Goods distribution method, system, electronic device and storage medium

Abstract

The application discloses a kind of parcel distribution method, system, electronic equipment and storage medium, the parcel distribution method includes: obtaining the business characteristic data of current waybill and historical waybill data;Determine the target parcel distribution scene type corresponding to current waybill based on business characteristic data;Determine target parcel distribution strategy based on target parcel distribution scene type and historical waybill data;Using target parcel distribution strategy determines the parcel distribution identifier of each to-be-shipped package of current waybill, parcel distribution identifier includes waybill main identifier and package flow direction identifier.The application determines the target parcel distribution scene type corresponding to current waybill based on business characteristic data, determines target parcel distribution strategy based on target parcel distribution scene type and historical waybill data, and using target parcel distribution strategy determines the parcel distribution identifier of each to-be-shipped package of current waybill, can determine parcel distribution identifier using target parcel distribution strategy adapted to the business characteristic data of waybill, can improve parcel distribution accuracy.

Description

Technical Field

[0001] This application relates to the field of logistics technology, specifically to a sorting method, system, electronic device, and storage medium. Background Technology

[0002] In the field of logistics sorting, parcel sorting refers to the process of allocating parcels to corresponding slots or next stations based on attributes such as destination and transit route. With the growth of express delivery volume, parcels awaiting sorting in the same batch often mix different situations such as direct pickup from the branch, consolidated pickup, one parcel per shipment, and multiple parcels per shipment. Moreover, the business characteristics such as the destination distribution of waybills and the proportion of multiple parcels per shipment show obvious dynamic fluctuations, which places higher demands on the flexibility and adaptability of the sorting system.

[0003] Existing sorting methods generally employ fixed rules for sorting, which are pre-set to sort by item and assign each package in the waybill to its own destination independently, and remain unchanged throughout the entire batch operation.

[0004] However, existing sorting methods are ill-suited to complex scenarios such as fluctuating destination consistency within the same shipment, varying proportions of multiple items per shipment, and mixed pickup methods. For example, when a shipment contains multiple items and all sub-items have the same destination, forcibly using piece-by-piece sorting can easily lead to multiple sub-items from the same shipment being split into different compartments or even different shipments, resulting in mis-sorting of multiple items within a single shipment. Furthermore, it may cause issues such as scattered shipments along the same route and repeated sorting during transit, significantly increasing the number of operations and logistics costs. Therefore, it is clear that existing sorting methods lack sufficient accuracy in complex and mixed scenarios. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application provides a sorting method, system, electronic device, and storage medium. This application determines the target sorting scenario type corresponding to the current waybill based on business characteristic data, determines the target sorting strategy based on the target sorting scenario type and historical waybill data, and uses the target sorting strategy to determine the sorting identifier for each parcel to be shipped in the current waybill. It can use a target sorting strategy adapted to the business characteristic data of the waybill to determine the sorting identifier, which can improve the accuracy of sorting and enable front-line staff to intuitively obtain the flow and ownership information of the parcels through the sorting identifier.

[0006] To address the above problems, this application provides the following technical solution:

[0007] In a first aspect, embodiments of this application provide a cargo sorting method, including: acquiring business characteristic data of the current waybill and historical waybill data, wherein the business characteristic data includes pickup information of the current waybill, destination information of all parcels to be shipped in the current waybill, destination route type information and customer information; Based on the business characteristic data, the target cargo distribution scenario type corresponding to the current waybill is determined, wherein the target cargo distribution scenario type is direct delivery type, transit cargo distribution type, or multiple-item cargo distribution type. The step of determining the target cargo distribution scenario type corresponding to the current waybill based on the business characteristic data includes: Based on the pickup information, the packaging method feature information of the current waybill is determined; The destination consistency feature information of the current waybill is determined based on the destination information; The transit link characteristic information of the current waybill is determined based on the destination route type information; The target cargo distribution scenario type is determined based on at least one of the following features: the packaging method feature information, the destination consistency feature information, and the transit link feature information; The target cargo allocation strategy is determined based on the target cargo allocation scenario type and the historical waybill data. The target sorting strategy is used to determine the sorting identifier for each parcel to be shipped in the current waybill. The sorting identifier includes the waybill master identifier and the parcel flow identifier. Both the waybill master identifier and the parcel flow identifier are uniquely bound to the current waybill.

[0008] Optionally, determining the target cargo distribution scenario type based on at least one of the packaging method feature information, the destination consistency feature information, and the transit link feature information includes: When the first preset judgment condition is met, the target cargo distribution scenario type is determined to be direct flow type. The first preset judgment condition includes at least two of the following conditions: the packaging method feature information indicates that the packages are packaged according to the flow direction; the destination consistency feature information indicates that all the packages to be shipped have the same destination; the transit link feature information indicates that the destination line type is a dedicated line; and the customer information and the historical waybill data determine that the current customer's historical waybill has no transit record. When the second preset judgment condition is met, the target cargo distribution scenario type is determined to be the transit cargo distribution type. The second preset judgment condition includes at least two of the following conditions: the packaging method feature information does not indicate packaging according to flow direction, the destination consistency feature information indicates that the destinations of the parcels to be shipped are inconsistent, the transit link feature information indicates that the destination route type is not a dedicated line, and the proportion of transit sorting records of the current customer's historical waybill is greater than the first preset ratio, based on the customer information and the historical waybill data. When the third preset judgment condition is met, the target cargo distribution scenario type is determined to be the multi-item cargo distribution type. The third preset judgment condition includes determining that the proportion of packages with the same destination among all the packages to be shipped is greater than the second preset proportion based on the destination consistency feature information, determining that multiple packages in the current customer's historical waybill data are mixed in different places based on the customer information and the historical waybill data, determining that there are cargo distribution errors in the current customer's historical waybill data, and that the packaging method feature information does not indicate at least two of the following conditions in the whole shipment packaging.

[0009] Optionally, determining the sorting identifier for each parcel to be shipped in the current waybill using the target sorting strategy includes: The preset algorithm corresponding to the target cargo allocation strategy is used to determine at least part of the transportation process for each parcel to be shipped in the current waybill based on the business feature data; Determine the target identification rule corresponding to the target distribution strategy in the preset identification rule base; The sorting identifier is determined based on the transportation process of each package to be shipped, the mapping relationship information of the identifier field in the target identifier rule, and the encoding rule.

[0010] Optionally, determining the target cargo allocation strategy based on the target cargo allocation scenario type and the historical waybill data includes: Based on the target cargo distribution scenario type, the historical waybill data, and the logistics links corresponding to the preset multiple cargo distribution scenario types, a directed graph of target logistics link cost propagation corresponding to the target cargo distribution scenario type is determined. The nodes in the directed graph of target logistics link cost propagation represent the logistics nodes of the current waybill. Each node corresponds to the operation cost of the associated cargo distribution granularity. The edge weights in the directed graph of logistics link cost propagation are the conditional probabilities of the propagation of anomalies from upstream nodes to downstream nodes, and the conditional probabilities are functions of the cargo distribution granularity, which is used to represent the fineness of cargo distribution. Based on the directed graph of the target logistics link cost propagation and multiple preset, distinct sorting granularities, the sorting cost of the current waybill under the constraint of each sorting granularity is determined. The multiple distinct sorting granularities include regional level, provincial level, city level, district level, branch level, and delivery person designated level. The target granularity of the current waybill is determined based on the allocation cost of the current waybill under the constraints of each allocation granularity. The target allocation strategy is determined based on the target allocation granularity.

