A method for logistics demand forecasting and capacity allocation that integrates multi-source data

Abstract

This invention discloses a method for logistics demand forecasting and capacity allocation that integrates multi-source data, relating to the field of cross-border logistics information processing and capacity allocation technology. This invention unifies order basic data, three-document verification data, declaration and inspection processing data, release result data, warehouse operation data, port shift cut-off data, and carrier capacity surplus data into an order status object. Based on the status transition chain, historical release wave distribution, and current declaration progress, inspection progress, and release progress, it determines the estimated time bucket for entering the pending configuration state. Then, it applies an entry control measure to the outbound demand unit, including the results of the three-document verification, release results, warehouse operation results, and cargo handling requirements. It writes capacity occupancy tags according to port identification consistency, carrier mode consistency, cut-off window connection, and cargo handling requirements compatibility. This method can prevent orders that have not yet been released, have not completed warehouse operations, or do not meet the current port processing conditions from being included in the capacity allocation set.

Description

Technical Field

[0001] This invention relates to the field of cross-border logistics information processing and capacity allocation technology, and in particular to a method for logistics demand forecasting and capacity allocation that integrates multi-source data. Background Technology

[0002] In cross-border bonded warehouse-port shipping scenarios, from order generation to actual shipment, orders typically undergo the following steps: verification of transaction, payment, and logistics information; declaration, inspection, or release; and in-warehouse picking, verification, sealing, and outbound handover before being connected to airport belly hold, pallet space, or cross-border truck space. Whether an order can enter the actual shipment stage depends not only on the port release status but also on the completion of in-warehouse operations and the carrier's cut-off window.

[0003] In existing technologies, one type of solution predicts future logistics demand based on historical orders, cargo volume fluctuations, wave planning, or external factors, and then reserves or allocates capacity accordingly. Another type of solution confirms customs clearance based on the consistency verification of transaction information, payment information, and logistics information, as well as the results of declaration processing, inspection processing, or release. These methods address demand forecasting and customs clearance confirmation respectively, and can support capacity reservation and release determination to a certain extent.

[0004] However, under the condition of rolling shipments within the day, orders do not continuously and evenly form executable shipment demands. Instead, they are affected by the release results, inspection return status, warehouse completion progress, and changes in the cut-off window. If capacity occupancy is directly implemented based on static forecast results or reserved capacity, orders that have not yet been released or whose warehouse operations have not been completed may enter the capacity pool in advance, causing the cargo space, pallet space, or truckload space in the current cut-off window to be occupied prematurely. When subsequent inspections are reopened, releases are delayed, container sealing is returned, or outbound handover is interrupted, the occupied capacity is difficult to release in time, and orders that are ready for shipment cannot be replenished with the corresponding capacity in time, thus creating a situation where one side is occupied and the other side misses the cut-off window.

[0005] Furthermore, existing processing methods typically execute release confirmation, warehouse operations, and port shipment in separate segments, resulting in a lack of continuous correlation between order status changes and capacity adjustment actions. Consequently, during rolling shipments, changes in order status from release and warehouse completion to actual shipment are difficult to promptly transmit to capacity utilization results. Consistent processing records are also difficult to establish between release, replacement, and subsequent window connections, leading to a lack of unified correspondence between the status and capacity data for the same order during shipment allocation adjustments.

[0006] Therefore, in the cross-border bonded warehouse-port shipment scenario, how to identify executable shipment demands and drive capacity occupancy, release, and replacement based on the status changes of orders at the release, warehouse completion, and cut-off windows has become a technical problem that needs to be solved. Summary of the Invention

[0007] This application provides a method for logistics demand forecasting and capacity allocation that integrates multi-source data, solving the problems of vacant order capacity, delayed release, and missed loading in the daily rolling shipment of cross-border bonded warehouses.

[0008] This invention provides a method for logistics demand forecasting and capacity allocation that integrates multi-source data, comprising: Collect basic order data, three-document verification data, declaration and inspection processing data, release result data, warehouse operation data, port schedule cut-off data, and carrier capacity surplus data; The data is associated with order objects to form order status objects and a state transition chain is constructed. The status change events and status retention time of each order status object are statistically analyzed according to the preset time bucket. Combined with the historical release wave distribution, current declaration progress, inspection progress and release progress, the estimated time bucket for each order status object to enter the pending configuration status is determined. Order status objects with the same port identification, carrier method, cut-off window and cargo handling requirements are aggregated into shipping demand units. Based on port schedule cut-off data and carrier capacity surplus data, a capacity supply unit is generated; Based on the verification results of the three documents, the release result data, the warehouse operation data, and the cargo handling requirements, the outbound demand unit is subject to access control. The outbound demand unit that passes the access control is written into the capacity configuration set, and matched with the capacity supply unit according to the consistency of port identification, the consistency of the mode of transport, the connection of the cut-off window, and the compatibility of cargo handling requirements. The capacity occupancy mark is written into the corresponding capacity supply unit. During the rolling shipping cycle, when an order status object with a capacity occupancy mark receives at least one of the following events: release rollback, inspection reopening, box sealing rollback, outbound handover interruption, or load cut-off failure, the corresponding capacity occupancy mark is released, and other shipping demand units that meet the access control are written into the capacity supply unit corresponding to the released capacity occupancy mark. By associating records of status changes, gate control criteria, and capacity adjustment results, a shipping processing chain corresponding to the order status object is formed.

[0009] In some embodiments, the order object is a single order or a batch of orders with the same port identification, the same mode of transport and the same cut-off window; The basic order data includes order identifier, product identifier, country or region of destination, port identifier, mode of transport, and shipping time limit; The three-item verification data includes the consistency verification results of the transaction order, payment order, and logistics order.

[0010] In some embodiments, the order status object in the status transition chain includes at least the pending declaration status, declaration processing status, inspection processing status, released but not completed status, pending configuration status, configured and ready for shipment status, and shipped status, and records the status change events, status change times, and event source identifiers between each status.

[0011] In some embodiments, the access control includes the following joint determination criteria: The three-item verification result is passed; the release result data corresponding to the order status object indicates that the goods have been released and there is no unresolved inspection return status; the warehouse operation data corresponding to the order status object indicates that the sealing of the box or the outbound handover has been completed. The cargo processing requirements corresponding to the order status object are compatible with the port requirements and the mode of transport; the shipment demand unit enters the capacity configuration set only when all joint judgment conditions are met simultaneously.

[0012] In some embodiments, the capacity supply unit includes a carrier identifier, a port identifier, a schedule identifier or vehicle number identifier, a cut-off window, an allocable capacity, cargo receiving boundary information, and a capacity occupancy marker; The capacity supply generation unit includes splitting the remaining capacity data according to port, mode of transport, schedule or train number and cut-off window, and associating the splitting results with the corresponding port schedule cut-off data.

[0013] In some embodiments, the matching includes: Candidate transport capacity supply units that are compatible with the shipping demand units in terms of port identification, mode of transport, cut-off window and cargo handling requirements; The candidate transport capacity supply units are sorted according to the release time, in-warehouse completion time, and remaining time before the cut-off point. Write the candidate capacity supply units that are ranked first and whose allocable capacity meets the conditions into the corresponding capacity occupancy flag.

[0014] In some embodiments, when the order status object receives at least one of the following events: release rollback event, inspection reopening event, box sealing rollback event, outbound handover interruption event, and load cut-off failure event, the corresponding capacity occupancy mark is removed based on the event timestamp, and the reason for removal, the time of removal, and the corresponding capacity supply unit identifier before removal are recorded.

[0015] In some embodiments, during the allocation of substitutes, the capacity supply unit corresponding to the released capacity occupancy mark is used as an index to select candidate units that are compatible with the capacity supply unit in terms of port identification, mode of transport, cargo handling requirements and cut-off window connection from the outbound demand units that currently meet the access control and have not yet been written with the capacity occupancy mark. After re-sorting according to release time, warehouse completion time and remaining cut-off time, they are written with the capacity occupancy mark. For order status objects that have not completed the replacement allocation within the current cut-off window, retain the original status change record and gating record, and write the new cut-off window, the corresponding candidate capacity supply unit set, and the rematching time into the order status object.

