Livestock breeding risk intelligent evaluation method and system based on multi-source data fusion
By fusing multi-source data to form root risk evidence, the problem of confirming the risk status after the migration of members in livestock farming has been solved. This has enabled the accurate positioning of the risk-affected objects and the traceability of the management basis, ensuring the accurate updating of the operational risk status.
Patent Information
- Application Number
- CN202611576563.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-13
- Publication Date
- 2026-11-10
AI Technical Summary
In livestock farming, during the risk assessment process, it is difficult to accurately determine whether the target of the risk has changed after the relocation of members, resulting in a lack of basis for the confirmation and release of risk status, which affects subsequent management decisions.
By fusing multi-source data, operational data of the target aquaculture unit is obtained, root risk evidence is formed, historical members and their affiliation are fixed, and risk source separation conditions are determined after members migrate. The current constraint objects are confirmed by utilizing the response transition period and routine data collection, and the operational risk status is updated.
Effectively distinguish between member migration and actual separation of operating conditions, reduce the interference of data changes before and after migration on risk assessment, provide traceable risk management basis, and ensure that the risk constraint objects correspond to the actual operating conditions.
Smart Images

Figure CN123070741A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of livestock farming operation risk analysis and intelligent management technology, and in particular to a method and system for intelligent assessment of livestock farming risks based on multi-source data fusion. Background Technology
[0002] As large-scale livestock farming develops towards intensive and digital management, data records covering changes in livestock herds, pen operations, feeding and water supply, production status, and daily management activities have gradually formed during the farming process. Existing livestock management systems can usually identify abnormal changes based on the above-mentioned operational data and form risk levels, management tips, or subsequent handling suggestions by combining historical records. For livestock that have been moved, isolated, or regrouped, their subsequent status can also be tracked through individual identification and transfer records. Such technologies provide a data foundation for risk analysis, herd management, and operational decisions such as slaughtering and moving of livestock in farms, and are an important data processing method in risk management of livestock farming activities.
[0003] In actual operation, risk assessment results often directly lead to changes in the livestock being raised and the management conditions. For example, if a breeding unit experiences persistent operational risks, the managers may transfer some of the abnormal livestock to other pens for continued observation. After the transfer, the composition of the original breeding unit changes, and the transferred livestock enters new operating conditions. At this point, the subsequent risk performance of the original breeding unit decreases, which may be related to the change in the original operating conditions or simply to the departure of the original abnormal members. If the transferred livestock continue to maintain an abnormal state, it may also be due to the persistent influence that was formed before the transfer. If the two breeding units before and after the transfer still use the same feed source, water supply conditions, or other common operating conditions, although the positions of the members have changed, the actual risk conditions may not have truly been separated. Therefore, it is difficult to directly determine whether the original risk should continue to constrain the original breeding unit and relocate members, or whether it should remain in a state of pending confirmation, based solely on the current group status after the transfer.
[0004] The aforementioned issues will further extend into the ongoing management of operational risks. After a risk is triggered, relocated members typically undergo more intensive manual review or enhanced observation, while members remaining in the original breeding unit continue to be routinely observed according to their original production rhythm. This changes the conditions under which different operating entities obtain data. At the same time, even after livestock leave their original operating conditions, their existing risk manifestations may continue for a certain period. The relocated breeding unit itself may also undergo concurrent operational adjustments. The resulting subsequent changes in status may be influenced by factors such as the continuation of historical impacts, changes in the destination, and changes in observation conditions. Simply comparing risk values before and after relocation is insufficient to directly serve as the basis for a change in the target of risk.
[0005] In livestock farming management, the identification of the target of risk directly affects whether subsequent management statuses such as relocation, herding, slaughter, and resource allocation can be adjusted. If the same batch of status changes used to judge the current target of risk are directly used as the basis for risk resolution, then there is a lack of independent operational basis between the identification of the risk source and the release of the risk status. This may cause the risk status that still needs to be managed to exit the current risk set prematurely. Therefore, in the case of continuous changes in the relationships among farming members and operating conditions, how to maintain an effective correspondence between the risk formation stage and the subsequent operating status, distinguish whether the migration of members actually changes the conditions under which the risk is exerted, and determine the operating targets that continue to exert risk under the condition that subsequent data is comparable, and then control the closure and release of the corresponding operating risk status accordingly, are data processing issues that still need further research in the process of intelligent risk assessment and management decision-making in livestock farming. Summary of the Invention
[0006] This application proposes a method and system for intelligent risk assessment of livestock farming based on multi-source data fusion, in order to solve the problems mentioned in the background technology.
[0007] To achieve the above objectives, this application adopts the following technical solution: Intelligent risk assessment methods for livestock farming based on multi-source data fusion include: Acquire multi-source aquaculture operation data of the target aquaculture unit and determine its operational risk status; When the operational risk status first meets the risk management judgment criteria, the historical members and their affiliation, the operational conditions for the risk formation, and the routine data collection conditions corresponding to the operational risk status are fixed to form root risk evidence. When some members of the historical members migrate, the migrating members and the remaining members are determined; If an operating condition exists that affects both the migrating and remaining members before the migration, continues to affect the remaining members after the migration, and ceases to affect the migrating members, then the risk source separation condition is determined, and the reference member for the target aquaculture unit is identified. The response transition period is determined based on the historical response records in the multi-source aquaculture operation data. After the response transition period ends, subsequent data of the migrated members, retained members and reference members are obtained according to routine data collection conditions. The current constraint objects of the root risk evidence are determined based on the differences in risk status of the migrated members and retained members and the concurrent status of the reference members, and the cutoff time of the data used is recorded. Only new data that is later than the deadline and meets the routine data collection conditions is used to confirm that the current constrained object has exited the operational risk status. The operational constraint status of the root risk evidence is updated based on the exit confirmation result, and the operational risk management result is output.
[0008] A smart risk assessment system for livestock farming based on multi-source data fusion includes: The root risk evidence formation module acquires multi-source aquaculture operation data of the target aquaculture unit and determines the operation risk status. When the operation risk status first meets the risk management judgment conditions, it fixes the historical members and member affiliation relationships, the operation conditions for risk formation, and routine data collection conditions corresponding to the operation risk status to form root risk evidence. The risk source separation determination module identifies the migrating members and the remaining members when some members in the historical members migrate. When there is an operating condition that jointly affects the migrating members and the remaining members before the migration, continues to affect the remaining members after the migration and then stops affecting the migrating members, the module determines the risk source separation condition and identifies the reference members for the target aquaculture unit. The current constraint object determination module determines the response transition period based on historical response records in multi-source aquaculture operation data. After the response transition period ends, it obtains subsequent data of migrating members, retained members, and reference members according to routine data collection conditions. Based on the differences in risk status of migrating members and retained members and the concurrent status of reference members, it determines the current constraint object of root risk evidence and records the cutoff time of the data used. The operational constraint status update module only uses new data that is later than the deadline and meets the routine data collection conditions to confirm that the current constraint object has exited the operational risk status. Based on the exit confirmation result, it updates the operational constraint status of the root risk evidence and outputs the operational risk management result.
[0009] The beneficial effects of this invention are as follows: This invention establishes historical members and their affiliations, the operating conditions involved in the formation of the risk, and routine data collection conditions when an operational risk first arises, forming root risk evidence. After members migrate, the historical interaction relationship is maintained along the same root risk evidence. As a result, subsequent changes in the member structure, such as relocation or merging, will not reverse the member scope at the time of risk formation. At the same time, by verifying the boundary of the influence of the same actual source operating conditions before and after migration, it is possible to distinguish between the spatial migration of members and the actual separation of operating conditions, thus avoiding premature changes in the operational risk status of the original breeding unit simply due to the migration of abnormal members.
[0010] After establishing the risk source separation conditions, this invention determines the response transition period based on the historical response records of the corresponding operating conditions. Under unified routine data collection conditions, it separately establishes the subsequent states of retained members, migrating members, and reference members. Combining the risk status differences between migrating members and retained members with the concurrent state of the target aquaculture unit, it determines the current constraint object. Thus, it can reduce the interference of the continuation of risk response before migration, changes in collection conditions, and concurrent changes in the target aquaculture unit on the risk attribution judgment, and ensure that the operating objects subject to continued risk constraint have data basis corresponding to the actual changes in operating conditions.
