A method for hierarchical closed-loop calibration of pig herd diet conversion efficiency

Abstract

The application provides a pig group daily ration conversion efficiency hierarchical closed-loop calibration method, and belongs to the field of pig precise feeding and daily ration regulation. The technical scheme is as follows: obtaining pig group production data and growth stages, determining daily ration conversion efficiency and efficiency state; establishing direction-amplitude calibration unit according to historical daily ration calibration data, determining response state, available daily ration calibration action space and overall calibratability; determining control level and daily ration calibration permission according to efficiency state and overall calibratability, determining target calibration unit and executing calibration when the permission is allowed; updating the response state and use restriction of the corresponding calibration unit according to the actual execution result and actual efficiency response, reconstructing the available daily ration calibration action space of the next control evaluation period and updating the control state. The application has the beneficial effects that the local response, action space and hierarchical permission are closed-loop linked, and the pig group daily ration calibration pertinence, stability and safety are improved.

Description

Technical Field

[0001] This invention relates to the field of precision pig feeding and diet regulation, and in particular to a method for hierarchical closed-loop adjustment of pig herd diet conversion efficiency. Background Technology

[0002] With the development of large-scale, intensive pig farming, feeding management is gradually evolving from experience-based formulations and phased quantitative supply to precision nutrition and dynamic regulation. Current pig herd diet management typically determines diet formulations based on production indicators such as growth stage, weight range, feed intake, weight gain, and feed conversion ratio, and makes phased adjustments based on nutritional parameters such as net energy, protein, amino acids, and minerals. With the application of automatic weighing, precision feeding, and livestock data acquisition equipment, pig herd feed intake, weight changes, and diet execution records can be continuously acquired at regular intervals, providing a data foundation for evaluating diet suitability based on actual production responses and implementing dynamic nutritional regulation. Therefore, using feed conversion efficiency as an evaluation indicator and adjusting dietary nutritional parameters based on the pig herd's production status has become an important technological direction for improving feed utilization efficiency and the level of refined feeding.

[0003] However, existing diet control methods still largely rely on the current feed conversion ratio, weight gain level, or their deviation from preset target values ​​as adjustment criteria. They make relatively crude use of historical diet adjustment response information, often failing to distinguish the local efficiency responses corresponding to different nutrient adjustment directions and magnitudes. Therefore, when a small adjustment to the same nutrient parameter fails, it is easily interpreted as an overall ineffectiveness of that nutrient direction, failing to accurately reflect the remaining adjustable space in the pig herd. Furthermore, there may be discrepancies between the target adjustment amount of the diet formulation and the actual changes in nutrient parameters after formulation and supply. If planned adjustments are still attributed to subsequent efficiency responses, it can easily lead to mismatches in learning from historical responses. On the other hand, pig herd production indicators themselves have phased differences and natural fluctuations. Controlling solely based on current efficiency values ​​or single thresholds can easily lead to frequent switching of adjustment actions. When an adjustment results in a weak response, no significant response, or an unfavorable response, existing methods lack a closed-loop mechanism to establish local restrictions on specific adjustment directions and magnitudes, re-evaluate the remaining available adjustment space, and further update the overall adjustment capacity and subsequent adjustment authority. Therefore, existing technologies are insufficient to simultaneously achieve reliable learning of local responses, attribution of actual execution results, evaluation of remaining safe action space, and hierarchical linkage control of pig herd adjustment authority. They also fall short in ensuring the safety of diet adjustments while continuously utilizing adjustment actions that still have positive response potential. Summary of the Invention

[0004] The purpose of this invention is to provide a method for hierarchical closed-loop adjustment of pig herd diet conversion efficiency, which realizes closed-loop linkage of local response, action space and hierarchical permissions, and improves the pertinence, stability and safety of pig herd diet adjustment.

[0005] This invention is achieved through the following measures: A method for stratified closed-loop calibration of diet conversion efficiency in pig herds, characterized by the following steps: S1. Obtain pig herd production data and growth stage in each evaluation period, determine the diet conversion efficiency based on the pig herd production data, and determine the efficiency status by combining the reference diet conversion efficiency corresponding to the growth stage. S2. Obtain historical diet adjustment data of the pig herd, establish direction-amplitude adjustment units based on the direction and magnitude of nutritional adjustment, determine the response state based on the historical efficiency response corresponding to each direction-amplitude adjustment unit, determine the available diet adjustment action space based on the response state, and determine the overall adjustability based on the available diet adjustment action space. S3. Based on the efficiency status and overall adjustability, the pig herd is stratified to determine the control level and the corresponding daily feed adjustment authority. S4. When the ration adjustment permission allows adjustment, determine the target adjustment unit based on the response status of each direction-amplitude adjustment unit in the available ration adjustment action space, execute the ration adjustment corresponding to the target adjustment unit, and determine the actual direction-amplitude adjustment unit to be executed based on the actual execution result. S5. Obtain the actual efficiency response corresponding to the ration adjustment, update the corresponding response status according to the actual efficiency response and the direction-amplitude adjustment unit corresponding to the actual execution, and when the actual efficiency response does not meet the preset positive response conditions and can be reliably attributed, form the usage restriction of the direction-amplitude adjustment unit corresponding to the actual execution in the next control evaluation cycle after the response evaluation is completed. S6. Based on the updated response status and the usage restrictions, redetermine the available diet adjustment action space for the next control evaluation cycle, update the overall adjustability based on the redetermined available diet adjustment action space, and update the control level and diet adjustment permissions for the next control evaluation cycle in conjunction with the efficiency status determined when the response evaluation is completed.

[0006] The invention also has the following specific features: Step S1 includes: Obtain the feed intake, body weight, and growth stage of the pig herd during the evaluation period corresponding to the actual diet implementation period; when a major diet switch occurs and available body weight data is available at the switch time, the evaluation window is redefined based on the switch time. Pigs that belong to the pig herd and have effective weight records at the beginning and end of the evaluation period are identified as the effective pig set. The effective weight gain is determined based on the total effective weight at the beginning and end of the effective pig set. The effective feed intake is determined by deducting the amount of leftover feed, the amount of recycled feed, and the amount of abnormal loss from the actual feed supply. Data validity is judged based on the completeness of pig herd production data, the continuity of the effective pig set, the correspondence between pig herd production data and the actual ration execution period, and the effective feed intake threshold. The ration conversion efficiency is determined based on the effective weight gain and effective feed intake. The reference diet conversion efficiency is determined based on the historical stable evaluation window of the corresponding growth stage, and the efficiency status is determined based on the ratio of the diet conversion efficiency to the reference diet conversion efficiency.

[0007] Step S2: Obtain historical diet adjustment data for the pig herd; establish direction-amplitude adjustment units based on the direction and magnitude of nutrient adjustments; determine the response status based on the historical efficiency response corresponding to each direction-amplitude adjustment unit, including: Historical diet adjustment events are obtained, including the target diet, the actual diet before adjustment, the actual diet after adjustment, and the corresponding efficiency response evaluation results. The direction of nutrient adjustment is determined based on the actual amount and direction of change of each nutrient parameter. The intensity of the direction's own action is determined based on the actual change in the nutritional parameters corresponding to the nutritional adjustment direction and the maximum allowable adjustment range per single cycle of the corresponding growth stage. The intensity of the direction's own action is then divided into a preset continuous range and a direction-amplitude adjustment unit is formed. Eligibility for response learning is determined based on the validity of historical diet adjustment data, the identifiability of nutritional adjustment directions, whether the response to previous major diet adjustments has been completed, and the consistency between the target diet and the actual implemented diet. For historical ration adjustment events that meet the response learning qualifications, the net efficiency response is determined based on the efficiency change trend before adjustment, the unit amplitude response is determined based on the net efficiency response and the direction's own motion intensity, and the response state of the corresponding direction-amplitude adjustment unit is updated accordingly.

[0008] In step S2, the available diet adjustment action space is determined based on the response state, and the overall adjustability is determined based on the available diet adjustment action space, including: The response reliability is determined based on the number of effective historical samples and the consistency of the unit amplitude response of each direction-amplitude calibration unit. The response state is then fused with the historical prior response of the corresponding growth stage based on the response reliability to obtain the effective response state. Finally, the response evaluation value is determined based on the historical response distribution of the corresponding growth stage. The remaining range between the current actual diet and the nutritional security boundary, the maximum allowable adjustment range per cycle for each nutritional adjustment direction, and the overall dietary security requirements determine the safe diet adjustment action space, and the available diet adjustment action space is determined based on the response evaluation value and the usage restrictions of the direction-amplitude adjustment unit. The remaining available adjustment space coverage is determined based on the safe and available ranges of each nutrient adjustment direction. The representative response for each nutrient adjustment direction is determined, and the available action response level is determined based on the representative responses of different nutrient adjustment directions. The overall adjustability is determined in combination with the remaining available adjustment space coverage.

[0009] Step S3 includes: Set an increase threshold and a decrease threshold for the efficiency state and the overall adjustability respectively. Perform hysteresis judgment based on the current evaluation result and the discrete state of the previous evaluation period to determine the high and low states of the efficiency state and the overall adjustability. The following are defined as the first control level, the second control level, the third control level, and the fourth control level: the first control level is characterized by high efficiency and high overall adjustability, the second control level is characterized by high efficiency and low overall adjustability, the third control level is characterized by low efficiency and high overall adjustability, and the fourth control level is characterized by low efficiency and low overall adjustability. The corresponding daily ration adjustment authority is also determined for each of these levels. The basic upper limit of the diet adjustment authority is determined based on the control level, and the actual executable diet adjustment authority is determined within the basic upper limit based on the available diet adjustment action space, nutritional safety boundary and existing usage restrictions; the minimum retention period of the control level is set, wherein safety tightening is not limited by the minimum retention period.

[0010] Step S4: When the ration adjustment permission allows adjustment, determine the target adjustment unit based on the response status of each direction-amplitude adjustment unit in the available ration adjustment action space, including: Based on the aforementioned diet adjustment authority, the remaining range between the currently implemented diet and the nutritional security boundary, and the maximum permissible adjustment range per single cycle for each nutritional adjustment direction, the current maximum permissible action intensity for each nutritional adjustment direction is determined, and the direction-amplitude adjustment units currently allowed to be executed are selected from the available diet adjustment action space to form a candidate diet adjustment action set. The efficiency improvement requirement is determined based on the current efficiency status and the efficiency status rise threshold. The action intensity required to achieve the efficiency improvement requirement is determined based on the response status of each candidate direction-amplitude adjustment unit. The action intensity to be executed is determined by combining the corresponding preset continuous amplitude range and the current maximum allowable action intensity. The expected efficiency response is determined based on the effective response status and the intensity of the action to be performed for each candidate direction-amplitude calibration unit. For candidate direction-amplitude calibration units whose expected efficiency response meets the efficiency improvement requirements, the direction-amplitude calibration unit with the smaller intensity of the action to be performed is selected first, and the target calibration unit is determined by combining the response evaluation value and response reliability of each candidate direction-amplitude calibration unit. When there is no candidate direction-amplitude calibration unit whose expected efficiency response meets the efficiency improvement requirements, the target calibration unit is determined from the candidate direction-amplitude calibration units with positive expected efficiency responses.

