Cross-border logistics order management method and cross-border logistics order management system

Abstract

The invention provides a cross-border logistics order management method and a cross-border logistics order management system, and the method comprises the steps: carrying out the similarity comparison of package feature data and historical package feature data, and obtaining a similarity comparison result; if the preset similarity threshold is not exceeded, determining that the to-be-sent parcel is a target parcel; inputting the parcel feature data of the target parcel into a pre-trained digital twinborn model to obtain a digital twinborn body; carrying out analog simulation on the digital twinborn body based on the digital twinborn model, and generating a standard logistics track and a standard transportation cost; after the target parcel is delivered, an actual logistics track, actual transportation cost and current logistics environment information are acquired; and when the logistics deviation value exceeds a logistics abnormal threshold value or the cost deviation value exceeds a cost abnormal threshold value, generating alarm information. In the mode, automatic identification of repeated delivery, logistics abnormity and transportation cost abnormity can be realized, so that the intelligent and precise level of cross-border logistics order management is improved.

Description

Technical Field

[0001] This application relates to the technical field of e-commerce logistics management, and in particular to a cross-border logistics order management method and a cross-border logistics order management system. Background Art

[0002] With the rapid development of cross-border e-commerce, merchants need to manage a large number of order packages sent overseas and effectively monitor their logistics status and transportation costs. However, the current cross-border logistics management system still has obvious deficiencies in terms of intelligence and automation.

[0003] In the prior art, duplicate shipments usually rely on manual verification of order information, which is inefficient and prone to missed inspections, resulting in cost waste; the monitoring of logistics trajectories mostly uses a timed polling method, lacking real-time performance. Some abnormal packages are not identified until the logistics is interrupted or detained for more than 72 hours, making it difficult to take timely intervention measures; in addition, transportation costs are mostly estimated based on empirical templates or fixed routes, and cannot accurately reflect the actual costs of different packages under different paths, resulting in a large deviation in cost accounting. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a cross-border logistics order management method and a cross-border logistics order management system, which can realize the automatic identification of duplicate shipments, logistics anomalies, and transportation cost anomalies, thereby improving the intelligence and precision of cross-border logistics order management, and effectively reducing operating costs and improving the efficiency of anomaly handling.

[0005] In the first aspect, the present invention provides a cross-border logistics order management method, including: obtaining package feature data of a package to be shipped; comparing the package feature data with historical package feature data in historical shipping records to obtain a similarity comparison result; if the similarity comparison result does not exceed a preset similarity threshold, determining the package to be shipped as a target package; inputting the package feature data of the target package into a pre-trained digital twin model to obtain a digital twin corresponding to the target package; performing simulation on the digital twin based on the digital twin model to generate a standard logistics trajectory and a standard transportation cost corresponding to the target package; after the target package is shipped, obtaining the actual logistics trajectory, actual transportation cost, and current logistics environment information of the target package; when the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics anomaly threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost anomaly threshold, generating an alarm message; wherein, both the logistics anomaly threshold and the cost anomaly threshold are dynamic thresholds adaptively determined according to the current logistics environment information.

[0006] In an alternative embodiment, the step of obtaining the package feature data of the package to be shipped includes: collecting the image information of the package to be shipped through an image acquisition device; the image information includes the packaging form, the barcode area, and the waybill information; obtaining the order information corresponding to the package to be shipped from an e-commerce platform; the order information includes the order number, the product identifier, the shipping address, the shipping time, the transportation method, and the logistics mode; obtaining the physical attribute information and processing requirements corresponding to the package to be shipped from a product management platform; and integrating the image information, the order information, the physical attribute information, and the processing requirements into the package feature data.

[0007] In an alternative embodiment, after the step of comparing the similarity between the package feature data and the historical package feature data in the historical shipping records to obtain a similarity comparison result, the method further includes: if the similarity comparison result exceeds a preset similarity threshold, determining that the package to be shipped is a duplicate shipment package and generating a duplicate shipment alarm message.

[0008] In an alternative embodiment, the step of generating the standard logistics track and the standard transportation cost corresponding to the target package by simulating and emulating the digital twin based on the digital twin model includes: inputting the digital twin into the digital twin model to output multiple candidate logistics tracks; the candidate logistics tracks are the feasible paths for simulating the target package under different transportation plans; obtaining the logistics correlation data corresponding to the target package through a preset external interface; the logistics correlation data includes the current logistics network status information, the transportation rule information, and the economic cost index information; calculating the scoring results of each candidate logistics track based on the logistics correlation data and a preset scoring rule; the scoring results include the transportation timeliness score, the transportation risk score, the compliance score, and the transportation economic cost score; generating a candidate path set including the path with the optimal compliance, the path with the optimal safety, and the path with the optimal economy based on the scoring results; selecting an optimal path from the candidate path set as the standard logistics track of the target package according to a preset path selection rule; and calculating the standard transportation cost corresponding to the standard logistics track through a transportation cost estimation model based on the standard logistics track.

[0009] In an alternative embodiment, the current logistics environment information includes the logistics network load rate, the historical transportation delay probability of the logistics path, the volatility of the transportation fuel cost, and the destination customs clearance stability index; after the step of obtaining the current logistics environment information of the target package after the target package is shipped, the method further includes: calculating a logistics risk coefficient based on the logistics network load rate, the historical transportation delay probability of the logistics path, the destination customs clearance stability index, and a preset logistics risk calculation method, and generating a logistics anomaly threshold based on the logistics risk coefficient and an initial logistics anomaly threshold; calculating a cost risk coefficient based on the volatility of the transportation fuel cost, the destination customs clearance stability index, and a preset cost risk coefficient calculation method, and generating a cost anomaly threshold based on the cost risk coefficient and an initial cost anomaly threshold.

[0010] In an optional embodiment, after the target package is shipped, after the step of obtaining the actual logistics trajectory of the target package, the method also includes: upon receiving the logistics node update information returned by the logistics service provider system, or when a preset first time interval is reached, updating the actual logistics trajectory to obtain the current actual logistics trajectory; outputting the current standard logistics trajectory corresponding to the current actual logistics trajectory through the digital twin model; constructing a trajectory matching distance matrix based on the current actual logistics trajectory and the current standard logistics trajectory; calculating the logistics deviation value between the current actual logistics trajectory and the current standard logistics trajectory based on the dynamic time warping algorithm, and generating the best matching path; the best matching path is used to describe the correspondence between the standard node in the current standard logistics trajectory and the actual node in the current actual logistics trajectory.

[0011] In an optional embodiment, if the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics anomaly threshold, the step of generating an alarm message includes: if the logistics deviation value exceeds the logistics anomaly threshold, determining that the target package has a trajectory anomaly, and obtaining the best matching path; if there is a standard node in the best matching path that does not match the actual node, determining that the trajectory anomaly is a jump point anomaly; if there is an actual node in the best matching path that does not match the standard node, determining that the trajectory anomaly is a detour anomaly; if the stay time of any actual node exceeds the sum of the standard stay time of its corresponding standard node and a preset stay threshold, determining that the trajectory anomaly is a detention anomaly; if the geographical direction of the actual node is inconsistent with the delivery direction of the target package, determining that the trajectory anomaly is a reverse anomaly; based on at least one trajectory anomaly, generating an alarm message.

