A cross-border logistics compliance automatic verification and risk early warning system

Abstract

The application provides a cross-border logistics compliance automatic checking and risk early warning system, relates to the technical field of logistics and supply chain management, and comprises: a collection module used for constructing a waybill collection dataset according to the delivery records generated by cross-border logistics outbound nodes and allocation outbound nodes and actual billing data; a matching module used for extracting a logistics company, a destination country partition and a declared charging weight from the waybill collection dataset, searching the corresponding fee rules according to the logistics company and the destination country partition, obtaining a parallel line segment family constituted by a charging weight interval and a maintained unit price mapping, constructing a fee judgment circle based on the parallel line segment family, combining the declared charging weight and a preset tolerance threshold, and calculating the rule matching freight through an effective tangent constraint condition. The application improves the timeliness, accuracy and overall control efficiency of cross-border logistics compliance risk investigation.

Description

Technical Field

[0001] This invention relates to the field of logistics and supply chain management technology, and in particular to an automated verification and risk warning system for cross-border logistics compliance. Background Technology

[0002] In the business data processing of cross-border trade and supply chain management, the settlement and compliance verification of logistics costs are crucial to ensuring the safety of corporate funds and operational efficiency. With the increase in cross-border business volume, companies need to process a large amount of outbound and transfer waybills as well as comprehensive billing data provided by logistics providers.

[0003] Taking monthly reconciliation of cross-border air and sea freight logistics as an example, logistics providers typically formulate tiered or segmented fee settlement rules based on different destination country zones and chargeable weight ranges. In the existing data processing flow, after receiving the logistics bill, simple table matching or manual retrieval is usually relied upon to verify the waybill details against the bill fees. However, this existing data processing method has some limitations in practical applications. Faced with complex fee rules mapped by destination country zones and chargeable weight ranges, existing systems may struggle to automatically build accurate freight calculation models. Due to the lack of automated parsing capabilities for the mapping relationship between chargeable weight ranges and unit prices, the system finds it difficult to quickly and accurately calculate the estimated matching freight for each waybill, making it difficult to identify discrepancies between the actual bill fees and the estimated fees in a timely manner. Furthermore, for waybills that have not been reconciled for more than a preset period, and for situations where there is a difference between the system's pre-stored packaging weight and the actual billable weight, the existing verification dimensions are relatively independent, and most lack comprehensive correlation analysis of data such as fee deviations, reconciliation delays, and weight anomalies. Summary of the Invention

[0004] This invention provides an automated verification and risk warning system for cross-border logistics compliance, which improves the timeliness, accuracy, and overall management efficiency of cross-border logistics compliance risk assessment.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: Firstly, an automated cross-border logistics compliance verification and risk warning system includes: The aggregation module is used to build a waybill aggregation dataset based on the shipping records generated by cross-border logistics outbound nodes and transfer outbound nodes, as well as actual billing data. The matching module is used to extract the logistics provider, destination country zone, and declared chargeable weight from the waybill aggregation dataset. Based on the logistics provider and destination country zone, it retrieves the corresponding cost rules and obtains a family of parallel line segments consisting of the mapping between chargeable weight range and maintenance unit price. Based on the family of parallel line segments, it constructs a cost determination circle by combining the declared chargeable weight and preset tolerance threshold, and calculates the rule matching freight through effective tangency constraints. The verification module is used to calculate the difference between the rule-matched shipping cost and the actual invoice shipping cost to obtain the cost deviation value; at the same time, it compares the invoice date with the preset reconciliation cycle to generate a reconciliation status identifier; and combines the cost deviation value and the reconciliation status identifier to form the cost and timeliness verification result. The comparison module is used to identify and mark waybills marked as overdue for reconciliation based on the reconciliation status identifier, and obtain the reconciliation timeliness anomaly identifier; at the same time, it extracts the pre-stored packaging weight and the billing weight from the waybill collection dataset, calculates the difference between the two, and obtains the weight deviation verification result. The early warning module is used to generate cost deviation early warning, reconciliation delay early warning and weight abnormality early warning according to the reconciliation timeliness anomaly identifier, cost deviation value in cost and timeliness verification results and weight deviation verification results, and obtain graded risk early warning information.

[0006] In a second aspect, a computing device includes: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, enable the one or more processors to implement the system.

[0007] Thirdly, a computer-readable storage medium storing a program that, when executed by a processor, implements the system.

[0008] The above-described solution of the present invention has at least the following beneficial effects: The system unifies shipping records and actual invoice data generated by cross-border logistics outbound and transfer outbound nodes into a unified waybill aggregation dataset. Based on the logistics provider, destination country region, and declared chargeable weight, it retrieves corresponding fee rules. Utilizing a family of parallel line segments formed by mapping chargeable weight ranges to maintenance unit prices, and combining declared chargeable weight with preset tolerance thresholds, it constructs a fee determination circle. Through effective tangency constraints, it calculates matching freight costs according to rules, improving the automation and accuracy of freight matching under complex fee rules. By calculating the difference between the matching freight cost and the actual invoice freight cost, and comparing the invoice date with the preset reconciliation period, it generates fee deviation values ​​and reconciliation status indicators, facilitating timely identification of fee anomalies and overdue reconciliation situations. Simultaneously, by extracting pre-stored packaging weight and invoice chargeable weight to calculate weight deviation, it generates corresponding tiered risk warning information based on preset risk classification rules for fee deviation, reconciliation timeliness anomalies, and weight deviation, improving the efficiency of cross-border logistics reconciliation and compliance verification, and enhancing the timeliness of identifying risks related to fee deviation, reconciliation delays, and weight anomalies. Attached Figure Description

[0009] Figure 1This is a flowchart illustrating an automated verification and risk warning system for cross-border logistics compliance provided by an embodiment of the present invention.

[0010] Figure 2 This is a schematic diagram illustrating the calculation rules for matching freight charges provided in an embodiment of the present invention.

[0011] Figure 3 This is a schematic diagram illustrating the cost and timeliness verification results provided in an embodiment of the present invention. Detailed Implementation

[0012] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0013] like Figure 1 As shown, an embodiment of the present invention proposes an automated verification and risk warning system for cross-border logistics compliance, comprising: The aggregation module is used to build a waybill aggregation dataset based on the shipping records generated by cross-border logistics outbound nodes and transfer outbound nodes, as well as actual billing data. The matching module is used to extract the logistics provider, destination country zone, and declared chargeable weight from the waybill aggregation dataset. Based on the logistics provider and destination country zone, it retrieves the corresponding cost rules and obtains a family of parallel line segments consisting of the mapping between chargeable weight range and maintenance unit price. Based on the family of parallel line segments, it constructs a cost determination circle by combining the declared chargeable weight and preset tolerance threshold, and calculates the rule matching freight through effective tangency constraints. The verification module is used to calculate the difference between the rule-matched shipping cost and the actual invoice shipping cost to obtain the cost deviation value; at the same time, it compares the invoice date with the preset reconciliation cycle to generate a reconciliation status identifier; and combines the cost deviation value and the reconciliation status identifier to form the cost and timeliness verification result. The comparison module is used to identify and mark waybills marked as overdue for reconciliation based on the reconciliation status identifier, and obtain the reconciliation timeliness anomaly identifier; at the same time, it extracts the pre-stored packaging weight and the billing weight from the waybill collection dataset, calculates the difference between the two, and obtains the weight deviation verification result. The early warning module is used to generate cost deviation early warning, reconciliation delay early warning and weight abnormality early warning according to the reconciliation timeliness anomaly identifier, cost deviation value in cost and timeliness verification results and weight deviation verification results, and obtain graded risk early warning information.

[0014] In this embodiment of the invention, shipping records and actual billing data generated by cross-border logistics outbound nodes and transfer outbound nodes are uniformly constructed into a waybill aggregation dataset. Based on the logistics provider, destination country region, and declared chargeable weight, corresponding fee rules are retrieved. A family of parallel line segments formed by mapping chargeable weight ranges to maintenance unit prices is used, combined with the declared chargeable weight and preset tolerance thresholds to construct a fee determination circle. Fee matching is calculated using effective tangency constraints, improving the automation and accuracy of fee matching under complex fee rules. By calculating the difference between the rule-matched freight and the actual billing freight, and comparing the billing date with the preset reconciliation period, fee deviation values ​​and reconciliation status indicators can be generated, facilitating timely identification of fee anomalies and overdue reconciliation situations. Simultaneously, weight deviation is calculated by extracting the pre-stored packaging weight and the billing weight, and corresponding graded risk warning information is generated based on preset risk classification rules for fee deviation, reconciliation timeliness anomalies, and weight deviation. This improves the efficiency of cross-border logistics reconciliation and compliance verification, and enhances the timeliness of identifying risks related to fee deviation, reconciliation delays, and weight anomalies.