[0016] In some embodiments, the outbound processing chain is at least associated with the data source identifier, status change record, access control result, capacity occupancy write record, release record, substitute allocation record and final processing result of the stored order status object, and updates the order status object ownership result of each time bucket in the subsequent rolling outbound cycle according to the outbound processing chain.

[0017] In some embodiments, when there are timestamp conflicts, missing fields, or inconsistent sources in the release result data, warehouse operation data, or port shift interception data corresponding to the same order status object, the order status object is marked as an abnormal pending verification status and prevented from entering the capacity configuration set. After receiving the supplemented or corrected business data, the state transition chain update, access control, and capacity matching are re-executed.

[0018] Through the above technical solution, the present invention can achieve at least the following beneficial effects: This invention unifies order basic data, three-document verification data, declaration and inspection processing data, release result data, warehouse operation data, port shift cut-off data, and carrier capacity surplus data into an order status object. Based on the status transition chain, historical release wave distribution, and current declaration progress, inspection progress, and release progress, it determines the estimated time bucket for entering the pending configuration state. Then, it performs access control on the shipping demand unit, including the three-document verification results, release results, warehouse operation results, and cargo handling requirements. It writes the capacity occupancy mark according to the consistency of port identification, the consistency of carrier method, the connection of cut-off window, and the compatibility of cargo handling requirements. This can block orders that have not been released, have not completed warehouse operations, or do not meet the current port processing conditions from the capacity configuration set, so that the objects entering the capacity pool are transformed from static predicted demand to executable shipping demand. In this way, the current capacity occupancy within the cut-off window is synchronously constrained by the actual order release status, warehouse completion status, and shipping window conditions, reducing the empty occupancy caused by orders that have not yet met the shipping conditions occupying belly holds, pallet spaces, or truckload spaces in advance, and shortening the waiting chain between warehouse waiting to be dispatched and port waiting to be loaded.

[0019] By statistically analyzing status change events and status dwell times using time buckets, and combining this with historical release wave distribution, current application progress, inspection progress, and release progress to determine the expected time bucket, discrete business events in different states such as application processing, inspection processing, and released but not yet completed can be transformed into comparable time-assigned results. This allows order objects to no longer be entered into capacity allocation based on aggregated demand or fixed reservation values, but rather into the corresponding cut-off window based on the expected time of entry into the pending allocation state. In this way, when the pace of inspection processing, release, or warehouse operations fluctuates, the time assignment of order objects can be adjusted accordingly, reducing frequent reassignment caused by prematurely pushing orders that should have been postponed to subsequent windows into the current window.

[0020] By triggering the release of capacity occupancy markers for events such as release rollback, inspection reopening, container sealing rollback, outbound handover interruption, and cut-off failure within the rolling shipment cycle, and writing other shipment demand units that meet access control requirements into the capacity supply unit corresponding to the released capacity occupancy marker, it is possible to promptly release previously invalidated occupancy and replenish compatible capacity for subsequent orders that meet the conditions, thereby reducing the concurrent probability of one side being idle while the other side misses the cut-off. Furthermore, by combining the shipment handling chain with the correlation recording of status changes, gating criteria, capacity occupancy writing, release, and replacement allocation, a consistent traceability relationship can be established between release criteria, warehouse operation criteria, and capacity adjustment actions, maintaining a closed loop for the handling of the same order object during abnormal rollbacks, manual reassignment, and subsequent time bucket updates. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0022] Figure 1 This is a flowchart of the logistics demand forecasting and capacity allocation method that integrates multi-source data in the embodiment. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0024] All terms used in this application (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0025] To facilitate understanding, the relevant terms and concepts involved in the embodiments of this application will be introduced below: An order object is a processing entity formed by a single order or by multiple orders with the same port identification, the same mode of transport, the same cut-off window, and the same cargo handling requirements. An order status object is a data record formed by associating an order object with order basic data, three-document verification data, declaration and inspection processing data, release result data, warehouse operation data, port shift cut-off data, and carrier capacity surplus data.

[0026] A state transition chain is a data link that records the sequence, timing, and event source of order status objects transitioning between different business states. A time bucket is a continuous time interval divided according to a fixed time granularity. A shipping demand unit is a demand record formed by aggregating order status objects with the same port identifier, mode of transport, cut-off window, and cargo handling requirements. A capacity supply unit is an allocable capacity record split according to port identifier, mode of transport, shift identifier or train number identifier, and cut-off window.

[0027] Access control is a processing rule that determines whether a shipping demand unit should be included in the capacity allocation set based on the results of three verifications (verification of documents, release results, warehouse operation results, and cargo handling requirements). The cut-off window is the deadline time interval for receiving goods corresponding to a port shift or train number. Capacity occupancy markers are occupancy records written to the capacity supply unit and associated with the order status object. The shipping disposal chain is a closed-loop record surrounding the order status object, recording status changes, gate control results, capacity adjustment results, and final disposal results. The abnormal pending status is an intermediate state placed when the data corresponding to the order status object has timestamp conflicts, missing fields, or inconsistent sources.

[0028] Furthermore, the historical release wave distribution is a time distribution record obtained by aggregating historical release completion events according to port identification, mode of transport, and cargo handling requirements. It includes at least the statistical period, time bucket identifier, sample size, release quantity statistics, inspection trigger ratio, and rollback event ratio. The rolling shipment cycle is a continuous processing cycle for multi-source business data, involving receiving, associating, updating status, generating requirements, determining gating, matching capacity, and recording disposal. The engineering caliber can be 15-60 minutes, remaining fixed within the same natural day for the same port, but can be recalibrated when switching between days. Status dwell time is the duration of an order status object from entering a certain status to leaving that status, recorded as the difference in event timestamps, with units of minutes. Cargo handling requirements are processing constraint records formed by mapping the cargo attributes corresponding to the order object, used to characterize the port handling restrictions and transport restrictions corresponding to general cargo, temperature-controlled cargo, sensitive cargo, or cargo requiring verification. The capacity configuration set is a collection of outbound demand units after gate access control. Each record in the capacity configuration set includes at least the outbound demand unit identifier, port identifier, mode of transport, cut-off window, demand capacity, cargo handling requirements, entry time into the set, and current gate control result. To ensure that records from different sources can be written in an aligned manner, an order status table, a wave distribution table, a capacity supply table, and a disposal chain record table can be established. The order status table, wave distribution table, capacity supply table, and disposal chain record table are only used for data recording, indexing, alignment, and rollback, and do not constitute new processing flows. The order status table can use the order identifier as the primary index key, and in the case of order consolidation batches, a consolidation batch identifier can be added as a parallel index key. The wave distribution table uses the port identifier, mode of transport, cargo handling requirements, and time bucket identifier as composite index keys. The capacity supply table uses the capacity supply unit identifier as the primary index key. The disposal chain record table uses the order status object identifier and event sequence number as composite index keys. Example 1:

[0029] like Figure 1 As shown, this embodiment presents the order-level shipping demand forecasting and capacity allocation process in a cross-border bonded warehouse-port shipping scenario. This process correlates multi-source business data to form order status objects and constructs a state transition chain. It then generates shipping demand units based on the estimated time bucket, and through access control, capacity matching, capacity occupancy updates, and shipping disposal chain recording, obtains the shipping configuration result corresponding to the order status object. Specifically, it includes: Step S1: Collect basic order data, three-document verification data, declaration and inspection processing data, release result data, warehouse operation data, port shift cut-off data, and carrier capacity surplus data; Step S2: Use order objects as the granularity to associate data to form order status objects and construct state transition chains; Step S3: Statistically analyze the status change events and status retention time of each order status object according to the preset time bucket. Combine the historical release wave distribution, current declaration progress, inspection progress and release progress to determine the estimated time bucket for each order status object to enter the pending configuration state. Then, aggregate order status objects with the same port identification, carrier method, cut-off window and cargo handling requirements into a shipping demand unit. Step S4: Generate a capacity supply unit based on port schedule cut-off data and carrier capacity surplus data; Step S5: Based on the verification results of the three documents, the release result data, the warehouse operation data and the cargo handling requirements, the outbound demand unit is subject to access control. The outbound demand unit that passes the access control is written into the capacity configuration set, and it is matched with the capacity supply unit according to the consistency of port identification, the consistency of the mode of transport, the connection of the cut-off window and the cargo handling requirements. The capacity occupancy mark is written into the corresponding capacity supply unit. Step S6: Receive updated data within the rolling shipping cycle. When an order status object with a capacity occupancy mark receives at least one of the following events: release rollback, inspection reopening, box sealing rollback, outbound handover interruption, or load cut-off failure, release the corresponding capacity occupancy mark and write other shipping demand units that meet the access control requirements into the capacity supply unit corresponding to the released capacity occupancy mark. Step S7: Link the status change record, the gate control basis, and the capacity adjustment results to form the outbound processing chain corresponding to the order status object; In this embodiment, the order status object includes at least the following: order identifier set, port identifier, mode of transport, destination country or region, cargo handling requirements, three-document verification results, declaration submission time, declaration receipt time, inspection status, release time, picking completion time, review completion time, sealing completion time, outbound handover time, current status, current status start time, and event source identifier. Multi-source business data is associated and written using order identifier, port identifier, mode of transport, and business timestamp. When multiple source records correspond to the same business field, the valid value is determined in the order of customs receipt, warehouse operation receipt, transport receipt, and original order record. For records with the same timestamp but inconsistent field values, the original values ​​from each source are retained, the conflicting field name and conflict source identifier are written, and the order status object is set to an abnormal pending verification state. Data reception, status update, demand generation, capacity matching, and processing records are repeatedly executed within the rolling shipment cycle, ensuring that the order status object is continuously updated as business events arrive.

[0030] Specifically, to ensure stable alignment of multi-source business data, event reception time, business activation time, and source priority can be uniformly written to records from each source. Event reception time is the time a record arrives at the current processing cycle; business activation time is the time the business occurred and was written back by the source system; source priority is used to determine the valid value when multiple records exist in the same business field. The business timestamp prioritizes the business activation time; if missing, the event reception time is used, and a time source marker is written to the order status object. When multiple status progress records are received for the same order status object within the same refresh window, they are first sorted by business activation time, then by source priority, with the record ranked first used as the valid value for the current cycle; records ranked later are retained in the original field log for subsequent verification and rollback. The refresh window is a time window used for batch merging arriving records within the rolling shipping cycle; the engineering standard can be 1-5 minutes, with a default of 2 minutes, to reduce jitter caused by frequent rewriting of the same order status object. If no new records are received for the same business field within two consecutive refresh windows, the most recent valid value will be used, and a field freeze flag will be added. The field freeze flag indicates that a business field will continue to use historical valid values ​​within the current period. If the continuous freeze period exceeds 30-120 minutes, the corresponding order status object will be set to an abnormal pending status, and it will be prevented from entering the capacity configuration set.

[0031] In this embodiment, time buckets are continuously divided according to a uniform granularity, and a candidate mapping relationship is established with the cut-off window of port shifts or trains. When the same cut-off window spans multiple time buckets, the time bucket where the cut-off time is located is recorded as the main cut-off reference time bucket, and the preceding time bucket that is before the main cut-off reference time bucket and is still within the same cut-off window is recorded as a candidate time bucket that can be entered in advance. The main cut-off reference time bucket and the preceding candidate time bucket are only used to construct the candidate time bucket set and are not directly used as the final expected time bucket assignment result of the order status object. The historical release wave distribution is formed by the statistics of historical release completion events under the same port identification, the same carrier method, and the same cargo handling requirements, and includes at least the wave date, release completion time interval, release quantity statistics, inspection trigger ratio, and rollback event ratio. The process of determining the estimated time bucket is based on calculating the remaining processing links up to the pending configuration state according to the current state of the order status object: For the declaration processing state, the combined waiting time for declaration completion, the inspection processing time under the triggered inspection conditions, the waiting time for release completion, and the warehouse completion time after release; for the inspection processing state, the combined waiting time from inspection completion to release completion and the warehouse completion time after release; for the released but not completed state, the combined waiting time for sealing completion or outbound handover completion; for other states, the estimated arrival time is calculated according to the order of remaining events between the current state and the pending configuration state. The historical release wave distribution is updated once after the end of each rolling shipment cycle; among them, orders with failed loading cutoff, abnormal pending verification orders, and manually reassigned orders are recorded as independent sample categories and retained as independent statistical labels, and are not included in the updated sample of historical release wave weights.

[0032] For example, the statistical period used for the historical release wave distribution can be determined based on closed-loop samples of the most recent 3 to 14 calendar days, with a default of taking the most recent 7 calendar days. When holiday markers, temporary inspection upgrade markers, or concentrated delay markers coincide with the current date, samples of the same date are preferred. Closed-loop samples are order status object samples that simultaneously possess the actual release completion time, actual warehouse completion time, and actual shipment completion time. To reduce the impact of extreme abnormal events on the current estimated time bucket, the status retention time of samples in the same stratum can be truncated. The truncation ratio can be 5% to 15%, with a default of 10%, and the median, 70th percentile, or 80th percentile is preferred as the statistical value. When the number of samples in the same stratum is less than 5 to 20, sampling can be performed step-by-step back according to port marker, mode of transport, and cargo handling requirements. The default back-sampling order is to first maintain consistency between port marker and mode of transport, and then maintain consistency between port markers. The statistical unit for status dwell time can be minutes. Discrete write-backs of less than 1 minute can be combined and recorded as 1 minute to avoid frequent swings in wave distribution and expected time buckets due to excessively fine granularity. If the number of samples in the same time bucket is 0, the historical release wave distribution results of the corresponding time bucket of the previous rolling outbound cycle are retained and written into the version reuse mark.

[0033] In this embodiment, the order status object in the status transition chain includes at least the pending declaration status, declaration processing status, inspection processing status, released but not completed status, pending configuration status, configured and ready for shipment status, and shipped status, and records the status change events, status change time, and event source identifier between each status. The pending declaration status corresponds to an order object that has been created but has not yet formed a valid declaration submission record; the declaration processing status corresponds to a declaration submission record that has been formed but the final declaration processing result has not yet been received; the inspection processing status corresponds to a status that has received an inspection trigger record but has not yet received an inspection completion record; the released but not completed status corresponds to a status that has received a release result record but the sealing completion record and outbound handover record do not meet the completion criteria; the pending configuration status corresponds to a status that the order object is expected to complete release and warehouse completion within the corresponding time bucket and enter the access control judgment; the configured pending shipment status corresponds to a status that has written a capacity occupancy mark but has not yet formed a shipment completion record; the shipped status corresponds to a status that has formed a shipment completion record that matches the capacity supply unit. Status change events include at least declaration submission, declaration receipt, inspection trigger, inspection completion, release completion, picking completion, review completion, sealing completion, outbound handover completion, capacity occupancy writing, capacity occupancy release, and shipment completion. For rollback events, the previous state is written back according to the event timestamp, and the fields of the state before rollback, the state after rollback, and the rollback reason are retained in the state transition chain.

[0034] In this embodiment, the access control includes the following joint determination conditions: The three-item verification result is passed; the release result data corresponding to the order status object indicates that the goods have been released and there is no unresolved inspection return status; the warehouse operation data corresponding to the order status object indicates that the sealing of the box or the outbound handover has been completed. The cargo handling requirements corresponding to the order status object are compatible with port requirements and transportation methods; the shipping demand unit enters the capacity allocation set only when all joint judgment conditions are met simultaneously. The entry control criteria must include at least the following input fields: three-document verification result field, release result field, inspection return flag field, sealing completion field, outbound handover field, port restriction field, carrier restriction field, and cargo handling requirement field. The three-document verification result field indicates the consistency verification status of the transaction order, payment order, and logistics order; the release result field indicates whether customs release is complete; the inspection return flag field indicates whether there are any pending inspection return events; the sealing completion field and the outbound handover field indicate the warehouse completion status; and the port restriction field and the carrier restriction field indicate the restrictions on cargo handling requirements imposed by the corresponding port and carrier method. The entry control criteria are executed in the following order: three-document verification, release result, inspection return, warehouse completion, and restriction compatibility. If a previous judgment fails, the name of the failed field, the failure time, and the reason for failure are recorded, and subsequent judgments are stopped. When supplementary or corrected data arrives, the judgment for the failed field is re-executed, and the entry control results are updated.