[0012] In an alternative embodiment, if the logistics mode of the target package is the overseas warehouse delivery mode, and after the target package arrives at the overseas warehouse and is sent to its corresponding delivery address from the overseas warehouse, the method further includes: when a preset second time interval is reached, obtaining the outbound record of the overseas warehouse; the outbound record includes the outbound order number, the outbound commodity identifier, the outbound delivery address, and the outbound time; obtaining a set of shipped order information within the preset second time interval from the e-commerce platform; the set of shipped order information includes the shipped order number, the shipped commodity identifier, the shipped delivery address, and the shipped time; if any outbound order number in the outbound record corresponds to multiple shipped order numbers, determining that there is an abnormal repeated outbound in the overseas warehouse; if the outbound quantity in the records with the same outbound commodity identifier in the outbound record is inconsistent with the shipped quantity of the same shipped commodity identifier in the corresponding set of shipped order information, determining that there is an abnormal mis-delivery in the overseas warehouse; if the outbound time of any outbound order number in the outbound record is not within the preset delivery time range of its corresponding shipped time, or the outbound delivery address corresponding to any outbound order number is inconsistent with its corresponding shipped delivery address, determining that there is an abnormal delivery in the overseas warehouse; generating an alarm message based on at least one of the abnormal repeated outbound, the abnormal mis-delivery, and the abnormal delivery.

[0013] In an alternative embodiment, after the step of obtaining a set of shipped order information within the preset second time interval from the e-commerce platform, the method further includes: obtaining the warehousing charging data corresponding to the overseas warehouse; calculating the actual warehousing cost based on the set of shipped order information and the preset warehousing charging method; if the warehousing charging data is different from the actual warehousing cost, determining that there is a charging anomaly in the overseas warehouse and generating a charging anomaly alarm message.

[0014] Second aspect, the present invention provides a cross-border logistics order management system, including: a package feature data acquisition module for acquiring package feature data of a package to be shipped; a target package determination module for comparing the package feature data with historical package feature data in historical shipping records to obtain a similarity comparison result; the target package determination module is further configured to determine the package to be shipped as a target package if the similarity comparison result does not exceed a preset similarity threshold; a standard information generation module for inputting the package feature data of the target package into a pre-trained digital twin model to obtain a digital twin corresponding to the target package; the standard information generation module is further configured to perform simulation on the digital twin based on the digital twin model to generate a standard logistics trajectory and a standard transportation cost corresponding to the target package; a real-time monitoring module for acquiring the actual logistics trajectory, actual transportation cost and current logistics environment information of the target package after the target package is shipped; an alarm information generation module for generating alarm information when the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics anomaly threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost anomaly threshold; wherein, both the logistics anomaly threshold and the cost anomaly threshold are dynamic thresholds adaptively determined according to the current logistics environment information.

[0015] Embodiments of the present application provide a cross-border logistics order management method and a cross-border logistics order management system, including: obtaining package feature data of a package to be shipped; comparing the similarity between the package feature data and historical package feature data in historical shipping records to obtain a similarity comparison result; if the similarity comparison result does not exceed a preset similarity threshold, determining the package to be shipped as a target package; inputting the package feature data of the target package into a pre-trained digital twin model to obtain a digital twin corresponding to the target package; performing simulation on the digital twin based on the digital twin model to generate a standard logistics track and a standard transportation cost corresponding to the target package; after the target package is shipped, obtaining the actual logistics track, actual transportation cost, and current logistics environment information of the target package; when the logistics deviation value between the actual logistics track and the standard logistics track exceeds the logistics anomaly threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost anomaly threshold, generating an alarm message; where the logistics anomaly threshold and the cost anomaly threshold are both dynamic thresholds adaptively determined according to the current logistics environment information. In this way, by constructing a digital twin of the target package and simulating multiple candidate logistics paths in the digital twin model, the transportation timeliness, compliance, safety, and economic cost of the paths can be comprehensively evaluated before shipment, so as to generate the optimal standard logistics track and standard transportation cost, and improve the accuracy of path selection and the transportation cost control ability; by obtaining the actual logistics track, transportation cost, and logistics environment information after shipment, and performing deviation analysis on the actual data and the standard data based on the dynamic time warping algorithm and the risk scoring mechanism, the track anomaly and cost anomaly occurring during the transportation process can be discovered in time, so as to improve the real-time performance and accuracy of logistics process monitoring; by introducing adaptive thresholds calculated based on the current logistics network load rate, path delay probability, fuel cost volatility, and customs clearance stability, the anomaly judgment standard can be dynamically adjusted, so as to enhance the adaptability of the system to complex logistics environments; and then realize the full-process intelligent management of cross-border logistics orders in links such as path prediction, cost estimation, process monitoring, anomaly warning, and warehouse reconciliation, and improve the safety, stability, and management efficiency of the cross-border logistics system.

[0016] Other features and advantages of the present application will be described in the following specification, and some of them will become obvious from the specification, or be understood by implementing the present application.

[0017] To make the above objects, features, and advantages of the present application more obvious and understandable, the following specific preferred embodiments are given in conjunction with the accompanying drawings and described in detail as follows. Description of the Drawings

[0018] To more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 Flowchart of the cross-border logistics order management method provided by the embodiment of the present application;

[0020] Figure 2 Schematic diagram of the cross-border logistics order management system provided by the embodiment of the present application;

[0021] Figure 3 Schematic diagram of the structure of the electronic device provided by the embodiment of the present application.

[0022] Icons: 1 - Package feature data acquisition module; 2 - Target package judgment module; 3 - Standard information generation module; 4 - Real-time monitoring module; 5 - Alarm information generation module; 301 - Processor; 302 - Memory; 303 - Bus; 304 - Communication interface. Specific embodiments

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions of the present application with reference to the drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the scope of protection of the present application.

[0024] To facilitate the understanding of this embodiment, the following will introduce the embodiments of the present application in detail.

[0025] Embodiment 1:

[0026] Figure 1 Flowchart of the cross-border logistics order management method provided by the embodiment of the present application.

[0027] Step S101, obtain the package feature data of the package to be sent.

[0028] In one embodiment, the steps of step S101 include the following steps S201 - S204.

[0029] Step S201, collect the image information of the package to be sent through an image acquisition device; the image information includes the packaging form, barcode area, and waybill information.

[0030] Here, the image acquisition device can be a camera array, an industrial camera, a mobile scanning terminal, etc., and is deployed at a parcel sorting table, a packaging station or a shipping port.

[0031] The image information at least includes the packaging form, the barcode area and the waybill information. The packaging form includes the box type, whether there is a special label (fragile, cold chain), the pasting standardization, etc. The barcode area is a one-dimensional / two-dimensional code label on the automatically identified parcel, which is used for subsequent OCR (Optical Character Recognition) or code value extraction of the order number, SKU (Stock Keeping Unit), etc. The waybill information is the key fields such as the sending and receiving addresses, the logistics service provider information, the waybill number, the destination country, etc. in the recognized waybill.

[0032] Among them, the image acquisition device can also assist in judging whether the parcel status is abnormal, such as packaging damage, waybill missing, barcode blur, etc.

[0033] Step S202, obtain the order information corresponding to the parcel to be sent from the e-commerce platform; the order information includes the order number, the product identifier, the receiving address, the shipping time, the transportation method and the logistics mode.

[0034] Here, through a preset API interface, the order data is pulled from the bound e-commerce platforms (such as Shopee, Amazon, Taobao, etc.).

[0035] The order number is the unique identifier of the order and can be matched with the waybill number recognized by the image; the product identifier (SKU) is used to identify the specific product type, category, value level; the receiving address contains fields such as the country, province, city, and postal code, which are used for subsequent path simulation; the shipping time is the scheduled shipping time, which is used as the starting point for scheduling and digital twin simulation; the transportation method can be air transportation, sea transportation, land transportation, railway, etc.; the logistics mode can be overseas warehouse, direct mail, FBA, self-delivery, etc.

[0036] Step S203, obtain the physical attribute information and processing requirements corresponding to the parcel to be sent from the product management platform.

[0037] Here, the physical attribute information usually comes from the warehouse management system or the product master data platform, and the physical attribute information includes size parameters (length, width, height), weight, packaging materials (such as cartons, foam boxes, soft packages, etc.).

