time data.

[0072] Q2: Set up the initial feeding plan and generate adjustment data

[0073] In the timed and quantity-based mode, an initial feeding plan is set based on the feeding demand model data, generating initial timed and quantity-based plan data. The feeding periods in the plan are then prioritized, generating timed and quantity-based period adjustment data to ensure feeding priority during critical periods. In the dynamic adjustment mode, a dynamic feeding plan is set, generating initial dynamic adjustment plan data. The adaptation range of the feeding parameters in the plan is analyzed, generating dynamic adjustment adaptation range data to clarify the reasonable adjustable range of the parameters.

[0074] Q3: Construct a feeding control strategy

[0075] Based on the dynamic adjustment adaptation range data, feeding behavior patterns are predicted, generating dynamic feeding behavior prediction data. Based on this prediction data, feeding parameters in the initial dynamic adjustment plan data are controlled and adjusted, generating dynamic feeding parameter adjustment data. Combining the dynamic feeding parameter adjustment data and the timed and quantitative period adjustment data, a feeding pattern control strategy is constructed, forming a specific feeding execution plan.

[0076] Q4: Optimize feeding control strategies

[0077] Feedback data is collected during the execution of the feeding pattern control strategy to form a feeding execution feedback dataset. This dataset is analyzed to evaluate feeding efficiency and generate feeding efficiency evaluation data. Based on the evaluation results, the feeding pattern control strategy is adaptively optimized to generate a dynamic feeding optimization strategy, enabling continuous improvement of the feeding program.

[0078] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.

[0079] Example 1:

[0080] This embodiment primarily illustrates the process of acquiring biometric data of the feeding subject and generating feeding pattern data. The acquisition of biometric data can be achieved through various sensor devices and data acquisition terminals. For example, a weight sensor can be used to periodically collect the subject's weight data; a camera combined with image recognition technology can be used to acquire physical characteristic data; wearable physiological monitoring devices (such as smart collars) can collect real-time health indicator data such as heart rate and body temperature; and sensors built into the feeding device can record feeding behavior data such as feeding time, feed amount, and feeding frequency. This data can come from different acquisition terminals and is stored in the system database in structured or unstructured form, forming the original set of biometric data of the feeding subject.

[0081] Data preprocessing is performed on the biometric data of the animals being fed to generate standardized biometric data. The data preprocessing stage includes several key operations: First, data normalization. Since different types of biometric data have different dimensions and value ranges, normalization algorithms (such as min-max standardization, Z-score standardization, etc.) are used to convert the data into a uniform scale range. For example, weight data (unit: kg) and feeding time data (unit: min) are both converted to values ​​within the [0,1] interval to facilitate subsequent model processing. Second, outlier removal. Statistical methods (such as IQR method, Z-score method) or machine learning algorithms (such as isolated forest, local outlier factor algorithm) are used to identify outliers in the data. For example, data showing a sudden increase in feed intake or a sudden decrease in body temperature during a certain period may be caused by equipment malfunctions. Due to factors such as malfunctions, human error, or sudden illness in the animal being fed, data must be removed from the dataset to avoid interfering with subsequent modeling. Thirdly, feature dimension alignment is crucial. When data from different acquisition terminals has missing or inconsistent dimensions, interpolation methods (such as linear or polynomial interpolation) are used to supplement the missing dimensions, or redundant dimensions are filtered and merged to ensure all sample data have the same feature dimensions. For example, the acquisition time intervals and feature index types for each batch of data should be standardized. Fourthly, data denoising is necessary. Filtering algorithms (such as Gaussian filtering or median filtering) or signal processing techniques are used to filter out noise data generated by environmental interference during the acquisition process. For example, fluctuations in heart rate monitoring data caused by electromagnetic interference can be removed, improving the signal-to-noise ratio. Through these data preprocessing steps, the raw biometric data is converted into standardized biometric data with a uniform format and reliable quality.

[0082] After data preprocessing, biometric identification needs to be performed on the standard biometric data to obtain the growth stage data of the fed animals. The biometric identification process can be implemented using machine learning or deep learning models. For example, an image recognition model based on a convolutional neural network (CNN) can be built to analyze the physical characteristics of the fed animals and determine their growth stage (e.g., juvenile, growth, maturity, senescence); or a recurrent neural network (RNN) can be used to model time-series data such as weight and feeding behavior, identifying the growth stage by analyzing the trends in the data. Taking the distinction between juvenile and growth stages as an example, juvenile individuals typically have a faster rate of weight gain, a higher feeding frequency but smaller single-feeding amounts, while growth-stage individuals have a gradually stabilizing rate of weight gain, a lower feeding frequency but an increased single-feeding amount. The model can accurately classify the growth stage by learning these differences in characteristics. Based on the identified growth stage data, combined with professional knowledge in animal nutrition and historical feeding data, a feeding requirement model is constructed. The model needs to comprehensively consider factors such as energy requirements, nutrient ratios, and feeding patterns at different growth stages. For example, young animals need higher protein and energy intake to support growth and development. The model can generate corresponding feed type suggestions (such as high-protein young animal feed), daily feeding frequency (such as 4-6 times / day), and feeding amount per feeding (such as 5%-8% of body weight), forming a structured feeding demand model.

[0083] In the step of confirming the feeding pattern based on the feeding demand model data, the first step is to analyze the health indicators and feeding behavior of the feeding subjects. The health indicator analysis mainly includes the assessment of data such as body temperature, heart rate, and body condition scores (e.g., body condition score). For example, by calculating the mean and standard deviation of body temperature, it can be determined whether the feeding subjects are within the normal body temperature range. If the body temperature fluctuates significantly and exceeds the normal threshold, it may indicate a health problem. The feeding behavior analysis focuses on mining historical feeding behavior data, including feeding duration, fluctuation range of feed intake, and regularity of feeding intervals. For example, if the analysis finds that a certain group's feed intake is significantly lower at a certain time (e.g., at night) than at other times, it may indicate that the feeding arrangement during that time period is unreasonable.