[0011] In step S4, the ration adjustment corresponding to the target adjustment unit is performed, including: Based on the nutrient adjustment direction, the intensity of the action to be performed, and the maximum allowable adjustment range per cycle for the corresponding growth stage of the target adjustment unit, the adjustment amount of the target nutrient parameters is determined, and the target diet is formed based on the current actual diet. The target ration is formulated and supplied, the actual ration after adjustment is obtained, and the target adjustment unit, the target ration, the actual ration after adjustment and the execution time are recorded to form the ration adjustment record to be evaluated. Based on the actual changes in nutrient parameters between the actual ration before and after calibration, the actual nutrient adjustment direction and range are redefined, and the corresponding direction-range calibration unit is determined accordingly. When the direction-range calibration unit corresponding to the actual execution is inconsistent with the target calibration unit, the direction-range calibration unit corresponding to the actual execution is used as the basis for subsequent efficiency response attribution.

[0012] Step S5 includes: The efficiency status within the response evaluation period corresponding to the calibration record of the diet to be evaluated is obtained. The efficiency change after calibration is corrected according to the efficiency change trend before calibration to determine the actual net efficiency response. The natural fluctuation scale is determined according to the historical stability evaluation window of the corresponding growth stage, and the response classification result is determined according to the actual net efficiency response and the natural fluctuation scale. The response classification result includes effective positive response, weak positive response, no obvious response and adverse response. The degree of predicted mismatch is determined based on the predicted efficiency response and the actual net efficiency response, and the eligibility for response learning is determined based on the data validity of the diet adjustment records to be evaluated, the identifiability of the direction of nutritional adjustment, whether the previous major diet adjustment response has been evaluated, and the consistency between the target diet and the actual implemented diet. When the response learning qualification is met, the unit amplitude response is determined based on the directional action intensity and actual net efficiency response of the directional-amplitude calibration unit corresponding to the actual execution, and the response state and response reliability of the directional-amplitude calibration unit corresponding to the actual execution are updated accordingly; when the directional-amplitude calibration unit corresponding to the actual execution is inconsistent with the target calibration unit, the response state of the target calibration unit is not updated using the actual net efficiency response.

[0013] Step S5 also includes: When the actual efficiency response can be reliably attributed to the direction-amplitude adjustment unit corresponding to the actual execution and meets the response learning qualification, the use priority of the corresponding direction-amplitude adjustment unit is reduced for weak positive responses, the repeated execution of the corresponding direction-amplitude adjustment unit with no obvious response limitation is allowed, the use of the corresponding direction-amplitude adjustment unit is suspended for adverse responses, and the direction-amplitude adjustment unit that expands to a higher actual amplitude along the same nutrient adjustment direction is temporarily prohibited. When the response learning qualification is not met, the response status of the specific direction-amplitude adjustment unit is not updated or local usage restrictions are not formed based on the actual efficiency response that cannot be reliably attributed; when the actual efficiency response is an unfavorable response, the actual executable ration adjustment authority is tightened and the current ration adjustment is stopped from being expanded. When the actual efficiency response is a valid positive response, the corresponding actual implemented diet is determined as a candidate safe diet, and the safe diet is updated based on the subsequent stability evaluation results; when the actual efficiency response is an unfavorable response and diet reversal is required, the nutritional parameters involved in the actual diet adjustment and the nutritional parameters of the formula linkage changes are reversed based on the safe diet.

[0014] Step S6 includes: The safe diet adjustment action space is redefined based on the current actual diet, the nutritional safety boundary of the corresponding growth stage, and the maximum allowable adjustment range per cycle for each nutritional adjustment direction. The available diet adjustment action space is also redefined based on the updated response status and the usage restrictions. Specifically, for directional-amplitude adjustment units whose executable range is reduced due to the usage restrictions, only the executable range is retained, while directional-amplitude adjustment units that are prohibited from use due to the usage restrictions are excluded. Based on the redefined available dietary adjustment action space, the remaining available adjustment space coverage and representative responses of each nutrient adjustment direction are redefined. Based on the representative responses, the available action response level is determined, and the overall adjustability is updated based on the available action response level and the remaining available adjustment space coverage. Based on the efficiency status determined at the time of response evaluation and the updated overall adjustability, hysteresis judgment is performed according to the preset rise and fall thresholds of efficiency status and overall adjustability. The control level and basic permission limit of the next control evaluation cycle are updated, and the actual executable diet adjustment permissions are further determined based on the available diet adjustment action space, nutritional safety boundary and usage restrictions.

[0015] The beneficial effects of this invention are as follows: Achieving closed-loop linkage between local response, action space, and hierarchical permissions improves the targeting, stability, and safety of hog herd diet adjustment; This invention establishes a direction-amplitude adjustment unit based on the direction and magnitude of nutrient adjustment. It forms a local response state based on the actual executed diet and corresponding efficiency response, and determines the overall adjustability by combining the remaining available adjustment space. The control level and diet adjustment authority are jointly determined by the efficiency state and overall adjustability. After adjustment, the response is assigned based on the actual execution results. Local restrictions are imposed on weak positive, no obvious, and adverse responses, and the subsequent available diet adjustment action space is reconstructed accordingly. This forms a hierarchical closed-loop adjustment process driven by actual responses, improving the targeting, stability, and safety of diet adjustment. Attached Figure Description

[0016] Figure 1 This is an overall flowchart of the method for hierarchical closed-loop adjustment of pig diet conversion efficiency provided in Embodiment 1 of the present invention. Detailed Implementation

[0017] To clearly illustrate the technical features of this solution, the following detailed implementation method will be used to explain the solution.

[0018] Example 1 See Figure 1 This embodiment provides a method for stratified closed-loop calibration of pig herd diet conversion efficiency, applied to pig farming environments with group rearing, pig herd production data collection, and diet formulation and supply conditions. A group of pigs at the same or similar growth stage, using uniform diet execution conditions within the same evaluation period, and whose herd production data is collected according to uniform statistical standards, is designated as the basic calibration unit, denoted as [missing unit]. ,in, This indicates the basic calibration unit number. Within a single evaluation period, the pig herd composition and diet execution conditions remain relatively stable within the same basic calibration unit. Adjustments to dietary nutrient parameters are permitted between different evaluation periods based on the diet calibration results.

[0019] The continuous evaluation periods are denoted as follows: ,in, This indicates the evaluation cycle number. The evaluation cycle is determined based on the growth stage of the pig herd, the frequency of pig production data collection, and the feeding and management cycle, establishing a time correspondence between the pig production data collected within each evaluation cycle and the corresponding actual diet execution period. After a diet adjustment is completed, a response waiting period and a response evaluation period are set. The response waiting period and response evaluation period are allowed to cover one or more evaluation cycles to ensure a clear time correspondence between the diet adjustment event and the subsequent actual efficiency response. When the response evaluation spans multiple evaluation cycles, the initiation cycle of the diet adjustment event, the completion cycle of the response evaluation, and the subsequent control evaluation cycle are used to characterize the time position of action execution, response formation, and the next round of control state, respectively.

[0020] Furthermore, the pig herd production data is acquired through feed metering equipment, pig weighing equipment, and breeding management equipment, and includes at least data for determining feed intake, weight changes, and growth stages; the diet adjustment is performed through diet formulation and supply equipment. For each diet adjustment, the target diet and the actual implemented diet are recorded separately, and the nutrient parameters of the target diet are vectorized as follows. The nutrient parameter vector of the actual implemented diet is denoted as... .in, This indicates the target nutrient parameters determined and distributed to ration formulation and supply equipment based on the ration adjustment decision. The nutritional parameters of the diet actually received by the pig herd are determined based on the actual feed formulation and supply records.

[0021] It should also be noted that when there is a deviation between the target diet and the actual implemented diet, the identification of subsequent nutrient adjustment direction, adjustment range, and direction-range correction units are all based on the actual changes in the implemented diet before and after correction, rather than directly on the planned adjustment amount in the target diet. For necessary changes in linked nutrient parameters caused by diet formulation constraints, the corresponding changes are recorded and processed in subsequent response learning qualification judgments, taking into account the identifiability of the main nutrient adjustment direction and the consistency between the target diet and the actual implemented diet.

[0022] In this embodiment, the processing of swine herd production data, the updating of response status, and the decision-making regarding diet adjustment are performed by data processing equipment with data storage and computing capabilities. Each evaluation period first assesses the completeness of swine herd production data, the continuity of swine herd composition, and the correspondence between swine herd production data and the actual diet execution period. Data that does not meet the preset data validity conditions is not considered valid learning data for the direction-amplitude adjustment unit's response status update. It should be noted that data validity is only one of the basic conditions for response learning eligibility. Whether a specific historical diet adjustment event can be used for the response status update of the corresponding direction-amplitude adjustment unit also needs to be determined in conjunction with the identifiability of the nutritional adjustment direction, whether the previous major diet adjustment response has been evaluated, and the consistency between the target diet and the actual implemented diet.

[0023] Under the above implementation conditions, the following steps are completed sequentially: efficiency status determination, response status and overall adjustability determination of the direction-amplitude adjustment unit, pig herd stratification and diet adjustment authority determination, target adjustment unit selection and diet adjustment execution, actual efficiency response feedback and local response status update, and updates of available diet adjustment action space, overall adjustability, control level and diet adjustment authority for the next control evaluation cycle, as detailed below: S1. Obtain pig herd production data and growth stages for each evaluation period; determine diet conversion efficiency based on the pig herd production data; and determine the efficiency status by combining the reference diet conversion efficiency corresponding to the growth stage. Specifically, this includes: For basic calibration unit According to the evaluation cycle Obtain feed intake, body weight, and growth stage data corresponding to the actual period of ration implementation. Indicates the basic calibration unit number. Indicates the evaluation cycle number. The evaluation cycle... The effective feed intake within is recorded as When using feed records to determine effective feed intake, the actual feed intake during the evaluation period is reduced by the amount of surplus feed, recovered feed, and abnormal losses that can be determined; when the actual net feed intake can be obtained directly, the corresponding actual net feed intake is used as the effective feed intake.