[0011] Optionally, determining the directed graph of the target logistics link cost propagation corresponding to the target distribution scenario type based on the target distribution scenario type, the historical waybill data, and the logistics links corresponding to multiple preset distribution scenario types includes: Based on the target cargo distribution scenario type and the logistics links corresponding to the various cargo distribution scenario types, a first logistics link cost propagation directed graph is determined. The node and edge weights of the first logistics link cost propagation directed graph are adjusted based on the business characteristic data to obtain the second logistics link cost propagation directed graph. Based on the historical waybill data, the target cargo distribution scenario type, the destination route type information, and the customer information, determine the normal operation cost and abnormal operation cost of each node in the second logistics link cost propagation directed graph, and obtain the target logistics link cost propagation directed graph.

[0012] Secondly, embodiments of this application provide a sorting system, the sorting system comprising: The acquisition module is used to acquire the business characteristic data of the current waybill and historical waybill data. The business characteristic data includes the pickup information of the current waybill, the destination information of all parcels to be shipped in the current waybill, the destination route type information, and the customer information. The cargo distribution strategy determination module is used to determine the target cargo distribution scenario type corresponding to the current waybill based on the business characteristic data, wherein the target cargo distribution scenario type is direct delivery type, transit cargo distribution type, or one shipment with multiple items cargo distribution type. The target cargo allocation strategy is determined based on the target cargo allocation scenario type and the historical waybill data. The identification module is used to determine the sorting identifier of each parcel to be shipped in the current waybill using the target sorting strategy. The sorting identifier includes the waybill master identifier and the parcel flow identifier, and both the waybill master identifier and the parcel flow identifier are uniquely bound to the current waybill.

[0013] Thirdly, embodiments of this application provide an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the sorting method as described in the first aspect.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing an executable program, which is executed by a processor to implement the sorting method as described in the first aspect.

[0015] This application provides a sorting method, system, electronic device, and storage medium. This application determines the target sorting scenario type corresponding to the current waybill based on business characteristic data, determines the target sorting strategy based on the target sorting scenario type and historical waybill data, and uses the target sorting strategy to determine the sorting identifier of each parcel to be shipped in the current waybill. It can use a target sorting strategy adapted to the business characteristic data of the waybill to determine the sorting identifier, which can improve the sorting accuracy and enable front-line staff to intuitively obtain the flow and ownership information of the parcels through the sorting identifier. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the sorting method provided in the embodiments of this application.

[0017] Figure 2 This is a schematic diagram of the structure of the sorting system provided in the embodiments of this application.

[0018] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0019] Figure 4 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0022] This application provides a sorting method, system, electronic device, and storage medium. It determines the target sorting scenario type corresponding to the current waybill based on business characteristic data, determines the target sorting strategy based on the target sorting scenario type and historical waybill data, and uses the target sorting strategy to determine the sorting identifier for each parcel to be shipped in the current waybill. It can use a target sorting strategy adapted to the business characteristic data of the waybill to determine the sorting identifier, which can improve the sorting accuracy and enable front-line staff to intuitively obtain the flow and ownership information of the parcels through the sorting identifier.

[0023] The sorting method provided in this application will be described in detail below with reference to the accompanying drawings.

[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating the sorting method provided in an embodiment of this application. For example... Figure 1 As shown, the sorting method includes steps S100 to S400.

[0025] Step S100: Obtain the business characteristic data of the current waybill and the historical waybill data.

[0026] The business characteristic data includes the current waybill's pickup information, the destination information of all parcels to be shipped under the current waybill, the destination route type information, and the customer information.

[0027] The current waybill's pickup information may include pickup point number, pickup timestamp, packing method, whether on-site loading was required, pickup personnel's skill level, and declared value of the goods. Packing methods include loose packages, packing according to destination, and full shipment packing.

[0028] In addition, pickup information may also include parcel dimension information, which includes at least one of the following: total number of parcels on the waybill, weight and volume of each parcel, total weight and volume, and cargo type of each parcel.

[0029] In some embodiments, the destination information of the parcel to be shipped may include at least one of the following: the administrative code of the destination, latitude and longitude coordinates, the delivery area to which the destination belongs, and the last delivery station number.

[0030] Destination route type information may include dedicated lines, large-volume LTL, general LTL, and shared freight. Customer information may include at least one of the following: customer identifier, historical average monthly shipment volume, waybill timeliness requirements, percentage of insured value, historical abnormal complaint rate, and historical percentage of multiple items in a single shipment. Cargo type may include high-value goods, general cargo, fragile items, and dangerous goods.

[0031] In order to determine the standardized distribution scenario type based on complex and disordered business characteristic data, it is necessary to extract features and classify the above data. Therefore, after performing step S100, this method performs step S200.

[0032] Step S200: Determine the target cargo distribution scenario type corresponding to the current waybill based on business characteristic data.

[0033] The target cargo distribution scenarios are categorized as direct shipment, transit distribution, or multi-item distribution within a single shipment. Different scenarios correspond to different cargo transportation processes and labeling management requirements. For example, direct shipments do not require secondary sorting at transit points, while multi-item distribution requires close attention to the logical relationships between multiple items to ensure simultaneous delivery of all packages within a single shipment.

[0034] The purpose of step S200 is to determine standardized cargo allocation scenario types based on complex and disordered business characteristic data, thereby providing clear cargo allocation scenario type information for the subsequent generation of target cargo allocation strategies. Existing technologies neglect the business characteristic attributes of the waybill itself, while this step achieves automatic classification of cargo allocation scenario types through comprehensive analysis of multi-dimensional features.

[0035] Step S200 includes steps S210 to S240.

[0036] Step S210: Determine the packaging method feature information of the current waybill based on the pickup information.

[0037] The pickup process is the starting point of the transportation process, and the pickup method directly determines whether the package needs to be unpacked and sorted later.

[0038] The packaging method characteristics of the current waybill are obtained by parsing the packaging method field in the pickup information.

[0039] For example, an enumeration type is used to represent the packaging method. When the packaging method field value is 1, the packaging method characteristic information is determined to be packaging by flow direction, that is, the collection point has sorted and loaded the packages according to their destination. When the packaging method field value is 2, the packaging method characteristic information is loose packaging. When the packaging method field value is 3, the packaging method characteristic information is whole package packaging.

[0040] Since the packaging method only reflects the loading status at the origin of the waybill, a single waybill may contain multiple parcels, and whether these parcels have the same destination will determine the subsequent transportation process. To assess the concentration of parcel flow, step S220 is executed.

[0041] Step S220: Determine the destination consistency feature information of the current waybill based on the destination information.

[0042] A single waybill may contain multiple parcels. Whether these parcels have the same destination is the key criterion for determining whether it is a scenario of multiple items being shipped in one shipment.

[0043] For example, firstly, the destination address field of all parcels to be shipped is extracted from the destination information. Then, keywords representing the province, city, and district levels of administrative divisions are extracted from the destination address field. Subsequently, the intersection of the destination administrative division keyword sets of all parcels is performed. If the intersection size equals the total number of parcels, the destination consistency feature is determined to be consistent; if the intersection size is less than the total number of parcels, the destination consistency feature is determined to be inconsistent. When the destination consistency feature is inconsistent, the number of parcels with the same destination in the current waybill is counted, and their proportion of the total number of parcels is calculated.

[0044] By using destination consistency information, the degree of dispersion of parcel flow can be quantified, providing data support for subsequent identification of multi-item splitting scenarios.

[0045] Whether a package can truly achieve direct delivery within a physical network is also constrained by the type of objective route.

[0046] Step S230: Determine the transit link characteristic information of the current waybill based on the destination route type information.

[0047] Specifically, this step first requires extracting the destination route type field from the destination route type information. If the destination route type field indicates a dedicated line or direct large-ticket shipment, the transit link characteristic information is determined to be direct and requires no transit. If the destination route type field indicates less-than-truckload (LTL) or carpooling, the transit link characteristic information is determined to be that transit and sorting are required.