[0084] Based on health status data and historical feeding behavior data, a pre-set algorithm calculates feeding pattern matching data. This algorithm comprehensively considers multiple dimensions of indicators; for example, health status is weighted at 0.6, and feeding behavior at 0.4. Each dimension includes several sub-indicators (such as body temperature, heart rate, and body condition score in the health status, and fluctuations in feed intake and regularity of feeding intervals in the feeding behavior). A weighted summation method is used to calculate the overall matching score. Specifically, for health status data, a higher score is obtained if all indicators are within the normal range; if abnormal indicators exist, the score is correspondingly lower. For feeding behavior data, a higher score is obtained if the feeding pattern is stable and the feed intake fluctuation is small; if there are obvious feeding abnormalities (such as overeating or prolonged refusal to eat), the score is lower.

[0085] The calculated feeding pattern matching data is compared with a preset standard pattern matching threshold. The standard pattern matching threshold can be preset based on factors such as the type and growth stage of the feeding object. For example, for young feeding objects in good health with regular feeding behavior, the threshold can be set to 0.7. That is, when the matching score is greater than 0.7, their feeding needs and health status are considered to have a high probability of dynamic changes, making a dynamic adjustment mode suitable. When the matching score is less than or equal to 0.7, it indicates that their needs are relatively stable, and a timed and quantitative mode can be used. If the dynamic adjustment mode is confirmed, corresponding parameter adjustment rules and data monitoring frequencies need to be configured for this mode. If the timed and quantitative mode is used, a fixed feeding plan needs to be generated based on the model data. Finally, the dynamic adjustment mode and the timed and quantitative mode are integrated into feeding pattern data. This data contains information such as the triggering conditions and parameter setting rules for both modes, providing a basis for pattern selection in subsequent initial feeding plan settings.

[0086] Throughout the implementation process, the timeliness and accuracy of data collection are crucial. For example, for dairy cows in raw milk production, it is necessary to collect data such as their weight, body temperature, and milk yield in real time to dynamically adjust feeding plans. Simultaneously, model construction and algorithm selection need to be optimized for specific feeding scenarios. For instance, for large-scale pig farms, distributed computing frameworks can be used to process massive amounts of biometric data, improving data processing efficiency and model training speed. Furthermore, the system must have data security protection mechanisms to ensure the security of biometric data of the fed animals during collection, storage, and transmission, preventing data leakage or malicious tampering. Through the coordinated operation of these various stages, intelligent processing of the entire process from biometric data collection to feeding pattern confirmation is achieved, laying the foundation for subsequent feeding plan setting and strategy optimization.

[0087] Example 2:

[0088] This embodiment further refines the relevant processes for setting the initial feeding plan and generating adjustment data in the overall implementation scheme, covering the plan setting and adjustment process under both timed and quantitative modes and dynamic adjustment modes. First, when setting the initial feeding plan in the timed and quantitative mode, it is necessary to divide a day or a feeding cycle into initial time periods based on the generated feeding demand model data and factors such as the growth stage and physiological state of the feeding target. For example, for poultry in the growth stage, according to their daily feeding patterns, 24 hours can be divided into three main feeding periods: morning, noon, and evening, with an 8-hour interval between each period. A supplementary feeding period at night is also set to meet their high energy needs, generating initial timed and quantitative plan data that includes information such as time period division, expected feed amount, and feed type. This plan data must initially meet the basic nutritional needs of the feeding target and consider the operating capacity of the feeding equipment (such as feeding speed and maximum feed amount) to avoid situations where the equipment load is too high or too low.

[0089] Based on the initial plan data, the feeding demand model data is synchronized with time-segment priority to generate time-segment priority data. Determining time-segment priorities requires comprehensive consideration of multiple factors. For example, young animals have weaker digestive abilities and require smaller, more frequent meals; their nighttime supplementary feeding time can have a higher priority than regular feeding times. Dairy cows in their peak milk production period have a more urgent nutritional need during breakfast, and their priority should be set to the highest. Priority allocation can be achieved by establishing an Analytic Hierarchy Process (AHP). First, influencing factors (such as the urgency of nutritional needs, the physiological state of the animal being fed, and equipment operating costs) are identified. Then, a judgment matrix is ​​constructed to calculate the weight of each factor. Finally, priority levels (such as Level 1, Level 2, and Level 3) are assigned to each feeding time segment based on the weights. For example, if the analysis shows that the weight of urgency of nutritional needs is 0.5, the weight of physiological state is 0.3, and the weight of equipment operating costs is 0.2, then if a certain time segment has high urgency of nutritional needs (score 4 points), high physiological needs (score 3 points), and low equipment operating costs (score 5 points), its comprehensive priority score can be calculated as 0.5×4 + 0.3×3 + 0.2×5 = 4.1 points, corresponding to Level 1 priority.

[0090] After prioritizing feeding time slots, the time slot priority data is used to perform time slot conflict detection on the feeding demand model data, generating time slot conflict detection data. Time slot conflicts can manifest in various forms, such as high-priority and low-priority time slots being too close in feeding time, causing the feed recipients to be unable to fully digest the feed from the previous feeding; or multiple high-priority time slots simultaneously requiring the use of the same feeding equipment, leading to competition for equipment resources. The conflict detection process can be implemented through the following steps: First, arrange the feeding time slots in the initial time-quantity plan data in chronological order to form a time slot sequence; then, calculate the time interval between adjacent time slots. If the interval is less than the preset minimum digestion time (e.g., 2 hours for poultry), it is determined to be a time conflict; at the same time, count the number of feeding equipment that needs to be activated within the same time slot. If it exceeds the total number of equipment, it is determined to be a resource conflict. For example, in an initial plan, two high-priority time slots are scheduled at 7:00 AM and 8:00 AM, with an interval of only 1 hour, which is less than the minimum digestion time of 2 hours required by poultry. In this case, time conflict detection data is generated; or, for example, 3 feeding equipment needs to be activated within the same time slot, but only 2 equipment are actually available, resource conflict detection data is generated.