[0038] The processing requirements include whether it is cold chain transportation, whether it is fragile, whether it needs to be insured, whether it is charged, etc.

[0039] Step S204, integrate the image information, the order information, the physical attribute information and the processing requirements into parcel feature data.

[0040] Here, the image information, order information, physical attribute information, and processing requirements are structurally encoded to form a unified package feature data object.

[0041] Step S102: Compare the package feature data with the historical package feature data in the historical shipping records to obtain a similarity comparison result.

[0042] Specifically, first, structurally encode the package feature data of the package to be shipped to construct a feature vector for similarity comparison. The package feature data includes fields such as product identification, shipping method, logistics mode, receiving address area, shipping time, size, weight, etc. Among them, categorical fields such as product identification and shipping method are processed by one-hot encoding. After numerical fields such as size and weight are normalized, they are combined with other fields to form the feature vector of the package.

[0043] Then, extract the corresponding field information of each historical package from the historical shipping records, and generate a set of historical feature vectors using the same encoding method as the package to be shipped. Calculate the similarity between the feature vector of the package to be shipped and each historical feature vector. The calculation can use distance metric algorithms such as cosine similarity, weighted Euclidean distance, or Mahalanobis distance. The feature weight coefficients can be configured according to the importance of different fields.

[0044] The result of the comparison calculation outputs a similarity score value, usually in the range of 0 to 1. The higher the similarity value, the more similar the two packages are at the feature level.

[0045] Step S103: If the similarity comparison result does not exceed the preset similarity threshold, determine the package to be shipped as the target package.

[0046] Further, after the steps of Step S102, the method further includes:

[0047] If the similarity comparison result exceeds the preset similarity threshold, determine the package to be shipped as a duplicate shipment package and generate a duplicate shipment alarm message.

[0048] Here, if the similarity score of any historical package and the current package to be shipped exceeds the preset similarity threshold, the system initially determines that the current package may be a duplicate shipment package. Otherwise, determine the current package as the target package and enter the subsequent digital twin modeling and path simulation process.

[0049] Specifically, to further reduce false alarms, additional judgment conditions can also be introduced to compare the following fields: whether the current order number is different from the historical order number; whether the current shipping time is close to the historical shipping time (for example: the shipping interval is less than 24 hours); whether the logistics mode of the current package is self-operated / agent-shipped resulting in duplicate submissions.

[0050] If all the above additional judgment conditions are met, it is determined that the package to be shipped is a duplicate shipment package.

[0051] Once it is determined that there is a duplicate shipment behavior, an alarm message is automatically generated. The duplicate shipment alarm message includes: the order number of the current package and the matched historical order number, the similarity score value, the possible duplicate shipment reason identifier (such as: interface duplicate order pushing, human error operation, system return and reissue without clearing the mark), and the recommended operation items (such as: manual confirmation, automatic interception, notify the customer service, etc.).

[0052] The alarm message is synchronously pushed to the abnormal order processing module, the warehousing system or the operator interface to intercept misdeliveries, reduce losses and optimize the fulfillment process.

[0053] Step S104, input the package feature data of the target package into the pre-trained digital twin model to obtain the digital twin corresponding to the target package.

[0054] Here, the structured package feature data is used as the input and loaded into the digital twin model platform. The digital twin model constructs the corresponding digital twin based on the package feature data of the target package and simulates the transportation process of the target package in the current environment.

[0055] Specifically, first, a large amount of sample data is extracted from the historical shipment records, and a training sample set is constructed according to the sample data. The sample data includes the basic features of the package (such as size, weight, product identifier, delivery address, transportation method, etc.), the actual logistics path, transportation timeliness, transportation cost, in-transit delay records and path risk events.

[0056] The digital twin model adopts a multi-input multi-output neural network structure. The input is the standardized package feature vector, and the output is multiple feasible logistics paths and the corresponding estimated time, cost and risk scores. In the training process, the model integrates information from multiple dimensions, including: parameters such as the volume and weight of the package in the physical dimension; path nodes, transfer frequencies, and transportation method structures in the path dimension; the frequency of customs clearance policy changes in the destination country in the policy dimension; fuel price and exchange rate fluctuations in the economic dimension; path congestion probability and historical delay ratio in the risk dimension, etc.

[0057] After training is completed, the digital twin model is deployed to the online prediction module for real-time processing of package data to be shipped.

[0058] After the target package is identified, the package feature data of the target package is encoded and standardized to form an input vector, which is sent to the digital twin model for processing. The digital twin model outputs multiple candidate transportation paths of the target package in the current logistics environment according to the current logistics network status, external cost index and customs clearance policy information, and each path is attached with the corresponding timeliness prediction, cost prediction and risk score.

[0059] The digital twin model generates a digital twin of the target package based on the package feature data, and the digital twin serves as the core simulation object for subsequent scoring, path screening, standard trajectory generation, and deviation comparison.

[0060] Step S105: Based on the digital twin model, perform simulation on the digital twin to generate the standard logistics trajectory and standard transportation cost corresponding to the target package.

[0061] Here, the digital twin model first generates multiple candidate logistics paths, simulates different transportation routes that the package may take, synchronously obtains logistics-related data from the external system interface, and calculates the transportation timeliness score, transportation risk score, compliance score, and economic cost score for each candidate path. It uses multi-objective optimization (such as Pareto front analysis) to screen out compliance paths, safety paths, and economic paths, and determines one of them as the standard logistics trajectory according to the preset path selection rule, and calculates the standard transportation cost of this path through the transportation cost estimation model.

[0062] In one embodiment, the steps of step S105 include the following steps S301 - S306.

[0063] Step S301: Input the digital twin into the digital twin model to output multiple candidate logistics trajectories; the candidate logistics trajectories are the feasible paths for simulating the target package under different transportation scenarios.

[0064] Here, input the digital twin of the target package that has been constructed (i.e., the virtual model containing information such as package size, weight, destination, and transportation mode) into the digital twin model platform. The digital twin model generates multiple feasible transportation paths based on package attributes, historical data distribution, current transportation resources, and time periods. Each candidate trajectory may include transfer nodes along the way, estimated arrival time, and combinations of transportation modes involved.

[0065] Step S302: Through a preset external interface, obtain the logistics-related data corresponding to the target package; the logistics-related data includes current logistics network status information, transportation rule information, and economic cost index information.

[0066] Here, obtain the logistics-related data from the logistics platform, third-party data providers, or industry regulatory platforms through the preset external interface API.

[0067] The current logistics network status information includes the current load rate, congestion level, and traffic efficiency of each transportation route, etc.

[0068] The transportation rule information includes destination country customs restrictions, embargoed categories, transportation mode restrictions, and holiday impacts, etc.

[0069] The economic cost index information includes the current exchange rate, transportation fuel price, international freight index, etc.

[0070] Step S303: Calculate the scoring results of each candidate logistics track based on the logistics correlation data and the preset scoring rules; the scoring results include transportation timeliness score, transportation risk score, compliance score, and transportation economic cost score.

[0071] Here, the transportation timeliness score represents the expected time consumption and timeliness volatility of the path under the current network state.

[0072] The transportation risk score represents the risks such as possible delays, jams, and interruptions of the path.

[0073] The compliance score represents whether the path complies with the policies, regulations, and customs clearance requirements of the destination, etc.

[0074] The transportation economic cost score represents the expected cost corresponding to the path (including freight, surcharges, fuel surcharges, etc.).

[0075] Each score can adopt a unified standard score, and different weights can be set for different scores according to the actual situation.

[0076] Step S304: Generate a candidate path set including the path with the best compliance, the path with the best safety, and the path with the best economy based on the scoring results.

[0077] Here, according to the scoring results, perform multi-objective clustering or Pareto analysis on the candidate paths, and extract the candidate path set, which includes at least one candidate path.