[0015] In a preferred embodiment of the present invention, the aggregation module is used to construct a waybill aggregation dataset based on the shipping records generated by cross-border logistics outbound nodes and transfer outbound nodes, as well as actual billing data. This dataset may include: In this embodiment, the cross-border logistics compliance automated verification and risk warning system is deployed in the enterprise's logistics management platform. This platform maintains data communication connections with multiple cross-border logistics outbound nodes, such as shipping warehouses located in different cities, and transfer outbound nodes, such as regional distribution centers. When the outbound nodes and transfer outbound nodes complete the outbound operation, they automatically generate shipping records. Each shipping record includes fields such as waybill number, logistics provider identifier, destination country zone identifier, declared chargeable weight value, pre-stored packaging weight value, and outbound timestamp.

[0016] A waybill number is a coded string automatically generated by the system to uniquely identify an independent waybill. Each waybill number is unique within the global scope of the system and serves as a unique primary key for associating shipping records, billing records, and verification results.

[0017] The logistics provider identifier is a unique code of the logistics service provider that carries the waybill, used to distinguish the pricing rules and service agreements of different logistics providers.

[0018] Destination country zoning refers to the business areas that are divided into destination countries or regions according to factors such as geographical location, transportation routes, and customs clearance policies, based on the actual operational needs of cross-border logistics. The same logistics provider may adopt different pricing strategies and rate levels for different destination country zoning. Therefore, the identification of destination country zoning is a necessary prerequisite for searching fee rules.

[0019] The declared chargeable weight refers to the weight of the goods declared by the shipper at the outbound stage for calculating freight charges. The unit of measurement is kilogram (kg), and its value is determined based on the actual weighing result or the volumetric weight conversion result. The volumetric weight conversion method is to multiply the length, width, and height dimensions of the outer packaging of the goods, then divide the product by the volumetric weight conversion factor specified by the logistics provider. The quotient is taken as the volumetric weight, and the larger value between the volumetric weight and the actual weighed weight is taken as the declared chargeable weight.

[0020] Pre-stored packaging weight refers to the weight data of the goods packaging that the shipper weighs and enters into the system during the outbound process. The unit of measurement is kilograms, and it comes from the weighing process after the goods are packaged at the outbound node. The pre-stored packaging weight serves as the shipper's self-assessment benchmark for the weight of the goods. It is compared with the billing weight in the logistics provider's invoice and is the basis for weight deviation verification.

[0021] Outbound timestamp refers to the precise moment when goods leave the outbound or transfer outbound node. The record format includes year, month, day, hour, minute, and second information, and is used for reconciliation timeliness calculation and logistics timeliness analysis.

[0022] At the same time, the system regularly obtains actual billing data from the billing interfaces provided by various logistics providers or through manual import. The actual billing data is organized by logistics provider, and each billing record includes fields such as billing waybill number, billing freight amount, billing weight amount, and billing date.

[0023] The billing freight refers to the amount of money that the logistics provider determines and formally issues to the shipper according to its internal pricing rules. Its currency unit is consistent with the system's base currency. The billing freight may include the total of multiple sub-items such as basic freight, fuel surcharge, remote area surcharge, and insurance.

[0024] Billing weight refers to the actual weight of the goods used by the logistics provider in the billing process. This value is the billing weight determined by the logistics provider after receiving the goods, either by weighing them a second time or by re-determining the billing weight according to its own volumetric weight conversion rules. There may be differences between the billing weight and the declared billing weight. The sources of these differences include, but are not limited to, the accuracy deviation between the logistics provider's weighing equipment and the shipper's weighing equipment, the logistics provider using a different volumetric weight conversion factor than the shipper, and the redistribution of weight due to splitting or merging of goods during transportation and transshipment. The unit of measurement for billing weight is kilograms.

[0025] The billing date refers to the date on which the logistics provider issues the bill. The record format includes year, month, and day information, and this date is the benchmark starting point for subsequent calculation of reconciliation time.

[0026] The system uses the waybill number as the primary key to associate and match shipping records from outbound and transfer outbound nodes with billing data from logistics providers, constructing a waybill aggregation dataset. The specific construction process is as follows: The system iterates through all shipping records generated by each outbound and transfer outbound node in the current batch, extracting the waybill number field from each shipping record; using the extracted waybill number as the search key, it retrieves billing records with the same waybill number from the imported billing data set. The billing data set is indexed by the waybill number, that is, the hash value of the waybill number is used as the index key and the complete billing record is used as the index value, so that the search operation based on the waybill number can be completed in constant time complexity.

[0027] If a matching billing record is found, the logistics provider field, destination country zone field, declared chargeable weight field, pre-stored packaging weight field, and outbound timestamp field from the shipping record are merged with the billing freight field, billing chargeable weight field, and billing date field from the billing record into the same waybill aggregation record. This forms a waybill aggregation record containing a total of nine fields: waybill number, logistics provider, destination country zone, declared chargeable weight, pre-stored packaging weight, outbound timestamp, billing freight, billing chargeable weight, and billing date. If no matching invoice record is found in the invoice data for a shipment record, meaning the logistics provider has not yet issued an invoice for the waybill or the invoice data has not yet been imported into the system, the invoice freight, invoice chargeable weight, and invoice date fields in the corresponding waybill collection record will be temporarily set to empty values. The waybill will be marked as pending reconciliation in the reconciliation status determination process. For the reconciliation status identifier, pending reconciliation means that the invoice data for the waybill has not yet arrived or has arrived but has not yet been reconciled and confirmed, and the actual elapsed time since the invoice date has not exceeded the standard time limit specified in the reconciliation cycle. Overdue reconciliation status means that the invoice data for the waybill has arrived, but the actual elapsed time since the invoice date to the current system processing time has exceeded the standard time limit specified in the reconciliation cycle, and the reconciliation and confirmation operation has not yet been completed.

[0028] like Figure 2 As shown, in a preferred embodiment of the present invention, the matching module is used to extract the logistics provider, destination country zone, and declared chargeable weight from the waybill aggregation dataset; retrieve the corresponding cost rules based on the logistics provider and destination country zone to obtain a family of parallel line segments consisting of a mapping between the chargeable weight range and the maintenance unit price; construct a cost determination circle based on the family of parallel line segments, combined with the declared chargeable weight and a preset tolerance threshold; and calculate the rule-matching freight rate through effective tangency constraints, which may include: After constructing the waybill aggregation dataset, the system extracts the logistics provider field, destination country partition field, and declared chargeable weight field from each waybill aggregation record for use in cost rule matching and freight calculation. First, the system constructs a parallel line segment family. Based on the value of the logistics provider field in the waybill aggregation dataset, the system locates the dedicated pricing strategy library corresponding to that logistics provider. The dedicated pricing strategy library is a set of cost pricing rules maintained separately for each logistics provider in the system. This set is organized into a hierarchical rule dictionary structure, using the logistics provider identifier as the first-level index key, the destination country partition as the second-level index key, and the mode of transport or service type as the third-level index key. The dedicated pricing strategy library stores all the tiered rate rules, surcharge rules, and pricing parameters formulated by the logistics provider for different destination country partitions and different modes of transport. The establishment of the dedicated pricing strategy library ensures that the billing logic between different logistics providers is isolated. Each logistics provider can independently adjust its own rate table without affecting the pricing rule system of other logistics providers. Simultaneously, the system can perform add, delete, modify, and query operations on rules according to the logistics provider dimension.

[0029] Based on the value of the destination country partition field in the waybill aggregation dataset, the system indexes the tiered rate table adapted to the current waybill flow in the dedicated pricing strategy library, thereby obtaining the logistics provider's original cost rule set under the destination country partition. Each rule in the original cost rule set defines three core elements: the lower limit of chargeable weight, the upper limit of chargeable weight, and the unit freight maintenance price.

[0030] A tiered tariff table divides the continuous range of chargeable weight into several discrete weight intervals. Each weight interval corresponds to a fixed unit freight maintenance price. The freight is calculated according to the maintenance price corresponding to the interval in which the chargeable weight falls. For example, a typical tiered tariff table may contain the following rule rows: the maintenance price per kilogram is the first unit price for chargeable weights from zero to the first boundary value (inclusive); the maintenance price per kilogram is the second unit price for chargeable weights from the first boundary value (exclusive) to the second boundary value (inclusive); and the maintenance price per kilogram is the third unit price for chargeable weights above the second boundary value (exclusive).

[0031] The lower and upper limits of chargeable weight represent the starting and ending boundary values ​​of a certain chargeable weight range, respectively. Together, they define a continuous weight range, both in kilograms. Between two adjacent weight ranges, the upper limit of chargeable weight in the previous range is equal to the lower limit of chargeable weight in the next range, ensuring that the entire chargeable weight range is continuously covered without gaps. In other words, any possible chargeable weight value can be found in the corresponding range in the tiered rate table.

[0032] The maintenance unit price refers to the unit price of transportation costs corresponding to each chargeable weight. The unit of measurement is the monetary amount per kilogram. Its specific value is configured and updated in the system by the logistics provider based on business considerations such as market conditions, transportation costs, and seasonal factors. The maintenance unit price can be an integer or contain decimal places, and the system maintains its original precision in the calculation.