[0091] Based on time-slot conflict detection data, the initial time-slot and quantity-based planning data is adjusted in terms of time slot priority, generating adjusted time-slot and quantity-based planning data. The adjustment strategy must be differentiated according to the conflict type and priority level:

[0092] When a time conflict is detected, and the conflict involves a high-priority time period (such as insufficient interval between nighttime supplementary feeding and morning feeding), time period shift adjustments are made to the low-priority time period first. For example, the low-priority midday feeding period is postponed from 12:00 to 14:00 to increase the interval between the morning and midday feeding periods to 3 hours to meet digestion needs, generating the first time period adjustment data.

[0093] When the conflict does not involve high-priority time periods or the time interval does not exceed the threshold, the time period priority data is optimized for time period density. For example, if the feeding amount is unevenly distributed within a certain time period (e.g., 80% of the feed is fed in the first 10 minutes and no feeding is done in the last 50 minutes), the feeding speed can be adjusted to distribute the feed evenly throughout the time period, generating second time period adjustment data.

[0094] If resource conflicts exist, the priority of conflicting time periods needs to be reassessed. Priority should be given to allocating equipment resources during high-priority time periods, and feeding tasks during low-priority time periods should be reassigned to other available equipment or time periods. For example, the low-priority midday time period can be allocated to standby feeding equipment or postponed to equipment idle time periods, generating corresponding adjustment data.

[0095] Finally, all kinds of adjustment data are integrated into timed and quantitative time period adjustment data. This data includes optimized time period division, priority order, equipment allocation scheme and other information to ensure the rationality and feasibility of the timed and quantitative feeding plan.

[0096] In dynamic adjustment mode, dynamic nutrient allocation is first set based on feeding demand model data to generate initial dynamic adjustment plan data. Dynamic nutrient allocation needs to consider real-time environmental parameters (such as temperature, humidity, and light duration) and changes in the physiological state of the feeding subjects (such as pregnancy or disease recovery). For example, when the ambient temperature drops below 10°C, mammals' energy demand increases to maintain body temperature. In this case, the fat content in the feed needs to be increased (e.g., from 5% to 8%), and the total daily feed intake needs to be increased (e.g., by 10%-15%) in the initial dynamic adjustment plan. Nutrient allocation settings can be achieved by establishing a dynamic nutrient model. This model takes environmental parameters and physiological state data as input and outputs the ratio of each nutrient component and feeding amount recommendations. For example, a regression model can be used to fit the relationship between temperature and energy demand. The formula can be expressed as: Energy demand (kJ / day) = α × temperature (°C) + β, where α and β are model parameters determined through training with historical data.

[0097] Based on the dynamic adjustment initial plan data, feeding simulations are performed on the feeding demand model data to generate dynamic feeding simulation data. The simulated feeding process is based on historical feeding data and the operating parameters of the feeding equipment. A simulation model is built using discrete event simulation technology (such as Arena and AnyLogic software) to simulate the feeding process under different nutrient distribution schemes. For example, if a dynamic adjustment initial plan is set as follows: feeding high-protein feed (protein content 20%) during breakfast, the feeding amount is 6% of body weight, and the feeding rate is 50g / min; the simulation model can simulate the changes in the feeding rate, feed intake completion rate, and feeding time distribution of the feeding object under this plan, generating dynamic feeding simulation data containing time series.

[0098] Nutritional requirement fluctuation analysis is performed on dynamic feeding simulation data to generate nutritional requirement fluctuation data. The analysis process can employ time series analysis methods (such as moving average and exponential smoothing) to identify the fluctuation trends and periodic characteristics of nutritional requirements in different time periods. For example, analysis revealed that a certain group experiences a significant peak in energy demand, with energy requirements 20% higher than the average level between 3 PM and 5 PM due to increased activity; conversely, energy demand drops to 80% of the average level during nighttime sleep, forming a trough. Based on these fluctuation characteristics, combined with the health indicators and feeding behavior data of the feeding subjects, the adaptive range of feeding parameters is determined, generating dynamic adjustment range data. For instance, for the afternoon energy demand peak, the feed amount is set to be adjusted within 110%-130% of the average amount, and the feeding speed within 40-60 g / min, ensuring that nutritional supply is still met during demand fluctuations while avoiding overfeeding and feed waste.

[0099] Throughout the implementation process, it is crucial to ensure the initial plan matches the actual feeding scenario. For example, in livestock farms implementing group feeding, the feeding needs of different groups may vary significantly, requiring the generation of separate initial plans and adjustment data for each group. Simultaneously, time-segment priority division and conflict detection must be dynamically updated. When the physiological state or environmental conditions of the feeding subjects suddenly change (such as sudden illness or extreme weather), the system should be able to recalculate priorities and detect conflicts in real time, ensuring the flexibility of the feeding plan. Furthermore, the accuracy of the simulated feeding model is paramount. Continuous collection of actual feeding data is necessary to calibrate the simulation model, such as comparing the error between simulated and actual feed intake, and adjusting model parameters to improve prediction accuracy. Through these refined steps, the initial plan can be scientifically set and adaptively adjusted under both timed and dynamic adjustment modes, providing a reliable basis for subsequent feeding behavior prediction and control strategy construction.