[0024] Each evaluation period should, in principle, correspond to a stable actual diet execution state. Furthermore, when a major diet switch occurs within an evaluation period and available body weight data is available at the switch point, the evaluation window is redefined based on the switch point, ensuring that the effective feed intake and body weight changes after the redefined window correspond to the same actual diet execution state. When available body weight data is unavailable at the switch point, the corresponding evaluation period is only used as a record of the pig herd's production status and is not broken down to form efficiency response samples corresponding to specific actual diet execution states. The evaluation period should, in principle, not cross growth stage boundaries. When a growth stage transition occurs within an evaluation period, the evaluation period is redefined based on the stage transition point.

[0025] Furthermore, both the beginning and end of the evaluation period belong to the basic calibration unit. Furthermore, pigs with valid weight records are identified as the valid pig set, and the initial total valid weight is determined based on the total valid weight of the valid pig set at the beginning and end of the evaluation period. and total effective weight at the end of the period And determine the effective weight gain:

[0026] in, Indicates the basic calibration unit During the evaluation period Effective weight gain within the group. When the average weight of the pigs is obtained from the farm site, the average weight is converted into the corresponding total effective weight according to the number of pigs in the effective pig set, so as to ensure that the same statistical caliber of pigs is used at the beginning and end of the evaluation period, and to avoid the impact of pig transfers on effective weight gain.

[0027] It should also be noted that the validity of pig herd production data during the evaluation period should be assessed, and the basic calibration unit should be used. During the evaluation period The valid state of the data is recorded as :

[0028] The preset data validity conditions include at least the following: the completeness of the pig herd production data meets preset requirements; the effective pig set meets continuous statistical requirements; the pig herd production data can establish a correspondence with the actual daily feed execution period; and the effective feed intake. Not less than the preset minimum effective feed intake .

[0029] when At that time, the data corresponding to the evaluation period will not be used as valid learning data for subsequent direction-amplitude adjustment unit response state updates; when At that time, it only indicates that the corresponding evaluation period meets the basic data validity conditions. Whether it can be further used for response learning of a specific direction-amplitude adjustment unit still needs to be determined according to step S2, in combination with the identifiability of the nutritional adjustment direction, whether the previous major diet adjustment response has been evaluated, and the consistency between the target diet and the actual implemented diet.

[0030] for The evaluation cycle is determined based on effective weight gain and effective feed intake to establish dietary conversion efficiency.

[0031] in, Indicates the basic calibration unit During the evaluation period Effective weight gain corresponding to effective feed intake per unit To prevent numerical calculations from having a denominator of zero, the value of this protection term is less than the preset minimum effective feed intake. This is not used as a substitute for effective feed intake. The higher the dietary conversion efficiency, the higher the effective weight gain per unit of effective feed intake.

[0032] It should also be noted that pig herds at different growth stages have different baseline dietary conversion efficiency (CFE) levels. Therefore, the CFE level without growth stage correction should not be used directly to determine the efficiency status of pig herds at different growth stages. (Regarding growth stage...) Historical stability evaluation windows were selected from historical evaluation data using the same statistical caliber, meeting the data validity criteria, having stable pig herd composition, and having not undergone major dietary adjustments, thus forming a set of historical stability evaluation windows. And based on the dietary conversion efficiency within the historical stable evaluation window, the reference dietary conversion efficiency for the corresponding growth stage is determined:

[0033] in, Indicates the historical basic adjustment unit number. Indicates the historical evaluation period number. This indicates that the median value is used. Priority is given to using historically stable evaluation windows under the same breeding environment, at the same growth stage, and with consistent statistical caliber. When the corresponding historically stable evaluation data is insufficient, it is supplemented by reference diet conversion efficiency pre-established according to the same statistical caliber for the corresponding growth stage or comparable historical batch data. The update cycle of the reference diet conversion efficiency is longer than the closed-loop adjustment cycle of the diet, making it a benchmark for interim efficiency comparison.

[0034] Furthermore, based on the basic calibration unit During the evaluation period Growth stage The corresponding reference diet conversion efficiency is called, and the current diet conversion efficiency is normalized to determine the efficiency status.

[0035] in, Indicates the basic calibration unit During the evaluation period The efficiency state; when When, it indicates that the current diet conversion efficiency is comparable to the reference diet conversion efficiency at the corresponding growth stage. When it is below the reference level for the corresponding growth stage, The time indicates a level higher than the reference level for the corresponding growth stage. This efficiency status is used to eliminate the impact of differences in baseline dietary conversion efficiency at different growth stages on subsequent herd stratification, but... It only indicates that it is comparable to the stage reference level and is not directly used as the threshold for dividing the control level; the subsequent control level is determined according to the hysteresis judgment results of efficiency status and overall adjustability in step S3.

[0036] S2. Obtain historical diet adjustment data for the pig herd, establish direction-amplitude adjustment units based on the direction and magnitude of nutritional adjustments, determine the response state based on the historical efficiency response corresponding to each direction-amplitude adjustment unit, determine the available diet adjustment action space based on the response state, and determine the overall adjustability based on the available diet adjustment action space, specifically including: Obtain basic calibration unit Historical ration adjustment data defines a complete record of a historical ration adjustment event as a single instance, including a clearly defined target ration, the actual ration implemented before adjustment, the actual ration implemented after adjustment, and the corresponding efficiency response evaluation results. This record is denoted as […]. The historical diet adjustment events mentioned above include at least the nutrient parameter vector of the target diet, the nutrient parameter vectors of the actual implemented diets before and after adjustment, the execution time of the diet adjustment, and the efficiency status determined according to step S1. It should be noted that the direction and magnitude of nutrient adjustment are determined based on the changes in the actual implemented diets before and after adjustment, and the target diet is mainly used to evaluate the consistency of the diet adjustment execution.

[0037] Furthermore, regarding the historical grain adjustment incident... The Middle For each nutrient parameter, the standardized variation is determined based on the actual values ​​before and after calibration, and the allowable range of nutrient parameters for the corresponding growth stage:

[0038] in, and These respectively represent historical food ration adjustment events. The Middle The actual values ​​of each nutritional parameter before and after calibration. and These represent the growth stages. The upper and lower limits of the corresponding nutritional parameters are specified. This is a value protection item. Based on the sign of the standardized change, the nutritional parameter, along with the direction of increase or decrease, is used to determine the direction of nutritional adjustment. The standardized change measures are used to compare the relative changes in different nutrient parameters during the same historical dietary adjustment event, and are not directly used as the basis for dividing the amplitude range of the direction itself.

[0039] For the determined direction of nutritional adjustment The intensity of the directional action is determined based on the actual changes in the corresponding nutrient parameters before and after calibration, as well as the maximum allowable adjustment range per cycle for the corresponding growth stage.

[0040] in, Indicates growth stage Nutritional adjustment direction The corresponding maximum allowable adjustment range per single period, This is a numerical protection item. The directional action intensity is set according to a preset continuous amplitude range. Divide and determine the amplitude range And adjust the direction of nutrition and amplitude range Combined to form direction-amplitude adjustment unit Therefore, local response states are established separately for the same nutritional adjustment direction under different movement intensities, and the response results of larger amplitude movements are not directly attributed to smaller amplitude movements in the same direction.

[0041] It should also be noted that not all historical ration adjustment events are used for the response status update of the corresponding directional amplitude adjustment unit. First, the validity status of the event-level data is determined based on the data validity judgment result of step S1. When the baseline window before the historical ration adjustment event, the actual execution period, and the response evaluation window all meet the preset data validity conditions, it is determined that... .

[0042] Furthermore, candidate major nutrient adjustment directions are determined based on the standardized changes in each nutrient parameter, and the proportion of candidate major nutrient adjustment directions in this dietary change is calculated:

[0043] in, This indicates the number of nutritional parameters included in the comparison. This is a numerical protection item. When the proportion of the stated direction reaches the preset direction identification requirement, the nutritional adjustment direction is determined to be identifiable and valid; when the preceding major diet adjustment has completed the corresponding response waiting period and response evaluation process, the preceding response isolation state is determined to be valid; when the standardized deviation between the target diet and the actual implemented diet is within the preset allowable implementation deviation range, the diet implementation consistency state is determined to be valid.

[0044] The eligibility for response learning of historical diet adjustment events is determined based on the validity status of event-level data, the identifiable status of nutritional adjustment direction, the isolation status of preceding responses, and the consistency status of diet implementation. Only when all of the above conditions are met will the corresponding historical diet adjustment event be used as a valid response learning sample for a specific direction-amplitude adjustment unit. Historical diet adjustment events that do not meet the eligibility criteria can still be retained as records of pig herd production status, but will not be used for updating the response status of the corresponding direction-amplitude adjustment unit.

[0045] For historical diet adjustment events that meet the eligibility criteria for response learning, a stability evaluation window is selected before adjustment that meets the data validity conditions and has not experienced major changes in the diet. The average efficiency status and efficiency change trend before adjustment are determined based on this stability evaluation window. After the response waiting period ends, the average efficiency status after adjustment is determined based on the efficiency status that meets the data validity conditions during the response evaluation period. The original change trend before adjustment is then extended to determine the net efficiency response of this historical diet adjustment event.

[0046] in, and These represent the average efficiency status of the stability evaluation window before calibration and the response evaluation window after calibration, respectively. This indicates the trend of efficiency changes before calibration. This represents the time span from the baseline point before calibration to the response evaluation point. When the pre-calibration stable history of the basic calibration unit itself is insufficient, it is supplemented by the historical stable trend of the same growth stage and similar efficiency status. The net efficiency response is used to improve the comparability of responses between different historical diet calibration events and does not represent a strict biological causal relationship between a single nutrient parameter and changes in herd efficiency.

[0047] Furthermore, the unit amplitude response of this historical diet adjustment event is determined based on the net efficiency response and the directional action intensity of the corresponding nutrient adjustment direction:

[0048] in, This is a numerical protection item. For newly added valid response samples belonging to the same direction and amplitude calibration unit, the response status is updated recursively based on the original response status and the current unit amplitude response:

[0049] in, In response to updating coefficients, and These represent the response status before and after the inclusion of the current valid response sample. The response update is represented using both old and new status formats to avoid confusing the occurrence period of historical ration adjustment events with the time index of the current control evaluation period.

[0050] Furthermore, the response reliability is determined based on the cumulative number of valid historical samples of the direction-amplitude calibration unit and the consistency of the historical unit amplitude response. The range of values ​​for the response reliability is: The response reliability increases when the cumulative number of valid historical samples increases and the dispersion of the historical unit amplitude response decreases, and vice versa. When the number of valid historical samples for the amplitude calibration unit in the corresponding direction of the basic calibration unit is insufficient, the response state is fused with the historical prior response of the corresponding growth stage to obtain the valid response state.

[0051] in, Indicates growth stage Lower direction amplitude adjustment unit The historical prior response. As the number of valid historical samples of the basic calibration unit increases and the reliability of the response improves, the proportion of its own historical response in the valid response state increases accordingly.