[0011] This invention further records the cutoff time of the data used when the current constraint object is determined, and only uses the new data formed after the cutoff time to perform the operational risk status exit confirmation, so that the process of determining the risk object and the process of removing operational constraints adopt mutually separate data intervals; thus, the recovered data after migration will not simultaneously bear the dual functions of risk attribution and risk removal, and the operational constraint status can be updated according to subsequent independent data, providing a traceable risk management basis for column transfer, grouping, column exit and operational resource allocation. Attached Figure Description
[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort: Figure 1 This is an overall flowchart of the present invention; Figure 2 Define the flowchart for the current constraint object. Detailed Implementation
[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0014] like Figure 1 and Figure 2 As shown, this invention discloses an intelligent risk assessment method for livestock farming based on multi-source data fusion, including the following specific steps: The present invention will be described below in the context of large-scale livestock farming. In this embodiment, the farming unit is taken as the object of business risk management. The system reads the member identifier and member affiliation change record through the existing farming management system, reads the actual source and effective period of the operating conditions through the feeding and water supply management system, and obtains the production status record through the existing farming monitoring equipment. The data from each source are established according to the member identifier, farming unit identifier and effective time to establish a corresponding relationship. Then, the root risk evidence is formed, the risk source separation condition verification after the member migration is performed, the current constraint object is determined and the business constraint status is updated in sequence.
[0015] The operational risk status indicates the management status of a livestock farming entity when it meets the existing risk management judgment conditions. The operational risk judgment rule adopts the configured production status analysis rule and outputs whether the corresponding entity meets the operational risk status within the specified collection period. The system saves the rule version and its actual input record index when it is first triggered. The same rule is used for retained members, migrated members and reference members in the future. The current constraint object indicates the scope of the farming unit or member that needs to be reviewed again based on the root risk evidence. Source separation and status difference are used to configure the management object. The obtained object allocation result is not used as proof of animal disease diagnosis or single operating condition cause.
[0016] The target aquaculture unit is identified by the affiliation record of its corresponding members. Each member is identified by its member identifier and is associated with its effective affiliation range. Migration events are associated with member details by event identifiers and the affiliation records include the affiliation unit from which the migration occurred, the destination aquaculture unit, and the actual time of completion of the migration. Operating condition records use both the source system identifier and the actual source identifier to locate the source. The actual supply object and effective period of the source are stored according to the condition type. Production status records point to the observed object through member identifiers or aquaculture unit identifiers and store the original observation time and its coverage period. Root risk evidence is subsequently reused by indexing these records to reuse the original facts. Member affiliation and source supply constitute the historical member boundaries and the operating condition effect boundaries, respectively.
[0017] The operational risk assessment rules are configured using a versioned approach. The rule configuration record saves the input field definitions, units and conversion methods, as well as the aggregation method, input time span and assessment boundaries generated from the original records. The assessment boundaries are the thresholds and state demarcation conditions that the rule has been enabled before the risk is triggered. The applicable conditions record the evaluable member types and input integrity requirements.
[0018] The system reads the saved configuration by rule version number and generates inputs with consistent field structures for the three types of members. The single-cycle calculation within the original rule and the status confirmation for two consecutive collection cycles in this method are executed separately. When the original rule contains cross-cycle inputs, the original records involved in the input are saved along with the calculation results. Before the rule is called, the system constructs the input according to the calculation method of the saved fields, and does not substitute the same field name for the same field meaning.
[0019] The summation field reads and accumulates valid values within the specified member range. The average field uses the number of valid members within the same range as the denominator. The ratio field records both the member identifier corresponding to the numerator and the member identifier of the denominator. If the units of the fields are different, they are first unified according to the saved conversion relationship. When the denominator is 0 or necessary records are missing, the system returns to the pending completion state. Risk output codes and operational constraint states are saved separately. Periodic results that are not risky are not directly written as constraint release.
[0020] After all source records are converted to a unified timeline, the original timestamps are retained. Member migration and operating condition adjustments use the actual effective time. The record creation time only indicates the time when the system receives the data. The valid interval is represented by the start time being valid and the end time being invalid. When the same member is transferred to a new column consecutively, the end time of the original interval is the same as the start time of the new interval. The operating conditions retain the supply interruption and source co-operation situation according to the actual execution records. Time with missing records is not automatically filled into continuous supply. The system confirms the range of data that has been received from the complete batch records of the source system and formally judges to use the data submitted within that range.
[0021] In this embodiment, multi-source aquaculture operation data of the target aquaculture unit is obtained and the operation risk status is determined. After the operation risk status meets the risk management judgment conditions for the first time, the historical members and member affiliation relationships, the operation conditions for risk formation and routine data collection conditions corresponding to the operation risk status are fixed, and the root risk evidence is formed with the fixed historical relationships.
[0022] When the operational risk status first meets the risk management judgment conditions, the system records the risk trigger time and obtains the index of the input record actually called in this judgment. When the input field is calculated from multiple original records, the system continues to trace back to these original records along the field index and reads the earliest and latest coverage times. The end point of the risk formation period is the risk trigger time, and the starting point is determined based on the coverage of the records actually involved in the formation of this risk. The fixed rule configuration and reference value before the trigger are saved as rule configuration, and their training data is not used as the source of the historical members in this case.
[0023] The system queries backward from the most recent complete collection period corresponding to the risk triggering time, expanding by one period each time, and summarizing the found member identifiers and the actual source identifiers of the operating conditions. If the summarization ranges are the same for two consecutive expansions, that location is recorded as a candidate stable position. To finally determine the risk formation period, the candidate window must cover every original record corresponding to the current input index. If there are earlier called records, the system continues to trace back to that record. The consecutive identical member list and source identifiers are only used to verify the scope of the object and cannot replace the time coverage of the actual called records.
[0024] The historical boundary verification can be performed 2 to 3 times, with 2 times being preferred. The selected number of times is written into the processing configuration before the root risk evidence is formed. The 3 times are used for repeated verification when the source system returns records in batches. This number is the number of confirmations for the record query and does not represent the physiological time required for risk formation. After the system reads the original records actually called this time, it determines the risk formation period from the earliest coverage start point to the risk trigger time and uses the complete original input replay saved rules. The verification result is consistent with the trigger record. Even if the object range in an earlier window is the same, the effective period of the called record is still completely preserved. The verification of the complete coverage of the actual input record is carried out by comparing the index set. The system uses the original record index saved by the risk trigger record as the range to be read and marks the indexes obtained by the query as read one by one. When the derived field contains the cumulative period, the original index corresponding to the period is verified again. After all the indexes to be read have been obtained, the earliest time is taken according to the original record coverage start point, instead of taking the first record according to the database return order. Based on this, the records that actually participated in the input can be found back during the risk formation period, and there is no situation where the same member in the current database is used to replace the actual input object this time.
[0025] Historical members are determined by the overlap between their affiliation interval and the operating condition effect interval. For each candidate member, the intersection of the effective interval to which they belong to the target aquaculture unit and the risk formation period is first taken. Then, the intersection of this intersection with the effective operating condition effect period that actually covers the member is taken. The common starting point is the latest time among the starting points, and the common ending point is the earliest time among the ending points. When the common ending point is later than the common starting point, an actual effect interval is obtained. The system records the member identifier that exists in the effect interval into the historical members and member affiliation relationship, and at the same time saves the actual source identifier corresponding to the interval. The overlap calculation is performed in groups according to the member identifier and the actual source identifier.
[0026] The intervals within a group are first arranged in ascending order of their start times. When the start time of a subsequent segment is not later than the end time of the current merged segment, the end time of the merged segment is the later of the two end times. When the start time is later than the current end time, a new segment is started. Only the left-closed and right-open intervals that meet at the boundary times do not generate repeated durations. The system saves the correspondence between the merged intervals and the original source record indexes. Subsequently, it can both calculate the actual coverage and locate the original record. The judgment of historical members only requires the existence of a verified positive-length effective interval.
[0027] When forming root risk evidence, the risk formation period is determined based on the data actually accessed when the operational risk status was formed, and the historical members and their affiliation are determined based on the members' actual presence during the risk formation period and the actual impact of the operating conditions on the members.
[0028] When a member enters the same breeding unit multiple times during the risk formation period, the system saves each actual presence interval separately. Overlapping intervals with the same source of influence are merged first, while intervals that do not connect are saved separately. Therefore, a member only appears once in the historical member range, but can be associated with multiple actual influence records. When verifying whether a member has experienced a certain operating condition, the system reads the actual influence interval corresponding to that member, instead of extrapolating the entire historical period based on the current member list at the time of risk triggering.
[0029] The operating conditions that participate in risk formation are recorded separately for their actual source identification and effective period of action; members who are actually in the target aquaculture unit and affected by the operating conditions that participate in risk formation during the risk formation period are recorded as historical members, thereby fixing the relationship between historical members and corresponding operating conditions in the risk formation stage.
[0030] The actual source identifier of the operating conditions follows the unique business identifier of the source system. Identifiers with the same name across systems are distinguished by the source system identifier. The condition type points to the actual supply content. The supply intensity and operation settings follow the corresponding execution version of the source system. When it is subsequently determined that the operating conditions maintain their original function, both the source identifier and the setting fields that have been called by the risk rules are compared. When the equipment identifier is the same but the actual supply settings have been adjusted, it is recorded as an adjustment of the operating conditions to avoid substituting the same equipment name for the same actual function.