[0035] It is understandable that all joint judgment conditions for access control are written to the gate control record using field-level Boolean results, forming the current gate control result. The gate control record is a data entry that records the access control judgment process for the same order status object within a certain rolling shipping cycle. It includes at least the order status object identifier, gate control time, gate control version, judgment results for each field, failure field name, failure reason, and current gate control result. The gate control version is the version identifier of the access control caliber within the current rolling shipping cycle, used to maintain consistency within the same cycle when rule fields, port restriction fields, or carrier restriction fields are updated. For order status objects that have completed sealing but not yet completed outbound handover, if the current port's receiving caliber allows sealing to occupy transportation capacity first, then sealing completion can be used as the warehouse completion criterion; for ports that require outbound handover before receiving, outbound handover completion is used as the warehouse completion criterion. The warehouse completion time is the business time corresponding to when the order status object meets the current port warehouse completion criterion. When sealing completion is used as the warehouse completion criterion, the warehouse completion time is taken as the sealing completion time; when outbound handover completion is used as the warehouse completion criterion, the warehouse completion time is taken as the outbound handover completion time. If any key field on which the gating depends is frozen in the current period and the freeze duration exceeds one cutoff window width, the current gating result will remain unpassed until a new valid business response is received.

[0036] In this embodiment, the capacity supply unit includes a carrier identifier, a port identifier, a schedule identifier or vehicle number identifier, a cut-off window, an allocable load capacity, cargo receiving boundary information, and a capacity occupancy marker. The generation of transport capacity supply units includes splitting the remaining transport capacity data according to port, mode of transport, schedule or vehicle number and cut-off window, and associating the splitting results with the corresponding port schedule cut-off data. A capacity supply unit includes at least a capacity supply unit identifier, carrier identifier, port identifier, shift identifier or train identifier, cut-off window, cargo acceptance start time, cargo acceptance end time, allocable capacity, occupied capacity, remaining capacity, cargo handling requirements, and capacity occupancy marker. Allocable capacity includes at least one or more of weight capacity, volumetric capacity, and piece capacity. When multiple capacity constraints exist for the same capacity supply unit, the capacity constraint that first reaches its limit is considered the valid constraint for the current capacity supply unit. For situations where the same shift identifier or train identifier corresponds to multiple cargo acceptance boundaries, multiple capacity supply units are formed, and they are recorded and sorted according to the cargo acceptance end time.

[0037] Specifically, demand capacity refers to the capacity demand record corresponding to the shipment demand unit occupying the transportation capacity supply unit. It may include at least one or more of the following: demand weight, demand volume, and demand number of pieces. The source of this demand capacity is primarily the latest valid values ​​from the order base data and warehouse operation data. If the same shipment demand unit has multiple capacity constraints, the capacity satisfaction condition means that the corresponding transportation capacity supply unit is not less than the demand capacity in each constraint dimension. When any dimension is insufficient, the transportation capacity supply unit will not be included in the candidate transportation capacity supply unit set. Receiving boundary information includes the receiving start time, receiving end time, and allowed receiving port operation boundary records, used to determine whether the shipment demand unit can still complete the handover within the corresponding cut-off window. Remaining capacity, occupied capacity, and allocable capacity are all recorded using the same measurement caliber. When weight, volume, and number of pieces coexist, the dimension most likely to reach the upper limit in historical occupancy can be used as the valid constraint dimension, and the remaining dimensions can be retained as parallel verification conditions. To avoid the problem of scattered loads being difficult to utilize due to multiple small-batch replacements, situations where the remaining load after allocation is less than 5% to 15% of the current capacity supply unit's allocable load can be recorded as low remaining load status, and its priority will be reduced in subsequent sorting.

[0038] In this embodiment, matching includes: Candidate transport capacity supply units that are compatible with the shipping demand units in terms of port identification, mode of transport, cut-off window and cargo handling requirements; The candidate transport capacity supply units are sorted according to the release time, in-warehouse completion time, and remaining time before the cut-off point. Write the candidate capacity supply units that are ranked first and whose allocable capacity meets the conditions into the corresponding capacity occupancy flag. The candidate capacity supply unit set is a set of capacity supply units that meet the following conditions: consistent port identification, consistent mode of transport, connectable cut-off windows, and compatible cargo handling requirements. Each record in the candidate capacity supply unit set includes at least the capacity supply unit identifier, remaining cut-off time, remaining capacity, receiving cut-off time, and current occupancy status. The allocation order is sorted according to the release time, the warehouse completion time, and the remaining cut-off time; when the sorting results are the same and there are multiple candidate capacity supply units, the capacity supply unit with the earlier receiving cut-off time is selected first; when the receiving cut-off time is still the same and the available capacity is not synchronized, the capacity supply unit with the remaining capacity closest to the occupancy demand of the outbound demand unit is selected first. The capacity occupancy marker includes at least the order status object identifier, capacity supply unit identifier, write time, occupied capacity, and current occupancy status.

[0039] Furthermore, the current occupancy status includes at least written, released, replaced, and expired. The write time prioritizes the business time when capacity matching is completed and the data is stored; if missing, the processing time of the current rolling shipping cycle is used. To ensure that the same order status object is not repeatedly allocated within the same cutoff window, a unique occupancy constraint can be established between the order status object identifier and the cutoff window. When the same order status object already has a capacity occupancy mark in the written state within the current cutoff window, subsequent capacity matching only allows updating the occupancy status corresponding to that mark, and no second written state record is added. When the same capacity supply unit receives write requests from multiple shipping demand units within the same refresh window, occupancy is first performed according to the sorting result, and then the capacity supply unit is updated according to the remaining capacity. If the remaining capacity after the update is insufficient to meet subsequent requests, the subsequent requests are written to the unallocated reason field. The unallocated reason field records the direct reason why the candidate capacity supply unit has not completed the write operation, and can include at least one of insufficient capacity, receiving boundary mismatch, gating failure, or occupancy conflict.

[0040] In this embodiment, the outbound disposal chain is associated with at least the data source identifier, status change record, access control result, capacity occupancy write record, release record, substitute allocation record and final disposal result of the stored order status object, and updates the order status object ownership result of each time bucket in the subsequent rolling outbound cycle according to the outbound disposal chain; The outbound processing chain uses the order status object identifier as the primary key for association and includes at least the event sequence number, event type, event time, event source identifier, pre-change status, post-change status, access control result, outbound demand unit identifier, capacity supply unit identifier, capacity occupancy marker, release reason, replacement reason, and final processing result. The final processing result includes at least completed outbound shipment, missed loading, manual reassignment, abnormal termination, and pending subsequent window processing. The order status object attribution result for each time bucket is updated based on the actual release completion time, actual warehouse completion time, actual time of writing capacity occupancy marker, and actual outbound completion time in the outbound processing chain; for order status objects that are rolled back, released, or replaced, both the original time bucket attribution record and the adjusted time bucket attribution record are updated simultaneously.

[0041] Specifically, the outbound processing chain can be further written with the pre-time bucket assignment value, post-time bucket assignment value, statistical version identifier, gating version, original capacity supply unit identifier, and new capacity supply unit identifier to support replay verification and version alignment. The pre-time bucket assignment value is the expected time bucket record of the order status object before this update; the post-time bucket assignment value is the expected time bucket record of the order status object after this update. For order status objects that have undergone manual reassignment, the expected time bucket, gating result, and candidate capacity supply unit set before manual reassignment are retained, and the capacity supply unit identifier after manual reassignment is written separately into the manual reassignment record field to ensure that manual adjustment and automatic processing chains coexist and are traceable. If there are two or more consecutive releases followed by rewriting in the outbound processing chain, the order status object is marked as a high-frequency adjustment sample at the end of the current rolling outbound cycle; high-frequency adjustment samples are sample categories used for independent observation during subsequent statistical version updates, and they can be retained separately without being included in the update samples of regular wave distribution and regular substitute weights.