[0048] Step S240: Determine the target cargo distribution scenario type based on at least one of the following features: packaging method feature information, destination consistency feature information, and transit link feature information.

[0049] When the first preset judgment condition is met, the target cargo distribution scenario type is determined to be direct flow type. The first preset judgment condition includes at least two of the following conditions: packaging method feature information indicates packaging according to flow direction, destination consistency feature information indicates that all parcels to be shipped have the same destination, transit link feature information indicates that the destination line type is dedicated line, and determining that the current customer's historical transit records have no transit records based on customer information and historical waybill data.

[0050] When the second preset judgment condition is met, the target cargo distribution scenario type is determined to be transit cargo distribution type. The second preset judgment condition includes at least two of the following conditions: the packaging method feature information does not indicate packaging according to flow direction, the destination consistency feature information indicates that the destination of the parcel to be transported is inconsistent, the transit link feature information indicates that the destination route type is not a dedicated line, and the proportion of transit sorting records of the current customer's historical waybill is greater than the first preset ratio, based on customer information and historical waybill data.

[0051] The reason for setting a first preset ratio is that if the proportion of historical transfer sorting records is lower than the first preset ratio, it means that most packages can be delivered directly, and forcibly marking them as transit-type shipments would lead to a waste of sorting resources. If the proportion of historical transfer sorting records is higher than the first preset ratio, then secondary sorting at the transit center is necessary. The first preset ratio can be flexibly adjusted according to the actual transit capacity of different branches.

[0052] When the third preset judgment condition is met, the target cargo distribution scenario type is determined to be a multi-item cargo distribution type. The third preset judgment condition includes determining that the proportion of packages with the same destination among all parcels to be shipped is greater than the second preset proportion based on the destination consistency feature information, determining that multiple parcels in the current customer's historical waybill data are mixed in different places based on customer information and historical waybill data, determining that there are cargo distribution errors in the current customer's historical waybill data, and at least two of the following conditions are not indicated by the packaging method feature information:

[0053] The reason for setting a second preset ratio is that it is economically efficient and feasible to transport packages to be shipped simultaneously only when the proportion of packages to the same destination is greater than the second preset ratio.

[0054] In the above judgment process, in order to prevent rule conflicts, the judgment priority of the target distribution scenario type is set.

[0055] For example, the priority for determining the target distribution scenario type is that the single-shipment multi-item distribution type is higher than the direct-shipment type, and the direct-shipment type is higher than the transit distribution type.

[0056] By using the multi-dimensional feature joint judgment steps described above, the accuracy of the target distribution scenario type can be improved, providing an accurate foundation for the generation of subsequent target distribution strategies.

[0057] In step S240 above, a pre-trained machine learning-based scenario classification model can be used to determine the target cargo distribution scenario type based on at least one of the following features: packaging method feature information, destination consistency feature information, and transit link feature information.

[0058] Step S300: Determine the target cargo allocation strategy based on the target cargo allocation scenario type and historical waybill data.

[0059] Specifically, in terms of technical implementation, it is first necessary to preset multiple sorting strategies, determine the sorting accuracy of each sorting strategy for the target sorting scenario type based on historical waybill data, and determine the sorting strategy with the highest sorting accuracy for the target sorting scenario type as the target sorting strategy.

[0060] Based on the pre-defined correspondence between sorting scenario types and sorting strategies, and the pre-defined multiple sorting strategies, the initial sorting strategy corresponding to the target sorting scenario type is determined. The initial sorting strategy is then updated based on historical waybill data to improve sorting accuracy, thus obtaining the target sorting strategy.

[0061] Specifically, first determine the target allocation granularity for the current waybill, and then determine the target allocation strategy based on the target allocation granularity.

[0062] Existing sorting systems generally use a fixed sorting granularity to perform sorting, that is, they are pre-set to sort by item and assign each package in the waybill to its own destination independently, and this remains unchanged throughout the entire batch operation.

[0063] However, a fixed granularity of sorting is difficult to adapt to complex scenarios such as fluctuating destination consistency, varying proportions of multiple items per shipment, and mixed pickup methods within the same batch of waybills. For example, when a shipment contains multiple items and all sub-items have the same destination, forcibly sorting by item can easily lead to multiple sub-items from the same shipment being split into different compartments or even different trains, resulting in missorting of multiple items per shipment. Furthermore, it may cause issues such as scattered shipments along the same route and repeated sorting during transit, significantly increasing the number of operations and logistics costs. Therefore, it is evident that existing sorting methods lack sufficient accuracy in complex and mixed scenarios.

[0064] This application can determine the cascading impact of sorting granularity on downstream transit and distribution, last-mile delivery, and anomaly handling, ensuring that the target sorting granularity determined at the sorting end is optimal across the entire supply chain. This improves sorting accuracy while reducing overall logistics and transportation costs. For the specific algorithm, please refer to steps S310 to S340.

[0065] Step S310: Based on the target cargo distribution scenario type, historical waybill data, and the logistics links corresponding to various preset cargo distribution scenario types, determine the directed graph of the target logistics link cost propagation corresponding to the target cargo distribution scenario type.

[0066] In the directed graph of the target logistics link cost propagation, each node represents a logistics node of the current waybill, and each node corresponds to the operation cost of the associated allocation granularity. The edge weights in the directed graph of the logistics link cost propagation are the conditional probabilities of anomalies from upstream nodes propagating to downstream nodes, and the conditional probabilities are functions of the allocation granularity. The allocation granularity is used to represent the fineness of the allocation.

[0067] When the target distribution scenario is direct-flow type, the logistics link includes pickup node, sorting node, trunk node, terminal node and exception node.

[0068] When the target cargo distribution scenario is a transit cargo distribution type, the logistics link adds transit to port nodes, transit re-distribution nodes and transit shipping nodes compared to the direct delivery type logistics link.

[0069] When the target distribution scenario is a multi-item distribution type, the logistics link adds splitting nodes and item combination flow nodes compared to the direct-flow logistics link.

[0070] Understandably, the finer the sorting granularity, the more sorting slots are required, the longer the sorting time per waybill, and the higher the workload for personnel. If a single sorting granularity is used, there may be problems such as some waybills being sorted too coarsely, leading to an increased missorting rate, and some waybills being sorted too finely, leading to excessively high sorting costs.

[0071] Through the above methods, firstly, the directed graph of logistics link cost propagation can represent the logistics link corresponding to the type of sorting scenario, thereby enabling adaptive determination of the logistics link based on the type of sorting scenario. This avoids problems such as mis-sorting of multiple items in a single shipment, dispersion of scattered orders in the same direction, and repeated sorting during transit, thus improving sorting accuracy. Secondly, the directed graph of logistics link cost propagation can determine the cascading impact of sorting granularity on downstream transit and distribution, last-mile delivery, and anomaly handling, ensuring that the target sorting granularity determined at the sorting end is optimal across the entire link, improving sorting accuracy while reducing overall logistics and transportation costs. Furthermore, the directed graph of logistics link cost propagation can mathematically quantify the invisible anomaly propagation risks and operational costs in the physical network, providing a computational model for subsequent cost optimization.

[0072] Optionally, step S310 includes steps S311 to S313.

[0073] Step S311: Determine the first logistics link cost propagation directed graph based on the target distribution scenario type and the logistics links corresponding to the preset multiple distribution scenario types.

[0074] Specifically, it is necessary to determine the logistics link corresponding to the target distribution scenario type from the logistics links corresponding to the various preset distribution scenario types, and determine the dependency relationship of logistics nodes and the basic edge weight of the propagation of anomalies from upstream nodes to downstream nodes based on the logistics links, thereby determining the nodes and directed edges in the directed graph of the first logistics link cost propagation.