[0078] The path with the best compliance is the path with the highest compliance score and is applicable to policy-sensitive scenarios.

[0079] The path with the best safety is the path with the lowest risk score and is applicable to scenarios with high requirements such as fragile and cold chain.

[0080] The path with the best economy is the path with the lowest cost score and is applicable to cost-oriented orders.

[0081] Step S305: Select an optimal path from the candidate path set as the standard logistics track of the target package according to the preset path selection rules.

[0082] Here, automatically select one of the paths as the standard path according to the preset path selection rules.

[0083] Specifically, the path selection rule can be the weighted scoring method, that is, corresponding weights are set for the transportation timeliness score, transportation risk score, compliance score, and transportation economic cost score, for example, 40%, 10%, 15%, and 35% respectively. The weighted total score is calculated based on the scoring results of each candidate path, and the path with the highest score is selected as the standard logistics track.

[0084] It can also be the priority decision-making method, that is, the dominant decision-making dimension is set according to the package attributes or business types. For example, for high-value or fragile goods, the path with the lowest transportation risk score is preferentially selected; for price-sensitive orders, the path with the lowest transportation cost is preferentially selected; for promotional orders that require quick fulfillment, the path with the highest transportation timeliness score is preferentially selected.

[0085] The constraint screening and sorting method can also be adopted, that is, first, hard thresholds such as compliance score, risk score, and cost ceiling are set, and the paths that do not meet the constraint conditions are excluded. Then, the remaining paths are sorted according to the main scoring dimension, and the optimal path is selected.

[0086] Step S306, based on the standard logistics track, calculate the standard transportation cost corresponding to the standard logistics track through the transportation cost estimation model.

[0087] Here, the transportation cost estimation model is a cost prediction model trained based on historical data. Its inputs include the path node sequence, transportation mode, package size and weight, estimated transportation duration, and fuel price, etc.; the output is the estimated total cost of the standard path, which can be split into sub-items such as transportation cost, additional cost, and platform commission.

[0088] Step S106, after the target package is shipped, obtain the actual logistics track, actual transportation cost, and current logistics environment information of the target package.

[0089] In one embodiment, the current logistics environment information includes the logistics network load rate, the historical transportation delay probability of the logistics path, the volatility of transportation fuel cost, and the destination customs clearance stability index.

[0090] Here, the logistics network load rate represents the congestion degree or resource utilization rate of each main transportation network, and can be obtained from the logistics service provider through a preset interface API.

[0091] The historical transportation delay probability of the logistics path is used to count the frequency of delays on this path or similar paths within the current time window.

[0092] The volatility of transportation fuel cost is based on fuel futures or public freight rate data, and is used to calculate the fluctuation range of the current fuel price relative to the stable interval.

[0093] The destination customs clearance stability index represents the recent policy adjustment frequency, inspection rate, or return rate in the destination country, etc., and is used to measure the uncertainty of the destination customs clearance environment.

[0094] After the step of obtaining the current logistics environment information of the target package after the target package is shipped in step S106, the method further includes the following steps S401 - S402.

[0095] Step S401, based on the logistics network load rate, the historical transportation delay probability of the logistics path, the destination customs clearance stability index, and a preset logistics risk calculation method, calculate the logistics risk coefficient, and generate a logistics anomaly threshold based on the logistics risk coefficient and the initial logistics anomaly threshold.

[0096] Here R w = α logistics network load rate + β historical transportation delay probability of the logistics path + γ destination customs clearance stability index

[0097] Among them, R w is the logistics risk coefficient, and α, β, γ are preset weight parameters, which can be pre - configured according to the actual situation.

[0098] Perform weighted adjustment on the logistics risk coefficient and the preset initial logistics anomaly threshold to generate a logistics anomaly threshold that takes effect in real - time, T w = T w0 ×(1 + R w )

[0099] Among them, T w is the logistics anomaly threshold, and T w0 is the initial logistics anomaly threshold.

[0100] Step S402, based on the transportation fuel cost volatility, the destination customs clearance stability index, and a preset cost risk coefficient calculation method, calculate the cost risk coefficient, and generate a cost anomaly threshold based on the cost risk coefficient and the initial cost anomaly threshold.

[0101] Specifically, based on the transportation fuel cost volatility and the customs clearance stability index, calculate the current cost risk coefficient, R f = θ transportation fuel cost volatility + δ(1 - customs clearance stability index).

[0102] Among them, R f is the cost risk coefficient, and θ, × are preset weight coefficients.

[0103] Generate a cost anomaly threshold according to the cost risk coefficient and the initial cost anomaly threshold, T f = T f0 ×(1 + R f )

[0104] Among them, Tf is the cost anomaly threshold, T f0 is the initial cost anomaly threshold.

[0105] If the logistics deviation value is greater than T w , a path anomaly alarm is triggered.

[0106] If the cost deviation value is greater than T f , a cost deviation alarm is triggered.

[0107] In one embodiment, after the step of obtaining the actual logistics track of the target package after the target package is shipped in step S106, the method further includes the following steps S501 - S504.

[0108] Step S501: When receiving the logistics node update information returned by the logistics service provider system or when reaching a preset first time interval, update the actual logistics track to obtain the current actual logistics track.

[0109] Here, the logistics nodes include node name (such as: departure center, customs port, overseas warehouse transfer point, delivery station), timestamp (node scanning or completion time), node status (arrival, outbound, customs clearance completed, in transit, etc.) and optional fields (latitude and longitude, responsible service provider, transportation mode, etc.).

[0110] When receiving new node update information or when reaching a preset time interval (such as every 12 hours), append the new node data to the track record of the target package to construct the current actual logistics track.

[0111] Step S502: Output the current standard logistics track corresponding to the current actual logistics track through the digital twin model.

[0112] Here, the package feature data at the current time point and the real - time logistics environment information are used as inputs, the digital twin model is called, the state that the target package should be in at this time is re - simulated in the simulation environment, and its current standard logistics track is output. The current standard logistics track is used to represent the sequence of nodes that the package should pass through under normal or optimal transportation conditions.

[0113] Among them, the standard track nodes and the actual track nodes have the same format.

[0114] Step S503: Construct a track matching distance matrix based on the current actual logistics track and the current standard logistics track.

[0115] Here, for the current actual logistics track A = {a1, a2,..., a m} and the current standard logistics track B = {b1, b2,..., b n}Perform node-level matching. Construct a trajectory matching distance matrix D(m×n), where each element D(i,j) in the matrix represents the matching distance (or degree of difference) between the actual node a i and the standard node b j The matching distance (or degree of difference) between them.

[0116] The distance definition can include whether the node names are the same, whether the node states are the same, the time difference in logistics status, the difference in spatial location information, route deviation, or cross-node jump. The smaller the matching distance, the closer the two nodes are.

[0117] Step S504, calculate the logistics deviation value between the current actual logistics trajectory and the current standard logistics trajectory based on the dynamic time warping algorithm, and generate the optimal matching path; the optimal matching path is used to describe the corresponding relationship between the standard nodes in the current standard logistics trajectory and the actual nodes in the current actual logistics trajectory.

[0118] Here, based on the constructed matching distance matrix, use the dynamic time warping algorithm to find the shortest path from D(0,0) to D(m,n) (i.e., the optimal matching path), and the shortest path describes the optimal alignment relationship between the actual trajectory and the standard trajectory.

[0119] The DTW cost function is used to accumulate the deviation of each pair of nodes, and the total matching cost output is the logistics deviation value at this time point. At the same time, the path alignment sequence is obtained (such as a2 corresponding to b1, a3 corresponding to b2).

[0120] Cache the above optimal matching path as a structured result, which clearly identifies which standard nodes are successfully matched (indicating normal), which standard nodes are not matched (possibly jump points), which actual nodes have no corresponding standard nodes (possibly detours), the time difference, spatial direction difference, etc. of each pair of matching nodes.