[0042] In an optional implementation of Example 1, when there are timestamp conflicts, missing fields, or inconsistent sources in the release result data, warehouse operation data, or port shift interception data corresponding to the same order status object, the order status object is marked as an abnormal pending status and prevented from entering the capacity configuration set. After receiving the supplemented or corrected business data, the state transition chain update, access control, and capacity matching are re-executed; In this embodiment, the exception record corresponding to the exception pending status includes at least the exception field name, exception field value, source identifier, timestamp, conflict type, and processing result. Conflict types include at least timestamp conflict, field missing, field value inconsistency, and source mismatch. Order status objects in the exception pending status retain the last valid status, the last valid time, and any incomplete subsequent processing links. After receiving supplemented or corrected business data, the exception fields are first checked for consistency, and then the subsequent status transition chain is reconnected based on the last valid status. Gate control failure records, capacity release records, and load cut-off failure records generated during the exception pending status are all retained in the outbound handling chain and stored in association with the recovered processing result.

[0043] For example, timestamp conflicts occur when multiple source records corresponding to the same business field give contradictory business effective times within adjacent refresh windows; missing fields occur when the key field required for current processing is null or not written back; inconsistent field values ​​occur when the same business field corresponds to different values ​​in multiple source records; and source mismatch occurs when the field mapping relationship between the business field and the source system does not conform to the preset criteria. In addition to the abnormal field name, abnormal field value, source identifier, timestamp, conflict type, and processing result, abnormal records can also include the most recent valid value, the most recent valid time, and the duration of the freeze. If the abnormal field is not a field required for current access control, estimated time bucket calculation, or capacity matching, the order status object is allowed to continue using the most recent valid value to participate in the current rolling shipment cycle processing, and a downgrade processing mark is written to the shipment disposal chain; the downgrade processing mark is used to characterize the record generated based on the most recent valid value rather than the latest written back value. If the abnormal field is a key field, the abnormal pending status remains unchanged, and only disposal records are allowed to be written; the state transition chain, access control, and capacity occupancy writes are not allowed.

[0044] In a preferred embodiment of Example 1, in order to determine the estimated time bucket for each order status object to enter the configuration state, the baseline dwell time for each status object is maintained according to the port identifier, carrier method and cargo type layer identifier, respectively, for the order status objects in the pending declaration state, declaration processing state, inspection processing state and released but not completed state.

[0045] Cargo type stratification is mapped from cargo handling requirements and is used to distinguish between general cargo, temperature-controlled cargo, sensitive cargo, or cargo requiring verification. The baseline dwell time for each status is updated using closed-loop samples within the rolling shipping cycle, with extreme abnormal dwell times truncated to avoid individual abnormal documents lengthening the overall estimate. The update method is as follows: , in, For the first Status within each rolling shipping cycle The port of entry is marked as The mode of transport is Cargo type stratification identification is as follows Baseline dwell time under the given conditions; For state The corresponding smoothing coefficient; The baseline dwell time for the previous rolling shipping cycle; This refers to the statistical value of the observation dwell time within the current rolling statistical window. The statistical value of the observation dwell time is preferentially taken from the truncated median or 70th percentile value of the same stratum samples from the past three to seven days; The value can be calibrated from 0.65 to 0.85. When the recent sample size is lower than the preset lower limit, the result of the previous period remains unchanged, and the value is retrieved step by step back from the port identifier, the mode of transport, the cargo type identifier to the port identifier, the mode of transport, and then back to the port identifier.

[0046] For a single order status object, the remaining status set consists of the processing steps that have not yet been completed between the current status and the pending configuration status. The current application progress, inspection progress, and release progress are converted into the completion percentage of the corresponding steps. For steps that have started but not yet completed, the baseline dwell time of that step is reduced by the remaining percentage; for steps that have not yet started, they are counted in full; for steps that have been completed, they are not counted again.

[0047] The estimated remaining time before entering the configuration state is calculated using the following formula: , in, For order status objects The estimated remaining time from the current moment until entering the configuration state; For order status objects The corresponding set of remaining states; For order status objects In state The completion rate; For the current rolling shipping cycle and the order status object The port identification, mode of transport, and cargo type classification are consistent. The baseline dwell time; , , Retrieve order status objects respectively Corresponding port identification, mode of transport, and cargo type classification; To verify the rollback correction duration; This is the time allotted for manual review and correction. Adjust the duration for port congestion.

[0048] The completion ratio can be determined as follows: In the "Application Processing" state, it can be calculated as the ratio of the number of fields that have been acknowledged to the number of fields that should be acknowledged; in the "Inspection Processing" state, it can be calculated as the ratio of the number of completed inspection nodes to the total number of inspection nodes; and in the processing stage where the release result has not yet been written back, it can be calculated as the ratio of the number of release acknowledgments returned to the number of release acknowledgments that should be returned. The completion ratio is limited to between 0 and 1 and is updated based on the latest valid business write-back result. When there are timestamp conflicts, missing fields, node withdrawals, or inconsistent sources, the completion ratio for the corresponding state will use the most recent valid value, and the estimated time bucket will not be advanced according to the abnormal write-back result. When the same order status object has the above-mentioned abnormalities in two consecutive refresh windows, the order status object will only retain the original estimated time bucket in the current cutoff window and will not be rolled back to an earlier time bucket. The inspection rollback correction time is the median of the historical additional time for the same layer when there are unresolved inspection reopening, supplementary data, or secondary control markers. The manual review correction time is the 80th percentile of the additional time for the same port and declaration type when a manual review marker is hit. When the inspection rollback correction time and the manual review correction time correspond to the same event source marker and their effective time intervals overlap, the larger one is used and not accumulated repeatedly. The port congestion correction time is counted separately as a port public disturbance item. The port congestion correction time is determined based on holiday markers, port queue length, and the degree of compression before the cut-off time. All three corrections are non-negative values. A single correction cannot exceed three times the width of the current time bucket, and the total correction cannot exceed the remaining time of the current valid cut-off window. If it exceeds this, the current cut-off window is determined to be unreachable and extended to the next valid cut-off window.

[0049] Specifically, the port queue length can be calculated from the port receiving queue record, handover queue record, or the corresponding business write-back time difference, and the unit can be minutes or batch number; the pre-cut-off compression degree is the compression ratio of the remaining time of the current cut-off window relative to the standard cut-off window width, and the standard cut-off window width can be determined based on the median value of historical samples of the same port and the same carrier. The default value of the port congestion correction time can be set to 0~180 minutes, preferably 0~120 minutes; when holiday indicators exist, the port queue length exceeds the historical median value, and the pre-cut-off compression degree increases, the port congestion correction time will be set to the higher level within its adjustable range. To prevent the estimated time bucket within the current cutoff window from repeatedly swinging back and forth, a window shift freeze threshold can be set. The window shift freeze threshold is an engineering parameter used to limit the reverse rollback of the estimated time bucket within two adjacent refresh windows. Its default value can be 1 time bucket width. When the change of the new estimated arrival time relative to the previous refresh window is less than the window shift freeze threshold, the estimated time bucket of the previous refresh window remains unchanged, and only the sorting position in the candidate time bucket set is updated.