[0075] Step S312: Adjust the node and edge weights of the first logistics link cost propagation directed graph according to the business characteristic data to obtain the second logistics link cost propagation directed graph.

[0076] Optionally, step S312 includes steps S3121 to S3124.

[0077] Step S3121: Determine the packing method and cargo type of the current waybill based on the pickup information in the business characteristic data.

[0078] As mentioned above, packaging methods include loose packages, packaging according to flow direction, and full shipment packaging. Goods types include high-value goods, general goods, fragile items, and dangerous goods.

[0079] Step S3122: Add or delete split nodes in the directed graph of the first logistics link cost propagation according to the packaging method and the type of goods.

[0080] When the packaging method is full shipment packaging, the entire package is transported together, and the split nodes in the first logistics link cost propagation directed graph are deleted.

[0081] When the packaging method is loose packaging and there is no splitting node in the first logistics link cost propagation directed graph, add a splitting node to the first logistics link cost propagation directed graph.

[0082] When the goods are high-value goods, add a compensation node to the first logistics link cost propagation directed graph.

[0083] Step S3123: Determine the timeliness requirements of the waybill based on the customer information in the business characteristic data.

[0084] In some embodiments, waybill timeliness requirements include same-day delivery, next-morning delivery, next-day delivery, standard delivery, economy delivery, etc.

[0085] Step S3124: Adjust the edge weights of the connecting split nodes based on the timeliness requirements of the waybill to obtain the directed graph of the second logistics link cost propagation.

[0086] In some embodiments, step S3124 may further include steps S31241 to S31243.

[0087] Step S31241: Determine the promised absolute delivery time of the current waybill based on the waybill's timeliness requirements and the current time, and calculate the difference between the promised absolute delivery time and the current system time to obtain the timeliness margin.

[0088] Step S31242: Determine the time sensitivity coefficient based on the time margin using a preset calculation method.

[0089] Among them, the stricter the timeliness requirements of the waybill, the greater the timeliness sensitivity coefficient; the more lenient the timeliness requirements of the waybill, the smaller the timeliness sensitivity coefficient.

[0090] Optionally, the formula for calculating the time sensitivity coefficient is: ,in, Indicates the time sensitivity coefficient. This indicates the promised absolute delivery time for the current waybill. Indicates the current time. This indicates the preset standard time limit. This represents the time sensitivity adjustment constant. Immediate availability margin, when When the value approaches 0 (about to time out), It increased dramatically.

[0091] In the above calculation formula, when the timeliness margin is sufficient, the risk coefficient increases slowly; when the timeliness margin approaches 0, the risk coefficient rises exponentially and sharply. In this way, the extreme sensitivity of waybills with high timeliness requirements to delays can be reflected.

[0092] Step S31243: In the first directed graph of logistics link cost propagation, for each incoming edge connected to the splitting node and the outgoing edge derived from the splitting node, multiply the weight of the basic edge by the time sensitivity coefficient to obtain the second directed graph of logistics link cost propagation.

[0093] By employing the above method, the timeliness requirements of waybills can be transformed into dynamic penalties for the edge weights of splitting nodes. This allows the directed graph of logistics cost propagation to reflect not only the logistics chain but also the implicit impact of time constraints on allocation decisions. In subsequent steps, the allocation costs for different allocation granularities are calculated. If a waybill with high timeliness requirements adopts a subdivided allocation scheme that leads to splitting, the amplified edge weights will result in extremely high abnormal propagation costs at the splitting nodes. This mathematically forces the optimization algorithm to automatically select a coarser allocation granularity that does not result in splitting for waybills with high timeliness requirements, thus ensuring the service quality of these orders.

[0094] Step S313: Based on historical waybill data, target cargo distribution scenario type, destination route type information, and customer information, determine the normal operation cost and abnormal operation cost of each node in the second logistics link cost propagation directed graph to obtain the target logistics link cost propagation directed graph.

[0095] In some embodiments, step S313 includes steps S3131 to S3135.

[0096] Step S3131: Based on historical waybill data, target cargo distribution scenario type and destination route type information, determine the baseline values ​​of normal operation cost and abnormal operation cost for each node in the directed graph of the second logistics link cost propagation at each cargo distribution granularity.

[0097] In some embodiments, step S3131 includes steps S31311 to S31313.

[0098] Step S31311: For each sorting granularity, the historical waybill samples are cross-grouped in two dimensions according to the preset sorting scenario types and destination route type information in the historical waybill data to form multiple homogeneous historical sample subsets.

[0099] By using the above methods, the benchmark values ​​for normal operating costs and abnormal operating costs can be made to have the same distribution scenario background, thereby reducing statistical bias caused by differences in cost structure under different distribution scenarios.

[0100] Step S31312: For each node in the directed graph of the second logistics link cost propagation, calculate the actual operating cost data per unit waybill for that node under normal and abnormal conditions in each homogeneous historical sample subset.

[0101] The actual operating cost data per unit waybill includes benchmark values ​​for normal operating costs and abnormal operating costs.

[0102] In some embodiments, considering that logistics operation cost data typically exhibits a long-tailed skewed distribution, a log-normal distribution is used to fit the sample distributions of the two cost classes respectively. The log-domain mean and standard deviation parameters of the sample distribution are obtained through maximum likelihood estimation. The original domain mean of each type of operation cost is then determined.

[0103] Step S31313: Determine the original domain mean of the cost of each type of operation as its sample distribution, and obtain the benchmark values ​​of normal operation cost and abnormal operation cost for each node of the second logistics link cost propagation directed graph at each sorting granularity.

[0104] The above methods can effectively suppress the interference of extreme outliers (such as occasional large-scale accident compensation) on the benchmark values ​​of normal operating costs and abnormal operating costs, and calculate the benchmark values ​​of normal operating costs and abnormal operating costs that reflect the actual resource consumption of nodes in specific distribution scenarios.

[0105] Step S3132: Determine the direction of the impact of sorting granularity on the normal operating cost of the node based on the node type, and determine that the abnormal operating cost of all nodes decreases as the sorting granularity is refined.

[0106] Among them, the normal operating cost of splitting nodes increases with the refinement of the sorting granularity, while the normal operating cost of transit nodes and terminal nodes decreases with the refinement of the sorting granularity.

[0107] Understandably, for splitting nodes, finer granularity of sorting increases the complexity of scanning and grid scheduling in normal operations, leading to a monotonically increasing cost for normal operations; conversely, lower missorting probability results in a monotonically decreasing cost for abnormal operations. For transit nodes, finer granularity of sorting reduces the workload of secondary sorting during transit, resulting in a monotonically decreasing cost for normal operations; higher abnormal interception rate leads to a monotonically decreasing cost for abnormal operations. For last-mile nodes, finer granularity of sorting increases the efficiency of last-mile route planning, resulting in a monotonically decreasing cost for normal operations; lower delivery anomaly rate leads to a monotonically decreasing cost for abnormal operations.

[0108] Step S3133: Determine the path anomaly cost elasticity coefficient for each node under each cargo distribution granularity based on historical waybill data.

[0109] In some embodiments, step S3133 includes steps S31331 to S31336.

[0110] Step S31331: For each sorting granularity, a pre-trained artificial intelligence model is used to determine the anomaly cost index tensor of each node based on historical waybill data.

[0111] Optionally, the anomaly cost index tensor includes dimensions of time delay cost, cargo damage cost, and human resource cost. Further, the component of each dimension in the initial anomaly cost index tensor is divided by the sum of all components, so that the sum of all components of the anomaly cost index tensor is 1, and this sum is used for subsequent calculations.