[0100] Example 3:

[0101] This embodiment describes the process of setting a dynamic feeding plan based on feeding demand model data under dynamic adjustment mode, specifically involving environmental parameter correlation analysis, feeding behavior prediction, nutrient allocation optimization, and simulation verification. First, in dynamic adjustment mode, environmental parameter correlation analysis needs to be performed on the feeding demand model data to generate environmental correlation parameter data. Environmental parameters cover multiple dimensions, such as temperature and humidity (e.g., indoor temperature 22±2℃, relative humidity 55%-70%), light duration (e.g., laying hens require 16 hours of light during the egg-laying period), air quality indicators (e.g., ammonia concentration ≤15ppm), and seasonal changes (e.g., high temperatures in summer, low temperatures in winter). These parameters can be collected in real time through a sensor network deployed in the breeding environment, such as temperature and humidity sensors, light intensity sensors, and gas detectors, with data transmitted to the system database at a frequency of minutes or hours. The system uses correlation analysis algorithms (such as Pearson correlation coefficient and Spearman rank correlation) to calculate the degree of correlation between each environmental parameter and the feeding behavior and nutritional requirements of the feeding objects. For example, it found that when the temperature in the barn exceeds 28°C, the feed intake of pigs decreases by an average of 15%-20%, thus determining that temperature is a highly correlated environmental parameter and generating environmental correlation parameter data that includes parameter name, correlation coefficient, and direction of influence (positive correlation / negative correlation).

[0102] Feeding behavior prediction is generated based on environmental parameter data, producing dynamic feeding prediction data. The prediction process can employ time-series prediction models in machine learning, such as Long Short-Term Memory (LSTM) networks and Gated Recurrent Units (GRUs). The model inputs historical environmental parameter data (such as temperature and humidity sequences from the past 7 days) and corresponding feeding behavior data (such as feed intake and feeding duration), and the output is the predicted feeding behavior value for future periods (such as the next 24 hours). Taking dairy cows as an example, by learning historical data from high-temperature periods in summer, the model can predict that when the maximum temperature reaches 30℃ within the next 24 hours, the feeding duration from 20:00 to 22:00 at night will increase by 30 minutes compared to the average, while the feeding duration from 10:00 to 12:00 during the day will decrease by 20 minutes, with feed intake increasing by 10% and decreasing by 15% respectively, forming dynamic feeding prediction data in time-series form. Furthermore, the prediction model can be corrected by incorporating weather warning data (such as high-temperature warnings and typhoon warnings) to improve the accuracy of predictions under extreme weather conditions.

[0103] Nutrient allocation is optimized based on dynamic feed intake prediction data and feeding demand model data to generate dynamic nutrient allocation data. Nutrient allocation optimization must adhere to the principle of "supply on demand," adjusting the proportions and amounts of nutrients in the feed according to predicted changes in feed intake behavior. For example, when a decrease in feed intake is predicted during periods of high temperature, to ensure sufficient energy and protein intake, the nutrient concentration of the feed needs to be increased. This could involve increasing the energy level of lactating sow feed from 3.2 MJ / kg to 3.5 MJ / kg and the protein content from 14% to 16%, while simultaneously increasing the addition of vitamin C and electrolytes to alleviate heat stress. Specific optimization strategies can be implemented using a linear programming model, with nutritional requirement objectives (such as maintaining energy requirements and growth and development requirements) as constraints and minimizing feed costs as the objective function, to solve for the optimal nutrient ratio. For example, for fattening pigs, if a 10% reduction in feed intake is predicted, the model calculates that the proportion of soybean meal in the feed needs to be increased from 18% to 22%, the proportion of corn needs to be reduced from 65% to 60%, and 0.3% of oil needs to be added to ensure that the daily weight gain target remains unchanged. This generates dynamic nutrient allocation data that includes the proportion of each nutrient, the type and amount of additives.

[0104] The system simulates and sets feeding parameters for the feeding equipment using dynamic nutrient distribution data, generating initial dynamic adjustment plan data. Feeding parameters include feeding amount, feeding speed, feeding time, and feed type switching, which must precisely match the dynamic nutrient distribution data. For example, based on the requirement in the dynamic nutrient distribution data that "high-protein feed (20% protein content) should be fed during breakfast, with a feeding amount of 6% of body weight," the system simulates and sets the parameters for the feeding equipment: setting the breakfast period to 07:00-07:30, the feeding speed to 40g / min, and the total feeding amount to 60kg (calculated based on the total weight of the herd), and starting the feed delivery system 10 minutes in advance to ensure that the equipment completes the feed type switching before feeding begins. During the simulation, the feasibility of the equipment parameters needs to be verified, such as checking whether the feeding speed is within the equipment's rated range (e.g., 20-50g / min) and whether the feeding amount exceeds the maximum capacity of the feed trough, generating initial dynamic adjustment plan data that includes parameter setting details and equipment operating status estimates.

[0105] The stability of feeding parameters is verified by dynamically adjusting the initial plan data, generating dynamic feeding simulation data. Stability verification is achieved through simulation testing, simulating equipment operation under different parameter settings based on the physical model of the feeding equipment and historical operating data. For example, when the feeding speed is set to 40g / min, the simulation calculates parameters such as the flow velocity of feed particles in the conveying pipeline and the load current of the equipment motor to determine whether there is a risk of pipeline blockage or motor overload. If the simulation results show that the peak current is within 80% of the motor's rated current and the feed flow rate is uniform, the parameter settings are considered stable; if the pressure in the pipeline exceeds the safety threshold, it indicates that the feeding speed needs to be adjusted (e.g., reduced to 35g / min) and re-verified. The dynamic feeding simulation data generated during the verification process includes equipment operating parameter curves (such as current and pressure change curves over time) and feed delivery error ranges (e.g., ±2%), used to evaluate the feasibility and reliability of the initial plan.