[0031] Routine data collection conditions are derived from the collection configuration that was enabled before the risk was triggered. The system saves the start and end rules of the collection cycle corresponding to the configuration. For each input field, the system saves the data source type, relative collection time, and collection method. The relative collection time is based on the actual time of the recorded business event, plus the time offset in the original configuration. When different breeding units use the same type of equipment, the meaning, units, and collection methods of the fields are unified. The physical equipment identifier is still retained by each unit. The corresponding records can be queried through the equipment identifier later.
[0032] Once a fixed association is established between the period during which the risk is formed, historical members and their affiliations, the operating conditions for participating in the risk formation, and the routine data collection conditions, subsequent member migrations and the addition of new members only continue to record the subsequent relationships, without changing the scope of historical members and the boundaries of the operating conditions that have already been formed.
[0033] The system generates a unique root risk evidence identifier, which permanently links the period of risk formation with historical members and their affiliation. It then links this identifier to the actual operating conditions and routine data collection conditions that participated in the formation of this risk. This identifier also points to the operating risk status and the saved rule configuration. The fixed record reflects the historical facts that were confirmed when the risk was triggered. Subsequent member migrations are recorded as new valid intervals in the independent member affiliation record. The correspondence between historical and current records is established with the same root risk evidence identifier. The fixed association of the root risk evidence is achieved by connecting the member details and source details through the master record identifier.
[0034] When querying the historical relationships of a migration, the member identifier is first obtained using the root risk evidence identifier, and then the source details are located using the member identifier and the historical scope of action. The process does not reverse the process from the current member list of the target aquaculture unit. If the same member experiences multiple operating conditions, the scope of action and record version of each source are saved separately. When the business records of the original source are subsequently revised, the fixed details still retain the version used at the time of the trigger, and the revised corresponding relationship is added for re-verification. The input basis of the original judgment remains traceable.
[0035] Root risk evidence is stored in the master record associated with detailed records. The master record stores the root risk evidence identifier, target aquaculture unit identifier, operational risk status code, risk formation period, and rule version number. Historical member details locate members using the root risk evidence identifier and member identifier, and locate the member's historical role using the actual presence interval and condition record index. Operational condition details store the version and role period of the actual source record. Routine data collection condition details store the version that was enabled and configured before the risk was triggered.
[0036] Subsequent status records reference the master record identifier and the identifier of this migration event. The member scope and duration of action in the historical details remain unchanged. The joint query of the historical member details takes the root risk evidence identifier as the entry point. The member identifier is used to connect the member ownership record, and the source record index is used to connect the operating condition execution record. The time intersection of the three query paths is the saved actual action interval. When reading subsequent source separation records, the root risk evidence identifier of the original details is referenced again. The current business record only provides the new actual action endpoint and the ownership after migration. According to this association, when a member migrates from the target breeding unit to the destination breeding unit, the old action interval can still be recalculated by the original index, and the current effective interval is located by the new event identifier. It does not rely on matching the names of the two current member lists to inherit the risk.
[0037] The routine data acquisition conditions also preserve the existing task completion time limits and time matching tolerances of the source system. Each subsequent acquisition task generates a target time based on the cycle start point and relative acquisition time position, and limits the available time range of the record according to the original tolerance. Fields without configured tolerances are matched according to the original timed acquisition task identifier. When a task corresponds to multiple duplicate uploaded records, the original record identifier is used to remove duplicates first, and then the valid record with the smallest difference from the target time is selected. When the differences are the same, the earlier original record is used. This selection is completed before reading the observation values. Supplementary acquisition does not change the number of routine tasks or the opportunity for formal comparison.
[0038] For members who were only actually present for a portion of the risk formation period, the system retains their confirmed period of influence. A shorter period of presence does not mean that the member did not experience the current operating conditions. Members participating in the source comparison later must also be within the scope of the applicable rules that have been saved and have sufficient records to reconstruct the corresponding inputs. The same risk evidence represents a common historical event and does not automatically mean that each member's exposure duration, status level, and migration selection reasons are the same. The system verifies the comparison scope based on the actual influence record and limits the output to the current operating review object.
[0039] Once the root risk evidence is formed, the historical members involved in subsequent member migrations are obtained from the historical members and member affiliation relationships corresponding to the root risk evidence. The operating conditions involved in the determination of risk source separation conditions are obtained from the operating conditions already associated in the root risk evidence. Subsequent data are determined according to routine data collection conditions, which determine their corresponding data sources, relative collection time, and collection methods.
[0040] When some members of the historical members migrate, the migrating members who actually left the target aquaculture unit and the remaining members who stayed in the target aquaculture unit are identified from the historical members associated with the root risk evidence. After the migrating members enter the target aquaculture unit, they continue to be associated with the same root risk evidence. When migration occurs again, the member affiliation relationship is updated along the root risk evidence.
[0041] When a member migration occurs, the system reads the member details that have actually been submitted for the migration event. The system takes the intersection of the historical member identifiers saved by the root risk evidence and the member details of the event, and records the members in the intersection as the migrated members. Members who continue to be validly affiliated in the current aquaculture unit and belong to the same historical member range are recorded as retained members. When an event includes multiple destination aquaculture units, the migration member range is saved separately according to the destination aquaculture unit identifier, and the corresponding reference members are matched separately. Common retained member records can be referenced by these relationships, but each subsequent data still retains its original record identifier.
[0042] The scope of migrated and retained members is fixed at the beginning of this comparison. Member identification details are recorded in the processing record corresponding to the migration event. Subsequently, each time a status input is generated, the record is read according to the details. When a member migrates again, the affected subsequent relationships are updated with a new migration event. When only the group name or business tag is adjusted, the actual member affiliation range remains unchanged, and the risk handling continues to use the original relationship. The original members of the target breeding unit are retained as existing members of that unit and do not acquire the historical member status due to the merger management.
[0043] After the system completes the migration member division, it points each historical member in the migration details to one actual destination breeding unit and verifies that its old affiliation range has ended at the actual migration time. The scope of retained members is the members who are still present in the same historical relationship within the unit that is being migrated out this time. Members who have migrated out in an earlier event will continue to be inherited by their original migration relationship. According to this association method, even if multiple destination breeding units are registered at the same time in one business submission, the current comparison can still clearly identify the retention range and reference range corresponding to each migration range, and member records will not be duplicated due to repeated consumption of the same migration event.
[0044] When migrating again along the same risk evidence line, the system reads the existing member scope directly involved in this migration, associates the members actually transferred out with the members that continue to be retained with new migration events, and saves the previous member relationship identifier pointed to by this relationship. The split relationship is used for historical tracing, and the current judgment references its subsequent valid relationship. When a member migrates continuously between multiple destination breeding units, the migration chain is restored according to the order in which the events take effect, and the historical root record remains unchanged.
[0045] Within the same business transaction, the system reads and locks the current affiliation record of the member to be changed, verifies that the relocated breeding unit is consistent with the recorded affiliation, and then writes the end time of the old affiliation, the start time of the new affiliation, and the migration event details. After writing, the transaction is committed, and subsequent processing only reads the committed version. The two valid affiliation intervals are connected at the time of migration completion. If the affiliation record of a member after leaving cannot be determined, it is a member to be supplemented. Comparable migration relationships are not formed by filling in the inferred destination breeding unit.
[0046] The risk source separation condition takes the same actual source of business conditions involved in the risk formation as the object. The system first verifies that the source covers both the migrating member and the retained member before the migration, and then verifies that it continues to cover the retained member after the migration and ends its effect on the migrating member. When there are business conditions that satisfy this complete relationship, the actual source identifier is written into the risk source separation condition record. When other business conditions still work together, the corresponding records are retained. The joint effect itself does not cancel the verified separation relationship, nor does it trigger the release of business constraints on its own.
[0047] Before migration, the joint effect is verified one by one for each member. The system queries the actual effect range of the root risk evidence stored by member identifier to confirm that the migrated members and the retained members in this comparison range experienced the source during the risk formation period. Only one source record with the same name was found in each of the two sets of records. Therefore, the entire range of members cannot be identified as having jointly experienced the risk. The effective attribution range of the members, the source execution record, and the actual effect range are used to determine the member details covered by this condition.
[0048] After migration, the system prioritizes verification of one complete routine data collection cycle. After the source system completes the data submission for this cycle, the system reads the execution records of the retained members who continue to receive data from the original source, and reads the actual end record of the original source's effect on the migrated members and their execution records from the source after migration. When submitting in batches across cycles, two cycles can be verified continuously, with the first cycle still verified independently under the same conditions. The waiting time for data to arrive and the actual time when the member is affected by the source are recorded separately. If a complete submission covers a longer period, it is processed according to the complete cycle covered by the submission, and waiting for two cycles is not automatically regarded as the data being complete.