[0112] For example, a pre-trained artificial intelligence model can be a neural network model or a Transformer model, etc.

[0113] Specifically, this includes: First, determining the origin and destination of the current waybill based on business characteristic data; then, extracting historical waybill records for the past preset number of days for all transportation paths between the origin and destination at each cargo granularity from historical waybill data; constructing a multimodal time series containing numerical and discrete type features based on the historical waybill records; and then inputting the multimodal time series into a pre-trained artificial intelligence model to obtain an anomaly cost index tensor.

[0114] Numerical features may include delay duration and compensation amount. Discrete features include complaint level and type of transportation anomaly.

[0115] Step S31332: Determine the insurance compensation amount coefficient and historical complaint intensity coefficient for each node based on historical waybill data.

[0116] Among them, the coefficient for insured compensation amount and the coefficient for historical complaint intensity have been normalized.

[0117] In this step, the ratio of the average insured value to the average node base freight cost for each node in the current cargo allocation granularity of the current customer's historical waybill data when a transportation abnormality occurs and compensation is required can be calculated, and then compressed to the (0,1) interval using the Sigmoid function to obtain the insured value compensation coefficient for that node.

[0118] Furthermore, based on historical waybill data, the system determines the complaint type and the number of complaints for each complaint type at each node within the current cargo allocation granularity when a transportation anomaly requires compensation. The system then determines the complaint weight for each complaint type based on a pre-defined correspondence between complaint types and complaint weights. Finally, it calculates the customer's historical weighted complaint frequency based on the complaint weight and the number of complaints for each type. This historical weighted complaint frequency is then logarithmically transformed and normalized using Min-Max to obtain the historical complaint intensity coefficient.

[0119] Step S31333: Determine the scaling matrix of each node at the current cargo allocation granularity based on the insured compensation amount coefficient and historical complaint intensity coefficient.

[0120] In some embodiments, the scaling matrix is: ,in, Represents the scaling matrix. This represents the first preset calculation coefficient. This represents the coefficient for insured compensation. This represents the second preset calculation coefficient. This indicates the intensity coefficient of historical complaints. This represents the third preset calculation coefficient. It is also the Riemannian metric tensor.

[0121] Furthermore, the component of each dimension in the initial scaling matrix is ​​divided by the sum of all components, so that the sum of all components in the scaling matrix is ​​1, and this is used for subsequent calculations.

[0122] In this way, the nonlinear shrinkage characteristics of the anomalous cost space can be reflected, and the exponential transformation causes the geodesic distance of the highly sensitive dimension to be amplified exponentially, that is, the anomalous cost space shrinks dramatically.

[0123] Step S31334: For each node, determine the conditional probability of the node causing timeliness anomalies, cargo damage, and additional human resource service expenses anomalies in the historical waybill data, and generate an anomaly perturbation distribution tensor based on the conditional probability of the node causing timeliness anomalies, cargo damage, and additional human resource service expenses anomalies.

[0124] Among them, the conditional probabilities of triggering timeliness anomalies, cargo damage anomalies, and additional human resource service expense anomalies in the abnormal disturbance distribution tensor are normalized, and the sum of the probabilities is 1.

[0125] Step S31335: Calculate the cross-modal optimal transmission cost between the node's anomaly cost index tensor and the anomaly perturbation distribution tensor based on the scaling matrix of each node.

[0126] In some embodiments, a scaling matrix can be used to calculate the squared Mahalanobis distance between standard basis vectors, thereby constructing a transmission cost matrix. The transmission cost matrix is ​​used to penalize mismatches between node anomaly types and the dimensions of insurance compensation amount coefficients and historical complaint intensity coefficients. Then, using the node's anomaly perturbation distribution tensor as the supply side and the anomaly cost index tensor as the demand side, an entropy regularization term is introduced, and the optimal transmission matrix and cross-modal optimal transmission cost are iteratively solved using the Sinkhorn algorithm. Step S31336: Determine the path anomaly cost elasticity coefficient of the node under the current cargo allocation granularity based on the cross-modal optimal transmission cost, and then determine the anomaly cost index tensor of each node under each cargo allocation granularity.

[0127] In some embodiments, the cross-modal optimal transmission cost after Min-Max normalization can be determined as the path anomaly cost elasticity coefficient.

[0128] In the above approach, the nonlinear tolerance space shrinkage of the high anomaly cost space is characterized by the Riemannian metric tensor, and the optimal transport theory is used to quantify the cross-modal resistance caused by the mismatch between node anomaly types, insurance compensation coefficients, and historical complaint intensity coefficients. The final output path anomaly cost elasticity coefficient enables personalized dynamic pricing and cost prediction for transport routes at each cargo granularity. For example, if a node frequently experiences cargo damage, but its insurance compensation coefficient is low, then the path anomaly cost elasticity coefficient will be low; conversely, if a node experiences minor delays, but its historical complaint intensity coefficient is high, then the path anomaly cost elasticity coefficient will be high.

[0129] Step S3134: For each node, the abnormal operation cost of the node at each sorting granularity is determined based on the baseline value of the abnormal operation cost of the node at each sorting granularity and the elasticity coefficient of the path abnormal cost, thereby obtaining the directed graph of the target logistics link cost propagation.

[0130] Optionally, the abnormal operation cost of a node at the current distribution granularity is the baseline value of the abnormal operation cost at that distribution granularity × (1 + path abnormal cost elasticity coefficient).

[0131] Step S320: Based on the directed graph of the target logistics link cost propagation and multiple preset different sorting granularities, determine the sorting cost of the current waybill under the constraints of each sorting granularity.

[0132] Among them, the various different levels of sorting include regional, provincial, municipal, district, branch, and deliveryman designated levels.

[0133] In some embodiments, step S320 may include steps S321 to S323.

[0134] Step S321: Determine the normal operating cost and abnormal operating cost of each node at each sorting granularity based on the directed graph of target logistics link cost propagation.

[0135] Step S322: For each sorting granularity, the normal operating cost and abnormal operating cost of the node under each sorting granularity are weighted by using the conditional probabilities of the upstream node under normal and abnormal conditions as weights, respectively, to obtain the expected operating cost per waybill of the node.

[0136] The edge weights corresponding to the edges between the current node and the upstream node are the conditional probabilities under abnormal conditions, while the conditional probabilities under normal conditions are 1 minus the conditional probabilities under abnormal conditions.

[0137] Step S323: Based on the expected operating cost per unit waybill of all nodes in the directed graph of the target logistics link cost propagation, determine the allocation cost of the current waybill under the constraint of the current allocation granularity, and then determine the allocation cost of the current waybill under the constraint of each allocation granularity.

[0138] Optionally, for each sorting granularity, the expected unit order operation cost of all nodes in the directed graph of the target logistics link cost propagation is added together to obtain the sorting cost of the current order under the constraint of the current sorting granularity, and then the sorting cost of the current order under the constraint of each sorting granularity is determined.

[0139] In some embodiments, step S323 may include steps S3231 to S3234.

[0140] Step S3231: For each sorting granularity, sum the expected operating costs per unit waybill of all nodes in the directed graph of the target logistics link cost propagation to obtain the basic sorting cost of the current waybill under the constraint of the current sorting granularity.

[0141] Step S3232: In the preset digital twin logistics model, determine the waybill distribution information of each area at the current sorting granularity.

[0142] Step S3233: Calculate the area load imbalance penalty based on the waybill distribution information under the current cargo allocation granularity.

[0143] Among them, the greater the deviation of the area's order volume from the average, the higher the value of the area's load imbalance penalty item, which is used to reflect additional costs such as grid congestion and manual dispatch.