[0106] During implementation, the real-time nature and accuracy of environmental parameters are crucial. For example, in aquaculture scenarios within greenhouses, distributed sensors are needed to monitor parameters such as water temperature and dissolved oxygen in real time. An emergency response mechanism should be triggered when parameters become abnormal (e.g., dissolved oxygen below 3 mg / L) to adjust the feeding plan and reduce feed waste and water pollution. Simultaneously, the feeding behavior prediction model needs to be updated with training data regularly to adapt to changes in feeding patterns caused by variations in the growth stages of the feed. For instance, the feeding patterns of broilers differ significantly from the brooding to the growing stage, requiring retraining of the model every two weeks. Furthermore, nutrient allocation optimization must consider the supply of feed ingredients and cost fluctuations. For example, when soybean meal prices rise sharply, the system can automatically switch to an optimized scheme using rapeseed meal and cottonseed meal as alternative protein sources, reducing feeding costs while ensuring nutritional needs are met.

[0107] Through a closed-loop operation involving environmental parameter correlation analysis, feeding behavior prediction, nutrient allocation optimization, parameter simulation settings, and stability verification, the dynamically adjusted feeding plan can respond in real time to environmental changes and fluctuations in the feeding population's needs, achieving a shift from "fixed-pattern feeding" to "dynamic precision feeding." This process not only improves feed utilization and reduces resource waste but also provides scientific and flexible decision support for intelligent production management, ensuring the effectiveness and reliability of feeding programs in different scenarios.

[0108] Example 4:

[0109] This embodiment describes the process of constructing a feeding control strategy, specifically involving feeding rate trend prediction, environmental impact analysis, dynamic adjustment of feeding time periods, and parameter optimization. First, feeding rate trend prediction is performed on the dynamic adjustment adaptation range data to generate dynamic feeding rate prediction data. The dynamic adjustment adaptation range data includes multi-dimensional information such as historical feeding rates, feeding time periods, and environmental parameters. It needs to be subjected to time series clustering to identify the periodic patterns of feeding behavior. Time series clustering can use DBSCAN (Density Peak Clustering) or K-means++ algorithms to divide feeding rate data within similar time periods into the same cluster. For example, dividing the feeding data of a certain livestock population by hour, clustering reveals that the feeding rates are higher during three time periods: 8:00-10:00, 14:00-16:00, and 20:00-22:00, forming three significant clusters, generating feeding behavior time period clustering data.

[0110] Based on time-series clustering results, feeding rate gradient analysis is performed to generate feeding rate gradient change data. Gradient analysis requires calculating the slope of feeding rate change within each clustering period and the rate difference between adjacent periods. For example, during the 8:00-10:00 clustering period, the feeding rate gradually increases from an initial 30 g / min to 50 g / min, with a slope of 10 g / (min·h); while during the 10:00-14:00 non-clustering period, the rate remains below 10 g / min, forming a clear boundary between high and low rate intervals. By analyzing the gradient differences at different growth stages or physiological states (e.g., the feeding rate gradient is gentler in pregnant animals and steeper in growing animals), feeding behavior characteristics can be further refined, generating gradient change data that includes information such as time period identifiers, initial rate values, slope of change, and time period duration.

[0111] By fitting the trend of dynamic adjustment range data to the gradient change data of feeding rate, dynamic feeding rate prediction data can be generated. Trend fitting can be performed using multinomial regression, exponential smoothing, or regression models in machine learning (such as random forest regression and support vector machine regression). Taking exponential smoothing as an example, a smoothing coefficient α=0.3 is set, and a weighted average is calculated on historical feeding rate data to predict the rate value for future periods. For example, based on the feeding rate sequence [45g / min, 48g / min, 50g / min] from 8:00 to 10:00 for the previous 3 days, the predicted rate for the next day in that period is calculated to be 45×0.3+48×0.3×(1-0.3)+50×(1-0.3)²≈48.2g / min. The predicted data should include a timestamp and the corresponding predicted rate value to provide a time-dimensional reference for subsequent adjustments to feeding parameters.

[0112] Based on dynamic feeding rate prediction data, an impact analysis of environmental parameters is conducted to generate environmental impact data. The correlation between environmental parameters and feeding rate can be determined through Granger causality tests or regression analysis to identify the impact path. For example, the analysis revealed that for every 1°C increase in indoor temperature, the feeding rate decreases by an average of 2 g / min, and the effect of temperature change on feeding rate has a lag effect (lag time approximately 1-2 hours). Simultaneously, sudden changes in light intensity (such as the instant lights are turned off or on) may cause brief fluctuations in feeding rate (fluctuation range ±10 g / min). These correlations are quantified into impact coefficients and time lags, generating environmental impact data that includes parameter name, direction of influence, intensity of influence, and lag time. For example, the correlation between temperature and feeding rate can be expressed as: feeding rate change = -2 g / (min・°C) × temperature change, lag time 1.5 hours.

[0113] The feeding schedule is dynamically adjusted based on environmental impact data to generate dynamic feeding parameter adjustment data. The adjustment process consists of the following steps: First, the feeding schedule priority is rearranged based on the environmental impact data. For example, if the indoor temperature is predicted to rise to 30℃ (exceeding the upper limit of comfort temperature 28℃) between 2:00 PM and 4:00 PM, the feeding rate may decrease during this period due to the negative correlation between temperature and feed intake rate. Therefore, high-priority feeding tasks originally planned for this period (such as feeding high-protein feed) need to be adjusted to the cooler early morning period (such as 6:00 AM to 8:00 AM). This generates time slot priority rearrangement data, which includes information such as the original time slot, the adjusted time slot, and the change in priority level.