[0050] Furthermore, the current declaration progress, inspection progress, and release progress are all recorded using a normalized caliber of 0 to 1, retaining 2 to 4 decimal places by default. In case of missing values, the most recent valid value is used and a progress freeze flag is added. The number of fields requiring a receipt, the total number of inspection nodes, and the number of nodes requiring a release receipt can be pre-maintained based on the current port declaration template, inspection template, and release write-back template, and updated synchronously with changes in the regulatory template. If the template version changes within the current rolling shipment cycle, both the old and new templates are retained, and the template version bound when the order status object first enters the corresponding processing stage is used as the basis for this progress calculation. To avoid extreme values ​​in the completion ratio causing the estimated time bucket to move forward too early, when the completion ratio is greater than 0 and less than 1, a minimum remaining processing ratio of 5% to 15% can be set, with a default of 10%. When the completion ratio is 1 but no valid progress record for the next status has been received, a remaining processing ratio of 5% to 10% is still retained, with a default of 5%, until the next status record arrives. If the inspection progress or release progress regresses within the current rolling shipment cycle, the latest valid progress after the regression will overwrite the previous progress, and the regression count will be written to the order status object. The regression count is a counter field that records the number of times the same order status object undergoes a reverse status change during the declaration, inspection, or release processing stages within the current rolling shipment cycle.

[0051] Historical release waves do not use a fixed template, but are updated on a rolling basis according to port identification, mode of transport, and cargo type. The wave weight of each time bucket is adjusted synchronously with the latest closed-loop results, so that the recent release rhythm has a greater impact on the expected time bucket assignment. The update method is as follows: , in, For the first The port identification is within each rolling shipment cycle. The mode of transport is Cargo type stratification identification is as follows In the time bucket Historical release wave weighting; Port signage The corresponding wavelet smoothing coefficient; The weight of historical release waves from the previous rolling shipping cycle; Add the current time bucket to the current scrolling statistics window. The number of closed-loop samples; Port signage The corresponding set of candidate time buckets; This serves as the identifier for traversing candidate time buckets.

[0052] when Alternatively, if the number of valid closed-loop samples in the current rolling statistics window is lower than the preset number, the historical release wave weights of each time bucket in the current rolling shipment cycle will not be updated, and the results of the previous rolling shipment cycle will be used instead.

[0053] To maintain consistency in the classification of estimated time buckets, candidate sorting, and window shifting criteria within the same rolling shipment cycle, the sub-state baseline dwell time and historical release wave weights used in the current rolling shipment cycle are bound to the same statistical version identifier. ;Statistical version identifier It is only used for recording, alignment, rollback, and version binding. It remains frozen within the current rolling shipping cycle and switches to the new version at the start of the next rolling shipping cycle. If the number of valid samples on which the new version is based is less than the preset number or there is a timestamp conflict, the previous version will continue to be used.

[0054] The candidate time bucket set is limited to the current and next open cut-off windows of the port. The time bucket width can be set to 15 minutes, 30 minutes, or 60 minutes, and remains fixed within the same rolling shipment cycle for the same port. The wave smoothing coefficient can be calibrated from 0.75 to 0.90. When holidays, temporary port inspection upgrades, or concentrated delays cause the actual arrival distribution of the two most recent rolling statistical windows to deviate from the historical distribution by more than a preset proportion, the wave smoothing coefficient will be lowered to 0.45 to 0.60 to improve the speed of following new fluctuations. When the latest sample is less than the preset number, the results of the previous cycle will be used. When the sample is insufficient and the holiday indicator, temporary inspection upgrade indicator, or concentrated delay indicator corresponding to the current port is inconsistent with the previous rolling shipment cycle, the results of the previous cycle will be used, but a shorter baseline dwell time and a earlier time bucket weight than the previous cycle will not be used as replacement values. For the expected time bucket assignment, only the original assignment is allowed to be maintained or postponed to the subsequent valid time bucket; it will not be pushed to an earlier time bucket.

[0055] For example, the smoothing coefficients corresponding to the status and the wave smoothing coefficients corresponding to the port are implementation parameters, which can be determined based on historical sample quantiles, validation set adjustments, or historical playback calibers. The smoothing coefficient corresponding to the status can be set to 0.65~0.85 by default, and the wave smoothing coefficient corresponding to the port can be set to 0.75~0.90 by default. Neither of them should be 0 or 1, to avoid completely ignoring the latest closed-loop samples or being overly sensitive to single-cycle fluctuations. If the number of effective closed-loop samples in the current rolling statistical window is less than 5~20, the statistical version of the previous rolling shipment cycle remains unchanged; if the number of effective closed-loop samples reaches more than 20 and the actual arrival distribution of the last two rolling statistical windows continues to deviate from the historical distribution, the wave smoothing coefficient can be adjusted down by a step size of 0.05~0.15. The statistical version identifier is used to record the binding version of the status-based benchmark dwell time and historical release wave distribution used in the current rolling shipment cycle. It is frozen and unchanged by default within a rolling shipment cycle; when the version switch fails or the sample is insufficient, the previous version is used and the reason for the version freeze is written.

[0056] When determining the expected time bucket, the estimated time when the order status object reaches the configuration state is calculated to be aligned with the center time of each candidate time bucket, and the historical release wave weight of the corresponding time bucket is added. The time bucket with the highest alignment is selected as the expected time bucket. The calculation method is as follows: , in, For order status objects Assign to time bucket The fit; The historical release wave weights for the current rolling shipment cycle; For order status objects The current time; For order status objects The estimated remaining time; For time buckets The central moment; For time buckets The time-distance attenuation scale; For order status objects The corresponding estimated time bucket; This is the candidate time bucket set. The time-distance decay scale is taken as half to one time bucket width. When the fit difference between two adjacent time buckets does not exceed the preset micro-difference threshold, the time bucket with the later end time is taken as the expected time bucket. When the expected time to reach the configuration state falls within the current cutoff window, but is less than half a time bucket width from the cutoff boundary, it is extended to the first time bucket of the next effective cutoff window to reduce premature configuration under the critical boundary.

[0057] Furthermore, the time bucket width is an implementation parameter that can be set to 15 minutes, 30 minutes, or 60 minutes, remaining fixed within the same rolling shipment cycle at the same port. When the cut-off window at the same port is generally less than 90 minutes, a time bucket width of 15 minutes or 30 minutes is preferred; when the cut-off window is longer and the business write-back frequency is lower, a time bucket width of 30 minutes or 60 minutes is preferred. The time interval attenuation scale can be set synchronously with the time bucket width, defaulting to 0.5 to 1.0 times the time bucket width to balance time fit and wave distribution stability. The differential threshold is an engineering parameter used to determine whether the fit between adjacent time buckets is sufficiently close, defaulting to 0.01 to 0.05, and can be adjusted based on the balance point between the expected number of time bucket switching and the actual number of missed cut-offs in historical playback. When the estimated time for the current order status object to reach the pending configuration status falls within the current cut-off window, and there is less than one safe handover margin before the receiving deadline, it will be extended to the next valid cut-off window. The safe handover margin is preferentially adopted as the 70th or 80th percentile value of the historical handover time under the same port, the same carrier, and the same cargo handling requirements. The default value can be 0.5 to 1.0 times the width of one time bucket.

[0058] Example 2: This embodiment, based on Embodiment 1, provides a process for the allocation of replacements after capacity release. This process uses the capacity supply unit corresponding to the released capacity occupancy marker as an index to perform filtering, sorting, and window shifting control on candidate objects that meet the entry control requirements, and then writes the replacement results into the order status object and the outbound disposal chain. Specifically, as follows: Orders that currently meet the gate access control requirements but have not yet been marked with capacity occupancy can be initially used to create a replacement index table. This table consists of candidate index records organized by port identifier, mode of transport, cut-off window, and cargo handling requirements, used to quickly filter compatible candidates after a release event occurs. If the cumulative replacement count for the same order status object reaches its limit within the current rolling shipping cycle, only the original gate control record and release record are retained, and the object is no longer moved to a later cut-off window. The maximum cumulative replacement count can be set to 2-3 times, with a default of 2. The observation lock duration, minimum adjustment interval, and maximum window shift duration are implementation parameters, which can be set to 0.3-1.0 times, 0.5-1.0 times, and 1-2 times the current time bucket width, respectively, and can be adjusted based on the secondary release rate, replacement success rate, and cut-off failure rate in historical replays. The default replacement threshold is 0.8 to 1.0 times the median priority value of historical successful replacement samples under the same port, carrier, and cargo handling requirements. When there are missing data, timestamp conflicts, or delayed write-back of cargo receiving boundary information in the real-time status, the replacement threshold can be increased by 5% to 15% to reduce the triggering frequency of high-risk replacements. If there are no available candidate objects within the same cut-off window and the allowed shift window, the current order status object will remain in the pending window processing state, and cross-port shift will not be performed, nor will it be rolled back to an earlier cut-off window.