[0144] In some embodiments, the formula for calculating the area load imbalance penalty is as follows: ,in, The identification coefficient represents the granularity of the sorting. This indicates the penalty for uneven load distribution within the current allocation granularity. This represents the scene-line adaptive penalty coefficient. This represents the real-time workstation idle attenuation factor. Indicates the number of districts. Represents a counting variable. Indicates the first The overall sorting load equivalent of the area This represents the average load equivalent of the area. This represents the global penalty saturation threshold.

[0145] In some embodiments, the average load equivalent of a region can be determined based on waybill distribution information. The global penalty saturation threshold is determined by the maximum daily scheduling loss of the sorting system. When calculated... Time to determine for .

[0146] in, ,in, Indicates the first Total number of waybills in the area Indicates the first The weighted average working hours for multiple shipments per area within the district. Indicates the first Total volume of packages in the area Indicates the first Total weight of packages in the area Indicates the first The volume conversion constant of the area Indicates the first Working hours conversion constant for the area.

[0147] in, , , , , and This information can be obtained from the waybill distribution information. ,in, This represents the preset baseline penalty constant. This represents the preset allocation scenario adjustment coefficient corresponding to the target allocation scenario type. This represents the preset route adjustment coefficient corresponding to the destination route type of the current waybill. This method allows for differentiated penalty levels based on the route type of the destination.

[0148] It should be noted that less-than-truckload (LTL) corresponds to Greater than the dedicated line . ,in, This represents a preset attenuation constant greater than 0. This indicates the idle rate of sorting stations; the more idle stations there are, the higher the average idle rate. When the sorting system is fully loaded and congested... The punishment was fully effective.

[0149] In a specific implementation, it is necessary to obtain the current idle status of all sorting stations and determine the sorting station idle rate based on the current idle status of all sorting stations. In addition, it is also necessary to normalize the area load imbalance penalty term to obtain the normalized area load imbalance penalty term.

[0150] Step S3234: Multiply the basic distribution cost of the current waybill under the current distribution granularity constraint by 1 and add the sum of the normalized area load imbalance penalty term to obtain the distribution cost of the current waybill under the current distribution granularity constraint, and then determine the distribution cost of the current waybill under each distribution granularity constraint.

[0151] The above method enables the granularity-sensitive costs of downstream nodes to be transmitted back to the sorting end, dynamically calibrating the utility boundary of different sorting granularities. When the current transit station is congested, the transit cost increases, and the algorithm automatically relaxes to a coarser sorting granularity to reduce transit and sorting pressure, thereby optimizing the overall logistics and transportation cost.

[0152] Step S330: Determine the target sorting granularity corresponding to the current waybill based on the sorting cost under the constraints of each sorting granularity.

[0153] Among them, the granularity of the lowest sorting cost is determined as the target sorting granularity for the current waybill.

[0154] Step S340: Determine the target allocation strategy based on the target allocation granularity.

[0155] For example, if the target allocation granularity is at the branch level, the target allocation strategy is a preset algorithm that allocates goods to the branch level. If the target allocation granularity is at the city level, the target allocation strategy is a preset algorithm that allocates goods to the city level.

[0156] Step S400: Use a target sorting strategy to determine the sorting identifier for each parcel to be shipped in the current waybill.

[0157] The cargo sorting identifier includes a waybill master identifier and a parcel flow identifier, both of which are uniquely bound to the current waybill. The parcel flow identifier is bound to each parcel awaiting shipment.

[0158] Step S400 includes steps S410 to S430.

[0159] Step S410: Using the preset algorithm corresponding to the target cargo allocation strategy, determine at least part of the transportation process for each parcel to be shipped in the current waybill based on business characteristic data.

[0160] Optionally, the origin, destination, and target sorting granularity of the destination information of the parcel to be shipped can be input into a preset route planning algorithm to calculate at least part of the transportation process of the parcel to be shipped.

[0161] For example, the preset path planning algorithm is Dijkstra's shortest path algorithm.

[0162] Step S420: Determine the target identification rule corresponding to the target distribution strategy in the preset identification rule library.

[0163] Step S430: Determine the sorting identifier based on the transportation process of each parcel to be shipped, the mapping relationship information of the identifier field in the target identifier rule, and the coding rule.

[0164] In some embodiments, step S430 may further include steps S431 to S435.

[0165] Step S431: Determine the value of the goods based on the pickup information in the business characteristic data.

[0166] The value of the goods is determined based on the declared value of the goods in the pickup information.

[0167] Step S432: Determine the mode of transport based on the value of the goods.

[0168] Specifically, the transportation method is determined based on the value of the goods and the pre-defined correspondence between the goods value and the mode of transport. When the value of the goods is higher than the pre-defined value, the goods are classified as high-value goods, and the corresponding transportation method is determined to be air transport or escorted transport. When the value of the goods is not higher than the pre-defined value, the goods are classified as ordinary goods, and the corresponding transportation method is conventional land transport.

[0169] Step S433: Determine the parcel flow identifier of each parcel based on its transportation process, transportation method, identification field mapping relationship information, and coding rules.

[0170] Specifically, the code corresponding to the destination node and the code corresponding to the transportation method in the transportation process are determined based on the mapping relationship information of the identification field, and all codes are concatenated according to the coding rules to obtain the parcel flow identifier of the parcel to be shipped.

[0171] In some embodiments, the content and display style of the parcel flow indicator can be configured according to the transportation process, transportation method, and preset identification rules for each parcel to be shipped. The dimensions for setting the identification display style may include font, font size, color, and symbols. Then, the content and display style of the parcel flow indicator are configured according to dimensions such as the originating airport, destination airport, and receiving point in the transportation process of the parcel to be shipped, and the preset identification rules.

[0172] Step S434: Determine the main identifier of the waybill based on the target cargo distribution scenario type.

[0173] Specifically, when the target cargo distribution scenario is a multi-item cargo distribution type, the main waybill identifier is determined to include the same cargo transportation identifier.

[0174] Step S435: Determine the sorting identifier of each parcel to be shipped based on the parcel flow identifier and waybill master identifier.

[0175] The distribution identifier is obtained by combining the package flow identifier and the main waybill identifier of each package to be shipped.

[0176] The above methods can be used to construct a multi-dimensional, hierarchical identification system, effectively avoiding the splitting and confusion of multiple goods under the same ticket, and guiding staff to use the correct transportation methods for operation.

[0177] In summary, the sorting method provided in this application has the following advantages: 1. By determining the target sorting scenario type corresponding to the current waybill based on business characteristic data, determining the target sorting strategy based on the target sorting scenario type and historical waybill data, and using the target sorting strategy to determine the sorting identifier for each parcel to be shipped in the current waybill, the target sorting strategy can be adapted to the business characteristic data of the waybill to determine the sorting identifier, which can improve the sorting accuracy and enable front-line staff to intuitively obtain the flow and ownership information of the parcel through the sorting identifier.

[0178] 2. By determining the target sorting scenario type corresponding to the current waybill based on business characteristic data, and by determining the target logistics link cost propagation directed graph corresponding to the target sorting scenario type based on the target sorting scenario type, historical waybill data, and the logistics links corresponding to various preset sorting scenario types, the logistics link cost propagation directed graph can represent the logistics link corresponding to the sorting scenario type, thereby realizing the adaptive determination of the logistics link according to the sorting scenario type, avoiding the problems of missorting multiple items in one shipment, scattered scattered orders in the same direction, and repeated sorting in transit, and improving the accuracy of sorting.