[0121] Step S107, when the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics anomaly threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost anomaly threshold, generate an alarm message; where the logistics anomaly threshold and the cost anomaly threshold are both dynamic thresholds adaptively determined according to the current logistics environment information.

[0122] In one embodiment, the step of generating an alarm message if the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics anomaly threshold in step S107 includes the following steps S601-S606:

[0123] Step S601, if the logistics deviation value exceeds the logistics anomaly threshold, determine that the target package has a trajectory anomaly and obtain the optimal matching path.

[0124] Here, first compare the logistics deviation value with the logistics anomaly threshold. If the logistics deviation value exceeds the logistics anomaly threshold, it is considered that the current deviation is within the acceptable range and no alarm is triggered; if the logistics deviation value does not exceed the logistics anomaly threshold, it is determined that there is a trajectory anomaly and the anomaly type analysis is started.

[0125] Step S602, if there is a standard node in the best matching path that is not matched with the actual node, determine that the trajectory anomaly is a jump point anomaly.

[0126] Here, the jump point anomaly means that a certain standard node fails to match any node in the actual trajectory, indicating that the key node has not been passed by. Its application scenarios include skipping scans, missed inspections, and direct shipments bypassing transfer stations, ports, and overseas warehouses, indicating that there may be behaviors such as uncustoms clearance, unbilling, and unhandover in the above application scenarios.

[0127] Specifically, if the standard trajectory is B = {b1, b2, b3, b4, b5}, the actual trajectory is A = {a1, a2, a4, a5}, and the best matching path is P = {b1 - a1, b2 - a2, b3 - NULL, b4 - a4, b5 - a5}, then determine that the trajectory anomaly is a jump point anomaly.

[0128] Step S603, if there is an actual node in the best matching path that is not matched with the standard node, determine that the trajectory anomaly is a detour anomaly.

[0129] Here, if there is a node in the actual trajectory that fails to match any standard node (such as a x - NULL), it means that the package has passed through an unexpected path, and there is a detour anomaly. Its application scenarios include warehouse transfer, long - way transfer, and wrong sorting, indicating that the above scenarios may lead to an extended transportation cycle, increased costs, or policy restrictions.

[0130] Step S604, if the residence time of any actual node exceeds the sum of the standard residence time of its corresponding standard node and the preset residence threshold, determine that the trajectory anomaly is a detention anomaly.

[0131] Here, through the time - stamp difference calculation, it is determined that the residence time of a certain actual node a i exceeds the standard residence time of its corresponding standard node b i plus the offset tolerance threshold ΔT, and determine that the trajectory anomaly is a detention anomaly. Its application scenarios include package backlog, customs clearance stagnation, and warehouse non - shipment. The detention anomaly is likely to cause customer complaints or disputes about lost packages.

[0132] Step S605, if the geographical direction of the actual node is inconsistent with the receiving direction of the target package, determine that the trajectory anomaly is a reverse travel anomaly.

[0133] Here, by comparing the longitude and latitude directions or the regional level path sequences, if it is determined that the geographical direction of a certain actual node significantly deviates from the direction of the region where the final receiving address of the package is located (such as moving in the opposite direction exceeding the set angle threshold or the geographical region level gap), then it is determined that the trajectory anomaly is a retrograde anomaly. Its application scenarios are dispatching errors, address information tampering, or return transit, indicating that there are suspected risks of misclassification, return, and misdelivery in the above application scenarios.

[0134] Step S606, generate an alarm message based on at least one type of trajectory anomaly.

[0135] Here, when any of the above trajectory anomalies is identified, an alarm message can be generated. The alarm message includes at least the package unique identifier, the current time point, the anomaly type, the anomaly node or path details, and the associated risk description and recommended handling actions (such as: abort, remind for manual review, notify the customer service, etc.).

[0136] In an embodiment, if the logistics mode of the target package is the overseas warehouse shipping mode, and after the target package arrives at the overseas warehouse and is sent from the overseas warehouse to its corresponding receiving address, the method further includes the following steps S701 - S706.

[0137] Here, if the logistics mode of the target package is the overseas warehouse shipping mode, after the package enters the overseas warehouse and is sent out from the overseas warehouse, a reconciliation verification will be performed based on the overseas warehouse outbound data and the shipping information of the e-commerce platform to detect possible duplicate outbound shipments, misdeliveries, and shipping anomalies in the outbound link, and generate an anomaly alarm message in a timely manner.

[0138] Step S701, when a preset second time interval is reached, obtain the outbound records of the overseas warehouse; the outbound records include the outbound order number, the outbound commodity identifier, the outbound receiving address, and the outbound time.

[0139] Here, after the target package is out of the warehouse and enters the stage of being delivered to the user, if the preset second time interval is met (which can be set in advance according to the actual situation, and can be set as a 12-hourly or daily scheduled task), the data reconciliation process will be automatically started.

[0140] The outbound order number is used to identify the operation behavior. The outbound commodity identifier refers to the SKU. The outbound receiving address is the recipient address recorded in the system. The outbound time is the scan or outbound processing timestamp.

[0141] Step S702, obtain the set of shipped order information within the preset second time interval from the e-commerce platform; the set of shipped order information includes the shipped order number, the shipped commodity identifier, the shipped receiving address, and the shipped time.

[0142] Here, pull the set of shipped order information within this time period from the e-commerce platform through the interface.

[0143] Compare the outbound record and the shipped order information set item by item by field.

[0144] Step S703, if any outbound order number in the outbound record corresponds to multiple shipped order numbers, determine that there is an abnormal duplicate outbound in the overseas warehouse.

[0145] Here, if the same outbound order number (i.e., the order number scanned by the warehouse) is associated with multiple shipped order numbers on e-commerce platforms, it indicates that the warehouse document has been used multiple times, or there are situations where the system fails to intercept and manual duplicate shipments occur. Determine that there is an abnormal duplicate outbound in the overseas warehouse.

[0146] Abnormal duplicate outbound refers to situations such as the platform order status not being refreshed in a timely manner due to network latency, and the customer service reissuing and manually generating a new order, which may cause duplicate package deliveries, additional transportation, and return costs.

[0147] Step S704, if the outbound quantity in the records with the same outbound commodity identifier in the outbound record is inconsistent with the shipped quantity of the same shipped commodity identifier in the corresponding shipped order information set, determine that there is an abnormal misdelivery in the overseas warehouse.

[0148] Here, count the outbound quantity of the same commodity identifier (SKU) in all outbound records and compare it with the shipped quantity of the same commodity identifier during the corresponding period on the e-commerce platform. If the quantities are inconsistent, it indicates that there are situations of misdelivery, missing delivery, or over-delivery. Determine that there is an abnormal misdelivery in the overseas warehouse.

[0149] Abnormal misdelivery includes picking errors, SKU mis-scanning, and wrong boxing in combined shipments, which may lead to customer complaints, refund requests, and imbalance in warehouse inventory.

[0150] Step S705, if the outbound time of any outbound order number in the outbound record is not within the preset outbound time range of its corresponding shipped time, or the outbound receiving address corresponding to any outbound order number is inconsistent with its corresponding shipped receiving address, determine that there is an abnormal shipment in the overseas warehouse.

[0151] Here, for each outbound order number in the outbound record, if the outbound time is not within the preset allowable range of its shipped time (e.g., the deviation shall not exceed 12 hours before or after), or its outbound receiving address is inconsistent with the receiving address of the corresponding order on the platform, it is regarded as an abnormal shipment.

[0152] Abnormal shipment refers to shipment delays, address tampering, and wrong label pasting on packages, etc., which are likely to lead to complaints, mis-delivery, delay claims, or difficulties in logistics tracing.

[0153] Step S706, generate an alarm message based on at least one of the abnormal duplicate outbound, abnormal misdelivery, and abnormal shipment.