[0114] Nutrient allocation balance calculations are performed on the time-priority rearrangement data to generate dynamic nutrient balance data. The balance calculation must ensure that the total nutrient supply for each time period after adjustment is consistent with the daily nutrient requirement target, avoiding nutrient imbalance caused by time-period adjustments. For example, when adjusting the amount of high-protein feed given in the afternoon to the morning, the feed ratio for each time period needs to be recalculated to ensure that the daily protein intake remains unchanged. This can be achieved by establishing a nutrient balance equation, such as: Σ(feeding amount per time period × protein content) = daily protein requirement target. Assuming that the original plan was to feed 5 kg of feed with 20% protein content in the afternoon, and after adjusting to the morning, if the original plan was to feed 10 kg of feed with 16% protein content in the morning, then the protein content of the morning feed needs to be adjusted to (10 × 16% + 5 × 20%) / 15 ≈ 17.33% to maintain the total daily protein intake (10 × 16% + 5 × 20% = 2.6 kg), generating dynamic nutrient balance data that includes adjustments to the nutrient composition of the feed for each time period.

[0115] Feeding parameters are optimized using dynamic nutritional balance data to generate dynamic feeding parameter adjustment data. Parameter optimization needs to combine predicted feed intake rates with equipment operating parameters. For example, in the adjusted morning period, if the predicted feed intake rate is 50 g / min and the equipment's maximum feeding speed is 60 g / min, the feeding speed can be set to 55 g / min, meeting feed intake needs while allowing for a safety margin. Simultaneously, the feeding time and duration need to be adjusted to ensure that the feeding task is completed within the predicted high feed intake rate period. For example, the original planned 30-minute feeding period from 2:00 PM to 2:30 PM can be adjusted to 6:30 AM to 7:00 AM, maintaining the same 30-minute duration, generating dynamic feeding parameter adjustment data including parameters such as feeding time, speed, feed type, and feed amount.

[0116] During implementation, the accuracy of feed intake rate trend prediction depends on the completeness of historical data and the adaptability of the model. For example, for newly introduced feed species, at least two weeks of historical feed intake data need to be collected to train a customized prediction model. Environmental impact analysis requires continuous updating of the correlations between parameters. For instance, when the ventilation system of the aquaculture environment is modified, the correlation strength between temperature and feed intake rate may change, requiring recalculation of the impact coefficients. Furthermore, dynamic adjustments to feeding times must consider the physical limitations of the equipment, such as the maximum conveying distance of feed delivery pipes and the capacity limitations of feed silos, to avoid equipment malfunctions due to adjustments.

[0117] Through the orderly execution of steps such as feed intake rate trend prediction, environmental impact analysis, time period priority rearrangement, nutritional balance calculation, and parameter optimization, the system can dynamically adjust the feeding schedule and parameter settings based on real-time predicted feeding behavior and environmental changes, achieving a precise match between feeding control strategies and actual feeding needs. This process not only improves feeding efficiency but also reduces feed waste and insufficient nutrient supply caused by fixed plans, providing dynamic optimization technical support for intelligent production management.

[0118] Example 5:

[0119] This embodiment details the specific structure and functions of each module of an intelligent production integrated management system for timed feeding. This system executes the aforementioned methods, achieving fully automated management of the entire process from biometric data collection to feeding strategy optimization through the collaborative operation of hardware devices and software algorithms. The overall system architecture includes a feeding pattern recognition module, a timed feeding planning module, a dynamic feeding adjustment module, and a strategy optimization module. These modules interact through data interfaces and store and manage data based on a unified database.

[0120] ① Feeding pattern recognition module

[0121] This module is responsible for acquiring biometric data of the fed subjects and generating feeding pattern data. On the hardware side, it collects data in real time, including weight, body temperature, physical characteristics, feeding time, and feed intake, through weight sensors, physiological indicator monitoring devices (such as heart rate sensors), image acquisition devices (such as cameras), and the feeding behavior recording device built into the feeding equipment. On the software side, the raw data is first preprocessed, including data normalization (e.g., converting weight data and feeding time data of different dimensions into values ​​in the [0,1] range), outlier removal (identifying and removing unreasonable data points using statistical methods), feature dimension alignment (supplementing missing data or merging redundant dimensions), and data noise reduction (using filtering algorithms to remove environmental interference), generating standard biometric data.

[0122] Biometric identification algorithms (such as deep learning-based image classification models) are used to analyze standard data to identify the growth stages of the feeding objects (such as juvenile stage and growth stage), and feed demand model data is constructed by combining animal nutrition knowledge. This model includes nutritional requirement parameters (such as protein and energy intake) and feeding patterns (such as the number of feedings per day) for different growth stages.

[0123] Finally, the feeding pattern is determined using a feeding pattern matching algorithm: health indicators (such as body temperature and body condition score) and feeding behavior data (such as feed intake fluctuations) in the feeding demand model data are analyzed to calculate the pattern matching score. The calculation formula is:

[0124]

[0125] in, The health indicators are scored (range 0-100) to reflect the stability of the health status of the fed animals; Feeding behavior is scored (range 0-100) to reflect the fluctuation of feeding patterns; and The weighting coefficients for health indicators and feeding behavior are respectively ( Preset according to the type of animal being fed, such as young animals. Adult livestock ).when Greater than the preset threshold If the dynamic adjustment mode is not used, the timed and quantitative mode is used, and the two modes are integrated into feeding mode data.

[0126] ② Timed feeding plan module

[0127] This module generates and adjusts feeding plans based on a timed and quantitative feeding model. First, based on feeding demand model data and feeding pattern data, the feeding cycle is divided into multiple initial time periods (such as three main periods: morning, noon, and evening). The feeding amount, feed type, and priority for each time period are set, generating initial timed and quantitative feeding plan data. Prioritization is based on a hierarchical analysis model, comprehensively considering factors such as the urgency of nutritional needs and the physiological state of the animals being fed. For example, the nighttime supplementary feeding period for young animals has a higher priority than regular feeding periods.