[0033] The system performs structured parsing on the original set of cost rules, traversing each rule in the set one by one to extract the defined lower and upper limits of chargeable weight, as well as the corresponding unit freight maintenance price. For each rule element consisting of a chargeable weight range and a corresponding maintenance price, the system abstracts it into an independent line segment element. The line segment element is the geometric representation of the cost rule in a two-dimensional Cartesian coordinate system composed of a weight axis (horizontal axis) and a unit price axis (vertical axis). The horizontal coordinate of the line segment element extends from the lower to the upper limit of chargeable weight. The horizontal coordinate of the left endpoint corresponds to the lower limit of weight in the range, and the horizontal coordinate of the right endpoint corresponds to the upper limit of weight in the range. The vertical coordinate of the line segment element corresponds to the maintenance price value corresponding to the weight range. Since the maintenance price within the same weight range is constant, the line segment element appears as a horizontal line segment in the coordinate system, and its vertical coordinate remains unchanged throughout the entire horizontal coordinate range.

[0034] All the parsed line segment elements are linearly arranged in ascending order of the minimum chargeable weight. Since all line segment elements are horizontal line segments, they are parallel to each other (all are zero-slope line segments) and are arranged in ascending order of weight range on the horizontal axis. Therefore, the entire set of these line segment elements is defined as a family of parallel line segments.

[0035] A parallel segment family is a collection of line segments that visualizes all tiered tariff rules of a logistics provider under a specific destination country's zone in the form of horizontal line segments on a two-dimensional weight-unit price coordinate system. Each line segment in the parallel segment family independently corresponds to a chargeable weight range and its maintenance unit price. The line segments are arranged sequentially according to the weight range. The first line segment corresponds to the range with the smallest chargeable weight, and its left endpoint is located at the minimum value of the horizontal coordinate (usually zero). The last line segment corresponds to the range with the largest chargeable weight, and its right endpoint can theoretically extend to positive infinity, indicating that all chargeable weights greater than the lower limit of the range are subject to the maintenance unit price of that range. Adjacent line segments are connected end-to-end in the horizontal coordinate direction, meaning the horizontal coordinate of the right endpoint of the previous line segment is equal to the horizontal coordinate of the left endpoint of the next line segment, thus ensuring that the entire horizontal coordinate axis (from zero to positive infinity) is fully covered without omission. The vertical coordinates (maintenance unit price) of different line segments can be different. There may be a jump in the vertical coordinate at the junction of two adjacent line segments, that is, a sudden change in the maintenance unit price. Overall, the family of parallel line segments appears as a stepped broken line extending along the weight axis in the coordinate system. The height of each step is equal to the maintenance unit price of that interval, and the width is equal to the weight range of that interval.

[0036] After the parallel line segment family is constructed, the system begins to execute the rule matching freight calculation process. The rule matching freight refers to the theoretical freight value calculated by geometric matching based on the declared chargeable weight and the tiered rate rules of the logistics provider under the specific destination country. This value will serve as the benchmark reference value for cost deviation verification.

[0037] The system extracts the declared chargeable weight value of the waybill from the waybill aggregation dataset and uses this value as the x-coordinate of the center of the cost determination circle. At the same time, it reads the preset tolerance threshold parameter value of the system and uses this tolerance threshold as the radius of the cost determination circle. In the weight-unit price two-dimensional coordinate system, the x-coordinate of the center (i.e., the declared chargeable weight) and the y-coordinate of the center are either zero or the average value of all maintenance unit prices in the parallel line segment family. The specific value of the y-coordinate does not affect the subsequent geometric intersection detection results, because the intersection detection only focuses on the projection range of the circle in the x-coordinate direction as the center position. The cost determination circle is constructed with the tolerance threshold as the radius.

[0038] The cost determination circle is a geometric tool for matching intervals within a family of parallel line segments. In actual cross-border logistics operations, the declared chargeable weight may contain measurement errors or deviations within the business tolerance range. For example, there may be systematic differences in accuracy between different weighbridges, fluctuations in the weight of packaging materials due to changes in environmental humidity, and slight weight changes due to moisture evaporation that may occur between the goods leaving the warehouse and before the logistics provider receives them. If a point-to-point precise numerical matching method is used, that is, a strict numerical comparison is made between the precise value of the declared chargeable weight and the upper and lower limits of a specific chargeable weight interval, a tiny deviation in the declared chargeable weight, such as a deviation of only a few tenths of a kilogram, may lead to matching to the wrong chargeable weight interval. This risk is particularly prominent when the declared chargeable weight is located near the boundary between two adjacent intervals. The cost determination circle introduces a tolerance radius to construct a circular buffer area centered on the declared chargeable weight. This buffer area covers a continuous floating range of the declared chargeable weight in the horizontal direction, ranging from the horizontal coordinate of the center minus the radius to the horizontal coordinate of the center plus the radius. As long as the geometric representation of a certain chargeable weight range, i.e., a horizontal line segment in the family of parallel line segments, has a valid geometric intersection with the cost determination circle, that range is considered a possible candidate matching range.

[0039] The preset tolerance threshold is a system configuration parameter that represents the maximum tolerance range allowed for the declared chargeable weight to deviate from the actual chargeable weight. The unit of measurement is kilograms. The rules for determining the tolerance threshold are as follows: Grouping is based on logistics providers and destination country regions. A sample set of absolute values ​​of the deviations between the declared chargeable weight and the billed chargeable weight for all waybills within the past preset statistical period (e.g., a complete past quarter) is extracted from this group. This sample set is then sorted, and the value at the top of the sorted results is selected. The value at the specified location (i.e., the upper quartile) is used as the initial tolerance threshold. This tolerance threshold is not static. During operation, the system continuously monitors the accuracy of cost deviation warnings. If, under a fixed tolerance threshold, the false alarm rate of cost deviation warnings—that is, the proportion of warnings issued by the system but confirmed by manual verification to be without anomalies—exceeds the preset upper limit of the false alarm rate, the current tolerance threshold is increased by a preset step, for example, by 0.1 kg, to reduce the false alarm rate. Conversely, if the false alarm rate is lower than the preset lower limit, and there are missed cases confirmed by manual verification—that is, actual cost anomalies exist but the system did not trigger a warning—the current tolerance threshold is decreased by a preset step, for example, by 0.1 kg, to improve detection sensitivity. This adaptive adjustment mechanism of the tolerance threshold allows the system to dynamically optimize matching accuracy based on the actual service quality of the logistics provider.

[0040] The system translates the cost determination circle along the weight axis (horizontal coordinate axis), traversing each line segment element in the parallel line segment family. For each line segment element, it checks whether the circumference of the cost determination circle intersects with the geometric contour of that line segment element, thus filtering the candidate line segment set. The specific detection logic is as follows: For each line segment element in the parallel line segment family, its horizontal coordinate range is obtained. This ranges from the lower limit to the upper limit of the chargeable weight corresponding to the line segment. The horizontal coordinate of the center of the cost determination circle, i.e., the declared chargeable weight, is compared with the horizontal coordinate range of the line segment element.

[0041] In the first scenario, the abscissa of the center of the circle is within the abscissa range of the line segment, that is, the declared chargeable weight is greater than or equal to the lower limit of the chargeable weight and less than or equal to the upper limit of the chargeable weight. In this case, the center of the charge determination circle is itself within the abscissa range of the line segment. The coverage of the circle extends from the center of the circle to the left and right sides in the abscissa direction by a tolerance radius, which will inevitably cover at least a part of the geometric outline of the line segment. The two will intersect, and the line segment element will be included in the candidate line segment set.

[0042] In the second scenario, the abscissa of the center of the circle is located to the left of the abscissa range of the line segment, meaning the declared chargeable weight is less than the lower limit of the chargeable weight. In this case, the difference between the abscissa of the center of the circle and the abscissa of the left endpoint of the line segment (i.e., the abscissa of the lower limit of the chargeable weight) is calculated, which is the abscissa of the left endpoint of the line segment minus the abscissa of the center of the circle. If this difference is less than or equal to the tolerance threshold, it means that the coverage area of ​​the fee determination circle extending to the right in the abscissa direction has reached the left end area of ​​the line segment, and the circumference of the circle and the line segment intersect near the left endpoint. The line segment element is included in the candidate line segment set. If the difference is greater than the tolerance threshold, it means that the coverage area of ​​the fee determination circle is not sufficient to reach the leftmost end of the line segment, and the two do not intersect. The line segment element is excluded.