[0059] In this embodiment, the release process is triggered by an event timestamp. When an order status object receives at least one of the following events: release rollback, inspection reopening, box sealing rollback, outbound handover interruption, or cut-off failure, the corresponding capacity occupancy marker is removed based on the event timestamp. The reason for removal, the time of removal, and the corresponding capacity supply unit identifier before removal are recorded in the status transition chain and the outbound disposal chain. During replacement allocation, the capacity supply unit corresponding to the removed capacity occupancy marker is used as an index. From the outbound demand units that currently meet the access control requirements and have not yet been written with a capacity occupancy marker, candidate objects compatible with the capacity supply unit in terms of port identifier, mode of transport, cargo handling requirements, and cut-off window connection are selected. These candidates are then reordered according to release time, warehouse completion time, and remaining cut-off time before being written with the capacity occupancy marker. For order status objects that have not completed replacement allocation within the current cut-off window, the original status change records and access control records are retained, and the new cut-off window, the corresponding candidate capacity supply unit set, and the rematching time are written into the order status object and the outbound disposal chain.

[0060] In a preferred embodiment of Example 2, the method for selecting a compatible capacity supply unit from the capacity supply units corresponding to the released capacity occupancy marker for replacement allocation is as follows: When an order status object with a capacity occupancy flag experiences a release rollback, inspection reopening, box sealing rollback, outbound handover interruption, or cut-off deadline failure, the corresponding capacity occupancy flag is first removed, and the corresponding capacity supply unit is written to the capacity release pool. For newly added shipping demand units to the capacity configuration set, a replacement candidate queue is formed from the capacity release pool, using the shipping demand unit as an index. The replacement candidate queue is not directly open to all capacity supply units; instead, it only includes capacity supply units that are compatible with the shipping demand unit in terms of port identification, mode of transport, cargo handling requirements, and load capacity boundaries, and that still have cargo receiving capacity before the cut-off deadline. The selection method is as follows: , in, To release transportation capacity supply units For shipping demand units Candidate valid markers; This is an indicator function that takes the value 1 when the condition is true and 0 when the condition is false. To release transportation capacity supply units Port signage; For shipping demand units Port signage; To release transportation capacity supply units The mode of transport; For shipping demand units The mode of transport; For shipping demand units Cargo handling requirements level; To release transportation capacity supply units Supported cargo handling capacity levels; This indicates that the cargo handling request is supported by the corresponding transport capacity supply unit; To release transportation capacity supply units The current allocatable load; For shipping demand units The required load capacity; For shipping demand units The estimated time when the conditions for receiving the goods will be met; For shipping demand units Corresponding safety handover margin; To release transportation capacity supply units The cutoff time of the corresponding cutoff window; This is a substitute for the suppression marker.

[0061] The safety handover margin is calibrated based on the historical time taken for loading, handover, and release order rewriting at the port, and can be taken as half to one time the width of the current time bucket; for temperature-controlled goods, sensitive goods, or goods requiring secondary verification, the calibration level is increased by one level; when port congestion or handover rewriting delays increase, the upper limit value is used.

[0062] The alternate candidate queue is processed using a hard constraint screening, hierarchical sorting, and peer-to-peer scoring method. Compatibility with the same port of entry, mode of transport, and cargo handling requirements is considered a hard constraint; if these conditions are not met, the candidate is excluded from the queue. Same carrier and same cut-off window are considered priority constraints, but not unique constraints. During sorting, the cut-off window is first compared to the target cut-off window of the shipping demand unit; then, the carrier is compared to the reference carrier identifier. Candidates with the same hierarchical level are then sorted in descending order of comprehensive priority value, calculated as follows: , in, To release transportation capacity supply units For shipping demand units The overall priority value; Consistent weighting for the intercept window; Weighting for consistent use by the carrier; Weighting of the load cutoff margin; Weighted by waiting time; Penalty weight for wasted load capacity; Weighting for repeated substitution penalties; Capture window consistency flag; For consistent marking by the carrier, when releasing capacity supply units Carrier identification and shipping demand unit Corresponding reference carrier identifier If the values ​​match, set the value to 1; otherwise, set the value to 0. For shipping demand units The carrier identifier of the original capacity supply unit whose capacity occupancy has been lifted within the current target cutoff window; when the shipping demand unit... When there is no flag indicating that capacity has been released, When the same shipping demand unit corresponds to multiple released capacity occupancy markers, the one with the most recent release time and currently not expired shall be selected. Used only for substitute ranking and cross-carrier determination; This is the normalized value of the load shear margin. For shipping demand units Normalized value of waiting time since entering the capacity allocation set; This is a normalized value representing the percentage of remaining load capacity after allocation. For shipping demand units The cumulative number of replacements within the current rolling shipping cycle. , and The values ​​of all values ​​are in the range of 0 to 1; The time remaining from the current moment to the deadline for releasing the transport capacity supply unit is determined by the ratio of the maximum remaining time of the candidate under the same port identifier. The target load window width is determined by the ratio of the cumulative waiting time after the shipping demand unit enters the capacity allocation set. The load capacity is determined by the ratio of the remaining load capacity after the replacement is completed to the currently allocable load capacity of the released capacity supply unit. When the maximum remaining time of the candidate under the same port identifier is zero, the target cut-off window width is missing, or the currently allocable load capacity of the released capacity supply unit is zero, the corresponding normalized value is 0, and the same-level comparison is only performed in the order of consistent cut-off window, consistent carrier, earlier cut-off time, and smaller remaining load capacity after allocation.

[0063] Weights satisfy To prioritize carrier consistency over cut-off window consistency, and carrier consistency over waiting time and capacity utilization, when the difference in the combined priority value of two candidates is less than a preset minor difference threshold, the unit with the earlier cut-off time and smaller remaining capacity after allocation is selected. Weights are updated once per rolling shipping cycle, referencing the previous cycle's replacement success rate and secondary release rate after replacement; as the secondary release rate continuously increases, the weighting is increased. , and and reduce When data is missing, timestamp conflicts, or holiday congestion causes real-time instability, the system switches to conservative reorganization to lower the priority of cross-carrier substitutes.

[0064] When making a substitute, it is not required that the port of entry, carrier, and cut-off window be completely consistent at the same time. However, the port of entry and the mode of transport must remain unchanged. The carrier and the cut-off window are preferred conditions.

[0065] The substitute threshold is denoted as This is used to determine whether to continue executing the substitute within the same cutoff window; The benchmark is determined based on the median of the comprehensive priority value of historical successful substitute samples under the same port identification, mode of transport, and cargo handling requirements, and is set to 0.8 to 1.0 times this median value. When there is port congestion, timestamp conflicts, or delayed write-back of cargo receiving boundary information, the benchmark is adjusted upwards by one level. When the number of historical samples in the same stratum is lower than the preset lower limit, cross-window shifting for substitutes is not performed; only substitute paths within the same window are retained. When there are no available release capacity supply units in the same cut-off window, or when the highest comprehensive priority value of the candidate within the same cut-off window is lower than the substitute threshold, limited shifting to subsequent adjacent cut-off windows is allowed, but regression to earlier cut-off windows is not permitted. The candidate set after shifting is as follows: , in, For shipping demand units The set of capacity supply units allowed to be released after window translation; For shipping demand units The cutoff time of the corresponding target capture window; For shipping demand units The maximum allowed window shift duration; the meanings of the remaining parameters are the same as described above. The maximum window shift duration is determined according to cargo handling requirements and shipping time limits. For general cargo, it can be one cut-off window width or two time bucket widths. For temperature-controlled cargo, sensitive cargo, and cargo with remaining shipping time limits below a preset threshold, it is zero to one time bucket width. When the estimated time when the shipping demand unit meets the receiving conditions is less than one safe handover margin from the receiving boundary of the new window, shift substitution is not performed. When there are no candidates from the same carrier, but the receiving boundary information, port requirements, and cargo handling capacity of candidates from different carriers are compatible, cross-carrier substitution is allowed, and cross-carrier substitution only takes effect when the target window is consistent or the shift does not exceed the maximum window shift duration.