[0052] Furthermore, a monotonic mapping relationship is established based on the historical distribution of the effective response states of the amplitude calibration unit in different directions at the corresponding growth stage, thereby converting the effective response states into... Response evaluation value within the range The response evaluation value is used to characterize the relatively positive historical response level of the corresponding direction amplitude adjustment unit in the current growth stage; the response evaluation value and the response state are distinguished from each other, wherein the response state stores historical unit amplitude response information, and the response evaluation value is used for subsequent action space filtering and comparison between different direction amplitude adjustment units.

[0053] Furthermore, based on the remaining range between the current actual diet and the corresponding nutritional security boundary at the growth stage, the maximum permissible adjustment range per cycle for each nutritional adjustment direction, and the overall dietary security requirements, the adjustment space for safe diets is determined. It should be noted that the maximum allowable adjustment range per cycle is used to establish a standardized scale for the intensity of the directional action itself, while the remaining range of the current actual implemented diet from the nutritional security boundary is used to determine whether a specific directional-amplitude adjustment unit still has actual executable space in the current evaluation cycle. The two have different functions.

[0054] Within the safety ration adjustment action space, based on the response evaluation values ​​of each directional amplitude adjustment unit and the currently effective usage restrictions, directional amplitude adjustment units that meet the minimum response evaluation requirements and are not prohibited from use are identified as normally usable units. These units, combined with the currently allowed amplitude range, form the usable ration adjustment action space. For directional amplitude adjustment units that are limited by amplitude range, only the amplitude sub-intervals that are currently allowed to be executed are counted in the available amplitude area; for directional amplitude adjustment units that are prohibited from use, they are not counted in the available daily feed adjustment action space during the prohibited period.

[0055] It should also be noted that, in order to evaluate the remaining dietary adjustment space, only the directions of nutrient adjustment within the non-empty safety margin region are statistically analyzed. The set of these nutrient adjustment directions is denoted as... The remaining available adjustment space coverage is determined based on the proportion of the available range to the safe range for each nutritional adjustment direction:

[0056] in, and These represent the directions of nutritional adjustment. The safe range and the available range, Measure( ) represents the interval length of the corresponding standardized amplitude region. When there is no nutrient adjustment direction with a non-empty safety amplitude region, the remaining available adjustment space coverage is set to 0.

[0057] For each nutrition adjustment direction with available direction-amplitude adjustment units, first determine the representative response from the available direction-amplitude adjustment units of that nutrition adjustment direction, and then determine the available motion response level based on the representative responses of different nutrition adjustment directions. This avoids the same nutritional adjustment direction being repeatedly counted in the overall evaluation due to multiple amplitude ranges. Overall adjustability is determined by combining the available action response level and the remaining available adjustment space coverage:

[0058] in, When the available diet adjustment action space is empty, the overall adjustability is directly determined to be 0. Thus, the overall adjustability simultaneously reflects the historical response levels of different current nutritional adjustment directions and the actual remaining executable adjustment space within the nutritional safety range. When the response state of a direction-amplitude adjustment unit changes or is subject to usage restrictions, the response evaluation value or actual executable range of the corresponding local adjustment unit is first changed, and then the overall adjustability is affected by the available action response level and the coverage of the remaining available adjustment space. The overall adjustability, together with the efficiency state obtained in step S1, serves as the input for step S3 to perform herd stratification and determine diet adjustment permissions.

[0059] S3. Based on the efficiency status and overall adjustability, the pig herd is stratified to determine the control level and the corresponding feed adjustment authority, specifically including: Obtain the basic calibration unit determined in step S1 During the evaluation period Efficiency status and the overall adjustability determined in step S2 The efficiency status is used to characterize the current diet conversion efficiency relative to the reference diet conversion efficiency at the corresponding growth stage, and the overall adjustability is used to characterize the ability of the basic adjustment unit to continue diet adjustment based on historical direction-amplitude response and the current remaining executable adjustment space.

[0060] Furthermore, to reduce frequent switching of control levels caused by normal fluctuations in efficiency status or overall adjustability around the threshold, a decrease threshold is set for each efficiency status. and rising threshold Set a decrease threshold for overall adjustability. and rising threshold ,satisfy:

[0061] The thresholds for the decrease and increase of efficiency status are determined based on the distribution of efficiency status within the historical stable evaluation window of the corresponding growth stage; the thresholds for the decrease and increase of overall adjustability are determined based on the distribution of overall adjustability corresponding to the states where normal diet adjustments can be implemented and the states where adjustments are restricted during the historical closed-loop adjustment process. It should be noted that in step S1... It only indicates that the current diet conversion efficiency is comparable to the reference level for the corresponding growth stage, and is not directly used as the threshold for dividing the control level.

[0062] For basic calibration units that have entered the continuous evaluation process, hysteresis judgments are made based on the current efficiency state, overall adjustability, and the previous evaluation state. When the previous discrete state of efficiency was a higher state, it only transitions to a lower state when the current efficiency state drops to no more than a decreasing threshold; when the previous discrete state was a lower state, it only recovers to a higher state when the current efficiency state rises to no less than an increasing threshold; when the efficiency state is between the decreasing and increasing thresholds, the previous discrete state is maintained. The same hysteresis judgment method is used for overall adjustability, maintaining the previous discrete state between the corresponding decreasing and increasing thresholds.

[0063] It should also be noted that for the basic calibration unit that is being stratified for the first time and does not yet have a previous discrete state, initial judgment reference values ​​are set between the rising and falling thresholds of the efficiency state and between the rising and falling thresholds of the overall adjustability, respectively, and the initial discrete state judgment is completed based on the current evaluation results. After the initial judgment is completed, the above-mentioned hysteresis rule is used in subsequent evaluation cycles. Therefore, the terms "high efficiency state", "low efficiency state", "high overall adjustability" and "low overall adjustability" mentioned later in this embodiment are all discrete states formed after hysteresis judgment, rather than results obtained by directly comparing the current value with a single threshold.

[0064] Furthermore, based on the discrete state after hysteresis judgment of efficiency status and overall adjustability, the basic adjustment unit is divided into a first control level, a second control level, a third control level, and a fourth control level. The first control level is defined as the unit with both high efficiency and high overall adjustability. When the efficiency is high and the overall adjustability is low, it is determined to be the second control level. When the efficiency is low but the overall adjustability is high, it is determined to be the third control level. When efficiency is low and overall adjustability is low, it is determined to be the fourth control level. .

[0065] For the first control level The basic upper limit of the ration adjustment authority is set as the maintenance authority, which is to maintain the current actual ration and not to actively initiate ration adjustments aimed at improving ration conversion efficiency. Due to the high overall adjustability, there is still a sufficient basis for available ration adjustment actions when the efficiency status decreases and the control level conversion conditions are met.

[0066] For the second control level Similarly, the basic authority limit for ration adjustment is set at the maintenance level, prioritizing the current actual ration execution. The difference between the second and first control levels lies in their lower overall adjustability; that is, the remaining available directional-amplitude adjustment units or the executable range have decreased. Therefore, when efficiency declines subsequently, the available ration adjustment actions are relatively limited. The first and second control levels can have the same basic authority limit in the current evaluation period, but maintain different overall adjustability states.

[0067] Furthermore, regarding the third control level Due to the low efficiency and high overall adjustability, the basic upper limit of the ration adjustment authority is determined to be the active adjustment authority. The active adjustment authority means that it allows further selection of direction and amplitude adjustment units from the available ration adjustment action space formed in step S2, but does not predetermine the specific nutritional adjustment direction and adjustment amplitude in this step; the specific target adjustment unit and the intensity of the action to be performed are determined according to step S4.

[0068] For the fourth control level Due to the low efficiency and overall adjustability, the basic upper limit of the ration adjustment authority is set as a restricted adjustment authority. Adjustment is permitted only if there are available directional-amplitude adjustment units within the available ration adjustment action space that meet the amplitude requirements corresponding to the restricted adjustment authority. If no available directional-amplitude adjustment units meet the restricted adjustment authority requirements, or if the available ration adjustment action space is empty, no new ration adjustment is initiated, and the current permitted ration state is maintained. It should be noted that the fourth control level only indicates a low current efficiency and low overall adjustability based on existing historical responses and remaining action space; it is not used to directly determine disease, environmental abnormalities, feeding equipment abnormalities, or other non-rational factors.

[0069] It should also be noted that the control level is used to determine the basic upper limit of the dietary adjustment authority, and does not mean that dietary adjustment will necessarily be implemented according to the basic authority after reaching the corresponding control level. Based on the available dietary adjustment action space formed in step S2, the current nutritional security boundary, and the usage restrictions formed in the previous evaluation cycle that have not yet been lifted, the actual executable dietary adjustment authority is further determined within the basic upper limit; the actual executable dietary adjustment authority can only be maintained or tightened relative to the basic upper limit, and must not exceed the basic upper limit allowed by the corresponding control level.

[0070] Furthermore, to reduce frequent control level migrations caused by normal short-cycle fluctuations, a minimum hold period for the control level is set. After the basic calibration unit enters a new control level, in principle, it will not restore the control level, restore the permissions, or expand the permissions due to fluctuations in normal efficiency status or overall adjustability before the minimum holding period is reached; after the minimum holding period is reached, it will be determined whether to switch control levels based on the hysteresis judgment results of the latest efficiency status and overall adjustability.

[0071] However, the minimum maintenance period does not limit safety tightening. When the available ration adjustment action space shrinks due to usage restrictions, the current actual ration being implemented approaches the nutritional safety boundary, or other established safety conditions require a reduction in the current execution authority, the actual executable ration adjustment authority can be tightened before the minimum maintenance period is reached. This safety tightening does not require an immediate change to the current control level; when the subsequent efficiency status or overall adjustability meets the transition conditions of other control levels according to the aforementioned hysteresis rules, the control level and its basic authority upper limit are updated synchronously. This avoids mechanically changing the control level due to short-term safety authority tightening, and also avoids long-term inconsistencies between the control level and the established efficiency status and overall adjustability status.

[0072] Through the above processing, the efficiency state formed in step S1 and the overall adjustability formed in step S2 are respectively converted into discrete states with hysteresis characteristics. The control level and basic permission limit are determined based on the combination of the two. Then, the actual executable diet adjustment permissions are determined by combining the available diet adjustment action space, nutritional safety boundary and existing usage restrictions, so as to provide permission boundaries for the target adjustment unit screening and diet adjustment execution in step S4.

[0073] S4. When the ration adjustment permission allows adjustment, the target adjustment unit is determined based on the response status of each direction-amplitude adjustment unit in the available ration adjustment action space, the ration adjustment corresponding to the target adjustment unit is executed, and the direction-amplitude adjustment unit to be actually executed is determined based on the actual execution result, specifically including: Obtain the actual executable ration adjustment permissions determined in step S3. When the actual executable ration adjustment permission is a maintain permission, no new active ration adjustment action is generated, and the currently allowed actual executed ration is maintained; when the actual executable ration adjustment permission allows active adjustment or restricted adjustment, further filter the currently executable direction-amplitude adjustment units from the available ration adjustment action space determined in step S2.