[0043] In the third scenario, the abscissa of the circle's center is located to the right of the abscissa range of the line segment, meaning the declared chargeable weight exceeds the chargeable weight limit. In this case, the difference between the abscissa of the circle's center and the abscissa of the right endpoint of the line segment (i.e., the abscissa of the chargeable weight limit) is calculated, which is the abscissa of the circle's center minus the abscissa of the right endpoint of the line segment. If this difference is less than or equal to the tolerance threshold, it means that the coverage area of ​​the fee determination circle extending to the left in the abscissa direction has reached the right end area of ​​the line segment, and the circumference of the circle and the line segment intersect near the right endpoint. The line segment element is then included in the candidate line segment set. If the difference is greater than the tolerance threshold, there is no intersection, and the line segment element is excluded.

[0044] After evaluating each of the three scenarios above, the system filters out all line segment elements that meet the geometric intersection condition, forming a candidate line segment set. The candidate line segment set is a temporary set of line segment elements that intersect with the circumference of the cost determination circle after geometric intersection detection. The billing weight interval corresponding to each line segment element in this set may become the interval to which the declared billing weight belongs. The number of elements in the candidate line segment set depends on the size of the tolerance threshold. The larger the tolerance threshold, the wider the coverage of the cost determination circle, and the more elements in the candidate line segment set. In extreme cases, multiple intervals adjacent to the declared billing weight in the parallel line segment family may be included in the candidates. The smaller the tolerance threshold, the narrower the coverage and the fewer the candidates. In extreme cases, only the interval where the abscissa of the circle center is located is included.

[0045] After obtaining the candidate line segment set, the system applies effective tangency constraints to the candidate line segment set in order to select a unique target line segment from the candidate line segment set. The application of effective tangency constraints is divided into four sub-steps. Sub-step one is to construct the line segment topology structure. According to the candidate line segment set, the weight interval boundary coordinates of each candidate line segment element are extracted one by one, that is, the lower limit value of the billing weight and the upper limit value of the billing weight, as well as the corresponding maintenance unit price value, to construct the line segment topology structure used for geometric relationship determination.

[0046] The segment topology refers to the data structure representation of the spatial arrangement and connection relationships between segment elements in the candidate segment set. This structure is organized in the form of a linked list or tree structure. Each segment node records its own horizontal coordinate range (starting and ending positions), vertical coordinate value (maintaining unit price), and the connection relationship between the segment and its adjacent segments (whether the ending horizontal coordinate of the preceding segment is equal to the starting horizontal coordinate of this segment, and whether the ending horizontal coordinate of this segment is equal to the starting horizontal coordinate of the subsequent segment). The construction of the segment topology allows the determination of effective tangency to be based on the relationship between segments (rather than the isolated information of a single segment).

[0047] Sub-step two involves contact point determination and slope verification. Based on the line segment topology, it checks whether the circumference of the cost determination circle and the geometric contour of the candidate line segment elements form a unique contact point within the weight range, and verifies the continuity of the line segment slope at the contact point location. The specific logic for contact point determination is as follows: Calculate the difference between the horizontal coordinate of the center of the cost determination circle and the horizontal coordinate of the midpoint of the candidate line segment element, i.e., the absolute value of the horizontal coordinate of the center minus the horizontal coordinate of the midpoint of the line segment; divide half the horizontal coordinate range of the line segment (i.e., the difference between the upper and lower limits of the billing weight) by 2 and compare it with the tolerance radius; if the result of the difference between the horizontal coordinate distance from the center of the circle to the midpoint of the line segment plus the tolerance radius is greater than half the horizontal coordinate range of the line segment, it means that the overlapping area of ​​the cost determination circle and the line segment has penetrated into the interior area of ​​the line segment, rather than merely grazing the edge of the line segment's endpoint. In this case, the line segment is determined to have formed a valid contact within the range and is recorded as a valid line segment. If the x-coordinate of the center of the circle is located at the junction of two adjacent line segments, that is, the upper limit of the chargeable weight of the previous line segment is equal to the lower limit of the chargeable weight of the next line segment, and the x-coordinate of the center of the circle is exactly equal to the x-coordinate value of this junction, then there are two possibilities: which side the center of the circle is biased towards at the junction. At this time, the following comparison is performed: calculate the x-coordinate distance from the x-coordinate of the center of the circle to the midpoint of the x-coordinate range of the previous line segment, and the x-coordinate distance from the x-coordinate of the center of the circle to the midpoint of the x-coordinate range of the next line segment; the line segment with the shorter distance is determined to have formed a valid internal contact.

[0048] The specific logic of slope verification is as follows: Since all segments in the parallel segment family are horizontal (with a slope of zero), in most cases, as long as the contact point is not at the junction of two adjacent segments with different ordinates (i.e., different maintenance unit prices), the slope at the contact point is continuous (both equal to zero). However, if the contact point happens to be at the junction of two adjacent horizontal segments with different maintenance unit prices, at this junction, the left segment ends with the first maintenance unit price value, and the right segment starts with the second maintenance unit price value, resulting in a sudden change in the ordinate. In this case, the slope is discontinuous at this point (zero on the left and zero on the right, but the ordinate value jumps discretely). For this case of endpoint contact, the system determines that it does not meet the slope continuity condition, and only retains the segment that overlaps more deeply with the cost determination circle within the interval (if both segments only contact the circle at the endpoint, both segments are excluded, and the search for other segments is expanded forward and backward).

[0049] Sub-step three removes invalid line segments. Based on the combined results of contact point determination and slope verification, invalid line segments that only form contact at the endpoints of the line segments are removed. That is, those line segments whose cost determination circles merely brush past their endpoints but whose centers do not penetrate into the interior of the interval. Valid line segments that achieve effective tangency within the interval are retained.

[0050] Sub-step four determines the target line segment. If, after the elimination in sub-step three, only one line segment remains in the set of valid line segments, then that line segment is the target line segment. If there are still more than one line segment in the set of valid line segments (for example, the tolerance radius is set too large, and the cost determination circle covers multiple adjacent billing weight intervals simultaneously), the system further performs a final selection based on the membership relationship between the billing weight intervals mapped by the remaining valid line segments and the declared billing weight. That is, it compares the value of the declared billing weight with the horizontal coordinate range of each line segment one by one, and selects the line segment corresponding to the horizontal coordinate range into which the value of the declared billing weight directly falls as the unique target line segment. If the value of the declared billing weight is exactly equal to a certain junction position, that is, it belongs to the boundary of two adjacent intervals at the same time, then according to the convention of lower boundary inclusion and upper boundary exclusion, that is, the definition method of the billing weight interval is that the lower boundary of the interval includes the boundary value and the upper boundary of the interval does not include the boundary value, the line segment corresponding to the left interval, that is, the interval whose horizontal coordinate range includes the junction value, is determined as the target line segment.

[0051] The target line segment refers to a unique line segment element that has been selected after filtering by the effective tangency constraint. The maintenance unit price value mapped by this line segment element is the unit freight pricing standard applicable to the current waybill. The core design purpose of the effective tangency constraint is to exclude candidate line segments that are only marginally touched at the endpoints of the cost determination circle. This marginal contact does not mean that the declared chargeable weight actually belongs to the chargeable weight range corresponding to the line segment; it simply means that the tolerance radius of the cost determination circle just extends to the edge of the range. If the effective tangency constraint is not applied and all candidate line segments are directly included in the subsequent freight calculation, it may result in a declared chargeable weight being matched with multiple maintenance unit prices, causing ambiguity and confusion.

[0052] After determining the target line segment, the system reads the maintenance unit price value mapped to that target line segment, performs a multiplication operation on the declared chargeable weight value and the maintenance unit price value, that is, multiplies the declared chargeable weight by the maintenance unit price per kilogram, and determines the rule-matching freight for that waybill. The rule-matching freight is an estimated freight value calculated based on the tiered rate rules of the logistics provider under the destination country's zones, through a geometric cost determination circle and parallel line segment family matching method, with the dimension of monetary amount. The rule-matching freight represents the theoretical freight amount that the waybill should generate under the assumption that the declared chargeable weight is accurate and the logistics provider strictly follows its own published rate rules for pricing.

[0053] like Figure 3As shown, in a preferred embodiment of the present invention, the verification module is used to calculate the difference between the rule-matched shipping cost and the actual invoice shipping cost to obtain a cost deviation value; simultaneously, it compares the invoice date with a preset reconciliation period to generate a reconciliation status identifier; combining the cost deviation value and the reconciliation status identifier into a cost and timeliness verification result may include: For the current waybill collection record, the system retrieves the value of the bill freight field corresponding to that waybill in the waybill collection dataset. The calculated rule-matching freight value is used as a benchmark reference value, and a difference calculation is performed between it and the bill freight value. That is, the bill freight value and the rule-matching freight value are subtracted first. The rule-matching freight value is subtracted from the bill freight value. Then, the absolute value of the difference is taken. Regardless of whether the difference is positive or negative, the non-negative value is taken, and this absolute value is recorded as the cost deviation value.