[0154] Here, once any abnormal situation is identified, the alarm mechanism can be triggered to generate alarm information.

[0155] By reconciling the time periods and comparing the fields between the outbound records of the overseas warehouse and the shipping data of the e-commerce platform, this application can accurately identify various typical outbound abnormal behaviors in the cross-border warehousing environment, effectively reduce the operational losses and customer complaints caused by duplicate shipments, misdeliveries, and shipping errors, and significantly improve the warehousing collaboration efficiency and fulfillment accuracy.

[0156] In one embodiment, after the steps of step S702, the method further includes the following steps S801 - S803.

[0157] Step S801, obtain the warehousing billing data corresponding to the overseas warehouse.

[0158] Here, through an interface or batch import method, obtain the warehousing expense detail data related to the preset time period from the overseas warehouse management system or the warehousing billing module.

[0159] The warehousing billing data includes, but is not limited to, the corresponding outbound order number, product identifier (SKU), warehousing start and end times (inbound time, outbound time), billing days or hours, unit billing standard, total payable fees (which can be split into storage fees, operation fees, and additional service fees, etc.), as well as currency and settlement method.

[0160] Step S802, calculate the actual warehousing fees based on the set of shipped order information and the preset warehousing billing method.

[0161] Here, combine the set of shipped order information obtained from the e-commerce platform and the preset warehousing billing method rule table to estimate the warehousing fees for each order independently.

[0162] Specifically, according to the order number or SKU information, match each order with its inbound and outbound records in the warehouse.

[0163] Warehousing cycle = outbound time - inbound time.

[0164] If charged by the piece, the fee = unit price × number of days × number of pieces.

[0165] If charged by volume, the fee = unit price × number of days × volume per piece × number of pieces.

[0166] A multi-level billing rule can also be adopted, for example, X yuan per day for the first 5 days and Y yuan per day starting from the 6th day.

[0167] Output the calculated actual warehousing fee list by order dimension, SKU dimension, or billing period dimension.

[0168] In step S803, if the warehousing billing data is different from the actual warehousing cost, it is determined that there is a billing anomaly in the overseas warehouse, and a billing anomaly alarm message is generated.

[0169] Here, the calculated actual warehousing cost is compared with the warehousing billing data obtained from the warehouse system field by field. If any of the following situations exist, it is determined that there is a billing anomaly in the overseas warehouse.

[0170] There is no matching warehousing record for the corresponding order number; or, the cost value has a deviation exceeding the preset tolerance (such as ±5%); or, the charging dimensions are inconsistent (such as incorrect mixing of piece counting and cubic meter counting); or, the billing start and end times are inconsistent (such as the outbound time is advanced but the cost is not reduced); or, the same order is billed multiple times, or there are duplicate charging records.

[0171] Once an anomaly is identified, a structured billing anomaly alarm message will be automatically generated. The billing anomaly alarm message includes the abnormal order number, SKU or storage unit, the comparison difference between the calculated cost and the cost in the warehouse system, the determination of the abnormal reason (such as duplicate billing, undercounted time, inconsistent rules, etc.) and the recommended handling actions (such as notifying the finance department for review, freezing payment, pushing to the warehouse for confirmation, etc.).

[0172] By establishing a three-party comparison mechanism of order-warehousing record-cost, this application realizes the automated verification of the reasonableness, integrity and accuracy of warehousing costs without relying on manual review, effectively avoiding operational losses, customer disputes or financial disputes caused by billing errors, and significantly improving the transparency and settlement efficiency of overseas warehouse cost management.

[0173] A cross-border logistics order management method provided by an embodiment of the present application includes: obtaining package feature data of a package to be shipped; comparing the similarity between the package feature data and historical package feature data in historical shipping records to obtain a similarity comparison result; if the similarity comparison result does not exceed a preset similarity threshold, determining the package to be shipped as a target package; inputting the package feature data of the target package into a pre-trained digital twin model to obtain a digital twin corresponding to the target package; performing simulation on the digital twin based on the digital twin model to generate a standard logistics track and a standard transportation cost corresponding to the target package; after the target package is shipped, obtaining the actual logistics track, actual transportation cost, and current logistics environment information of the target package; when the logistics deviation value between the actual logistics track and the standard logistics track exceeds the logistics anomaly threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost anomaly threshold, generating an alarm message; where the logistics anomaly threshold and the cost anomaly threshold are both dynamic thresholds adaptively determined according to the current logistics environment information. In this method, by constructing a cross-border logistics order management method based on a digital twin model, combining package feature perception, standard track simulation, real-time path monitoring, and overseas warehouse outbound verification, dynamic simulation, anomaly identification, and cost auditing of the entire logistics process are realized. By introducing a similarity comparison mechanism between package feature vectors and historical records, high-precision identification of duplicate shipments can be achieved, so as to give an early warning in time before shipment and avoid waste of resources caused by repeated performance; by training the digital twin model to construct a twin body for the target package and generating a standard logistics track and a standard transportation cost in the current logistics environment, dynamic simulation of path prediction and cost estimation is realized; after the package is shipped, the actual track and the standard track are compared in real time through the dynamic time warping algorithm to identify various transportation anomalies such as jump points, detours, stays, and retrogrades, and the early warning conditions are adaptively judged according to the dynamic threshold generated by the logistics environment information, so as to improve the accuracy and robustness of anomaly identification. In addition, in the overseas warehouse shipping mode of the present application, by comparing the time periods of outbound records and platform shipping information, accurate identification of shipping anomaly behaviors such as repeated outbound, misdelivery, and inconsistent addresses is realized; and by constructing a warehousing cost accounting logic and a pre-designed fee rule, automatic bill checking and abnormal billing alarm after outbound are completed, significantly improving the accuracy and efficiency of overseas warehouse cost settlement. Based on this, the present application can realize the full-link intelligent control from pre-shipment warning, in-shipment monitoring to post-shipment settlement, thereby reducing logistics risks, optimizing the cost structure, and improving the cross-border performance quality and management intelligence level.

[0174] Embodiment 2:

[0175] Figure 2 It is a schematic diagram of a cross-border logistics order management system provided by an embodiment of the present application.

[0176] Referring to Figure 2 , the cross-border logistics order management system includes:

[0177] The package feature data acquisition module 1 is used to acquire the package feature data of the package to be shipped.

[0178] The target package judgment module 2 is used to compare the similarity between the package feature data and the historical package feature data in the historical shipping records to obtain the similarity comparison result.

[0179] The target package judgment module 2 is also used to determine the package to be shipped as the target package if the similarity comparison result does not exceed the preset similarity threshold.

[0180] The standard information generation module 3 is used to input the package feature data of the target package into a pre-trained digital twin model to obtain the digital twin corresponding to the target package.

[0181] The standard information generation module 3 is also used to perform simulation on the digital twin based on the digital twin model to generate the standard logistics track and standard transportation cost corresponding to the target package.

[0182] The real-time monitoring module 4 is used to acquire the actual logistics track, actual transportation cost and current logistics environment information of the target package after the target package is shipped.

[0183] The alarm information generation module 5 is used to generate alarm information when the logistics deviation value between the actual logistics track and the standard logistics track exceeds the logistics anomaly threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost anomaly threshold; wherein, both the logistics anomaly threshold and the cost anomaly threshold are dynamic thresholds adaptively determined according to the current logistics environment information.

[0184] In an optional implementation manner, the package feature data acquisition module 1 is further used for:

[0185] Collect the image information of the package to be shipped through an image acquisition device; the image information includes the packaging form, barcode area and waybill information.

[0186] Obtain the order information corresponding to the package to be shipped from an e-commerce platform; the order information includes the order number, product identifier, shipping address, shipping time, transportation method and logistics mode.