[0074] Furthermore, for the preset continuous amplitude range corresponding to the direction's own motion intensity. Let its lower limit and upper limit be respectively denoted as and Based on the actual executable diet adjustment authority determined in step S3, the remaining range of the current actual diet from the corresponding nutritional security boundary, and the direction of nutritional adjustment. The maximum permissible adjustment range in a single cycle at the corresponding growth stage determines the current maximum permissible intensity of the nutrient adjustment direction. The current maximum permissible motion intensity uses the same standardized scale as the directional motion intensity in step S2.

[0075] From the available diet adjustment action space, select directional-amplitude adjustment units whose amplitude range has an actual re-peeling range that is related to the current maximum allowable action intensity, to form a candidate diet adjustment action set:

[0076] It should be noted that the candidate ration adjustment action set is screened for permissions and safety ranges before the target adjustment unit is determined. It does not first select a large adjustment unit that exceeds the current execution permission and then mechanically cut off its action intensity to other ranges, thereby maintaining the correspondence between the direction-amplitude adjustment unit and the subsequent actual execution action.

[0077] Furthermore, based on the current efficiency state determined in step S1 Compared with the efficiency state rise threshold determined in step S3 Identify current efficiency improvement needs:

[0078] in, This indicates the amount of improvement required to reach the efficiency improvement threshold of the current efficiency state. The efficiency improvement requirement is used as a reference target to determine the scale of this round of actions and does not imply that a single ration adjustment must bring the efficiency state to the aforementioned efficiency improvement threshold within a single response evaluation period.

[0079] For each direction-amplitude adjustment unit in the candidate ration adjustment action set, invoke the effective response state formed in step S2. Response Evaluation and response reliability Among them, the effective response state is used to characterize the unit amplitude response level after combining its own historical response and the historical prior response of the corresponding growth stage; the response evaluation value is used to compare the relative positive response levels of the amplitude adjustment unit in different directions; and the response reliability is used to characterize the reliability of the corresponding response basis. The three are not interchangeable.

[0080] For candidate direction-amplitude adjustment units with a positive effective response status, the required action intensity to achieve the current efficiency improvement requirement is determined based on the following:

[0081] in, This is a value protection item. Further, considering the required action intensity, the corresponding preset continuous amplitude range, and the current maximum allowable action intensity, the proposed action intensity for this candidate direction amplitude adjustment unit is determined:

[0082] When the required action intensity is within the currently allowed range, the required action intensity is used; when the required action intensity is below the current allowed lower limit, the current allowed lower limit is used; when the required action intensity exceeds the current allowed upper limit, the current allowed maximum action intensity is used. This ensures that the proposed action intensity has a definite basis for generation, rather than being arbitrarily selected within the range corresponding to the direction-amplitude adjustment unit.

[0083] Furthermore, the expected efficiency response of each candidate direction amplitude calibration unit is determined based on the effective response status and the intensity of the action to be performed:

[0084] The predicted efficiency response is used for comparison among candidate actions in this round and for the predicted mismatch evaluation in step S5, but does not indicate that the corresponding efficiency change will necessarily occur.

[0085] For candidate direction-amplitude adjustment units whose expected efficiency response meets the current efficiency improvement requirements, priority is given to those with lower proposed action intensity. When multiple candidate direction-amplitude adjustment units have the same or similar proposed action intensity, priority is given to those with higher response evaluation values ​​and higher response reliability. When no candidate direction-amplitude adjustment unit is found whose expected efficiency response meets the current efficiency improvement requirements, within the current safety range and the actual executable ration adjustment authority, priority is given to positive candidate direction-amplitude adjustment units with higher expected efficiency responses; when expected efficiency responses are similar, priority is given to those with lower proposed action intensity.

[0086] It should also be noted that for directional-amplitude adjustment units that are in a state of reduced usage priority due to previous weak positive response, but have not yet been excluded from the available ration adjustment action space, when there are candidate directional-amplitude adjustment units that are not subject to the aforementioned priority restrictions and can meet the current efficiency improvement requirements, the target adjustment unit is preferentially determined from the candidate directional-amplitude adjustment units that are not subject to priority restrictions; when there are no candidate units that meet the conditions and are not subject to priority restrictions, the candidate units with reduced usage priority are then allowed to participate in the target selection.

[0087] Based on the above selection results, the target calibration unit for this ration calibration was determined. And the corresponding intensity of the actions to be executed. In principle, only one primary target adjustment unit is set up within a single response evaluation process. While the current primary diet adjustment has not yet completed its response waiting period and response evaluation process, no new primary diet adjustments are added unless there is a need for tightening security measures, in order to maintain the identifiability of the direction of subsequent actual efficiency responses. If the candidate diet adjustment action set is empty, no new target adjustment unit is formed in this round, and the currently allowed actual execution diet is maintained.

[0088] Furthermore, the adjustment symbol is determined based on the nutritional adjustment direction corresponding to the target calibration unit. Wherein, the direction of increase in nutrient parameters corresponds to positive values, and the direction of decrease in nutrient parameters corresponds to negative values; based on the intensity of the action to be performed and the maximum allowable adjustment range per cycle for the corresponding growth stage, the standardized action intensity is restored to the target nutrient parameter adjustment amount:

[0089] The target nutrient parameter adjustment is applied to the current actual diet to form the target diet, whose nutrient parameter vector is represented as follows:

[0090] in, This includes at least the adjustment amounts corresponding to the target nutrient parameters. Other nutrient parameters that require necessary linkage changes due to the implementation of the diet formulation are determined and recorded synchronously according to the formulation constraints when the target diet is formed, but these linkage changes are not pre-assigned as independent responses of the target adjustment unit before the diet is implemented.

[0091] According to the target diet, the diet is adjusted using the diet formulation and supply equipment, and the actual adjusted diet is determined based on the actual feed preparation and supply records. Its nutrient parameter vector is denoted as follows: The target diet represents the nutritional parameters required for the diet adjustment decision, while the actual implemented diet represents the nutritional parameters actually received by the pig herd. The two are recorded separately and are not pre-set to be completely consistent due to the completion of diet formulation and supply equipment.

[0092] Furthermore, based on the actual changes in nutrient parameters between the actual rations before and after calibration, and following the same nutrient adjustment direction and direction-intensity identification method as in step S2, the actual nutrient adjustment direction and actual range of this ration calibration are redefined, and the corresponding direction-range calibration unit is determined accordingly. ).when:

[0093] When the two are inconsistent, the actual execution result still belongs to the original target adjustment unit; when the two are inconsistent, the subsequent efficiency response is not mechanically attributed to the original target adjustment unit, but the direction-amplitude adjustment unit corresponding to the actual execution is used as the basis for the response learning qualification judgment and subsequent efficiency response attribution in step S5.

[0094] After the ration execution is completed, the target calibration unit, the intensity of the planned action, the target ration, the actual rations executed before and after calibration, the corresponding directional-amplitude calibration unit, and the ration calibration execution time are recorded to form a ration calibration record to be evaluated. After the ration calibration record to be evaluated obtains the corresponding actual efficiency response and completes the response evaluation in step S5, it, together with the corresponding response evaluation result, forms a complete historical ration calibration event, which is used for subsequent update of the response status of the directional-amplitude calibration unit.

[0095] S5. Obtain the actual efficiency response corresponding to the ration adjustment, update the corresponding response status based on the actual efficiency response and the direction-amplitude adjustment unit corresponding to the actual execution, and when the actual efficiency response does not meet the preset positive response condition and can be reliably attributed, form the usage restriction of the direction-amplitude adjustment unit corresponding to the actual execution in the next control evaluation cycle after the response evaluation is completed, specifically including: Obtain the ration adjustment record to be evaluated generated in step S4. After the completion of this ration adjustment, a preset response waiting period will follow, and the response evaluation period will begin. The efficiency status during the response evaluation period will be obtained using the same data validity judgment and efficiency status determination method as in step S1. The initiation evaluation period for this ration adjustment will be denoted as... The evaluation period corresponding to the completion of the response evaluation is denoted as . ,in, The local usage restrictions established in this step will take effect from the next control evaluation cycle after the response evaluation is completed.

[0096] Furthermore, based on the actual implemented diet before and after calibration in the calibration record to be evaluated, and following the same nutritional adjustment direction and direction-based action intensity identification method as in step S2, the corresponding direction-amplitude calibration unit is identified. ).when When the two are inconsistent, the actual execution in this instance still belongs to the target calibration unit; when the two are inconsistent, the subsequent actual efficiency response is not mechanically attributed to the original target calibration unit.

[0097] Using the same trend correction method as in step S2, the actual net efficiency response of this diet adjustment is determined based on the average efficiency status of the stability evaluation window before adjustment, the efficiency change trend before adjustment, and the average efficiency status during the response evaluation period:

[0098] in, This refers to the recent food reconciliation incident. and These represent the average efficiency states of the stability evaluation window and the response evaluation window before calibration, respectively. This indicates the trend of efficiency changes before calibration. This represents the time span from the baseline point before calibration to the response evaluation point. If the current basic calibration unit lacks sufficient historical stability, it is supplemented by historical stability trends from the same growth stage and with similar efficiency states as described in step S2.

[0099] It should also be noted that, in order to distinguish between the actual changes after diet adjustment and normal production fluctuations, efficiency status changes in adjacent evaluation periods were extracted from historical stability evaluation windows at the same growth stage where no major diet adjustment occurred, and the median absolute deviation of these changes was used to determine the scale of natural fluctuations. Based on the natural fluctuation scale, thresholds for no obvious response and effective positive response are set:

[0100] in, .

[0101] The response classification is based on the actual net efficiency response: a response is classified as effective positive if the actual net efficiency response is not lower than the effective positive response threshold; a response is classified as weakly positive if the actual net efficiency response is higher than the no-significant-response threshold but lower than the effective positive response threshold; a response is classified as no-significant-response if the absolute value of the actual net efficiency response does not exceed the no-significant-response threshold; and a response is classified as unfavorable if the actual net efficiency response is lower than a negative value of the no-significant-response threshold. Therefore, whether an actual response is considered effective positive depends solely on the actual net efficiency response, not on the accuracy of the predicted efficiency response.