[0054] It should be noted that when performing subtraction, the system first checks whether the billing freight field is empty. If the billing freight field is empty, that is, the logistics provider's bill for the waybill has not yet arrived or has not yet been imported, the cost deviation value is temporarily recorded as pending. The cost deviation verification for this waybill is temporarily skipped in the current batch and will be recalculated after the billing data is completed. If the billing freight field has a valid value, the above difference calculation process will be executed normally.

[0055] The cost deviation value refers to the absolute difference between the theoretical freight cost (rule-matched freight cost) and the actual invoiced freight cost. It has the same dimension as the freight cost (monetary amount). The magnitude of the cost deviation value directly reflects the consistency between the logistics provider's actual charging behavior and its published rate rules. The smaller the cost deviation value, the more standardized the logistics provider's billing behavior is and the closer it is to its promised tiered rate rules. A larger cost deviation value indicates one or more of the following situations: the logistics provider is not strictly following its published tiered rate table; the logistics provider's pricing system has calculation errors; there is a significant difference between the declared chargeable weight and the chargeable weight actually used by the logistics provider; or the logistics provider has charged additional fees not reflected in the tiered rate rules in addition to the basic freight cost. The purpose of taking the absolute value is to uniformly measure the severity of the cost deviation. Regardless of whether the deviation is positive (the logistics provider overcharges) or negative (the logistics provider undercharges), the absolute value of the deviation reflects the degree of deviation in billing accuracy. The positive and negative direction information of the cost deviation value can also be simultaneously retained in the system for distinguishing the type of deviation during verification.

[0056] The system extracts the value of the billing date field from the waybill aggregation dataset. If the billing date field is empty, meaning the logistics provider has not yet issued an invoice, the reconciliation timeliness check is skipped, and the waybill is automatically marked as pending reconciliation (because there is no invoice yet, there is no basis for judging overdue). If the billing date field has a valid value, the system calculates the actual elapsed time from the billing date to the current processing time. The actual elapsed time refers to the number of calendar days that have passed between the calendar day of the billing date and the calendar day of the current system processing time. It is calculated by dividing the date of the current system processing time (only the year, month, and day portions are taken, excluding hours, minutes, and seconds) by the billing date (only the year, month, and day portions are taken). The actual elapsed time is calculated by subtracting the date number from the date number of the billing date from the date number of the current date, and the difference is taken as the actual elapsed time in days. If the billing date and the current processing time are the same calendar day, the actual elapsed time is counted as zero days. The actual elapsed time reflects how many calendar days have passed since the logistics provider issued the bill and the company has not yet completed the reconciliation of the bill. The current processing time refers to the time indicated by the system's real-time clock when the system performs the verification operation. It is used in the calculation of the actual elapsed time in terms of date (year, month, day), that is, the date part of the current system time is taken out, ignoring the specific hour, minute, and second information, because the reconciliation cycle is usually based on days.

[0057] The system compares the actual elapsed time in days with the standard time limit specified in the preset reconciliation period. The preset reconciliation period is a system configuration parameter that specifies how many calendar days the company should complete the reconciliation and confirmation process after receiving the logistics provider's invoice. The rules for determining the preset reconciliation period are as follows: First, the system reads the agreed time limit (usually in working days) from the reconciliation clauses in the service agreements or contracts signed between the company and each logistics provider. Then, the number of working days is converted to calendar days based on a five-day work week (i.e., working days multiplied by...). Then, based on the statistical analysis of historical reconciliation data, examine the average number of days of delay in receiving bills from each logistics provider, and add a reasonable grace period based on the contractual agreement, such as adding one or two calendar days as a buffer. Different logistics providers can be configured with different reconciliation cycle parameter values, because the reconciliation time limits agreed in the service contracts of different logistics providers may be different. Alternatively, a default reconciliation cycle uniformly set by the enterprise at the system level can be used (as a fallback parameter for logistics providers that are not individually configured). When the system performs the reconciliation timeliness judgment, it first queries the exclusive reconciliation cycle configuration of the logistics provider corresponding to the current waybill; if the logistics provider has not configured it separately, the system-level default reconciliation cycle is used.

[0058] If the actual elapsed days are greater than the standard time limit, meaning the number of days since the billing date exceeds the maximum reconciliation period allowed by the system, the system generates an overdue reconciliation flag indicating an overdue status. If the actual elapsed days are less than or equal to the standard time limit, meaning the reconciliation period has not yet expired since the billing date, the system generates a pending reconciliation flag indicating a normal status.

[0059] The "Overdue Reconciliation" flag is a status label indicating that the logistics provider's invoice for this waybill has been issued, and the elapsed time since the invoice date has exceeded the company's stipulated reconciliation period, but the reconciliation and confirmation process has not yet been completed. This flag indicates that there is a risk of delay in reconciliation for this waybill. If it is not addressed for a long period, it may affect the company's normal settlement and payment process with the logistics provider, affect the integrity of monthly financial closing, and may cause the company to lose the right to object to the invoice expenses due to exceeding the objection period stipulated in the contract.

[0060] The "Pending Reconciliation" flag is a status label used to indicate that the logistics provider's invoice for this waybill has been issued but the reconciliation and verification process has not yet been completed, and the actual elapsed time since the invoice date has not exceeded the standard time limit stipulated in the preset reconciliation cycle. This flag indicates that the reconciliation of this waybill is within the normal time frame and does not require immediate processing, but finance personnel still need to complete the verification before the deadline.

[0061] The system logically encapsulates the numerical value of the expense deviation and the status value of the reconciliation status identifier to construct a composite data structure that simultaneously contains quantitative indicators of expense compliance and qualitative indicators of reconciliation timeliness. This composite data structure contains two sub-fields: the first sub-field stores the expense deviation value (numerical data, in base currency), and the second sub-field stores the reconciliation status identifier (enumerated data, with values ​​of pending reconciliation or overdue reconciliation). The system records the entire composite data structure as the expense and timeliness verification result.

[0062] The cost and timeliness verification result is a composite data structure aggregating two verification dimensions. The first dimension is cost compliance, which uses a quantitative value of the cost deviation to characterize the degree of deviation between the theoretical freight cost and the actual expenditure. The second dimension is reconciliation timeliness, which uses a qualitative label of the reconciliation status identifier to characterize the reconciliation completion progress of the waybill at the current moment. Information from these two dimensions will be extracted and used separately later. The cost deviation value is extracted to generate cost deviation warnings, and the reconciliation status identifier is extracted to identify waybills that have not been reconciled on time and to generate reconciliation delay warnings. The cost and timeliness verification result also implicitly records the waybill number of the current waybill. This waybill number serves as an implicit association key, allowing subsequent reverse tracing of the complete waybill information in the waybill aggregation dataset from the cost and timeliness verification result.

[0063] In a preferred embodiment of the present invention, the comparison module is used to identify and mark waybills marked as overdue for reconciliation based on the reconciliation status identifier, thereby obtaining a reconciliation timeliness anomaly identifier; simultaneously, it extracts the pre-stored packaging weight and the billing weight from the waybill aggregation dataset, calculates the difference between the two, and obtains a weight deviation verification result, which may include: After constructing the cost and timeliness verification results, the system enters the stage of identifying reconciliation timeliness anomalies and verifying weight deviations. This step includes two parallel processing lines: one line is responsible for identifying and marking overdue waybills, and the other line is responsible for verifying the consistency between the pre-stored packaging weight and the billed weight. In the first processing line, identifying and marking reconciliation timeliness anomalies, the system iterates through the cost and timeliness verification results of all waybill collection records in the current batch, extracting all records whose reconciliation status flag is overdue. For each extracted record, the system locates the waybill collection record with the same waybill number in the waybill collection dataset based on the waybill number associated with the cost and timeliness verification result, and embeds an overdue attribute tag data structure into the metadata layer of the located waybill collection record.

[0064] The overdue attribute tag is a metadata structure attached to the waybill aggregation record, containing the following three subfields: the first subfield is the waybill number (string type, uniquely identifying the abnormal waybill); the second subfield is the overdue days (numeric type, its value equals the actual elapsed time in days minus the standard time limit in days); and the third subfield is the logistics provider information (string type, identifying the carrier logistics provider for this waybill, facilitating the statistical management of reconciliation timeliness anomalies by logistics provider). The system outputs the entire data structure carrying the above complete overdue attribute tag information as a reconciliation timeliness anomaly identifier.

[0065] The reconciliation timeliness anomaly identifier is a type of tagged data containing structured overdue information. Its core value lies in elevating the status information of a waybill that has been overdue for reconciliation from a simple right or wrong judgment to structured information that includes the number of overdue days (quantifying the severity of the overdue) and the responsible logistics provider (pointing to the entity that needs to be urged). This allows subsequent risk classification and early warning notifications to differentiate between priorities based on the specific value of the overdue days and accurately point to the responsible logistics provider that needs to be urged. The number of overdue days is the most critical quantitative indicator in the reconciliation timeliness anomaly identifier. Its value directly determines the level of timeliness risk. The larger the number of overdue days, the longer the reconciliation of the waybill has been delayed, the higher the risk level, and the more priority it needs to be addressed.