[0049] The termination of the action is based on the records of executed source switching, supply end, or member leaving the supply range. The system intersects the original source execution record with the actual on-site interval of the migrated member to obtain its final actual action endpoint, and points the source switching completion record to this endpoint. If the source execution interval of the remaining member continuously covers the supply period during the verification period, the original source action corresponding to the migrated member has ended, and the actual supply details of the target breeding unit are complete, the risk source separation condition and the actual action end time are written.
[0050] If no record from the same source is found, keep the record pending verification to avoid mistaking an upload omission for termination. Continuous operation is verified according to the supply period to be executed. The system obtains the supply period to be executed for this cycle from the source system, merges the actual execution intervals from the same source, and covers the supply period segment by segment: the start of the actual interval should not be later than the start of the covered period, and the merged end point should reach the end point of that period. Intermittent feeding is verified according to each actual supply task, and continuous water supply is verified according to the continuous connection interval. Therefore, "continuous" adopts the original execution method of the operating conditions. The migration member side verifies the end record and the actual supply task after migration at the same time. Any unexplained task gaps are kept pending verification. Termination cannot be determined solely by the empty query result of the old source.
[0051] When operating conditions are stored in batches, the coverage area is determined by the actual supply to the breeding unit and the execution period of the batch. Then, the member's action area is obtained by intersecting with the member's affiliation interval. When there is a clearing or replacement process when the feed source is switched, the switching completion time confirmed by the execution record is used. The planned switching time does not replace the actual end time. When there are branches supplying water, the valid records of each branch are retained. When the original branch still forms a supply connection with the migrating member, it continues to be recorded as a joint action. The end of the operating condition record indicates the end of the verifiable external action. The state continuation caused by it is handled by the subsequent response transition period. When the source system record is a supply plan, the execution completion record is first connected through the task identifier, and then the actual source and completion time are read. When the original source is still in contact with the migrating member in the shared storage or transportation link, the record of the actual supply that has not yet ended is used. After the execution record confirms that the original source is disconnected, the actual endpoint is written into the source separation details corresponding to the migration event. The endpoint in this detail can be later than the member space migration time. The response transition period is calculated from the end of the actual action to avoid the supply switching delay being missed.
[0052] In the implementation of feeding and operating conditions, if the two member areas used the same batch of feed before the relocation, and the remaining member continued to receive the same batch after the relocation, and the execution record of the target breeding unit confirmed that the relocated member had switched to a different batch of feed, and the actual effect record of the original batch ended, the source formed a risk source separation condition. When the same water supply source still covers the two member areas, its relationship continues to be associated with the root risk evidence, and the interpretation of the common operating conditions is retained when bilateral risks occur subsequently.
[0053] After establishing the risk source separation conditions, reference members are identified from the target aquaculture unit that were continuously present in the aquaculture unit before the migration occurred, were not included in the historical members of the root risk evidence, and had available contemporaneous data. The contemporaneous status of the reference members is used to identify the contemporaneous changes of the target aquaculture unit itself.
[0054] The verification range of reference members before migration is 1 to 2 complete routine data collection cycles, preferably 2. In this implementation, the two most recent complete cycles before migration are used. First, it is confirmed that the candidate members continuously belong to the target breeding unit. Then, the member range is filtered according to the applicable conditions of the saved rules. Subsequently, the original records of the range are used to calculate the operating risk status of the same period. If the corresponding judgment conditions are not met for two consecutive cycles and the routine input is complete, the range is fixed as the reference member. The alternative implementation of 1 cycle is applicable when only 1 completed cycle can be obtained before migration. The status of the same period is still confirmed according to the same rules. The number of historical cycles actually used is clearly indicated in the record.
[0055] The system verifies the reference members according to the growth stage and production category recorded in the applicable rules, and retains the records of the actual operating conditions they received in the target breeding unit. The scope of reference members is determined before reading the status results after migration, and this scope of members is maintained for comparison with the same period. If the addition of a migrated member causes an anomaly in the reference member due to joint breeding, the same period data of the reference member reflects the changes that have occurred in the target breeding unit, and the system maintains the operating constraints; the specific source of the change is still further verified by the operating records.
[0056] The reference member's contemporaneous data is obtained according to the actual effective collection task of the target aquaculture unit, and the historical verification results and current period results of the member's scope are saved separately. When there are other recorded supply adjustments in the target aquaculture unit during the comparison period, the event and the effective period are associated with the current period to facilitate the identification of situations where the input background changes at the same time. The historical applicable objects and field configurations used for this comparison remain fixed. If the reference member undergoes a new migration, the affected scope is re-verified. The system limits the period that can be compared contemporaneously based on these existing records, and does not interpret the lack of similar risk signals as the complete absence of destination factors.
[0057] After verification, the system will associate the migration event with the actual source record of the separated operating conditions. The retained members correspond to the side where the original source continues to have an effect, while the migrated members correspond to the side where the actual effect of the original source has ended. This correspondence, together with the root risk evidence, provides the historical starting point and the difference in subsequent effects. The reference members only provide the observable status of the target aquaculture unit at the same time and do not participate in the redefinition of the original historical members' scope. Their failure to enter the operating risk status indicates that the same risk performance has not been observed, and this cannot be used to prove that the target aquaculture unit has no other impact.
[0058] This migration processing record is located using root risk evidence identifiers and migration event identifiers. The record points to the retained members, migrated members, and reference members through the member scope details, and then points to the operating conditions that have formed risk source separation conditions through the actual source details. At the same time, the operating conditions that still work together are preserved. The subsequent response calculation reads the type of separated operating conditions and the actual end time, and the status calculation reads the details of the three types of members.
[0059] Before entering the formal comparison, the system extends the verification scope of the source execution record to the end of the current comparison cycle to confirm that the source separation relationship used for comparison is still valid. When generating subsequent calculation tasks from the source separation details, the task references the root risk evidence identifier and migration event identifier, and the member scope uses the version number of the fixed details. The response transition period calculation references the actual effect of each condition to end the record.
[0060] When reading status data, first confirm that the version still corresponds to the fact that the migration has been completed. Then query the observations according to the units where the three types of members are located. Only when the historical members, separated sources and subsequent collection cycles can be linked through these identifiers will the results of this cycle enter the object mapping. The missing corresponding records are saved as tasks to be completed, and the missing inputs of the original task will be supplemented later.
[0061] The response transition period is determined based on the historical response records in the multi-source aquaculture operation data. This period is used to select the starting point of data that can be used for state comparison after the members leave the original operating conditions. The upper limit of the response duration is the empirical time value obtained by matching the historical records according to the specified sorting rules. Its function is to set a unified waiting boundary. It cannot be concluded from this value that the historical influence will necessarily disappear after exceeding this boundary. The judgment of the current constraint object still uses the source relationship and the measured subsequent state of each member's range at the same time.
[0062] For each operating condition that forms a risk source separation condition, historical condition adjustment events are queried using the condition type code and operating risk status code. Historical events must also meet the applicable objects and input configuration requirements of the saved rules. There are execution records for both the end of the original actual source and the completion of the adjustment. The most recent complete collection cycle before the adjustment should meet the corresponding operating risk status. After the adjustment, continuous records are obtained according to the same routine data collection conditions, and it is verified that other operating conditions that have been called by the historical risk judgment during the same period maintain their original status.
[0063] For historical events that meet the aforementioned conditions, starting from the first complete collection cycle after the actual adjustment completion time, the cycle status is recalculated sequentially using the same judgment rule saved this time. After finding the first two consecutive complete cycles in which both events exit the operational risk status, the completion time of the second cycle is recorded as the response confirmation completion time of the event. The response duration is obtained by subtracting the actual adjustment completion time from the response confirmation completion time.
[0064] This time includes the observation period required to complete the continuous exit confirmation. The start time of the first normal cycle is not used to replace the confirmed completion time in advance. When searching for a continuous exit cycle, the system reads the complete status results in ascending order of the cycle start point, and the continuous count starts from 0.
[0065] If a complete cycle fails to meet the risk assessment criteria, the count is incremented by 1; if the next complete cycle also fails to meet the criteria, the count reaches 2, and the end time of that cycle is recorded. If a risk state occurs or a specified cycle is missing during the period, the continuous count restarts from 0. Observed records are saved. The duration of historical responses is obtained by subtracting two defined times. The original calculation is completed in a uniform time unit. When displayed in hours, the total number of seconds is divided by 3600. The sorting uses the original calculated value without truncation.
[0066] The system uses historical event identifiers to deduplicate data, retaining completed adjustment records, original record indexes involved in calculations, and continuous period results. Repeated uploads of the same event are counted only once. Records with inverted timestamps and missing key fields are verified first. Response durations that are longer but valid are kept at their original values. For similar events that have not yet formed continuous exit confirmations, the observed period and current risk status are saved. Unfinished periods cannot be written as 0, nor can a shorter upper limit be obtained simply by deleting events that still pose a risk. The historical calculation scope is determined before this migration. If there are applicable events within the scope that have not yet been completed, the system continues to wait for their records to be completed, and current operational constraints remain valid.