[0187] Obtain the physical attribute information and processing requirements corresponding to the package to be shipped from a commodity management platform.

[0188] Integrate the image information, order information, physical attribute information and processing requirements into package feature data.

[0189] In an optional implementation manner, the target package judgment module 2 is further used for:

[0190] If the similarity comparison result exceeds the preset similarity threshold, determine the package to be dispatched as a duplicate shipment package and generate a duplicate shipment alarm message.

[0191] In an alternative embodiment, the standard information generation module 3 is further configured to:

[0192] Input the digital twin into the digital twin model to output multiple candidate logistics trajectories; the candidate logistics trajectories are the feasible paths of the simulated target package under different transportation plans.

[0193] Obtain the logistics association data corresponding to the target package through a preset external interface; the logistics association data includes the current logistics network status information, transportation rule information, and economic cost index information.

[0194] Based on the logistics association data and the preset scoring rules, calculate the scoring results of each candidate logistics trajectory; the scoring results include transportation timeliness scoring, transportation risk scoring, compliance scoring, and transportation economic cost scoring.

[0195] Based on the scoring results, generate a candidate path set including the optimal path for compliance, the optimal path for safety, and the optimal path for economy.

[0196] According to the preset path selection rules, select an optimal path from the candidate path set as the standard logistics trajectory of the target package.

[0197] Based on the standard logistics trajectory, calculate the standard transportation cost corresponding to the standard logistics trajectory through a transportation cost estimation model.

[0198] In an alternative embodiment, the current logistics environment information includes the logistics network load rate, the historical transportation delay probability of the logistics path, the volatility of transportation fuel costs, and the destination customs clearance stability index.

[0199] The real-time monitoring module 4 is further configured to:

[0200] Based on the logistics network load rate, the historical transportation delay probability of the logistics path, the destination customs clearance stability index, and the preset logistics risk calculation method, calculate the logistics risk coefficient, and generate a logistics anomaly threshold based on the logistics risk coefficient and the initial logistics anomaly threshold.

[0201] Based on the volatility of transportation fuel costs, the destination customs clearance stability index, and the preset cost risk coefficient calculation method, calculate the cost risk coefficient, and generate a cost anomaly threshold based on the cost risk coefficient and the initial cost anomaly threshold.

[0202] In an alternative embodiment, the real-time monitoring module 4 is further configured to:

[0203] When the logistics node update information returned by the logistics service provider system is received, or when the preset first time interval is reached, the actual logistics trajectory is updated to obtain the current actual logistics trajectory.

[0204] The digital twin model outputs the current standard logistics trajectory corresponding to the current actual logistics trajectory.

[0205] Construct a trajectory matching distance matrix based on the current actual logistics trajectory and the current standard logistics trajectory.

[0206] Based on the dynamic time warping algorithm, the logistics deviation value between the current actual logistics trajectory and the current standard logistics trajectory is calculated, and the best matching path is generated; the best matching path is used to describe the correspondence between the standard nodes in the current standard logistics trajectory and the actual nodes in the current actual logistics trajectory.

[0207] In an optional embodiment, the alarm information generating module 5 is further configured to:

[0208] If the logistics deviation value exceeds the logistics anomaly threshold, it is determined that the target package has a trajectory anomaly and the best matching path is obtained.

[0209] If there is a standard node in the best matching path that does not match the actual node, the trajectory anomaly is determined to be a jump point anomaly.

[0210] If there is an actual node in the best matching path that does not match the standard node, the trajectory anomaly is determined to be a detour anomaly.

[0211] If the residence time of any actual node exceeds the sum of the standard residence time of its corresponding standard node and the preset residence threshold, the trajectory anomaly is determined to be a residence anomaly.

[0212] If the geographical direction of the actual node is inconsistent with the delivery direction of the target package, the trajectory anomaly is determined to be a retrograde anomaly.

[0213] Based on at least one trajectory anomaly, an alarm message is generated.

[0214] In an optional embodiment, if the logistics mode of the target package is the overseas warehouse delivery mode, and after the target package arrives at the overseas warehouse and sends its corresponding delivery address from the overseas warehouse, the alarm information generation module 5 is further used to:

[0215] When the preset second time interval is reached, the outbound shipment record of the overseas warehouse is obtained; the outbound shipment record includes the outbound order number, outbound product identification, outbound delivery address and outbound time.

[0216] Obtain a set of shipped order information within a preset second time interval from the e-commerce platform; the shipped order information set includes a shipped order number, a shipped product identifier, a shipped delivery address, and a shipped time.

[0217] If any outbound order number in the outbound record corresponds to multiple shipped order numbers, it is determined that there is a duplicate outbound anomaly in the overseas warehouse.

[0218] If the outbound quantity in the record with the same outbound product ID in the outbound record is inconsistent with the shipped quantity of the same shipped product ID in the corresponding shipped order information set, it is determined that there is an error in the overseas warehouse.

[0219] If the shipment time of any shipment order number in the shipment record is not within the preset shipment time range of its corresponding shipped time, or the shipment delivery address corresponding to any shipment order number is inconsistent with its corresponding shipped delivery address, it is determined that there is a shipment abnormality in the overseas warehouse.

[0220] Generate an alarm message based on at least one of the duplicate shipment exception, wrong shipment exception and delivery exception.

[0221] In an optional embodiment, the alarm information generating module 5 is further configured to:

[0222] Get the storage billing data corresponding to the overseas warehouse.

[0223] Calculate actual storage fees based on the shipped order information set and the preset storage billing method.

[0224] If the warehousing billing data is different from the actual warehousing fee, it is determined that there is a billing anomaly in the overseas warehouse, and a billing anomaly alarm message is generated.

[0225] The cross-border logistics order management system provided in this application embodiment utilizes a modular design to not only enhance the flexibility and scalability of system functions, but also facilitates integration with logistics platforms, e-commerce systems, and warehouse management systems, thereby enabling efficient collaboration across different business processes for cross-border logistics orders. This system deployment significantly improves the level of automated management throughout the cross-border logistics process, reduces the cost of human intervention for exception handling, and enhances full-process visibility and intelligent decision-making capabilities.

[0226] The computer program product provided in the embodiments of the present application includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the previous method embodiments. For specific implementation, please refer to the method embodiments and will not be repeated here.

[0227] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems and devices can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0228] In addition, in the description of the embodiments of the present application, unless otherwise clearly defined and limited, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be a direct connection or an indirect connection through an intermediate medium, and it may be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific situations.

[0229] If the above-mentioned functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0230] The embodiments of the present application also provide an electronic device, as Figure 3 shown, which is a schematic structural diagram of the electronic device. Among them, the electronic device includes a processor 301 and a memory 302. The memory 302 stores computer-executable instructions that can be executed by the processor 301. The processor 301 executes the computer-executable instructions to implement the above-mentioned method for identifying the path to be planned.

[0231] In Figure 3 the shown embodiment, the electronic device further includes a bus 303 and a communication interface 304. Among them, the processor 301, the communication interface 304, and the memory 302 are connected through the bus 303.

[0232] Among them, the memory 302 may include high-speed random access memory (RAM), and may also include non-volatile memory (non-volatile memory), such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 304 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 303 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an E ISA (Extended Industry Standard Architecture) bus, etc. The bus 303 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0233] In the description of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate the description of this application and simplify the description. They do not indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0234] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection of the claims.

Claims

1. A cross-border logistics order management method, characterized in that: include: Obtain package feature data of the package to be sent; Comparing the package feature data with historical package feature data in historical shipping records to obtain a similarity comparison result; If the similarity comparison result does not exceed the preset similarity threshold, the package to be sent is determined to be the target package; Inputting the package feature data of the target package into a pre-trained digital twin model to obtain a digital twin corresponding to the target package; Simulating the digital twin based on the digital twin model to generate a standard logistics trajectory and standard transportation cost corresponding to the target package; After the target package is shipped, obtaining the actual logistics track, actual transportation cost and current logistics environment information of the target package; When the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics abnormality threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds the cost abnormality threshold, an alarm message is generated; wherein, the logistics abnormality threshold and the cost abnormality threshold are both dynamic thresholds adaptively determined based on the current logistics environment information.