[0102] Furthermore, when the actual execution of the corresponding direction-amplitude calibration unit matches the target calibration unit, the predicted efficiency response generated in step S4 is invoked to determine the degree of prediction mismatch between the predicted efficiency response and the actual net efficiency response:

[0103] in, This is a numerical protection item. The degree of prediction mismatch is used to evaluate the consistency of the existing effective response state with the prediction of the actual response, and is not used to directly change the classification result of the actual response; when the direction-amplitude calibration unit corresponding to the actual execution is inconsistent with the target calibration unit, the predicted efficiency response of the original target calibration unit is not used to evaluate the prediction consistency of the actual execution unit.

[0104] Furthermore, according to the response learning eligibility rules determined in step S2, the data validity, identifiability of the nutritional adjustment direction, evaluation of whether the preceding major diet adjustment response has been completed, and consistency between the target diet and the actual implemented diet are assessed for this diet adjustment event. When the response learning eligibility is met and the actual efficiency response can be reliably attributed to the corresponding direction-amplitude adjustment unit of the actual implementation, the unit amplitude response is determined based on the intensity of the actual direction's own action:

[0105] And update the response status of the direction-amplitude calibration unit corresponding to the actual execution:

[0106] in, The response reliability is updated synchronously based on the number of newly added valid samples and the consistency of historical unit amplitude responses. The valid response state and response evaluation value are then re-determined according to the prior fusion and mapping method described in step S2. The degree of prediction mismatch can be used as a basis for response consistency verification, but the response state should not be mechanically lowered due to a large degree of prediction mismatch.

[0107] It should be noted that when the response learning qualification is not met, the actual efficiency response is still recorded as the pig herd production status, but it is not used to update the response status of specific directional-amplitude adjustment units that cannot be reliably assigned, or to form local usage restrictions based on response evidence. When the actual net efficiency response is an unfavorable response, even if the response learning qualification is not met, the expansion of the current diet adjustment can still be stopped and the actual executable diet adjustment authority can be tightened in accordance with the safety control principle; the aforementioned safety tightening is not considered as historical failure evidence of the original target adjustment unit.

[0108] Furthermore, when the actual efficiency response does not meet the preset positive response conditions, can be reliably attributed to the corresponding direction-amplitude adjustment unit, and meets the response learning qualification, local usage restrictions are formed based on the response classification results. For weak positive responses, the usage priority of the corresponding direction-amplitude adjustment unit is reduced, but the response status of other amplitude ranges in the same nutrient adjustment direction is not modified accordingly; for no obvious response, the continuous repeated execution of the corresponding direction-amplitude adjustment unit is restricted, and the direction-amplitude adjustment unit is suspended when the preset number of consecutive no obvious responses is reached; for unfavorable responses, the corresponding direction-amplitude adjustment unit is suspended, and based on safety control principles, it is temporarily prohibited to expand to a direction-amplitude adjustment unit with a higher amplitude than the current actual amplitude along the same nutrient adjustment direction. The higher amplitude restriction is only a safety action restriction and does not change the existing response status, response reliability, and response evaluation value of the corresponding higher amplitude unit.

[0109] When the actual efficiency response is a valid positive response and the data used for response evaluation meets the validity criteria, the corresponding actual implemented diet will be designated as a candidate safe diet; the candidate safe diet will be continuously subjected to a preset number of cycles. After confirming that no significant efficiency deterioration has occurred within a valid and stable evaluation window, the system is updated to the most recently confirmed safe food ration on the current calibration path. .

[0110] It should also be noted that when an adverse response occurs and it is necessary to reduce the impact of this ration adjustment, the most recently confirmed safe diet on the current adjustment path will be used as the basis for withdrawal, and priority will be given to withdrawing the nutrient parameters involved in this actual ration adjustment and the nutrient parameters that have undergone necessary linkage changes due to the formulation implementation. The set of these parameters will be denoted as... For nutritional parameters belonging to the parameter set Determine the parameter values ​​after the pullback:

[0111] in, .when When this happens, the corresponding nutrient parameters are reverted to the values ​​corresponding to the most recently confirmed safe diet; when Partial reversal will be performed. For other nutrient parameters that are not part of the parameter set involved in this actual calibration and have already maintained a valid and stable state, the current execution state will be maintained in principle. The ration after reversal must still meet the nutritional safety boundary of the current growth stage, and the reversal action itself will not be used as a reverse effective response for learning the original target calibration unit.

[0112] After completing the above processing, the diet calibration record to be evaluated formed in step S4 is combined with the corresponding direction-amplitude calibration unit, actual net efficiency response, predicted mismatch degree, response learning qualification, response classification result, and the formed local usage restrictions to form a complete historical diet calibration event. The updated response status, response reliability, response evaluation value, and currently effective usage restrictions are used as inputs for step S6 to redetermine the available diet calibration action space for the next control evaluation cycle.

[0113] S6. Based on the updated response status and the aforementioned usage restrictions, redetermine the available diet adjustment action space for the next control evaluation cycle, update the overall adjustability based on the redetermined available diet adjustment action space, and update the control level and diet adjustment permissions for the next control evaluation cycle in conjunction with the efficiency status determined upon completion of the response evaluation. Specifically, this includes: After completing the response evaluation of this ration adjustment in step S5, the evaluation period corresponding to the completion of the response evaluation is recorded as . The evaluation cycle for the next round of closed-loop control is then recorded as... This will serve as the next control evaluation cycle. Obtain the evaluation period ending in... The established efficiency status The update includes the response status, response reliability, effective response status, response evaluation value, and currently effective usage restrictions for each direction amplitude adjustment unit after step S5. If step S5 implements a ration pullback, the actual executed ration that is stably implemented after the pullback will be used as the current ration basis for redetermining the ration adjustment action space in the next control evaluation cycle.

[0114] Furthermore, based on the nutrient safety boundary of the growth stage corresponding to the next control evaluation cycle, the maximum allowable adjustment range per cycle for each nutrient adjustment direction, the remaining range of the current actual diet from the corresponding nutrient safety boundary, and the overall dietary safety requirements, the safe diet adjustment action space is redefined in the same manner as in step S2. .

[0115] Based on the safety ration adjustment action space, the available ration adjustment action space is redefined according to the updated response evaluation value and current usage restrictions. For directional amplitude adjustment units whose response evaluation value meets the minimum response evaluation requirement of step S2 and are not in a prohibited state, they are retained in the available ration adjustment action space according to their currently allowed range of amplitude. For directional amplitude adjustment units that are subject to a reduction in usage priority but whose executable range has not been reduced, their available range is still retained, but their selection order is reduced in the next round of target selection in step S4. For directional amplitude adjustment units that are subject to range restrictions, only the currently actually allowed range sub-interval is included in the available range. For directional amplitude adjustment units that are in a prohibited state, they are excluded from the available ration adjustment action space during the prohibited period.

[0116] It should be noted that when a certain amplitude adjustment unit in a certain direction is restricted due to no obvious response or an unfavorable response, the existing response status, response reliability, and response evaluation value of other amplitude ranges in the same nutritional adjustment direction will not be modified accordingly. If there are lower amplitude adjustment units in the same nutritional adjustment direction that are not prohibited and meet the current response requirements and nutritional safety requirements, the lower amplitude unit that is closest to the restricted amplitude will be retained as an optional action in the next round; the lower amplitude unit only obtains candidate status, and whether it is actually executed still follows step S4 to re-form the candidate diet adjustment action set and determine the target adjustment unit.

[0117] Furthermore, a limit is set for the holding period of the direction-amplitude adjustment unit that is suspended in step S5. Reaching the specified limit maintenance period does not mean that the response evaluation value of the amplitude adjustment unit in that direction will automatically recover, but only that it has been qualified for re-evaluation. When the amplitude adjustment unit in the corresponding direction meets the nutritional safety requirements of the current growth stage again, it can conduct limited exploration with a smaller action intensity close to the lower limit within the corresponding amplitude range; after obtaining a new actual efficiency response, the response learning qualification is re-evaluated and the corresponding response status, response reliability and response evaluation value are updated according to step S5, and the normal usability is restored after the normal use requirements are met.

[0118] Furthermore, based on the newly determined available ration adjustment action space, the remaining available adjustment space coverage is re-determined using the same standardized amplitude area coverage method as in step S2. Among them, the directional amplitude adjustment unit that is limited by the amplitude range or is in a limited trial state calculates the available amplitude area only according to the amplitude sub-interval that is currently allowed to be executed; the directional amplitude adjustment unit that is only limited by the reduction of the use priority and does not reduce the executable amplitude range does not reduce its spatial coverage.

[0119] Representative responses are determined from the currently available direction-amplitude adjustment units for each nutritional adjustment direction, and the available motion response levels are redefined based on the representative responses for different nutritional adjustment directions. This allows for the determination of the overall adjustability corresponding to the next control evaluation cycle.

[0120] in, When there is no nutrient adjustment direction with a non-empty safety margin, the remaining available adjustment space coverage is set to 0; when the newly determined available diet adjustment action space is empty, the overall adjustability is directly set to 0. Thus, the local response changes and local usage limitations generated in step S5 first act on the response evaluation or actual executable range of the specific direction-amplitude adjustment unit, and then are transmitted to the overall adjustability through the available action response level and the remaining available adjustment space coverage.

[0121] Furthermore, the efficiency status achieved up to the completion of the response evaluation. With the redefined overall adjustability Based on the efficiency status and the dual threshold hysteresis rule for overall adjustability determined in step S3, discrete state judgment is performed to determine the control level for the next control evaluation cycle. And the corresponding basic permission limit. Then, based on the newly determined available diet adjustment action space, the current nutritional security boundary, and the still effective usage restrictions, determine the diet adjustment permissions that can actually be executed in the next control evaluation cycle within the basic permission limit.

[0122] It should also be noted that when usage restrictions or nutritional safety conditions reduce or exhaust the available space for dietary adjustments, or necessitate a reduction in actual execution authority, the actual executable dietary adjustment authority should be tightened promptly according to the safety tightening priority rule determined in step S3, without being subject to the minimum maintenance period. Restrictions on the restoration of control levels, permissions, or the expansion of permissions; when the efficiency state and overall adjustability that have been established subsequently meet the conversion conditions of other control levels according to the hysteresis rules, the control level and the upper limit of basic permissions are updated synchronously.

[0123] It should be further explained that the next control evaluation cycle At the outset, complete hog herd production data for the evaluation period was not yet available, therefore, data not yet obtained was not utilized. Determine the control authority for this evaluation cycle. During the evaluation cycle... After completion, the corresponding efficiency status is determined according to the actual pig herd production data obtained in step S1, and used for the subsequent closed-loop control status update.