[0067] When multiple operating conditions are adjusted simultaneously, historical events are marked with the end time of the original actual source and the combination of conditions involved. When there is a synchronous adjustment history that matches the current combination, the upper limit of the response duration corresponding to the combination is obtained using the same method and checked against the upper limit of each individual item. The combination record is used to supplement the existing individual historical basis and avoid directly interpreting the maximum value of an individual item as the physical upper limit of the combined effect of any multiple conditions. Multi-condition events that lack compatible combination records are retained as cases to be verified. The preferred implementation method is to use source changes with compatible historical records for object judgment.
[0068] After a similar historical event is completed, the effective response duration is sorted in ascending order of value. The preferred range of the finite sample coverage parameter is 0.05 to 0.1, preferably 0.05. In this embodiment, its name specifically refers to the selection parameter of the upper sorting position. When calculating, first add 1 to the number of effective historical records, then calculate 1 minus the finite sample coverage parameter, multiply the two results and round up. The integer obtained is used as the sorting position starting from 1. The response duration at this position is read as the upper limit of the response duration. This sequential statistics does not require a preset normal distribution, but still requires that the historical events participating in this comparison have the same calculation caliber and compatible action conditions.
[0069] When there are no fewer than 19 valid historical records, the preferred parameter 0.05 ensures that the sorting position falls within the sample range. This value prioritizes retaining longer response records. When there are 9 to 18 valid records, first calculate the number of valid records plus 1, divide 1 by the result, and then compare it with 0.05 and take the larger value.
[0070] At this point, the sorting position calculated above is exactly the last position, and the maximum effective response duration can be read directly to avoid rounding to the nearest whole number. If there are fewer than 9 items, the parameter range of 0.05 to 0.1 cannot form an effective position under the same rule. The system saves the state to be accumulated and continues to monitor. The selection of this parameter is not equivalent to a 90% or 95% safety probability guarantee for this individual response. The sorting calculation uses the original time value saved with seconds as the unified unit of calculation.
[0071] When the parameter is 0.05, the number of valid records can be increased by 1, multiplied by 95, and then divided by 100. If there is a remainder in the quotient, take the integer quotient and add 1; if there is no remainder, take the integer quotient. When the parameter is determined by the reciprocal of the number of records and the sample is 9 to 18 records, the last position can be used directly, which is equivalent to the complete calculation. This implementation avoids the offset caused by rounding the approximate decimal first and then rounding up. Response records with equal time are arranged in the order of event identifiers, and their values are the same. The sorting identifier order does not change the upper limit of the selected response duration.
[0072] Each separated operating condition forms a response duration upper limit and saves the corresponding historical event index. For a single change, the upper limit of the single change is used as the response transition period length. For multiple changes with compatible combination history, the maximum value between the single upper limit and the combination upper limit is used. The start time is the end time of the confirmed original actual source action. If the ends of multiple sources are not synchronized, the latest time is taken. The start time is added to the duration determined this time to obtain the end time of the response transition period.
[0073] This value remains unchanged in this judgment. New records only enter the next historical calculation version. When a migrated member is affected by the original source again, the comparison relationship is re-formed according to the updated effect record. The historical event list and its completed version used in this response calculation are saved. When the maximum value is taken, the event range that provides the value is also recorded. When a source actually terminates later in the migrated member, the public wait starts with the latest endpoint, which can ensure that each source gets a subsequent period of no less than the maximum wait time in this case. The cycle comparison starting point formed according to this rule is a conservative scheduling result and is not used as a new risk level score. The new upper limit formed after the historical record is supplemented is used for the next calculation version. If the current event needs to be recalculated, a new version is explicitly generated and the subsequent cycle is reselected from the new boundary.
[0074] After the response transition period ends, subsequent data for retained members, migrated members, and reference members are obtained according to the routine data collection conditions for the preservation of root risk evidence, and the same judgment rules used when the operational risk status was formed continue to be used to generate the contemporaneous status of the three types of members.
[0075] After the transition period ends, the first complete cycle is obtained according to the cycle start and end rules of the routine data collection conditions. The system queries the original records of the corresponding fields using three types of member ranges. First, it verifies the collection task to which the record belongs, and then determines the matching record based on the saved time tolerance. Candidate records of the same field are all processed according to the same selection rules. The source record identifier and the actual coverage period are written into the input record. If a field only has supplementary monitoring records but no corresponding routine task records, the cycle is kept in the pending state and the missing set is not replaced by the set with more observations.
[0076] The three types of members form independent inputs. For member-level quantities, the original field calculation method is first used to summarize the quantities within the validity period of each member, and then the group field required by the rules is generated according to the member range. The group average quantity uses the actual number of valid members in the range as the denominator. For common environmental quantities such as temperature and humidity at the breeding unit level, the contemporaneous records of the unit where the member range is actually located are referenced and the source identifier is retained. The total feed intake of the whole pen cannot be estimated by the proportion of the number of people in the mixed group to calculate the individual feed intake of the migrating members and the reference members. When such fields are involved, the original records that can be located to the members or the independent measurement records of the member range are used.
[0077] The rule is invoked only if the input source, dimension, and applicable quantity all conform to the saved rule. After the member range is fixed, the system verifies the member coverage of each input field. Member-level fields should be able to locate the valid records within the current range one by one, while the group common field should cover the specified period of the breeding unit in which the range is located. If only the lower limit of the group quantity is met but member-level records are missing, the input cannot be supplemented by allocating the total number of the whole pen according to the number of people. This verification is used simultaneously with the minimum member quantity condition of the original rule, and the result is noted in the periodic input record. The group field is calculated only after the condition is met. Therefore, the production status data of the original members of the target group will not be automatically obtained when the same member moves into a larger target group.
[0078] The current constraint object is determined to use two consecutive complete routine data collection cycles. Subsequent exit confirmations will also use two consecutive new cycles. The system first executes the original judgment rule on the independent input of each cycle, and then checks whether the results of two adjacent cycles are the same. This continuous confirmation layer is fixed at two cycles and is recorded separately from the historical input span that may be used within the rule. The number of confirmations in this layer will not be rewritten to 3 because three cycles have been used within a certain module.
[0079] Two consecutive cycles characterize the continuous state during the observation period. The risk judgment boundary of the original rule remains unchanged. Each complete acquisition cycle is uniquely determined by the start and end points in the routine configuration. The system generates cycle input at the end of the cycle and after the tasks corresponding to the required fields have been completed. Different member ranges share the same pair of cycle identifiers. After calling the same rule, they write their respective results. When the rule output meets the risk, the original risk code is saved. When the output does not meet the risk, the exit result corresponding to the code is saved. Interface errors and missing fields are not converted into exits. The confirmation results of the two cycles reference both cycle inputs. Therefore, the continuity of the state and the observation range used can be verified subsequently.
[0080] Each member's results for each period are saved as three categories: satisfying the operational risk status, not satisfying the operational risk status, and pending completion. If both are satisfied for two consecutive periods, the operational risk status is maintained; if neither is satisfied for two consecutive periods, the operational risk status is exited. Inconsistent results and pending completion data continue to be recorded. The same period boundaries and judgment rules are used for reference members, so the subsequent mapping compares the combination of states in the same time period.
[0081] The system first checks the status of the reference members in the same period. If the reference members have entered the operational risk status or are still waiting for a complete cycle, the current record remains pending judgment and the operational constraint status remains valid. If the reference members have not entered the operational risk status in the same period, the system then compares the retained members with the migrated members: if the retained members maintain the risk and the migrated members exit the risk, the breeding unit where the retained members are located is determined as the current constraint object; if the migrated members maintain the risk and the retained members exit the risk, the scope of the migrated members is determined as the current constraint object. This result is based on the verified source separation relationship and observation status to configure the current management object, and is not extrapolated to unobserved members.
[0082] When both retained and migrated members maintain an operational risk status, the shared operational conditions and operational constraints are preserved, and the concurrent status is continuously updated. When a reference member has not entered the corresponding risk and both retained and migrated members have exited concurrently, the target aquaculture unit and the migrated member are recorded as current constraint objects awaiting exit confirmation. Each of the two objects obtains subsequent new data to complete the confirmation, and the remaining unclear statuses remain pending judgment. This combination order is used for each object judgment: reference availability is verified first, and then single-sided retention and double-sided exit are processed to avoid the same status triggering both release and continuation judgments simultaneously.
[0083] When the state combination is written to the current constraint object record, the cycle identifier and member scope version used this time are also written. When the object type is a breeding unit, the identifier points to the unit where the current retained member is located, and the evidence member scope still points to the retained details. When the object type is a member scope, the identifier points to the current migration details. Two records to be confirmed are saved for the two exit scenarios, indicating that the recovery status of each side needs to be confirmed later, and a single source conclusion is not formed. Therefore, the same object type field will not represent both the physical breeding unit and the arbitrarily spliced current group.