[0066] The second processing line is weight deviation verification. It should be noted that the scope of weight deviation verification is different from the scope of reconciliation timeliness anomaly identification. Weight deviation verification covers all waybill collection records in the current batch (regardless of their reconciliation status). As long as there is billing weight data, the verification will be performed. The reason is that weight deviation is an independent verification dimension (involving the physical weight data of the goods), which is physically independent from cost deviation (involving monetary data) and reconciliation timeliness (involving time data). Even if the cost deviation of a waybill is within the normal range and the reconciliation status is normal, its weight data may still be abnormal.

[0067] The system extracts the pre-stored packaging weight field value and the billing weight field value from each waybill collection record in the waybill collection dataset. The system checks whether the billing weight field is empty. If it is empty, meaning the logistics provider has not yet issued an invoice or the invoice does not provide billing weight data, the weight deviation verification for that waybill is skipped (because there is no comparison benchmark), and the weight verification status of that waybill is marked as deferred. If the billing weight field has a valid value, the following verification process continues.

[0068] Subtract the billed weight from the pre-stored packaging weight. The difference (retaining the sign, not the absolute value) is recorded as the original weight difference. The original weight difference refers to the numerical difference between the billed weight on the logistics provider's invoice and the packaging weight recorded by the shipper at the time of shipment. The unit of measurement is kilograms. The sign of the original weight difference has a clear physical meaning: a positive value indicates that the logistics provider's billed weight is higher than the shipment weight (possible reasons include inflated readings on the logistics provider's weighing equipment). The following reasons may be attributable to different logistics providers: Firstly, the logistics provider may use a more conservative volumetric weight conversion factor, leading to an overestimation of the chargeable weight. Secondly, the actual weight may increase due to moisture absorption during transport. A negative value indicates that the logistics provider's chargeable weight is lower than the outbound weight (possible reasons include the logistics provider offering weight discounts or concessions, using a more lenient volumetric weight conversion factor, or using a lower-than-expected weighing device). The absolute value of the original weight difference reflects the degree of consistency between the shipper and the logistics provider in determining the weight of the goods; a larger absolute value indicates poorer consistency and requires further verification.

[0069] Next, the system inputs the original weight difference into the preset weight compliance judgment logic and performs a two-level judgment. The first-level judgment extracts the value of the original weight difference, takes its absolute value, that is, if the original weight difference is positive, it keeps the original value, and if it is negative, it takes its opposite value. The absolute value of the original weight difference is then compared with the preset weight tolerance threshold. Weight tolerance threshold is a system configuration parameter, measured in kilograms, used to define the upper limit of the acceptable weight difference between the shipper and the logistics provider. The value of the weight tolerance threshold is determined based on three levels: First, the equipment error level, which uses the greater of the nominal accuracy values ​​of the shipper's and logistics provider's commonly used weighing equipment (e.g., the nominal accuracy of a weighbridge is usually a certain kilogram value), as the baseline component for equipment error contribution. Second, the environmental factor level, which considers the weight change caused by changes in humidity and temperature during transportation, leading to moisture absorption by packaging materials or evaporation of moisture from the goods. This change is determined based on the characteristics of the goods and the climatic conditions of the transportation route. For example, goods packaged in cardboard boxes may gain weight by absorbing moisture in the high-humidity environment of sea transport, while some goods with high moisture content may lose weight by losing water in a dry environment. Third, the industry practice level, which refers to the generally accepted weight deviation range (usually a certain percentage of the billing weight) for similar goods in the cross-border logistics industry. The system weights and sums the deviation components contributing from the three levels according to their respective weights, and uses the weighted sum as the comprehensive weight tolerance threshold. The weighting coefficients are set with the highest weight for the equipment error layer (because the accuracy deviation of the weighing equipment is the primary source of weight difference), followed by the environmental factors layer (because changes caused by environmental factors are conditional and uncertain), and the lowest weight for the industry practice layer (only for supplementary reference). For example, the weight for the equipment error layer is 0.5, the weight for the environmental factors layer is 0.3, and the weight for the industry practice layer is 0.2. The weight tolerance threshold can be configured differently in the system according to the logistics provider and cargo category dimensions to adapt to the weighing equipment accuracy of different logistics providers and the physical characteristics of different goods.

[0070] The second-level judgment performs branch processing based on the comparison results of the first-level judgment. If the absolute value of the original weight difference is less than or equal to the weight tolerance threshold, that is, the weight difference between the shipper and the logistics provider is within an acceptable and reasonable range, then the weight data of the waybill is determined to be normal, the system marks the weight verification status of the waybill as normal, and no weight deviation verification result is generated.

[0071] If the absolute value of the original weight difference exceeds the weight tolerance threshold, meaning the weight difference exceeds an acceptable reasonable range, the system performs a structured transformation on the original weight difference to generate a weight deviation verification result. The weight deviation verification result is a data structure containing the following three pieces of information: The first piece of information, the direction of deviation, is determined by the sign of the original weight difference. If the original weight difference is positive, meaning the billed weight is greater than the pre-stored packaging weight, the deviation direction is marked as "positive high," indicating that the logistics provider's billed weight is higher than the outbound weight, suggesting that the logistics provider may have increased the billed weight without notifying the shipper. If the original weight difference is negative, meaning the billed weight is less than the pre-stored packaging weight, the deviation direction is marked as "negative low," indicating that the logistics provider's billed weight is lower than the outbound weight. Possible reasons include the logistics provider offering weight discounts, using different measurement standards, or a reduction in cargo weight during transportation. Clearly distinguishing the deviation direction is crucial for subsequent manual verification. A positive high usually requires verification of the weight measurement method and pricing basis with the logistics provider, while a negative low, although advantageous to the shipper from a cost perspective, still requires confirmation of whether the weight reduction was due to shortages or damage to goods.

[0072] The second piece of information, deviation magnitude, is calculated by first dividing the absolute value of the original weight difference by the billed weight (absolute deviation divided by the billed weight), and then multiplying the ratio by 100% to obtain a percentage value. This percentage value represents the proportion of the weight deviation relative to the billed weight recognized by the logistics provider. Deviation magnitude is a relative quantitative indicator. Its advantage lies in eliminating the comparability issues caused by differences in the absolute weight of goods between different waybills. A 10-kilogram absolute deviation has completely different severity for goods weighing 100 kilograms and goods weighing 1,000 kilograms. The deviation magnitude indicator normalizes the absolute deviation by dividing it by the billed weight, allowing weight deviations between waybills of different weight classes to be compared and classified on a uniform scale.

[0073] The third information is the anomaly level. Based on the numerical range of the deviation, the corresponding anomaly level code is matched with the preset weight anomaly grading standard. The weight anomaly grading standard is a set of preset mapping rules from range to level in the system. For example, a deviation range of 0% to a certain threshold is mapped to a slight anomaly, a deviation range of 0% to a moderate anomaly (such as between a certain threshold and another threshold) is mapped to a severe anomaly, and a deviation range of 0% to a high threshold is mapped to a severe anomaly. The code value of the anomaly level will be extracted and used to determine the risk level of the weight anomaly warning in the subsequent process.

[0074] In a preferred embodiment of the present invention, the early warning module is used to generate cost deviation early warning, reconciliation delay early warning, and weight abnormality early warning based on the reconciliation timeliness anomaly identifier, the cost deviation value in the cost and timeliness verification result, and the weight deviation verification result according to a preset risk classification rule, thereby obtaining graded risk warning information, which may include: The system extracts the cost deviation value corresponding to each waybill from all cost and timeliness verification results, and substitutes the cost deviation value into a preset cost risk classification matrix for matching. The cost risk classification matrix is ​​a two-dimensional decision table data structure, which is constructed as follows: the horizontal axis dimension is defined as the sequence of large and small ranges of cost deviation values. Based on the statistical distribution characteristics of the cost deviation values ​​of all historical waybills, the system uses the quantile method to divide the range of cost deviation values ​​into multiple continuous numerical intervals. Specifically, it collects cost deviation value samples of all waybills that have completed reconciliation and have been manually confirmed to be without abnormalities within a preset statistical period in the past (e.g., a complete quarter in the past), and calculates the distribution characteristic parameters of this sample set, including the cost deviation values ​​corresponding to the median, upper quartile, upper decimal, and other quantile positions. Using these quantile positions as interval boundaries, the cost deviation value is divided into multiple intervals from zero to the maximum value. The interval between zero and the upper quartile is set as the first interval (corresponding to low risk, because the cost deviation of most normal shipments in history falls within this interval); the interval between the upper quartile and the upper decimal is set as the second interval (corresponding to medium risk, where the cost deviation begins to exceed the normal fluctuation range); and the interval where the cost deviation value exceeds the upper decimal is set as the third interval (corresponding to high risk, where the cost deviation deviates significantly from the historical normal level). The vertical axis dimension is defined as a risk level sequence, including three levels: low risk, medium risk, and high risk. Each level has a unique risk level code value (e.g., represented by a numerical code, with codes from smallest to largest corresponding to risk levels from lowest to highest). Each row of the cost risk grading matrix consists of a mapping relationship between a cost deviation value range and a risk level code. When performing matching, the system reads the specific value of the cost deviation value, scans the interval boundaries in the cost risk grading matrix row by row, finds the interval in which the cost deviation value falls, extracts the risk level code corresponding to that interval, and combines the risk level code with the original value of the cost deviation value to generate a cost deviation warning. The cost deviation warning is a data unit containing two sub-fields: risk level code (indicating the severity of the cost deviation) and deviation amount (i.e., the cost deviation value itself, for reference during manual verification of the specific deviation amount).