[0179] 3. By determining the target sorting granularity for the current waybill based on the sorting cost under the constraints of each sorting granularity, and performing sorting operations according to the target sorting granularity, the sorting granularity can be adaptively and dynamically adjusted, thereby improving sorting accuracy while reducing overall logistics and transportation costs.

[0180] 4. By adjusting the edge weights connecting split nodes based on the timeliness requirements of the waybill, a second directed graph of logistics link cost propagation is obtained. This transforms the timeliness requirements of the waybill into a dynamic penalty for the edge weights of the split nodes, enabling the directed graph of logistics link cost propagation to reflect not only the logistics link but also the implicit impact of time constraints on allocation decisions. In subsequent steps, the allocation costs of different allocation granularities are calculated. If a high-time-requirement waybill adopts a subdivided allocation scheme that leads to splitting, the amplified edge weights will result in extremely high abnormal propagation costs at the split nodes. This mathematically forces the optimization algorithm to automatically select a coarser allocation granularity that does not result in splitting for high-time-requirement waybills, thus ensuring the service quality of high-time-requirement orders.

[0181] 5. By utilizing the Riemannian metric tensor to characterize the nonlinear tolerance space contraction of the high anomaly cost space, and employing optimal transport theory to quantify the cross-modal resistance caused by the mismatch between node anomaly types, insurance compensation coefficients, and historical complaint intensity coefficients, the final output path anomaly cost elasticity coefficient enables personalized dynamic pricing and cost prediction for transportation routes at each cargo granularity. For example, if a node frequently experiences cargo damage, but its insurance compensation coefficient is low, then the path anomaly cost elasticity coefficient will be low; conversely, if a node experiences minor delays, but its historical complaint intensity coefficient is high, then the path anomaly cost elasticity coefficient will be high.

[0182] 6. By using a penalty term for uneven load distribution in different areas, the granularity-sensitive costs of each downstream node can be transmitted back to the sorting end, dynamically calibrating the utility boundary of different sorting granularities. When the current transfer station is congested, the transfer cost increases, and the algorithm automatically relaxes to a coarser sorting granularity to reduce the transfer and sorting pressure, thereby optimizing the overall logistics transportation cost.

[0183] Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of the sorting system provided in the embodiments of this application. For example... Figure 2 As shown, the sorting system 300 includes an acquisition module 310, a sorting strategy determination module 320, and an identification module 330.

[0184] The acquisition module 310 is used to acquire the business characteristic data of the current waybill and historical waybill data. The business characteristic data includes the pickup information of the current waybill, the destination information of all parcels to be shipped in the current waybill, the destination route type information, and the customer information.

[0185] The cargo allocation strategy determination module 320 is used to acquire business characteristic data and historical cargo data for the current waybill. The business characteristic data includes the pickup information of the current waybill, the destination information of all parcels to be shipped in the current waybill, the destination route type information, and customer information. Based on the business characteristic data, the module determines the target cargo allocation scenario type corresponding to the current waybill, wherein the target cargo allocation scenario type is direct delivery type, transit cargo allocation type, or multi-item cargo allocation type. Based on the target cargo allocation scenario type and historical cargo data, the module determines the target cargo allocation strategy. The identification module 330 is used to determine the sorting identifier of each parcel to be shipped in the current waybill using the target sorting strategy. The sorting identifier includes the waybill master identifier and the parcel flow identifier, and both the waybill master identifier and the parcel flow identifier are uniquely bound to the current waybill.

[0186] Finally, the software architecture and hardware environment of the sorting system provided in this application embodiment are described. The sorting system of this application adopts a layered, decoupled, and microservice architecture at the software level, built on the Spring Cloud Alibaba microservice framework. The front-end presentation layer uses HTML5 to build the interface skeleton and CSS3 to achieve responsive layout, adapting to multiple devices such as industrial PDAs, mobile smartphones, and tablets. It combines Vue.js or React frameworks to develop a visual display module and interactive functions for sorting identifiers. The interface layer follows RESTful specifications to design standardized interactive interfaces, integrating Token authentication, TLS / SSL data encryption, and request rate limiting mechanisms. The back-end service layer uses Java (Spring Boot) as the core development framework, carrying core business logic such as automated generation of sorting identifiers, multi-dimensional classification, and cost propagation directed graph computation, and handles asynchronous tasks through message queues (such as RabbitMQ). The data storage layer adopts a hybrid storage architecture of MySQL and MongoDB. MySQL is used to store structured data such as waybill master identifiers and customer information, while MongoDB is used to store semi-structured data such as cost propagation directed graph computation history and exception logs, and Redis is introduced to cache hot data. The system is deployed on the enterprise's existing core logistics business system server and seamlessly integrates with the existing waybill management system, cargo allocation system, and printing equipment management system through a data synchronization interface.

[0187] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. For example... Figure 3 As shown, the electronic device 400 includes: one or more processors 410 and a memory 420. Figure 3 Take a processor 410 as an example.

[0188] In some embodiments, the processor 410 and the memory 420 can be connected via a bus or other means. Figure 3 Taking the example of a connection between China and Israel via a bus.

[0189] Processor 410 is used to acquire business characteristic data and historical waybill data for the current waybill. The business characteristic data includes the pickup information of the current waybill, the destination information, destination route type information, and customer information for all parcels to be shipped in the current waybill. Based on the business characteristic data, the processor determines the target distribution scenario type corresponding to the current waybill, wherein the target distribution scenario type is direct flow type, transit distribution type, or one-shipment-multiple-item distribution type. Based on the target distribution scenario type and historical waybill data, the processor determines the target distribution strategy. The processor uses the target distribution strategy to determine the distribution identifier for each parcel to be shipped in the current waybill. The distribution identifier includes the waybill master identifier and the parcel flow identifier, both of which are uniquely bound to the current waybill.

[0190] The memory 420, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules of the sorting method in the embodiments of this application. The processor 410 executes various functional applications and data processing of the electronic device 400 by running the non-volatile software programs, instructions, and modules stored in the memory 420, thereby implementing the sorting method of the above-described method embodiments.

[0191] The memory 420 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device 400. Furthermore, the memory 420 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. Optionally, the memory 420 may include memory remotely located relative to the processor 410, and these remote memories may be connected to the controller via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0192] One or more modules are stored in memory 420. When executed by one or more processors 410, they perform the sorting method in any of the above method embodiments, for example, the sorting method described above. Figure 1 The method steps S100 to S400.

[0193] Please refer to Figure 4 , Figure 4 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. The computer-readable storage medium 500 stores program code 510, which can be called by a processor to execute the sorting method described in the above method embodiments.

[0194] Computer-readable storage medium 500 may be an electronic storage device such as flash memory, electrically erasable programmable read-only memory (EEPROM), hard disk, or read-only memory (ROM). Optionally, the computer-readable storage medium includes a non-volatile computer-readable medium. The computer-readable storage medium 500 has storage space for program code that performs any of the method steps of the above-described dispensing method. This program code can be read from or written to one or more computer program products. The program code may, for example, be compressed in a suitable form.

[0195] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described sorting method.

[0196] In summary, this application provides a sorting method, a sorting system, an electronic device, and a storage medium. The sorting method includes: acquiring business characteristic data and historical waybill data for the current waybill; the business characteristic data includes the current waybill's pickup information, destination information, destination route type information, and customer information for all parcels to be shipped under the current waybill; determining the target sorting scenario type corresponding to the current waybill based on the business characteristic data, wherein the target sorting scenario type is direct shipment, transit sorting, or multi-piece sorting; determining a target sorting strategy based on the target sorting scenario type and historical waybill data; and using the target sorting strategy to determine the sorting identifier for each parcel to be shipped under the current waybill, wherein the sorting identifier includes a waybill master identifier and a parcel flow identifier, and both the waybill master identifier and the parcel flow identifier are uniquely bound to the current waybill. This application determines the target sorting scenario type corresponding to the current waybill based on business characteristic data, determines the target sorting strategy based on the target sorting scenario type and historical waybill data, and uses the target sorting strategy to determine the sorting identifier for each parcel to be shipped in the current waybill. It can use a target sorting strategy adapted to the business characteristic data of the waybill to determine the sorting identifier, which can improve the sorting accuracy and enable front-line staff to intuitively obtain the flow and ownership information of the parcels through the sorting identifier.