[0084] When the current constraint object is not yet determined, for each new complete cycle obtained, the cycle and the immediately preceding complete cycle are combined to form a new two-cycle window, and the same combination judgment is re-executed. When the current constraint object is determined, the member range, condition execution version and cycle result used this time are saved, and the process proceeds to the subsequent exit confirmation. The same record will no longer be moved continuously to the cutoff time. Only when the actual member migrates, the original source is re-applied, or the state of the same period changes in the opposite direction, will the corresponding record be re-judged and a new judgment version be generated.
[0085] Once the current constraint object is determined, record the cutoff time of the actual data used in this judgment, and use this cutoff time to divide the data range of the current constraint object formation stage and the subsequent exit confirmation stage.
[0086] The cutoff time is determined based on the original data actually used in this judgment. For fields where the input is an instantaneous value, the original observation time is read. For fields where the input is a periodic summary, the latest original observation time participating in the summary and the record coverage end point are read. The latest of these times is taken as the cutoff time for this data, and the end time of the completed period is saved at the same time. The input results are uploaded after the cutoff time without changing the period to which the original data belongs. The next stage obtains new data from the complete period starting after the completed period.
[0087] The exit confirmation query first requires that the original observation time is later than the cutoff time, and then requires that the source record coverage of the corresponding aggregated input is within the subsequent complete period. When the rule adopts the rolling average or cross-period cumulative amount, the original record is verified along the field index until the observation period used by the new input is all after the cutoff time before it is counted as exit confirmation. The fixed judgment rules and existing reference values in the root risk evidence continue to be read, but the observation records used in this source judgment remain in the original judgment stage.
[0088] This isolation refers to the fact that observation data are not reused, but does not mean that the animal states before and after are statistically independent. When using periodic summary input, the system saves the start and end points of the coverage for the new period. The start point of the new period can be connected to the end point of the period determined by the source. The new observation time included must still be later than the data cutoff time. Left-closed and right-open intervals are not repeatedly included in historical observations at the common boundary.
[0089] Even if the creation time of an old period record uploaded late is later than the deadline, its original coverage period still points to the old period, so it cannot enter the new period for confirmation. Cross-period fields within the rule are also verified according to the original coverage period until the entire calculation support period is in the new data range before it is recorded as a complete confirmation period.
[0090] Exit confirmation only uses new data that is later than the deadline and meets the routine data collection conditions. When the operating conditions corresponding to the current constraint object are adjusted, the corresponding response transition period is determined based on the historical response records, and new complete collection cycle data is obtained after the response transition period ends. When the operating conditions remain unchanged, new data is obtained directly from the complete collection cycle after the deadline.
[0091] The scope of members saved when the current constraint object was formed is used as the scope of exit confirmation in this round. When the current constraint object is a breeding unit, the retained historical member details used for this judgment are read. When the current constraint object is a migrating member, the corresponding migrating member details are read. Data is organized according to this fixed scope in both confirmation cycles. The records of members added later are used for their own management and are not included in the denominator of this round. When some historical members migrate out during the confirmation period, the new affiliation is continued and the judgment of the affected scope is rearranged. It is not possible to remove the original abnormal members from the exit confirmation by only retaining the members who are still present.
[0092] The system saves the current member range version and data cutoff time for each current constraint object. When the member range remains unchanged, two new cycles are obtained consecutively. When the member range changes, some observations of the old range are retained as historical records. The newly determined objects use the new range version and new cutoff time. The migration relationships corresponding to multiple target aquaculture units are processed separately. The normal results of one target unit do not replace the confirmation of another target unit. This correspondence method allows the exit confirmation to be traced back to the specific member, source conditions and original observations from the final results.
[0093] When the operating conditions of the current constraint object are actually adjusted after the deadline, the execution record of the conditions involved in the adjustment is read, and the response transition period is determined according to the aforementioned historical event matching and sequential statistics method. The current data deadline, the end time of the completed source judgment cycle, and the end time of the adjusted response transition period are compared, and the latest time is taken as the selection boundary for the subsequent cycle. New input is organized from the first complete collection cycle with a start time no earlier than the boundary. When the operating risk status is exited for two consecutive complete cycles, the exit confirmation result is written.
[0094] When the operating conditions remain in their original state, the boundary of the subsequent cycle selection is the later of the data cutoff time and the end time of the source judgment cycle. Two complete cycles are obtained consecutively from the cycle whose start time is no earlier than this boundary. The exit confirmation is completed according to the same rule. This result indicates that continuous exit has been observed under the fixed object and fixed judgment rule, and the previous historical judgment of the current constraint object is maintained. It does not prove that the operating conditions have been eliminated or that a certain adjustment has produced a causal effect. If the operating conditions are adjusted again during the confirmation period, the waiting boundary is recalculated based on the latest actual completion record. The uncompleted cycle is not spliced with the adjusted cycle.
[0095] Both the end of the response transition period and the data cutoff time are data selection criteria. If the adjustment of operating conditions has been completed during the source determination period, its execution record can continue to be used to determine the objective waiting endpoint, but the observation for exit confirmation still starts from the complete cycle after this cutoff time; the later of the two boundaries shall prevail.
[0096] If adjustments are made after the deadline, the process will restart from the time the adjustments are completed, and then enter two new cycles. This process will not pre-calculate the time of unexecuted plans, nor will it generate a risk exit result directly because the management action has been completed.
[0097] The exit confirmation result saves the current constraint object identifier, member scope version, and input index of two completed cycles, and records the confirmation completion time. If the current constraint object still meets the business risk status, the business constraint of the object remains valid, and then the confirmation continues in a new cycle. When the retained member and the migrated member are both the object to be confirmed, each generates an exit confirmation result, and then the business constraint status of the same risk evidence is updated accordingly.
[0098] For another member scope that has exited the risk when the source judgment is made, the observation association of the same root risk evidence is still retained. When the same type of risk reappears in this scope during the current object confirmation period, the current relationship is re-verified according to the new contemporaneous state. The original source judgment result can no longer be used as the basis for unchanging object allocation. Before the root risk evidence is updated to be released, the system checks that there are no pending relationships or newly appeared records of the same type of risk in the relevant records still in the current management scope. If multiple current constraint objects are confirmed separately, the results are registered one by one. Release is submitted only after the last pending object is completed.
[0099] When a historical member associated with the current constraint object migrates again, the system first continues to record the executed member affiliation change, then verifies the boundary of the operating conditions before and after the migration along the same risk evidence. The new member scope obtains the corresponding reference data and response transition period respectively, and determines the subsequent current constraint object according to the same state combination. After this re-judgment is completed, a new cutoff time and member scope version are written for the relationship, and subsequent confirmation is obtained from the new boundary to obtain complete observations again.
[0100] The processing records for the second migration retain the succession relationship through the previous relationship identifier. The old relationship that has been split is only used as a historical index. The final status summary uses its subsequent valid relationship. The breeding unit that needs to be reviewed before migration still retains the original operating risk record. The migration of historical members does not constitute a confirmation of the unit's exit. If there are original operating conditions that continue to play a role, the unit will continue to retain the corresponding review requirements.
[0101] When it is confirmed that the required members cannot obtain sufficient data, the current object and the reason for confirmation are saved, and the subsequent valid records continue to take over. When the previous source judgment pointed to the original breeding unit, the new migration relationship only changes the subsequent affiliation of the corresponding historical members. The record of the original breeding unit as the business object to be reviewed is still retained separately. When the previous judgment pointed to the scope of the migrating members, the new migration event breaks the scope into new actual destinations, and subsequent confirmations are formed separately according to the new destinations.
[0102] The original member relationship record and the new object record are linked by a migration event identifier. The current status summary only takes the latest valid version of each relationship and retains the original unit review requirements that have not been completed, so as to avoid losing a still valid business constraint in the record replacement due to repeated column switching.
[0103] When only some members migrate again for the same event, the valid attribution of other members in the original record and the completed observations are still preserved. Before re-summarizing the operational constraint status, the member range referenced by each currently valid relationship is verified to be consistent with the most recent actual migration record. The confirmation result of the old object only corresponds to the old range to avoid the same old result being copied to multiple new ranges. Members that have left the current data traceable boundary are recorded as exiting the range status, and their risk history is still retained. Exit confirmation results are not automatically generated if the data is unavailable.
[0104] Once the current constraint object has completed the exit confirmation, the operational constraint status of the root risk evidence is updated based on the result. Once all current constraint objects associated with the same root risk evidence have completed the exit confirmation, the operational constraint status is updated to be lifted, and an operational risk management result is generated accordingly.