2. The cross-border logistics order management method according to claim 1, wherein The step of obtaining the package characteristic data of the package to be sent includes: Capturing image information of the package to be shipped by an image acquisition device; the image information includes packaging form, barcode area and label information; Obtaining order information corresponding to the package to be shipped from the e-commerce platform; the order information includes order number, product identifier, delivery address, shipping time, transportation method and logistics mode; Obtaining physical attribute information and processing requirements corresponding to the package to be shipped from the product management platform; The image information, the order information, the physical attribute information and the processing requirements are integrated into the package feature data.

3. The cross-border logistics order management method according to claim 1, wherein After the step of performing a similarity comparison between the package feature data and historical package feature data in historical shipping records to obtain a similarity comparison result, the method further includes: If the similarity comparison result exceeds the preset similarity threshold, the package to be shipped is determined to be a duplicate shipment package, and a duplicate shipment alarm message is generated.

4. The cross-border logistics order management method according to claim 1, characterized in that: The step of simulating the digital twin based on the digital twin model to generate a standard logistics trajectory and standard transportation cost corresponding to the target package includes: Inputting the digital twin into the digital twin model, and outputting a plurality of candidate logistics trajectories; the candidate logistics trajectories are feasible paths for simulating the target package under different transportation schemes; Obtaining logistics-related data corresponding to the target package through a preset external interface; the logistics-related data includes current logistics network status information, transportation rule information, and economic cost indicator information; Calculate a scoring result for each candidate logistics trajectory based on the logistics-related data and preset scoring rules; the scoring result includes a transportation timeliness score, a transportation risk score, a compliance score, and a transportation economic cost score; Based on the scoring results, a candidate path set is generated, which includes a compliance-optimal path, a security-optimal path, and an economic-optimal path; Select an optimal path from the candidate path set as the standard logistics trajectory of the target package according to the preset path selection rules; Based on the standard logistics trajectory, calculate the standard transportation cost corresponding to the standard logistics trajectory through the transportation cost estimation model.

5. The cross-border logistics order management method according to claim 1, wherein The current logistics environment information includes the logistics network load rate, the historical transportation delay probability of the logistics path, the volatility of transportation fuel costs, and the destination customs clearance stability index; After the step of obtaining the current logistics environment information of the target package after the target package is shipped, the method further includes: Based on the logistics network load rate, the historical transportation delay probability of the logistics path, the destination customs clearance stability index, and the preset logistics risk calculation method, calculate the logistics risk coefficient, and generate a logistics anomaly threshold based on the logistics risk coefficient and the initial logistics anomaly threshold. Based on the volatility of transportation fuel costs, the destination customs clearance stability index, and the preset cost risk coefficient calculation method, calculate the cost risk coefficient, and generate a cost anomaly threshold based on the cost risk coefficient and the initial cost anomaly threshold.

6. The cross-border logistics order management method according to claim 5, characterized in that: After the step of obtaining the actual logistics trajectory of the target package after the target package is shipped, the method further includes: When receiving the logistics node update information returned by the logistics service provider system or reaching the preset first time interval, update the actual logistics trajectory to obtain the current actual logistics trajectory; Output the current standard logistics trajectory corresponding to the current actual logistics trajectory through the digital twin model; Construct a trajectory matching distance matrix based on the current actual logistics trajectory and the current standard logistics trajectory; Calculate the logistics deviation value between the current actual logistics trajectory and the current standard logistics trajectory based on the dynamic time warping algorithm, and generate the best matching path; the best matching path is used to describe the corresponding relationship between the standard nodes in the current standard logistics trajectory and the actual nodes in the current actual logistics trajectory.

7. The cross-border logistics order management method according to claim 6, characterized in that: If the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds the logistics anomaly threshold, the step of generating an alarm message includes: If the logistics deviation value exceeds the logistics anomaly threshold, determine that the target package has a trajectory anomaly, and obtain the best matching path; If there are standard nodes in the best matching path that are not matched with the actual nodes, determine that the trajectory anomaly is a skip point anomaly; If there are actual nodes in the best matching path that are not matched with the standard nodes, determine that the trajectory anomaly is a detour anomaly; If the residence time of any actual node exceeds the sum of the standard residence time of its corresponding standard node and the preset residence threshold, determine that the trajectory anomaly is a detention anomaly; If the geographical direction of the actual node is inconsistent with the receiving direction of the target package, determine that the trajectory anomaly is a reverse anomaly; Generate the alarm message based on at least one of the trajectory anomalies.

8. The cross-border logistics order management method according to claim 1, wherein If the logistics mode of the target package is the overseas warehouse shipping mode, and after the target package arrives at the overseas warehouse and is sent to its corresponding receiving address from the overseas warehouse, the method further includes: When a preset second time interval is reached, obtaining the outbound shipment record of the overseas warehouse; the outbound shipment record includes the outbound order number, outbound commodity identifier, outbound delivery address and outbound time; Obtaining a set of shipped order information within the preset second time interval from the e-commerce platform; the set of shipped order information includes the shipped order number, the shipped product identifier, the shipped delivery address, and the shipped time; If any of the outbound order numbers in the outbound record corresponds to multiple shipped order numbers, it is determined that there is a duplicate outbound anomaly in the overseas warehouse; If the outbound quantity in the record with the same outbound commodity identifier in the outbound record is inconsistent with the shipped quantity with the same shipped commodity identifier in the corresponding shipped order information set, it is determined that there is a wrong shipment exception in the overseas warehouse; If the shipment time of any shipment order number in the shipment record is not within the preset shipment time range of the corresponding shipment time, or the shipment delivery address corresponding to any shipment order number is inconsistent with the corresponding shipped delivery address, it is determined that the overseas warehouse has the shipment anomaly; Based on at least one of the duplicate shipment exception, the wrong shipment exception, and the delivery exception, an alarm message is generated.

9. The cross-border logistics order management method according to claim 8, characterized in that: After the step of obtaining the set of information about shipped orders within the preset second time interval from the e-commerce platform, the method further includes: Obtain storage billing data corresponding to the overseas warehouse; Calculate actual storage fees based on the shipped order information set and a preset storage billing method; If the warehousing billing data is different from the actual warehousing fee, it is determined that there is a billing anomaly in the overseas warehouse, and a billing anomaly alarm message is generated.

10. A cross-border logistics order management system, characterized in that, include: A package feature data acquisition module is used to obtain the package feature data of the package to be sent; a target package determination module, configured to perform a similarity comparison between the package feature data and the historical package feature data in the historical delivery records to obtain a similarity comparison result; The target package determination module is further configured to determine that the package to be sent is the target package if the similarity comparison result does not exceed a preset similarity threshold; A standard information generation module is used to input the package feature data of the target package into a pre-trained digital twin model to obtain a digital twin corresponding to the target package; The standard information generation module is further configured to simulate the digital twin based on the digital twin model to generate a standard logistics trajectory and standard transportation cost corresponding to the target package; A real-time monitoring module is used to obtain the actual logistics track, actual transportation cost and current logistics environment information of the target package after the target package is shipped; An alarm information generation module is used to generate an alarm message when the logistics deviation value between the actual logistics trajectory and the standard logistics trajectory exceeds a logistics abnormality threshold, or the cost deviation value between the actual transportation cost and the standard transportation cost exceeds a cost abnormality threshold; wherein the logistics abnormality threshold and the cost abnormality threshold are both dynamic thresholds adaptively determined based on the current logistics environment information.

Images