[0197] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for sorting goods, characterized in that, include: Obtain the business characteristic data of the current waybill and historical waybill data. The business characteristic data includes the pickup information of the current waybill, the destination information of all parcels to be shipped in the current waybill, the destination route type information, and the customer information. Based on the business characteristic data, the target cargo distribution scenario type corresponding to the current waybill is determined, wherein the target cargo distribution scenario type is direct delivery type, transit cargo distribution type, or multiple-item cargo distribution type. The step of determining the target cargo distribution scenario type corresponding to the current waybill based on the business characteristic data includes: Based on the pickup information, the packaging method feature information of the current waybill is determined; The destination consistency feature information of the current waybill is determined based on the destination information; The transit link characteristic information of the current waybill is determined based on the destination route type information; The target cargo distribution scenario type is determined based on at least one of the following features: the packaging method feature information, the destination consistency feature information, and the transit link feature information; The target cargo allocation strategy is determined based on the target cargo allocation scenario type and the historical waybill data. The target sorting strategy is used to determine the sorting identifier for each parcel to be shipped in the current waybill. The sorting identifier includes the waybill master identifier and the parcel flow identifier. Both the waybill master identifier and the parcel flow identifier are uniquely bound to the current waybill.

2. The sorting method according to claim 1, characterized in that, Determining the target cargo distribution scenario type based on at least one of the following features: the packaging method feature information, the destination consistency feature information, and the transit link feature information, includes: When the first preset judgment condition is met, the target cargo distribution scenario type is determined to be direct flow type. The first preset judgment condition includes at least two of the following conditions: the packaging method feature information indicates that the packages are packaged according to the flow direction; the destination consistency feature information indicates that all the packages to be shipped have the same destination; the transit link feature information indicates that the destination line type is a dedicated line; and the customer information and the historical waybill data determine that the current customer's historical waybill has no transit record. When the second preset judgment condition is met, the target cargo distribution scenario type is determined to be the transit cargo distribution type. The second preset judgment condition includes at least two of the following conditions: the packaging method feature information does not indicate packaging according to flow direction, the destination consistency feature information indicates that the destinations of the parcels to be shipped are inconsistent, the transit link feature information indicates that the destination route type is not a dedicated line, and the proportion of transit sorting records of the current customer's historical waybill is greater than the first preset ratio, based on the customer information and the historical waybill data. When the third preset judgment condition is met, the target cargo distribution scenario type is determined to be the multi-item cargo distribution type. The third preset judgment condition includes determining that the proportion of packages with the same destination among all the packages to be shipped is greater than the second preset proportion based on the destination consistency feature information, determining that multiple packages in the current customer's historical waybill data are mixed in different places based on the customer information and the historical waybill data, determining that there are cargo distribution errors in the current customer's historical waybill data, and that the packaging method feature information does not indicate at least two of the following conditions in the whole shipment packaging.

3. The sorting method according to claim 1, characterized in that, The step of determining the sorting identifier for each parcel to be shipped in the current waybill using the target sorting strategy includes: The preset algorithm corresponding to the target cargo allocation strategy is used to determine at least part of the transportation process for each parcel to be shipped in the current waybill based on the business feature data; Determine the target identification rule corresponding to the target distribution strategy in the preset identification rule base; The sorting identifier is determined based on the transportation process of each package to be shipped, the mapping relationship information of the identifier field in the target identifier rule, and the encoding rule.

4. The sorting method according to claim 3, characterized in that, The step of determining the sorting identifier based on the transportation process of each package to be shipped, the mapping relationship information of the identifier field in the target identifier rule, and the encoding rule includes: The value of the goods is determined based on the pickup information in the aforementioned business characteristic data; The mode of transport is determined based on the value of the goods. The parcel flow identifier of each parcel to be shipped is determined based on the transportation process of each parcel to be shipped, the transportation method, the mapping relationship information of the identification field, and the coding rules; The waybill master identifier is determined based on the target cargo distribution scenario type, wherein when the target cargo distribution scenario type is the one-shipment-multiple-item distribution type, the waybill master identifier is determined to include the same shipment transportation identifier; The sorting identifier of each parcel to be shipped is determined based on the parcel flow identifier and the waybill master identifier.

5. The sorting method according to claim 1, characterized in that, The determination of the target cargo allocation strategy based on the target cargo allocation scenario type and the historical waybill data includes: Based on the target cargo distribution scenario type, the historical waybill data, and the logistics links corresponding to the preset multiple cargo distribution scenario types, a directed graph of target logistics link cost propagation corresponding to the target cargo distribution scenario type is determined. The nodes in the directed graph of target logistics link cost propagation represent the logistics nodes of the current waybill. Each node corresponds to the operation cost of the associated cargo distribution granularity. The edge weights in the directed graph of logistics link cost propagation are the conditional probabilities of the propagation of anomalies from upstream nodes to downstream nodes, and the conditional probabilities are functions of the cargo distribution granularity, which is used to represent the fineness of cargo distribution. Based on the directed graph of the target logistics link cost propagation and multiple preset, distinct sorting granularities, the sorting cost of the current waybill under the constraint of each sorting granularity is determined. The multiple distinct sorting granularities include regional level, provincial level, city level, district level, branch level, and delivery person designated level. The target granularity of the current waybill is determined based on the allocation cost of the current waybill under the constraints of each allocation granularity. The target allocation strategy is determined based on the target allocation granularity.

6. The sorting method according to claim 5, characterized in that, The process of determining the directed graph of target logistics link cost propagation corresponding to the target distribution scenario type based on the target distribution scenario type, the historical waybill data, and the logistics links corresponding to multiple preset distribution scenario types includes: Based on the target cargo distribution scenario type and the logistics links corresponding to the various cargo distribution scenario types, a first logistics link cost propagation directed graph is determined. The node and edge weights of the first logistics link cost propagation directed graph are adjusted based on the business characteristic data to obtain the second logistics link cost propagation directed graph. Based on the historical waybill data, the target cargo distribution scenario type, the destination route type information, and the customer information, determine the normal operation cost and abnormal operation cost of each node in the second logistics link cost propagation directed graph, and obtain the target logistics link cost propagation directed graph.

7. A sorting system, characterized in that, The sorting system includes: The acquisition module is used to acquire the business characteristic data of the current waybill and historical waybill data. The business characteristic data includes the pickup information of the current waybill, the destination information of all parcels to be shipped in the current waybill, the destination route type information, and the customer information. The cargo distribution strategy determination module is used to determine the target cargo distribution scenario type corresponding to the current waybill based on the business characteristic data, wherein the target cargo distribution scenario type is direct delivery type, transit cargo distribution type, or one shipment with multiple items cargo distribution type. The target cargo allocation strategy is determined based on the target cargo allocation scenario type and the historical waybill data. The identification module is used to determine the sorting identifier of each parcel to be shipped in the current waybill using the target sorting strategy. The sorting identifier includes the waybill master identifier and the parcel flow identifier, and both the waybill master identifier and the parcel flow identifier are uniquely bound to the current waybill.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the sorting method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an executable program, which is executed by a processor to implement the sorting method as described in any one of claims 1 to 6.

Images