[0124] Furthermore, when a basic adjustment unit enters a new growth stage, for direction-amplitude adjustment units with the same nutrient adjustment direction and a corresponding relationship between standardized amplitude ranges, the historical response state formed in the previous growth stage is retained, and its response reliability is reduced based on the transferability of the historical response to the new growth stage. Subsequently, according to the prior fusion method in step S2, the historical response state is fused with the historical prior response corresponding to the new growth stage using the reduced response reliability. For direction-amplitude adjustment units where a corresponding relationship between nutrient adjustment direction and standardized amplitude range cannot be established between the old and new growth stages, the response state of the original growth stage is not directly used; instead, the corresponding local response state is re-established based on the historical prior response of the new growth stage and subsequent newly added effective responses. After completing the growth stage transition, the safe diet adjustment action space and the available diet adjustment action space are re-determined according to the nutrient safety boundary corresponding to the new growth stage.

[0125] Through the above processing, step S6 is formed sequentially as follows: Updated local response status and usage restrictions → Available ration adjustment action space → Overall adjustability → Control level → Feedback relationship of actual executable ration adjustment permissions.

[0126] When the actual executable diet adjustment authority in the next control evaluation cycle allows for continued adjustment, return to step S4, select the target adjustment unit from the newly determined available diet adjustment action space, and enter the next round of diet adjustment; when the actual executable diet adjustment authority does not allow the initiation of a new diet adjustment, maintain the currently allowed actual executed diet, and continue to acquire pig herd production data for subsequent evaluation cycles, thereby completing the hierarchical closed-loop adjustment of pig herd diet conversion efficiency.

[0127] Example 2: Closed-loop adjustment of pig diet under effective positive response conditions This embodiment is based on Embodiment 1, using a basic calibration unit. The following is an example of the operation of the present invention, which is illustrated by a ration adjustment process. The parameters below are exemplary operating parameters.

[0128] During the evaluation period, The initial total effective body weight was 6500 kg, and the final total effective body weight was 7200 kg. The effective feed intake was 1800 kg, resulting in a diet conversion efficiency of 0.3889. The corresponding reference diet conversion efficiency for the growth stage was 0.4000. The value is 0.9723. Let the efficiency state rise threshold be 0.995, and the current efficiency state remain at a low level.

[0129] Based on historical daily food supply adjustment data, the second range of net energy increase is expected. Set as Its effective response state is 0.0673, response reliability is 0.82, and response evaluation value is 0.84; lysine increases in the following direction. The effective response state is 0.0538. According to S2 of Example 1, the remaining available calibration space coverage is determined to be 0.78, and the available action response level is 0.80. The overall tunability is 0.668, which is higher than its rise threshold of 0.60. Therefore, it is determined to be the third control level according to S3. Active adjustments are allowed.

[0130] According to S4 calculations, the direction of net energy increase and the direction of lysine increase are... The required action intensities are 0.337 and 0.422 respectively, therefore the direction of net energy increase is chosen. As the target adjustment unit. The current actual net energy from the daily food intake is... To achieve the target daily net energy of approximately The actual preparation result is The corresponding actual action intensity is 0.327, which is still within the range of... Therefore, the actual direction and amplitude adjustment unit is consistent with the target adjustment unit.

[0131] In this embodiment, the response evaluation period is one evaluation cycle, and the efficiency status is 0.9970 when the response evaluation is completed. Combining the average efficiency status of 0.9710 before calibration and the trend of 0.0010 per evaluation cycle, the actual net efficiency response is 0.0250, which is higher than the effective positive response threshold of 0.020, and is therefore determined to be an effective positive response. The corresponding response status is updated from 0.070 to 0.072; after adding effective samples, the number of effective samples increases and the dispersion of the unit amplitude response decreases, and the response reliability is updated from 0.82 to 0.85. In this embodiment, we take... The actual daily ration is updated to a safe ration after being confirmed by two effective stability evaluation windows.

[0132] Recalculated according to S6, the remaining available adjustment space coverage for the next control evaluation cycle is 0.75, the available action response level is 0.82, and the overall adjustability is 0.666. Combined with the efficiency status of 0.9970 at the completion of the response evaluation, the control level is adjusted according to the hysteresis rule. Updated to The basic permission limit is adjusted to maintain permission, thereby completing a closed-loop adjustment of the daily ration conversion efficiency.

[0133] Example 3: Localized Diet Adjustment Restrictions and Action Space Reconstruction under Adverse Response Conditions This embodiment is based on Embodiment 1, using a basic calibration unit. An adverse response during a ration adjustment illustrates the process of local constraints and motion space reconstruction. This is determined according to S1. The efficiency state is 0.968; according to S2, the remaining available calibration space coverage is determined to be 0.82, and the available action response level is 0.78. The overall tunability was found to be 0.675. Let the threshold values ​​for decreasing and increasing overall tunability be 0.45 and 0.60, respectively. Given the current low efficiency and high overall tunability, it is determined to be the third control level according to S3. Active adjustments are allowed.

[0134] Current available room for adjustment in daily food intake includes the direction of net energy increase. and the direction of lysine increase The direction of net energy increase The amplitude range is According to S4, it was identified as the target calibration unit, with a planned action intensity of 0.62; the actual action intensity was 0.60, still falling within the range. Therefore, the actual efficiency response can be attributed to the amplitude adjustment unit in that direction.

[0135] After the response evaluation was completed, the actual net efficiency response obtained according to S5 was -0.018, with no obvious response threshold of 0.010; therefore, it was determined to be an unfavorable response. This event meets the requirements for response learning, and the direction of net energy increase will be updated based on the actual negative samples. The response status and reliability were assessed, and the unit was paused, while the direction of net energy increase was temporarily prohibited. Direction of net energy increase The existing response state will not be mechanically modified and will be retained as a low-amplitude candidate action while meeting response evaluation requirements and nutritional safety conditions; lysine increase direction Keep it available. If a rollback is necessary, use the most recently confirmed safe diet as the benchmark and roll back the nutritional parameters involved in this actual adjustment and formula-related changes.

[0136] After reconstructing the available diet adjustment action space for the next control evaluation cycle according to S6, the remaining available adjustment space coverage decreased to 0.60, the available action response level was adjusted to 0.70, and the overall adjustability was updated to 0.490. Since it is still above the decrease threshold of 0.45, according to the hysteresis rule, the overall adjustability remains high, and the control level is maintained. However, the actual authority to adjust the diets that can be executed has been further tightened. If the next round of adjustments continues in the direction of net energy increase, the target adjustment unit should be re-determined from the still available lower-amplitude adjustment units in accordance with S4.

[0137] Example 4: Response attribution correction when the target diet differs from the actual implemented diet Based on Example 1, this embodiment uses actual analytical data from a publicly available swine nutrition experiment as playback input to explain the process of re-identifying and responding to the deviation between the target diet and the actual diet.

[0138] The calculated value of FeedA's lysine formula from the publicly available data used is... Actual analysis value The calculated value of FeedB's lysine formulation is The actual analytical value is 1.27%.

[0139] This embodiment only uses the publicly available measured data to verify the actual execution mechanism of the present invention, and does not describe it as data from aquaculture experiments conducted by the applicant.

[0140] Using the actual analytical value of 0.92% in FeedA as the lysine parameter for the actual diet before calibration, and the calculated value of 1.30% in FeedB as the lysine parameter for the target diet, the target adjustment is 0.38 percentage points. The maximum allowable adjustment range per cycle for the direction of lysine increase corresponding to the growth stage is set to 0.40 percentage points, and the direction's own action intensity is also set. Values ​​less than 0.90 are classified as the second amplitude range. , Divided into the third amplitude range Therefore, the intensity of the motion in the direction corresponding to the target daily ration is:

[0141] Therefore, the target adjustment unit determined by S4 is the direction of lysine increase. .

[0142] After the actual formulation of the diet, the actual implemented diet after adjustment was characterized by the actual analytical value of FeedB, which was 1.27%. This indicates an actual increase of 0.35 percentage points in lysine, and the actual directional self-motion intensity is:

[0143] The actual motion intensity of 0.875 falls into the second amplitude range. Therefore, the actual direction-amplitude adjustment unit is the lysine-increasing direction. , with target calibration unit Inconsistent. Following S4 of Example 1, the actual execution results are recorded in the diet adjustment record to be evaluated, and the direction of lysine increase corresponding to the actual execution is used. This serves as the basis for attributing subsequent actual efficiency responses.

[0144] After the response evaluation is completed, according to S5, the data validity, identifiability of the nutritional adjustment direction, whether the previous major dietary adjustment response has been evaluated, and the consistency between the target diet and the actual implemented diet are determined. When this event meets the response learning qualification and the actual efficiency response can be reliably attributed, the unit amplitude response is determined based on the actual action intensity and the actual net efficiency response, and the lysine increase direction is updated. The response status and reliability; the original target calibration unit is not updated using the actual efficiency response. The response status. If the actual efficiency response does not meet the preset positive response conditions, the corresponding usage restrictions will also apply to the actual execution. Instead of mechanically acting on the original target calibration unit .

[0145] Therefore, when the deviation in nutritional parameters between the target diet and the actual diet is sufficient to cause the actual action to cross a preset range, this invention can re-identify the direction-range adjustment unit based on the actual diet, ensuring that the subsequent efficiency response corresponds to the diet adjustment actually received by the pig herd. This avoids incorrectly updating the historical response status of other range adjustment units due to response attribution based on the target diet. This actual execution attribution result further serves as the basis for updating the subsequent available diet adjustment action space and overall adjustability.

[0146] The technical features of this invention not described can be implemented by or using existing technology, and will not be repeated here. Of course, the above description is not a limitation of this invention, and this invention is not limited to the examples above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of this invention should also be within the protection scope of this invention.

Claims

1. A method for stratified closed-loop adjustment of diet conversion efficiency in pig herds, characterized in that, Includes the following steps: S1. Obtain pig herd production data and growth stage in each evaluation period, determine the diet conversion efficiency based on the pig herd production data, and determine the efficiency status by combining the reference diet conversion efficiency corresponding to the growth stage. S2. Obtain historical diet adjustment data of the pig herd, establish direction-amplitude adjustment units based on the direction and magnitude of nutritional adjustment, determine the response state based on the historical efficiency response corresponding to each direction-amplitude adjustment unit, determine the available diet adjustment action space based on the response state, and determine the overall adjustability based on the available diet adjustment action space. S3. Based on the efficiency status and overall adjustability, the pig herd is stratified to determine the control level and the corresponding daily feed adjustment authority. S4. When the ration adjustment permission allows adjustment, determine the target adjustment unit based on the response status of each direction-amplitude adjustment unit in the available ration adjustment action space, execute the ration adjustment corresponding to the target adjustment unit, and determine the actual direction-amplitude adjustment unit to be executed based on the actual execution result. S5. Obtain the actual efficiency response corresponding to the ration adjustment, update the corresponding response status according to the actual efficiency response and the direction-amplitude adjustment unit corresponding to the actual execution, and when the actual efficiency response does not meet the preset positive response conditions and can be reliably attributed, form the usage restriction of the direction-amplitude adjustment unit corresponding to the actual execution in the next control evaluation cycle after the response evaluation is completed. S6. Based on the updated response status and the usage restrictions, redetermine the available diet adjustment action space for the next control evaluation cycle, update the overall adjustability based on the redetermined available diet adjustment action space, and update the control level and diet adjustment permissions for the next control evaluation cycle in conjunction with the efficiency status determined when the response evaluation is completed.

2. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 1, characterized in that, Step S1 includes: The feed intake, body weight, and growth stage of the pig herd are obtained within the evaluation period corresponding to the actual diet implementation period; when a major diet switch occurs and available body weight data is available at the switch time, the evaluation window is redefined based on the switch time. Pigs that belong to the pig herd and have effective weight records at both the beginning and end of the evaluation period are identified as the effective pig set. The effective weight gain is determined based on the initial total effective weight and the final total effective weight of the effective pig set. The effective feed intake is determined by deducting the amount of surplus feed, the amount of recycled feed, and the amount of abnormal loss from the actual feed supply. Data validity is judged based on the completeness of pig herd production data, the continuity of the effective pig set, the correspondence between pig herd production data and the actual ration execution period, and the effective feed intake threshold. The ration conversion efficiency is determined based on the effective weight gain and effective feed intake. The reference diet conversion efficiency is determined based on the historical stable evaluation window of the corresponding growth stage, and the efficiency status is determined based on the ratio of the diet conversion efficiency to the reference diet conversion efficiency.

3. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 1, characterized in that, Step S2: Obtain historical diet adjustment data for the pig herd; establish direction-amplitude adjustment units based on the direction and magnitude of nutrient adjustments; determine the response status based on the historical efficiency response corresponding to each direction-amplitude adjustment unit, including: Historical diet adjustment events are obtained, including the target diet, the actual diet before adjustment, the actual diet after adjustment, and the corresponding efficiency response evaluation results. The direction of nutrient adjustment is determined based on the actual amount and direction of change of each nutrient parameter. The intensity of the direction's own action is determined based on the actual change in the nutritional parameters corresponding to the nutritional adjustment direction and the maximum allowable adjustment range per single cycle of the corresponding growth stage. The intensity of the direction's own action is then divided into a preset continuous range and a direction-amplitude adjustment unit is formed. Eligibility for response learning is determined based on the validity of historical diet adjustment data, the identifiability of nutritional adjustment directions, whether the response to previous major diet adjustments has been completed, and the consistency between the target diet and the actual implemented diet. For historical ration adjustment events that meet the response learning qualifications, the net efficiency response is determined based on the efficiency change trend before adjustment, the unit amplitude response is determined based on the net efficiency response and the direction's own motion intensity, and the response state of the corresponding direction-amplitude adjustment unit is updated accordingly.

4. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 3, characterized in that, In step S2, the available diet adjustment action space is determined based on the response state, and the overall adjustability is determined based on the available diet adjustment action space, including: The response reliability is determined based on the number of effective historical samples and the consistency of the unit amplitude response of each direction-amplitude calibration unit. The response state is then fused with the historical prior response of the corresponding growth stage based on the response reliability to obtain the effective response state. Finally, the response evaluation value is determined based on the historical response distribution of the corresponding growth stage. The remaining range between the current actual diet and the nutritional security boundary, the maximum allowable adjustment range per cycle for each nutritional adjustment direction, and the overall dietary security requirements determine the safe diet adjustment action space, and the available diet adjustment action space is determined based on the response evaluation value and the usage restrictions of the direction-amplitude adjustment unit. The remaining available adjustment space coverage is determined based on the safe and available ranges of each nutrient adjustment direction. The representative response for each nutrient adjustment direction is determined, and the available action response level is determined based on the representative responses of different nutrient adjustment directions. The overall adjustability is determined in combination with the remaining available adjustment space coverage.

5. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 1, characterized in that, Step S3 includes: Set an increase threshold and a decrease threshold for the efficiency state and the overall adjustability respectively. Perform hysteresis judgment based on the current evaluation result and the discrete state of the previous evaluation period to determine the high and low states of the efficiency state and the overall adjustability. The following are defined as the first control level, the second control level, the third control level, and the fourth control level: the first control level is characterized by high efficiency and high overall adjustability, the second control level is characterized by high efficiency and low overall adjustability, the third control level is characterized by low efficiency and high overall adjustability, and the fourth control level is characterized by low efficiency and low overall adjustability. The corresponding daily ration adjustment authority is also determined for each of these levels. The basic upper limit of the diet adjustment authority is determined based on the control level, and the actual executable diet adjustment authority is determined within the basic upper limit based on the available diet adjustment action space, nutritional safety boundary and existing usage restrictions; the minimum retention period of the control level is set, wherein safety tightening is not limited by the minimum retention period.

6. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 1, characterized in that, Step S4: When the ration adjustment permission allows adjustment, determine the target adjustment unit based on the response status of each direction-amplitude adjustment unit in the available ration adjustment action space, including: Based on the aforementioned diet adjustment authority, the remaining range between the currently implemented diet and the nutritional security boundary, and the maximum permissible adjustment range per single cycle for each nutritional adjustment direction, the current maximum permissible action intensity for each nutritional adjustment direction is determined, and the direction-amplitude adjustment units currently allowed to be executed are selected from the available diet adjustment action space to form a candidate diet adjustment action set. The efficiency improvement requirement is determined based on the current efficiency status and the efficiency status rise threshold. The action intensity required to achieve the efficiency improvement requirement is determined based on the response status of each candidate direction-amplitude adjustment unit. The action intensity to be executed is determined by combining the corresponding preset continuous amplitude range and the current maximum allowable action intensity. The expected efficiency response is determined based on the effective response status and the intensity of the action to be performed for each candidate direction-amplitude calibration unit. For candidate direction-amplitude calibration units whose expected efficiency response meets the efficiency improvement requirements, the direction-amplitude calibration unit with the smaller intensity of the action to be performed is selected first, and the target calibration unit is determined by combining the response evaluation value and response reliability of each candidate direction-amplitude calibration unit. When there is no candidate direction-amplitude calibration unit whose expected efficiency response meets the efficiency improvement requirements, the target calibration unit is determined from the candidate direction-amplitude calibration units with positive expected efficiency responses.

7. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 6, characterized in that, In step S4, the ration adjustment corresponding to the target adjustment unit is performed, including: Based on the nutrient adjustment direction, the intensity of the action to be performed, and the maximum allowable adjustment range per cycle for the corresponding growth stage of the target adjustment unit, the adjustment amount of the target nutrient parameters is determined, and the target diet is formed based on the current actual diet. The target ration is formulated and supplied, the actual ration after adjustment is obtained, and the target adjustment unit, the target ration, the actual ration after adjustment and the execution time are recorded to form the ration adjustment record to be evaluated. Based on the actual changes in nutrient parameters between the actual ration before and after calibration, the actual nutrient adjustment direction and range are redefined, and the corresponding direction-range calibration unit is determined accordingly. When the direction-range calibration unit corresponding to the actual execution is inconsistent with the target calibration unit, the direction-range calibration unit corresponding to the actual execution is used as the basis for subsequent efficiency response attribution.

8. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 7, characterized in that, Step S5 includes: The efficiency status within the response evaluation period corresponding to the calibration record of the diet to be evaluated is obtained. The efficiency change after calibration is corrected according to the efficiency change trend before calibration to determine the actual net efficiency response. The natural fluctuation scale is determined according to the historical stability evaluation window of the corresponding growth stage, and the response classification result is determined according to the actual net efficiency response and the natural fluctuation scale. The response classification result includes effective positive response, weak positive response, no obvious response and adverse response. The degree of predicted mismatch is determined based on the predicted efficiency response and the actual net efficiency response, and the eligibility for response learning is determined based on the data validity of the diet adjustment records to be evaluated, the identifiability of the direction of nutritional adjustment, whether the previous major diet adjustment response has been evaluated, and the consistency between the target diet and the actual implemented diet. When the response learning qualification is met, the unit amplitude response is determined based on the directional action intensity and actual net efficiency response of the directional-amplitude calibration unit corresponding to the actual execution, and the response state and response reliability of the directional-amplitude calibration unit corresponding to the actual execution are updated accordingly; when the directional-amplitude calibration unit corresponding to the actual execution is inconsistent with the target calibration unit, the response state of the target calibration unit is not updated using the actual net efficiency response.

9. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 8, characterized in that, Step S5 also includes: When the actual efficiency response can be reliably attributed to the direction-amplitude adjustment unit corresponding to the actual execution and meets the response learning qualification, the use priority of the corresponding direction-amplitude adjustment unit is reduced for weak positive responses, the repeated execution of the corresponding direction-amplitude adjustment unit with no obvious response limitation is allowed, the use of the corresponding direction-amplitude adjustment unit is suspended for adverse responses, and the direction-amplitude adjustment unit that expands to a higher actual amplitude along the same nutrient adjustment direction is temporarily prohibited. When the response learning qualification is not met, the response status of the specific direction-amplitude adjustment unit is not updated or local usage restrictions are not formed based on the actual efficiency response that cannot be reliably attributed; when the actual efficiency response is an unfavorable response, the actual executable ration adjustment authority is tightened and the current ration adjustment is stopped from being expanded. When the actual efficiency response is a valid positive response, the corresponding actual implemented diet is determined as a candidate safe diet, and the safe diet is updated based on the subsequent stability evaluation results; when the actual efficiency response is an unfavorable response and diet reversal is required, the nutritional parameters involved in the actual diet adjustment and the nutritional parameters of the formula linkage changes are reversed based on the safe diet.

10. The method for stratified closed-loop adjustment of pig herd diet conversion efficiency according to claim 1, characterized in that, Step S6 includes: The safe diet adjustment action space is redefined based on the current actual diet, the nutritional safety boundary of the corresponding growth stage, and the maximum allowable adjustment range per cycle for each nutritional adjustment direction. The available diet adjustment action space is also redefined based on the updated response status and the usage restrictions. Specifically, for directional-amplitude adjustment units whose executable range is reduced due to the usage restrictions, only the executable range is retained, while directional-amplitude adjustment units that are prohibited from use due to the usage restrictions are excluded. Based on the redefined available dietary adjustment action space, the remaining available adjustment space coverage and representative responses of each nutrient adjustment direction are redefined. Based on the representative responses, the available action response level is determined, and the overall adjustability is updated based on the available action response level and the remaining available adjustment space coverage. Based on the efficiency status determined at the time of response evaluation and the updated overall adjustability, hysteresis judgment is performed according to the preset rise and fall thresholds of efficiency status and overall adjustability. The control level and basic permission limit of the next control evaluation cycle are updated, and the actual executable diet adjustment permissions are further determined based on the available diet adjustment action space, nutritional safety boundary and usage restrictions.

Images