[0075] Cost deviation warnings are early alerts for abnormal billing accuracy. They are triggered when the actual charge from the logistics provider for a waybill differs from the theoretical freight calculated by the system based on the logistics provider's published tiered pricing rules, exceeding the normal fluctuation range. Cost deviation warnings are categorized based on the magnitude of the deviation. Low-risk warnings typically require routine checks by finance personnel. Medium-risk warnings suggest communication and confirmation with the logistics provider within the current reconciliation cycle. High-risk warnings require immediate initiation of a dedicated verification process, including checking the logistics provider's billing details, verifying the accuracy of cargo declaration information, and confirming with the logistics provider whether there are any anomalies in the pricing system or manual data entry errors.

[0076] From all the system's reconciliation timeliness anomaly markers, the system extracts the overdue days value from each anomaly marker and substitutes this overdue days value into a preset timeliness risk grading matrix for matching. The timeliness risk grading matrix is ​​a two-dimensional decision table data structure, constructed as follows: the horizontal axis is defined as a sequence of overdue days within time periods. Based on the urgency requirements of the company's reconciliation management, the overdue days are divided into multiple consecutive time periods from zero to the maximum possible overdue days. For example, overdue days between zero and the first time period boundary represent days that have just exceeded the standard time limit but have not yet exceeded it. A certain number of days is designated as the first period (corresponding to low risk, with slight delays in reconciliation but not yet urgent); the period between the boundary of the first period and the boundary of the second period (i.e., exceeding the standard time limit by a significant number of days) is designated as the second period (corresponding to medium risk, with significant delays in reconciliation, requiring finance or operations personnel to issue reminders to logistics providers); the period exceeding the boundary of the second period, i.e., with severe reconciliation delays, is designated as the third period (corresponding to high risk, with severe reconciliation delays potentially affecting month-end financial closing, quarterly audits, and contract performance assessments with logistics providers). The boundaries of each time period are set based on the company's internal financial management standard operating procedures, including monthly closing dates, quarterly financial audit dates, and contractual agreements with logistics providers regarding the period for disputed invoices. The vertical axis is defined as the timeliness risk level sequence, which uses the same risk level coding system as the cost risk grading matrix (to maintain consistency of the three-dimensional risk warning at the coding level). The system matches the overdue days with each time period interval in the timeliness risk grading matrix, finds the time period interval in which the overdue days fall, extracts the corresponding risk level code, and combines the risk level code with the responsible entity information carried in the reconciliation timeliness anomaly identifier, i.e., the logistics provider information, to generate a reconciliation delay warning.

[0077] Reconciliation delay warning is a signal indicating that the reconciliation process is overdue. The key information for reconciliation delay warning is the number of overdue days (quantifying the severity of the delay) and the responsible logistics provider (identifying the entity that needs to be urged). The responsible entity field directly identifies the specific logistics provider, allowing the warning information to be accurately routed to the finance or operations personnel responsible for liaising with that logistics provider. It supports targeted urging, that is, sending reconciliation urging notification emails or system messages to specific logistics providers, requiring them to complete the submission of invoices or supplementation of reconciliation data within a specified period.

[0078] In all weight deviation verification results of the system, the absolute value of the original weight difference is extracted from each result. This absolute value is then substituted into a preset weight risk classification matrix for matching. The weight risk classification matrix is ​​a two-dimensional decision table data structure. Its construction method is similar to the two classification matrices mentioned above, but the measurement method of the horizontal axis is different. The horizontal axis is defined as the sequence of multiples between the absolute value of the original weight difference and the weight tolerance threshold. The reason for using the multiple relationship rather than the absolute value of the weight difference itself as the classification basis is that the goods in different waybills naturally have different weight classes. For goods with a relatively light total weight, such as a document or a small package, a deviation of a few kilograms may be very significant. However, for goods with a relatively heavy total weight, such as a fully loaded container, the same weight deviation may be completely within the normal range. By dividing the weight difference by a tolerance threshold (which is differentiated according to cargo type and logistics provider) to obtain a multiple, the influence of absolute weight magnitude can be eliminated, achieving a unified grading standard across waybills and cargo types. A typical way to divide the multiple range is as follows: the range where the absolute value of the original weight difference is greater than one time but less than or equal to two times the weight tolerance threshold is mapped to a low-risk level, indicating that the weight deviation is slightly outside the normal range, possibly due to occasional weighing errors or fluctuations in the weight of packaging materials, and can be automatically monitored during the next reconciliation; the range where the absolute value of the original weight difference is greater than two times but less than or equal to five times the weight tolerance threshold is mapped to a low-risk level, indicating that the weight deviation is slightly outside the normal range, possibly due to occasional weighing errors or fluctuations in the weight of packaging materials, and can be automatically monitored during the next reconciliation; the range where the absolute value of the original weight difference is greater than two times but less than or equal to five times the weight tolerance threshold is mapped to a low-risk level. The threshold range is mapped to a medium-risk level, indicating a significant weight deviation. It is recommended to verify the weight measurement method with the logistics provider or request weighing certificates from them. The range where the absolute value of the original weight difference is greater than five times the weight tolerance threshold is mapped to a high-risk level, indicating a severely abnormal weight deviation. This is highly likely due to errors in the logistics provider's billing system, data entry errors, loss or damage of goods during transportation, or malicious misreporting of weight. An immediate special investigation is required, including obtaining the original weighing records from the logistics provider, verifying the weight handover records at each stage of transportation, and, if necessary, dispatching personnel to the logistics provider's warehouse for on-site re-weighing. The system divides the absolute value of the original weight difference by the weight tolerance threshold to obtain a multiple value. It then matches this multiple value within the multiple range in the weight risk grading matrix, extracts the corresponding risk level code, and combines this risk level code with the deviation direction information carried in the weight deviation verification result to generate a weight anomaly warning.

[0079] Weight anomaly alerts are abnormal signals indicating inconsistencies in cargo weight determination. Unlike cost deviation alerts (which focus on the accuracy of billing amounts), weight anomaly alerts focus on the consistency of the basic physical quantity and cargo weight on which billing is based. Weight anomalies can be the root cause of cost deviations. If the logistics provider changes the weight value used for billing without notifying the shipper, it will inevitably lead to cost deviations. Therefore, there is a certain correlation between weight anomaly alerts and cost deviation alerts. When generating tiered risk alert information, the system can associate and label situations where the same waybill triggers both cost deviation alerts and weight anomaly alerts simultaneously. For example, an associated alert label can be added to the alert information to indicate to manual verification personnel that these two types of alerts may have a common cause, namely, a problem with the logistics provider's weight determination.

[0080] The system encapsulates the three types of early warning data units generated in the aforementioned three steps—cost deviation early warning data unit (including risk level code and deviation amount), reconciliation delay early warning data unit (including risk level code and responsible logistics provider), and weight anomaly early warning data unit (including risk level code and deviation direction)—in a unified early warning message format. The unified early warning message format refers to a standardized data structure template preset in the system. This template adopts a nested data structure organization method. The outermost layer is the early warning package, which includes the waybill number (primary key identifier), the early warning generation timestamp (recording the time of early warning generation), and three early warning sub-packages (corresponding to the three dimensions of cost deviation, reconciliation delay, and weight anomaly, respectively). Each early warning sub-package adopts a unified triplet structure, with a risk level coding field (using the same coding system to facilitate cross-dimensional comparison and comprehensive sorting), a risk quantification indicator field (storing the quantification indicator specific to this dimension; the cost deviation dimension stores the deviation amount, the reconciliation delay dimension stores the overdue days, and the weight anomaly dimension stores the deviation percentage), and a risk-specific information field (storing the qualitative or auxiliary information specific to this dimension; the risk-specific information for the cost deviation dimension can be empty or store the deviation direction, the reconciliation delay dimension stores the responsible logistics provider identifier, and the weight anomaly dimension stores the deviation direction). For a specific waybill, if the verification result of a certain dimension is normal (i.e., no warning is triggered), the content of the warning sub-package corresponding to that dimension will be empty or marked as normal. The purpose of adopting a unified warning message format is to ensure that the warning information of different batches, different waybills, and different logistics providers has a completely consistent data representation (consistent field names, consistent field types, and consistent field order). This allows downstream warning display systems (such as dashboards and warning list pages), notification push systems (such as email notifications, instant message notifications, and system in-site notifications), and data analysis systems (such as risk trend analysis and logistics provider service quality assessment) to parse and process the warnings with a unified interface, without having to write different parsing logic for different types of warnings.