[0105] The operational constraint status of root risk evidence is set to valid and released. When root risk evidence is first formed, it is written as valid. Subsequently, the corresponding object determination status and exit confirmation result are read based on the current valid relationship. If there is a relationship to be determined, an object that is yet to be exited and confirmed, or a member scope that is being updated, it remains valid. Only when a clear current constraint object scope has been formed and each current valid relationship has an exit confirmation result that matches the scope is it written as released. When the current constraint object field is empty, the empty set is not checked.
[0106] The results of operational risk management are output in the form of records. These records locate historical events by identifying root risk evidence, locate the operational objects to be reviewed by the current constraint object type and object identifier, and save the operational constraint status and its effective time. When the business process such as column transfer and column exit reads the effective status, it enters the existing risk review process. When the release status is read, it resumes the judgment according to the original business conditions. The status output by this method only indicates whether the review requirement corresponding to the historical risk is still valid, and does not replace the judgment of other operational conditions.
[0107] The status update of root risk evidence is controlled by version number. The current constraint object, deadline, and exit confirmation result all refer to the current judgment version. Before submitting the release result, the system compares the member range version and source condition version saved in the result with the current valid version. If the versions are consistent, the exit confirmation result and the new business constraint status are written in the same transaction. When the version changes, the processing continues along the updated migration relationship. When a business object is associated with multiple root risk evidences, the latest status is read one by one. If any risk evidence is still valid, the object continues to be constrained by the corresponding review requirements. The normal and abnormal results of each root risk evidence are not offset equally.
[0108] After the status is submitted, the downstream reads the latest version of the same risk evidence. The read content includes the current constraint object and the operational constraint status. New objects are not mixed with old versions of status. If the same exit confirmation result is read repeatedly, only the submitted result is returned. Repeated tasks do not generate additional release operations. When the source record is subsequently revised after verification, the revision association is saved and the affected current judgment version is recalculated. Existing historical evidence and original results can still be traced. This version control is used for the continuous use of defined risk relationships and does not change the original judgment rules of operational risk status.
[0109] When no members migrate, this implementation method still preserves the original root risk evidence and the original business objects, and continues to monitor according to routine data collection conditions. In this case, the source separation mapping is not used to infer new objects. When some migration occurs but the reference members are unavailable, the same historical relationship and business constraints are maintained first. Processing continues after the available records are formed. If the necessary conditions are not met, the output of the record to be determined or confirmed is a clear operating result and cannot be equated with the technical proof that the risk has subsided.
[0110] The numerical values in this implementation method are determined according to different uses. Historical queries are preferably checked twice consecutively, and three checks are allowed as additional records. The final boundary is always based on the complete coverage of the actual call records. The verification of the role of operating conditions is preferably done in one complete cycle. When submitting across batches, the verification is supplemented according to the actual completed records. The reference members are preferably checked in two complete cycles before migration. The continuous judgment of risk status and exit confirmation are both done in two consecutive complete cycles. The limited sample coverage parameter only controls the sorting position of historical responses. It is preferably 0.05. When there are 9 to 18 records, the specific value is formed in the range of 0.05 to 0.1 according to the number of records. The cycle duration is taken from the actual collection configuration before the risk is triggered. No additional general hours with lack of sources are set.
[0111] The following uses hypothetical input records to illustrate the calculation order. The numerical values are only used to demonstrate recalculated data processing and do not represent actual field test results. It is assumed that the target breeding unit U01 uses rule version V3.2, the routine cycle is 2 hours, and the operational risk status is abnormal planned feeding completion.
[0112] This version of the configuration divides the actual feed intake recorded by each member within a cycle by the planned amount for that member in the same cycle to obtain the completion rate. Then, the completion rates of the members are added together and divided by the number of members in that range. The rule configuration uses 0.85 as the judgment boundary for this example. When the average completion rate is less than 0.85, the output shows that the operational risk status is met; when it reaches 0.85, the output shows that it is not met. The risk management judgment condition for this example is that the risk judgment of a single cycle is met for two consecutive cycles. Therefore, root risk evidence is formed after the second cycle is completed. The rule also reads the actual source records of feeding and water supply in the current period to verify the correspondence during the input period. When the records are complete, the completion rate judgment is performed. 0.85 is a hypothetical existing rule input and is not used as the physiological threshold uniformly recommended by this invention.
[0113] In this calculation example, the planned quantity is a positive number read from the rule configuration, and the actual feed intake is provided by the measurement record that corresponds to the member identifier. When the member's current planned quantity is 0, the input is converted to pending completion and the ratio is not calculated. 0.85 only limits the assumed program input here. The actual deployment adopts the risk rule configuration that has been implemented in the farm and its applicable scope. This implementation maintains the same judgment boundary for the same risk evidence, and demonstrates how this boundary is jointly invoked by each stage through the following specific values. The example duration and number of members are also limited to this set of calculation inputs.
[0114] Assuming that U01 triggers this operational risk status at 10:00 on June 18, 2026, the original records actually entered this time cover 06:00 to 10:00, involving 120 historical members. The sum of the completion rate of the members in each period is 96. 96 divided by 120 equals 0.8, which meets the risk conditions in this example. The system saves the risk formation period from 06:00 to 10:00 and associates it with the feed source F2308 and water source W03 actually supplied during the same period. The migration event M01 at 12:00 migrates 18 historical members to U02, and 102 historical members remain in U01. When the execution record confirms at 12:00, the effect of F2308 on the migrated members ends, and U02 is supplied by F2311 from this time. During the complete verification period from 12:00 to 14:00, F2308 continues to cover the remaining members, thus forming a separation of feed sources; W03 continues to be saved as a common water source.
[0115] Assuming there are 20 similar historical adjustment events that have been verified, the first value of their response duration in ascending order is 6.5 hours, adjacent values differ by 0.5 hours, and the last value is 16 hours. The finite sample coverage parameter is set to 0.05. Adding 1 to 20 gives 21, multiplying by 0.95 gives 19.95, rounding up to 20, and reading the 20th event gives 16 hours. The system takes the actual end time of action at 12:00 on June 18 as the starting point and obtains the end time of the response transition period at 04:00 on June 19. In this example, only the feeding conditions form a source separation, so the other separation waiting time is not calculated based on W03, which is still acting together. The reference members of U02 are the 30 original members that meet the applicable rules. Assuming that the sum of the completion rates of the two most recent cycles before the migration is 28.8, dividing by 30 gives 0.96, and neither has entered the corresponding operational risk state, these 30 members are fixed as reference members.
[0116] In the two complete cycles from 04:00 to 06:00 and from 06:00 to 08:00 on June 19th, assuming the sum of the completion rates of retained members is 81.6 and 83.64 respectively, dividing by 102 yields 0.8 and 0.82; the sum of the completion rates of migrating members is 17.1 and 17.28, dividing by 18 yields 0.95 and 0.96 respectively; and the sum of the completion rates of referral members is 29.1 and 28.8, dividing by 30 yields 0.97 and 0.96 respectively. Retention was achieved in both cycles. If a member meets the risk, a migrating member, or a reference member does not enter the risk, the system will write U01 as the current constraint object. The results of each cycle are formed based on the complete original records up to the end of the cycle. The end time of the source judgment cycle is 08:00. The data cutoff time is written as the latest overwrite endpoint of the actual usage record at 08:00. Assuming that U01 completes the feed source adjustment at 10:00 on the same day, and the compatibility history record still gives 16 hours, then the subsequent wait will be until 02:00 on June 20.
[0117] During the two new cycles from 02:00 to 04:00 and from 04:00 to 06:00, the sum of the completion rates of the 102 fixed historical members was 95.88 and 97.92, respectively, with averages of 0.94 and 0.96. Both met the non-risk conditions of the example rule. The system completed the exit confirmation at 06:00. During the verification period, the member scope and role record versions remained unchanged. Subsequently, the operational constraint status of the risk evidence was updated to "released." The source judgment cycle and exit confirmation cycle in the above calculations did not reuse the original observations, and the example did not change the confirmation denominator by adding new members. This example can be implemented with the same configuration, substituting the edges. If the original feed F2308 of the 18 migrating members continues to act after 12:00, the source separation verification returns to pending confirmation. The 16-hour waiting calculation does not directly lead to U01 as the current constraint object. If 20 historical records are reduced to 8, the number of records does not meet the same sorting implementation conditions, and the system continues to save the effective operational constraints. If the reference member enters the operational risk state in any formal comparison cycle, the object mapping continues to wait for the subsequent complete cycle. If the summary record used for exiting confirmation is generated in the new cycle but contains observations before 08:00, it remains in the original judgment stage according to the original coverage period.
[0118] Each processing step is determined by explicit fields and time relationships; the example does not interpret these branches as actual recognition performance.