[0128] The rationality of the initial plan is analyzed using a time-slot conflict detection algorithm. The detection criteria include time slot intervals (e.g., the interval between adjacent time slots must be greater than the minimum digestion time). ) and device resource usage (such as the number of devices called in the same period not exceeding the total number of available devices). If a conflict is detected (e.g., the time interval is less than...), or the number of devices exceeds If the initial plan is adjusted according to priority, high-priority time periods remain unchanged, while low-priority time periods are shifted to conflict-free times (generating the first time period adjustment data), or the feeding distribution within the time period is optimized (generating the second time period adjustment data), and finally integrated into time-based and quantity-based time period adjustment data.

[0129] ③ Dynamic feeding adjustment module

[0130] This module is responsible for the planning settings and parameter optimization in dynamic adjustment mode. First, it acquires real-time environmental data through environmental parameter acquisition devices (such as temperature and humidity sensors, light intensity sensors), combines it with feeding demand model data for correlation analysis, identifies key environmental parameters affecting feeding behavior (such as the negative correlation between temperature and feed intake), and generates environmental correlation parameter data.

[0131] Using time-series prediction models (such as LSTM neural networks) to train on environmental parameters and foraging behavior data, we can predict foraging behavior in future periods (such as the predicted amount of foraging in a certain period). The system generates dynamic feed intake prediction data. Based on the prediction results, the proportion of nutrients in the feed is adjusted through a nutrient allocation optimization algorithm (such as increasing vitamin C content during high-temperature periods) to generate dynamic nutrient allocation data. The feeding equipment parameters (such as feeding speed and feed type) are simulated and set to generate dynamic adjustment initial plan data.

[0132] The stability of parameters is verified through simulation models, such as simulating the flow rate of feed in a pipeline. With the load current of the equipment motor Relationship, ensure Within the rated flow rate range of the equipment Inner and Not exceeding the rated current of the motor This generates dynamic feeding simulation data that includes the range of parameter fluctuations.

[0133] ④ Strategy Optimization Module

[0134] This module optimizes control strategies by collecting feeding execution feedback data. At the hardware level, it collects execution data such as actual feeding amount, feed completion rate, and feed waste in real time through sensors (such as weight sensors and flow meters) and environmental monitoring equipment installed on the feeding equipment, forming a feeding execution feedback dataset.

[0135] On the software side, the feedback data is first statistically analyzed to calculate feeding efficiency indicators (such as feed intake). The feeding efficiency assessment data is generated. Then, based on the assessment results, the parameters of the feeding mode control strategy (such as the parameter adaptation range of the dynamic adjustment mode and the time period priority weight of the timed and quantitative mode) are adjusted through adaptive optimization algorithms (such as genetic algorithms) to generate a dynamic feeding optimization strategy and realize the iterative improvement of the feeding program.

[0136] The system's modules collaborate seamlessly through standardized data interfaces. For example, the feeding pattern recognition module transmits feeding pattern data to the timed feeding plan module and the dynamic feeding adjustment module as the basis for plan settings. The strategy optimization module feeds back the optimized strategy to the first three modules, updating model parameters and planning logic. The entire system is scalable, adapting to different types of feeding objects (such as livestock, poultry, and aquatic animals) and farming scenarios (such as large-scale farms and family farms) by adding sensor types or upgrading algorithm models, achieving intelligent and precise timed feeding management.

[0137] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0138] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent integrated production management that enables timed feeding, characterized in that, Includes the following steps: Q1: Obtain the biometric data of the feeding object; model the feeding demand based on the biometric data of the feeding object to generate feeding demand model data; confirm the feeding mode of the feeding equipment based on the feeding demand model data to obtain feeding mode data, which includes timed and quantitative mode and dynamic adjustment mode. Q2: Based on the timed and quantitative feeding mode, set the initial feeding plan for the feeding demand model data and generate the timed and quantitative initial plan data; The initial time-based and quantity-based feeding plan data is prioritized for different feeding periods to generate time-based and quantity-based period adjustment data. Based on the dynamic adjustment mode, dynamic feeding plan settings are made for the feeding demand model data, and dynamic adjustment initial plan data is generated. Analyze the adaptation range of feeding parameters based on the dynamically adjusted initial plan data to generate dynamically adjusted adaptation range data, including the following steps: Q21: Divide the feeding demand model data into initial time periods based on the timed and quantitative mode to generate initial timed and quantitative plan data; synchronize the feeding demand model data with time period priority based on the initial plan data to generate time period priority data. Q22: Use time period priority data to detect time period conflicts in the feeding demand model data and generate time period conflict detection data; use the time period conflict detection data to adjust the time period priority of the initial time and quantity plan data and generate time and quantity time period adjustment data. Q23: Based on the dynamic adjustment mode, dynamically allocate nutrients to the feeding demand model data to generate dynamic adjustment initial plan data; simulate feeding based on the feeding demand model data according to the dynamic adjustment initial plan data to generate dynamic feeding simulation data. Q24: Perform nutritional requirement fluctuation analysis on dynamic feeding simulation data to generate nutritional requirement fluctuation data; based on the nutritional requirement fluctuation data, perform feeding parameter adaptation range analysis on dynamic feeding simulation data to generate dynamic adjustment adaptation range data. Q22 includes the following steps: Q221: Use time period priority data to dynamically filter feeding time periods from feeding demand model data to obtain high-priority time period data; Q222: Perform time period overlap analysis on high-priority time period data and feeding demand model data to generate time period overlap detection data; Q223: When the overlap detection data of the time period exceeds the preset threshold, the time period offset adjustment is performed on the initial time-quantitative plan data to generate the first time period adjustment data; Q224: When the overlap detection data of the time period does not exceed the threshold, the time period priority data is optimized for time period density to generate the second time period adjustment data; Q225: Integrate the adjustment data for the first time period and the adjustment data for the second time period into timed and quantitative time period adjustment data; Q3: Predict feeding behavior patterns from dynamic adjustment adaptation range data to generate dynamic feeding behavior prediction data; adjust dynamic feeding parameters from the initial dynamic adjustment plan data based on the dynamic feeding behavior prediction data to generate dynamic feeding parameter adjustment data; construct feeding control strategies from the feeding pattern data using dynamic feeding parameter adjustment data and time-based quantitative adjustment data to generate feeding pattern control strategies. Q4: Collect feeding execution feedback data for the feeding mode control strategy to obtain a feeding execution feedback dataset; evaluate the feeding efficiency of the feeding execution feedback dataset to generate feeding efficiency evaluation data; and adaptively optimize the feeding mode control strategy using the feeding efficiency evaluation data to generate a dynamic feeding optimization strategy.