[0081] The system outputs the packaged early warning information as a tiered risk early warning message. Each early warning record in the tiered risk early warning message clearly indicates the risk dimension category (expense deviation, reconciliation delay or weight anomaly), the severity level of the risk (low risk, medium risk or high risk), and the quantitative indicators of the risk (deviation amount, overdue days or deviation percentage).

[0082] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the system as described above. All implementations in the above system embodiments are applicable to this embodiment and can achieve the same technical effects.

[0083] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the system as described above. All implementations in the above system embodiments are applicable to this embodiment and can achieve the same technical effects.

[0084] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An automated verification and risk warning system for cross-border logistics compliance, characterized in that, include: The aggregation module is used to build a waybill aggregation dataset based on the shipping records generated by cross-border logistics outbound nodes and transfer outbound nodes, as well as actual billing data. The matching module is used to extract the logistics provider, destination country zone, and declared chargeable weight from the waybill aggregation dataset. Based on the logistics provider and destination country zone, it retrieves the corresponding cost rules and obtains a family of parallel line segments consisting of the mapping between chargeable weight range and maintenance unit price. Based on the family of parallel line segments, it constructs a cost determination circle by combining the declared chargeable weight and preset tolerance threshold, and calculates the rule matching freight through effective tangency constraints. The verification module is used to calculate the difference between the rule-matched shipping cost and the actual invoice shipping cost to obtain the cost deviation value; at the same time, it compares the invoice date with the preset reconciliation cycle to generate a reconciliation status identifier; and combines the cost deviation value and the reconciliation status identifier to form the cost and timeliness verification result. The comparison module is used to identify and mark waybills marked as overdue for reconciliation based on the reconciliation status identifier, and obtain the reconciliation timeliness anomaly identifier; at the same time, it extracts the pre-stored packaging weight and the billing weight from the waybill collection dataset, calculates the difference between the two, and obtains the weight deviation verification result. The early warning module is used to generate cost deviation early warning, reconciliation delay early warning and weight abnormality early warning according to the reconciliation timeliness anomaly identifier, cost deviation value in cost and timeliness verification results and weight deviation verification results, and obtain graded risk early warning information.

2. The automated verification and risk warning system for cross-border logistics compliance according to claim 1, characterized in that, The waybill collection dataset includes waybill number, logistics provider, destination country zone, declared chargeable weight, pre-stored packaging weight, outbound timestamp, and data from the actual invoice, such as invoice freight, invoice chargeable weight, and invoice date; the reconciliation status indicators include no invoice issued, pending reconciliation, and overdue reconciliation.

3. The automated verification and risk warning system for cross-border logistics compliance according to claim 2, characterized in that, The family of parallel line segments includes: Based on the logistics provider field in the waybill aggregation dataset, locate the dedicated pricing strategy library corresponding to the logistics provider, and based on the destination country partition field in the waybill aggregation dataset, index the tiered rate table adapted to the flow direction in the dedicated pricing strategy library to obtain the original cost rule set of the logistics provider under the destination country partition. The original set of cost rules is structured and parsed to extract the lower limit of chargeable weight, the upper limit of chargeable weight, and the corresponding unit freight maintenance price defined in each rule. The correspondence between each set of chargeable weight intervals and maintenance prices is abstracted into an independent line segment element, where the horizontal axis of the line segment represents the weight interval and the vertical axis represents the maintenance price. The extracted line segment elements are linearly arranged in ascending order of the minimum chargeable weight, thus constructing a family of parallel line segments consisting of multiple parallel and sequentially distributed line segments.

4. The automated verification and risk warning system for cross-border logistics compliance according to claim 3, characterized in that, The calculation rules for matching shipping costs include: Using the declared chargeable weight extracted from the waybill collection dataset as the abscissa and the preset tolerance threshold as the radius, a cost determination circle is constructed in the weight-cost two-dimensional coordinate system. The coverage of the cost determination circle defines the allowable fluctuation range of the declared chargeable weight. The cost determination circle is translated and scanned along the weight axis, traversing each line segment element in the parallel line segment family, detecting whether the circumference of the cost determination circle intersects with the geometric contour of the line segment element, and filtering out all candidate line segments that meet the geometric intersection conditions. Apply a valid tangency constraint to the candidate line segment set. The valid tangency constraint requires that the cost determination circle and the line segment element form a unique contact point within the billing weight range and that the slope of the line segment at the contact point is continuous. Eliminate invalid line segments that only contact at the endpoints of the line segments and determine the unique target line segment that satisfies the valid tangency constraint. Read the maintenance unit price mapped to the target line segment, perform a multiplication operation between the declared chargeable weight and the maintenance unit price, and determine the result as the rule-matching freight.

5. The automated verification and risk warning system for cross-border logistics compliance according to claim 4, characterized in that, The determination of the unique target line segment that satisfies the effective tangency constraint includes: Based on the candidate line segment set, the weight interval boundary coordinates and corresponding maintenance unit price of each line segment element are extracted one by one to construct the line segment topology structure for geometric relationship determination; Based on the line segment topology, it is determined whether the circumference of the cost determination circle and the geometric contour of the line segment element form a unique contact point within the interval, and the continuity of the line segment slope at the contact point is verified to eliminate discontinuous contact caused by the truncation of the line segment endpoints. Based on the contact point determination and slope verification results, invalid line segments that only form contact at the endpoints of the line segments are eliminated, and valid line segments that achieve effective tangency within the interval are retained; Based on the membership relationship between the billing weight range mapped by the retained valid line segments and the declared billing weight, a unique target line segment that satisfies the valid tangency constraint is determined.

6. The automated verification and risk warning system for cross-border logistics compliance according to claim 5, characterized in that, The cost and timeliness verification results include: Retrieve the bill freight value from the waybill collection dataset, use the rule-matched freight as the benchmark reference value, and perform an absolute value subtraction operation with the bill freight value to obtain the cost deviation value, which represents the degree of deviation between the theoretical cost and the actual expenditure. Extract the billing date from the waybill collection dataset, calculate the actual elapsed time between the billing date and the current processing time, compare the actual elapsed time with the standard time limit specified in the preset reconciliation cycle, if the actual elapsed time exceeds the standard time limit, generate an overdue reconciliation flag indicating an overdue status, if the actual elapsed time does not exceed the standard time limit, generate a pending reconciliation flag indicating a normal status. By logically encapsulating the cost deviation value and the reconciliation status identifier, a composite data structure is constructed, which includes quantitative indicators of cost compliance and qualitative indicators of reconciliation timeliness, to obtain the cost and timeliness verification results.

7. The automated verification and risk warning system for cross-border logistics compliance according to claim 6, characterized in that, The weight deviation verification results include: Extract all records with billing weight data from the waybill collection dataset one by one, locate the corresponding waybill entity based on the waybill index number bound in the record, embed overdue attribute tags in the metadata layer of the waybill entity, and generate a reconciliation timeliness anomaly identifier carrying waybill number, overdue days and logistics provider information. Based on the waybill index number in the reconciliation timeliness anomaly identifier, reverse index to waybill collection dataset, extract the pre-stored packaging weight field value and the billing weight field value under the same waybill index number from the waybill collection dataset, calculate the billing weight value and the pre-stored packaging weight value to obtain the original weight difference value representing the deviation between the actual weight and the billing weight. The original weight difference is input into the preset weight compliance judgment logic. The absolute value of the original weight difference is compared with the preset weight fault tolerance threshold. If the absolute value of the original weight difference is greater than the weight fault tolerance threshold, the original weight difference is converted into a weight deviation verification result including deviation direction, deviation magnitude and abnormality level.

8. The automated verification and risk warning system for cross-border logistics compliance according to claim 7, characterized in that, The tiered risk warning information includes: Substitute the cost deviation value into the preset cost risk classification matrix, match the corresponding cost risk level according to the numerical range of the cost deviation value, and generate a cost deviation warning including risk level code and deviation amount. Substitute the number of overdue days carried in the reconciliation timeliness anomaly identifier into the preset timeliness risk classification matrix, match the corresponding timeliness risk level according to the time period interval where the overdue days are located, and generate a reconciliation delay warning including risk level code and responsible party; Substitute the absolute value of the original weight difference in the weight deviation verification result into the preset weight risk classification matrix, match the corresponding weight risk level according to the multiple relationship between the absolute value of the difference and the weight fault tolerance threshold, and generate a weight anomaly warning including risk level code and deviation direction. The cost deviation warning, reconciliation delay warning, and weight anomaly warning are encapsulated in a unified message format to construct a structured data packet, thereby obtaining tiered risk warning information.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the system as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the system as described in any one of claims 1 to 8.

Images