[0119] After this continuous processing, the root risk evidence maintains the basis for the formation of this risk through the actual range of historical members' roles. Member migration forms a verifiable source separation relationship through the execution records of the original actual source. Subsequently, the contemporaneous status of the three types of members is calculated according to fixed rules. The observed risk differences are converted into the current constraint object. Then, new data with complete source periods after the cutoff time is used to form the exit confirmation result. The effect is limited to blocking specific error update paths such as the current member list covering historical relationships, source record missingness being mistakenly treated as termination of role, mixed calculation of different member ranges, and repeated use of source judgment data for removal. The response transition period is the experience waiting given by historical data. The current constraint object is the business review object. The output is used to determine the requirements of this business review. The subsequent business risk status is still updated based on the continuously acquired new records.
[0120] This invention also discloses an intelligent risk assessment system for livestock farming based on multi-source data fusion, comprising: The root risk evidence formation module acquires multi-source aquaculture operation data of the target aquaculture unit and determines the operation risk status. When the operation risk status first meets the risk management judgment conditions, it fixes the historical members and member affiliation relationships, the operation conditions for risk formation, and routine data collection conditions corresponding to the operation risk status to form root risk evidence. The risk source separation determination module identifies the migrating members and the remaining members when some members in the historical members migrate. When there is an operating condition that jointly affects the migrating members and the remaining members before the migration, continues to affect the remaining members after the migration and then stops affecting the migrating members, the module determines the risk source separation condition and identifies the reference members for the target aquaculture unit. The current constraint object determination module determines the response transition period based on historical response records in multi-source aquaculture operation data. After the response transition period ends, it obtains subsequent data of migrating members, retained members, and reference members according to routine data collection conditions. Based on the differences in risk status of migrating members and retained members and the concurrent status of reference members, it determines the current constraint object of root risk evidence and records the cutoff time of the data used. The operational constraint status update module only uses new data that is later than the deadline and meets the routine data collection conditions to confirm that the current constraint object has exited the operational risk status. Based on the exit confirmation result, it updates the operational constraint status of the root risk evidence and outputs the operational risk management result.
[0121] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for intelligent risk assessment in livestock farming based on multi-source data fusion, characterized in that, include: Acquire multi-source aquaculture operation data of the target aquaculture unit and determine its operational risk status; When the operational risk status first meets the risk management judgment criteria, the historical members and their affiliation, the operational conditions for the risk formation, and the routine data collection conditions corresponding to the operational risk status are fixed to form root risk evidence. When some members of the historical members migrate, the migrating members and the remaining members are determined; If an operating condition exists that affects both the migrating and remaining members before the migration, continues to affect the remaining members after the migration, and ceases to affect the migrating members, then the risk source separation condition is determined, and the reference member for the target aquaculture unit is identified. The response transition period is determined based on the historical response records in the multi-source aquaculture operation data. After the response transition period ends, subsequent data of the migrated members, retained members and reference members are obtained according to routine data collection conditions. The current constraint objects of the root risk evidence are determined based on the differences in risk status of the migrated members and retained members and the concurrent status of the reference members, and the cutoff time of the data used is recorded. Only new data that is later than the deadline and meets the routine data collection conditions is used to confirm that the current constrained object has exited the operational risk status. The operational constraint status of the root risk evidence is updated based on the exit confirmation result, and the operational risk management result is output.
2. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 1, characterized in that, The specific steps to generate root risk evidence are as follows: The period during which the risk was formed is determined based on the data actually used when the operational risk status was formed; historical members and their affiliation are determined based on the actual presence of members during the period during which the risk was formed and the actual impact of the operating conditions on the members. Establish a fixed association between the period during which the risk arises, historical members and their affiliations, operating conditions, and routine data collection conditions and the status of the operational risk. This fixed association remains unchanged after members migrate or new members join.
3. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 2, characterized in that, The operating conditions are recorded separately for the actual source identification and the effective period of effect; members who were actually in the target aquaculture unit and were affected by at least one operating condition during the risk formation period are identified as historical members; The interaction between historical members and corresponding operating conditions is fixed in the root risk evidence, and the boundary of the operating conditions before the member's migration is determined by the interaction relationship.
4. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 1, characterized in that, When historical members migrate, identify the migrating members who actually left the target aquaculture unit and the remaining members who continued to stay in the target aquaculture unit from the historical members. After migrating members enter the target aquaculture unit, they continue to be associated with the root risk evidence. The scope of historical members corresponding to the root risk evidence remains the historical members determined when the root risk evidence was formed. When a member migrates again, the member's affiliation is updated based on the same risk evidence.
5. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 4, characterized in that, The specific steps to determine the separation criteria for risk sources are as follows: Verify the combined effect of operating conditions from the same actual source on both migrating and remaining members before the migration; and verify the continued effect of operating conditions on remaining members and the termination of their effect on migrating members within a collection cycle formed according to routine data collection conditions after the migration. When a joint effect exists before migration and continues after migration, or when the effect terminates, the risk source separation condition is determined; when the operating conditions remain effective for both the migrating members and the remaining members after migration, the operating constraint status of the root risk evidence is maintained.
6. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 1, characterized in that, The specific steps for determining the response transition period are as follows: Extract adjustment records corresponding to the operating conditions that form the risk source separation conditions from the historical response records; determine the response duration from when the operating conditions are adjusted until the corresponding operating object exits the operating risk state for two consecutive collection cycles formed according to the routine data collection conditions; The upper limit of the response duration for the corresponding operating condition is determined based on the response duration, and the response transition period is determined based on the upper limit of the longest response duration among the operating conditions involved in the risk source separation condition.
7. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 6, characterized in that, To determine the current constraint object, the specific operation is as follows: After the response transition period ends, subsequent data on retained members, migrated members, and reference members are obtained according to routine data collection conditions, and the status is determined using the same judgment rules used when the operational risk status is formed. If the subsequent data of the retained member meets the judgment conditions for the operational risk status for two consecutive collection cycles, and the subsequent data of the migrated member exits the operational risk status for two consecutive collection cycles and the reference member does not enter the operational risk status during the same period, the breeding unit where the retained member is located will be determined as the current constraint object. If the subsequent data of the migrated member meets the judgment criteria for the operational risk status for two consecutive collection periods, and the subsequent data of the retained member exits the operational risk status for two consecutive collection periods, and the reference member does not enter the operational risk status during the same period, the migrated member will be identified as the current constraint object. When members enter a state of operational risk at the same time, the operational constraints based on the evidence of risk should be maintained.
8. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 1, characterized in that, To confirm that the currently constrained object has exited the operational risk status, the specific steps are as follows: When determining the current constraint object, the cutoff time of the data is used as the starting point for subsequent state confirmation, and only new data that is later than the cutoff time and meets the routine data collection conditions is used; When the operating conditions corresponding to the current constraint object are adjusted, the corresponding response transition period is determined based on the historical response records, and the new data is processed after the response transition period ends. If the operating conditions corresponding to the current constraint object remain unchanged, new data will be processed continuously after the cutoff time. When the current constrained object is confirmed to have exited the operational risk status in two consecutive data collection cycles, an exit confirmation result is generated.
9. The intelligent risk assessment method for livestock farming based on multi-source data fusion according to claim 8, characterized in that, The specific steps to update the operational constraint status of the root risk evidence are as follows: When a historical member associated with the current constraint object migrates again, the membership relationship after migration is updated based on the root risk evidence, and the judgment of the risk source separation condition and the determination of the current constraint object are re-executed; If the current constraint object has not yet been determined, maintain the operational constraint status based on the root risk evidence; If the current constraint object has not yet formed an exit confirmation result, the operational constraint status based on the root risk evidence shall be maintained; Once all currently constrained objects associated with the root risk evidence have reached an exit confirmation, the operational constraint status of the root risk evidence is lifted, and the operational risk management result is output based on the root risk evidence that is still under effective operational constraint status.
10. A livestock farming risk intelligent assessment system based on multi-source data fusion, characterized in that, include: The root risk evidence formation module acquires multi-source aquaculture operation data of the target aquaculture unit and determines the operation risk status. When the operation risk status first meets the risk management judgment conditions, it fixes the historical members and member affiliation relationships, the operation conditions for risk formation, and routine data collection conditions corresponding to the operation risk status to form root risk evidence. The risk source separation determination module identifies the migrating members and the remaining members when some members in the historical members migrate. When there is an operating condition that jointly affects the migrating members and the remaining members before the migration, continues to affect the remaining members after the migration and then stops affecting the migrating members, the module determines the risk source separation condition and identifies the reference members for the target aquaculture unit. The current constraint object determination module determines the response transition period based on historical response records in multi-source aquaculture operation data. After the response transition period ends, it obtains subsequent data of migrating members, retained members, and reference members according to routine data collection conditions. Based on the differences in risk status of migrating members and retained members and the concurrent status of reference members, it determines the current constraint object of root risk evidence and records the cutoff time of the data used. The operational constraint status update module only uses new data that is later than the deadline and meets the routine data collection conditions to confirm that the current constraint object has exited the operational risk status. Based on the exit confirmation result, it updates the operational constraint status of the root risk evidence and outputs the operational risk management result.