[0066] To mitigate the jitter caused by frequent releases and refills, both the capacity supply unit and the outbound demand unit are subject to lockout duration, minimum adjustment interval, and repeated substitute suppression. After releasing the capacity occupancy marker, the capacity supply unit enters a short observation period, during which only substitute requests from the same cut-off window and the same carrier are accepted. After the observation period ends, it is then opened to the candidate set allowing window shifting. The substitute suppression rules are as follows: , in, To release transportation capacity supply units With shipping demand unit The substitution suppression flag is set between the two states; a value of 1 indicates that the state is suppressed, and a value of 0 indicates that participation in substitution is allowed. The current moment; To release transportation capacity supply units The most recent release moment; To release transportation capacity supply units Observation lock duration; To release transportation capacity supply units The most recent moment when substitute adjustments were completed; To release transportation capacity supply units The minimum adjustment interval; For shipping demand units The maximum number of substitutions allowed within the current target cutoff window. At that time, release the transportation capacity supply unit At the current determination time, it will not enter the shipping demand unit. The alternative candidate queue; when At that time, one is allowed to enter the substitute candidate queue.

[0067] The observation and locking duration can be set at one-third to one time the width of the current time bucket, the minimum adjustment interval can be set at half the width of the time bucket to one time bucket, and the maximum number of replacements can be 2 or 3. When a high proportion of secondary releases occurs at the same port in two consecutive rolling shipping cycles, the observation and locking duration and the minimum adjustment interval are increased, and the maximum window shift duration is decreased by one setting.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0069] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of this application and form different embodiments. For example, all the embodiments above can be used in any combination. The information disclosed in this background section is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art.

Claims

1. A method for logistics demand forecasting and capacity allocation that integrates multi-source data, characterized in that, include: Collect basic order data, three-document verification data, declaration and inspection processing data, release result data, warehouse operation data, port schedule cut-off data, and carrier capacity surplus data; The data is associated with order objects to form order status objects and a state transition chain is constructed. The status change events and status retention time of each order status object are statistically analyzed according to the preset time bucket. Combined with the historical release wave distribution, current declaration progress, inspection progress and release progress, the estimated time bucket for each order status object to enter the pending configuration status is determined. Order status objects with the same port identification, carrier method, cut-off window and cargo handling requirements are aggregated into shipping demand units. Based on port schedule cut-off data and carrier capacity surplus data, a capacity supply unit is generated; Based on the verification results of the three documents, the release result data, the warehouse operation data, and the cargo handling requirements, the outbound demand unit is subject to access control. The outbound demand unit that passes the access control is written into the capacity configuration set, and matched with the capacity supply unit according to the consistency of port identification, the consistency of the mode of transport, the connection of the cut-off window, and the compatibility of cargo handling requirements. The capacity occupancy mark is written into the corresponding capacity supply unit. During the rolling shipping cycle, when an order status object with a capacity occupancy mark receives at least one of the following events: release rollback, inspection reopening, box sealing rollback, outbound handover interruption, or load cut-off failure, the corresponding capacity occupancy mark is released, and other shipping demand units that meet the access control are written into the capacity supply unit corresponding to the released capacity occupancy mark. By associating records of status changes, gate control criteria, and capacity adjustment results, a shipping processing chain corresponding to the order status object is formed.

2. The method for logistics demand forecasting and capacity allocation that integrates multi-source data according to claim 1, characterized in that, The order object refers to a single order or a batch of orders with the same port identification, the same mode of transport, and the same cut-off window. The basic order data includes order identifier, product identifier, country or region of destination, port identifier, mode of transport, and shipping time limit; The three-item verification data includes the consistency verification results of the transaction order, payment order, and logistics order.

3. The method for logistics demand forecasting and capacity allocation that integrates multi-source data according to claim 1, characterized in that, The order status objects in the status transition chain include at least the pending declaration status, declaration processing status, inspection processing status, released but not completed status, pending configuration status, configured and ready for shipment status, and shipped status, and record the status change events, status change times, and event source identifiers between each status.

4. The method for logistics demand forecasting and capacity allocation that integrates multi-source data according to any one of claims 1, characterized in that, The access control includes the following joint determination conditions: The three-item verification result is passed; the release result data corresponding to the order status object indicates that it has been released and there is no unresolved inspection rollback status; The warehouse operation data corresponding to the order status object indicates that the box sealing or outbound handover has been completed. The cargo processing requirements corresponding to the order status object are compatible with the port requirements and the mode of transport; the shipment demand unit enters the capacity configuration set only when all joint judgment conditions are met simultaneously.

5. A method for logistics demand forecasting and capacity allocation that integrates multi-source data according to any one of claims 1, characterized in that, The capacity supply unit includes carrier identification, port identification, schedule identification or train number identification, cut-off window, allocable capacity, cargo receiving boundary information, and capacity occupancy marker; The capacity supply generation unit includes splitting the remaining capacity data according to port, mode of transport, schedule or train number and cut-off window, and associating the splitting results with the corresponding port schedule cut-off data.

6. The method for logistics demand forecasting and capacity allocation that integrates multi-source data according to any one of claims 1, characterized in that, The matching includes: Candidate transport capacity supply units that are compatible with the shipping demand units in terms of port identification, mode of transport, cut-off window and cargo handling requirements; The candidate transport capacity supply units are sorted according to the release time, in-warehouse completion time, and remaining time before the cut-off point. Write the candidate capacity supply units that are ranked first and whose allocable capacity meets the conditions into the corresponding capacity occupancy flag.

7. The method for logistics demand forecasting and capacity allocation based on multi-source data according to any one of claims 1, characterized in that, When the order status object receives at least one of the following events: release rollback event, inspection reopening event, box sealing rollback event, outbound handover interruption event, or load cut-off failure event, the corresponding capacity occupancy mark is removed based on the event timestamp, and the reason for removal, the time of removal, and the corresponding capacity supply unit identifier before removal are recorded.

8. The method for logistics demand forecasting and capacity allocation based on multi-source data according to any one of claims 7, characterized in that, When allocating replacements, the capacity supply unit corresponding to the capacity occupancy mark that has been removed is used as the index. From the current outbound demand units that meet the access control and have not yet been written with the capacity occupancy mark, candidate units that are compatible with the capacity supply unit in terms of port identification, mode of transport, cargo handling requirements and cut-off window connection are selected. After being sorted according to release time, warehouse completion time and remaining cut-off time, they are written with the capacity occupancy mark. For order status objects that have not completed the replacement allocation within the current cut-off window, retain the original status change record and gating record, and write the new cut-off window, the corresponding candidate capacity supply unit set, and the rematching time into the order status object.

9. A method for logistics demand forecasting and capacity allocation that integrates multi-source data according to any one of claims 1, characterized in that, The shipment disposal chain at least associates the data source identifier, status change record, access control result, capacity occupancy write record, release record, substitute allocation record and final disposal result of the stored order status object, and updates the order status object ownership result of each time bucket in the subsequent rolling shipment cycle based on the shipment disposal chain.

10. A method for logistics demand forecasting and capacity allocation that integrates multi-source data according to any one of claims 1, characterized in that, When there are timestamp conflicts, missing fields, or inconsistent sources in the release result data, warehouse operation data, or port shift cut-off data corresponding to the same order status object, the order status object will be marked as an abnormal pending status and prevented from entering the capacity configuration set. After receiving the supplemented or corrected business data, the state transition chain update, access control, and capacity matching are re-executed.

Images