2. The intelligent production management method for timed feeding according to claim 1, characterized in that, Q1 includes the following steps: Q11: Obtain biometric data of the animals being fed; Q12: Perform data preprocessing on the biological characteristic data of the feeding objects to generate standard biological characteristic data. The data preprocessing includes data normalization, outlier removal, feature dimension alignment, and data noise reduction. Q13: Perform biometric identification on standard biometric data to obtain growth stage data of the feeding object; based on the growth stage data, model feeding requirements to generate feeding requirement model data; Q14: Confirm the feeding mode of the feeding equipment based on the feeding demand model data, and generate feeding mode data, which includes timed and quantitative mode and dynamic adjustment mode.

3. The intelligent production management method for timed feeding according to claim 2, characterized in that, Q14 includes the following steps: Q141: Analyze the health indicators of the feeding subjects using the feeding demand model data to generate health status data of the feeding subjects; analyze the feeding behavior using the feeding demand model data to generate historical feeding behavior data. Q142: Calculate the feeding pattern matching degree based on health status data and historical feeding behavior data, and generate feeding pattern matching degree data; Q143: Compare the feeding pattern matching data with the preset standard pattern matching threshold. When the matching data is greater than the standard threshold, confirm the dynamic adjustment mode. Q144: When the matching degree data is less than or equal to the standard threshold, confirm the timed and quantitative mode; integrate the dynamic adjustment mode and the timed and quantitative mode into the feeding mode data.

4. The intelligent production management method for timed feeding according to claim 1, characterized in that, Q23 includes the following steps: Q231: Based on the dynamic adjustment mode, perform environmental parameter correlation analysis on the feeding demand model data to generate environmental correlation parameter data; Q232: Predict feeding behavior from environmental parameter data to generate dynamic feeding prediction data; optimize nutrient allocation from feeding demand model data based on dynamic feeding prediction data to generate dynamic nutrient allocation data. Q233: Simulate and set feeding parameters for feeding equipment using dynamic nutrient distribution data to generate dynamic adjustment of initial plan data; Q234: Verify the stability of feeding parameters based on dynamically adjusted initial plan data, and generate dynamic feeding simulation data.

5. The intelligent production management method for timed feeding according to claim 1, characterized in that, Q3 includes the following steps: Q31: Predict feeding rate trends from dynamically adjusted adaptation range data to generate dynamic feeding rate prediction data; Q32: Analyze the correlation impact of environmental parameters based on dynamic foraging rate prediction data, and generate environmental correlation impact data; Q33: Based on environmental impact data, dynamically adjust the feeding time period of the initial plan data to generate dynamic feeding parameter adjustment data; Q34: Construct a feeding pattern control strategy by adjusting data through dynamic feeding parameters and timed and quantitative time periods.

6. The intelligent production management method for timed feeding according to claim 5, characterized in that, Q31 includes the following steps: Q311: Perform time-series clustering on dynamically adjusted adaptation range data to generate foraging behavior time-period clustering data; Q312: Perform feeding rate gradient analysis based on time series clustering results to generate feeding rate gradient change data; Q313: By using the data on the gradient change of feeding rate to perform trend fitting on the data of dynamic adjustment adaptation range, dynamic feeding rate prediction data is generated.

7. The intelligent production management method for timed feeding according to claim 6, characterized in that, Q33 includes the following steps: Q331: Based on environmental impact data, the feeding time priority is rearranged from the dynamic adjustment initial plan data to generate time priority rearranged data; Q332: Perform nutritional allocation balance calculations on time-priority rearranged data to generate dynamic nutritional balance data; Q333: Optimize feeding time parameters using dynamic nutritional balance data to generate dynamic feeding parameter adjustment data.

8. An intelligent integrated production management system for realizing timed feeding, characterized in that, For executing the intelligent production integrated management method for timed feeding as described in claim 1, the system includes: The feeding pattern recognition module is used to acquire the biological characteristic data of the feeding object; to model the feeding requirements of the biological characteristic data and generate feeding requirement model data; and to confirm the feeding pattern based on the model data, generating a timed and quantitative pattern or a dynamic adjustment pattern. The timed feeding plan module is used to set the initial feeding plan based on the timed and quantitative mode, generate time period priority data, and generate timed and quantitative time period adjustment data through time period conflict detection and adjustment. The dynamic feeding adjustment module is used to set dynamic feeding plans based on dynamic adjustment modes and generate dynamic adjustment adaptation range data; it generates dynamic feeding parameter adjustment data through feeding behavior prediction and environmental correlation analysis. The strategy optimization module is used to collect feeding execution feedback data, evaluate feeding efficiency and optimize control strategies, and generate dynamic feeding optimization strategies to execute feeding operations.

Images