CN120806803B

CN120806803B - Customer waybill straightening methods, devices, computer equipment and storage media

Info

Publication number: CN120806803B
Application number: CN202511276433.4A
Authority: CN
Inventors: 鄢宇; 张兴; 张玉成
Original assignee: Shenzhen Leap New Technology Co ltd
Current assignee: Shenzhen Leap New Technology Co ltd
Priority date: 2025-09-08
Filing date: 2025-09-08
Publication date: 2025-11-18
Anticipated expiration: 2045-09-08
Also published as: CN120806803A

Abstract

The application relates to the technical field of logistics information processing, and provides a customer freight bill straightening method, a device, computer equipment and a storage medium, wherein the method obtains freight bill data to be straightened according to the dimensions of a mail company and a mail point part of the customer freight bill data; removing the waybill data with the number of times of allocation flow less than or equal to 1 from the to-be-straightened waybill data, obtaining the attribution allocation information corresponding to each piece of to-be-straightened waybill data, generating a distribution configuration cargo flow chart according to a plurality of attribution allocation information, extracting the to-be-straightened waybill data with the customer type being a consignment company or a payment company and the warp weight being greater than or equal to the preset weight from the distribution configuration cargo flow chart, obtaining the corresponding piece of un-straightened cargo volume information and piece of un-straightened cargo volume information, generating analysis result information according to the to-be-straightened waybill data, the piece of un-straightened cargo volume information and piece of un-straightened cargo volume information, and completing the straightening of the customer waybill data. The method is suitable for optimizing the transportation line and allocating the resources in the logistics transportation network.

Description

Customer bill straightening method, device, computer equipment and storage medium

Technical Field

The present application relates to the field of logistics information processing technologies, and in particular, to a method and apparatus for straightening a customer manifest, a computer device, and a storage medium.

Background

In logistics transportation management, the straightening treatment of a customer bill (namely, unnecessary allocation links are reduced by optimizing a transportation line, so that the transportation efficiency is improved) is a key link for reducing the transportation cost and shortening the transportation timeliness. In the prior art, the following problems are common in customer waybill processing:

1. the traditional method does not accurately filter specific transportation modes (such as whole vehicle transportation and direct flight/air-to-air transportation), so that subsequent analysis contains a large amount of invalid data, and the traditional method lacks unified data format and field mapping standard and has poor data compatibility;

2. The allocation flow direction configuration is blind, namely, the screening rule of the number of times of the allocation flow of the freight bill is not clear, the low-value allocation data cannot be effectively removed, the determination of the attribution allocation information depends on manual experience, and a systematic attribution rule based on an actual circulation log is not formed, so that the scientificity of the allocation flow direction table is insufficient;

3. The straightening strategy lacks quantitative basis, wherein the existing scheme does not combine the customer type (such as a consignment company and a payment company) with the cargo weight threshold value to analyze the cargo amount, and cannot accurately evaluate the difference between the un-straightened cargo amount and the cargo amount to be straightened, so that the priority ordering and effect evaluation of the straightening operation lack data support, and the optimal allocation of transportation resources is difficult to realize.

Accordingly, a need exists for a method that addresses at least one of the problems described above.

Disclosure of Invention

The application provides a method, a device, computer equipment and a storage medium for straightening a customer freight bill, and aims to solve the problems that in the prior art, the customer freight bill processing generally has rough data preprocessing, allocation flow direction configuration blindness and a straightening strategy lacks quantitative basis.

In a first aspect, the present application provides a method for straightening a customer manifest, including:

the method comprises the steps of carrying out general pre-filtering on acquired customer waybill data, removing the waybill of which the service mode is whole-vehicle transportation, removing the waybill of which the transportation mode is direct flight and air-air, and unifying data format and field mapping of the customer waybill data;

Removing the waybill data with the number of times of allocation flow less than or equal to 1 from the waybill data to be straightened, and obtaining the attribution allocation information corresponding to each piece of the waybill data to be straightened;

Extracting to-be-straightened bill data of which the customer type is a consignment company or a payment company and the weight is greater than or equal to the preset weight from the distribution configuration cargo flow list, and acquiring corresponding information of the quantity of the un-straightened cargo and information of the quantity of the un-straightened cargo;

And generating analysis result information according to the to-be-straightened bill data, the un-straightened cargo quantity information and the to-be-straightened cargo quantity information, and straightening the customer bill data according to the analysis result information.

In some embodiments, the method comprises the steps of carrying out general pre-filtering on acquired customer waybill data, removing the waybill with a service mode of whole-car transportation, removing the waybill with a transportation mode of direct-flight and empty-air transportation, unifying data format and field mapping of the customer waybill data, analyzing service mode fields and transportation mode fields in the customer waybill data, matching the values of the service mode fields with a preset whole-car transportation service identifier, matching the values of the transportation mode fields with a preset direct-flight transportation mode identifier and empty-air transportation mode identifier, carrying out logic deletion or marking filtering state on the waybill data with the service mode of matching whole-car transportation identifier or the transportation mode of matching direct-flight or empty-air identifier, carrying out field standardization processing on unfiltered waybill data, including mapping data items with the same meaning but different field names in different customer systems into unified standard fields, converting unstructured data into structured data format, and carrying out default value filling or abnormal data marking on the missing fields.

In some embodiments, the method comprises the steps of eliminating the ticket data with the number of times of allocation flow less than or equal to 1 from the ticket data to be straightened, obtaining the attribution allocation information corresponding to each piece of ticket data to be straightened, extracting an allocation node flow log of the ticket for the ticket data to be straightened which is screened according to the dimensions of a mail company and a mail point part, counting the number of allocation center nodes actually passing through each ticket in a transportation flow as the number of times of allocation flow, reserving the ticket data with the number of times of allocation flow greater than 1, eliminating the ticket data with the number of times of allocation flow less than or equal to 1, and determining the attribution allocation center corresponding to the ticket as attribution allocation information according to the last actually operated allocation center node information in the allocation node flow log for the reserved ticket data by combining with preset allocation center attribution rules, wherein the attribution allocation center rules comprise attribution matching according to the geographic area, operation main body or network level of allocation nodes.

In some embodiments, the obtaining corresponding undirected cargo amount information and the undirected cargo amount information includes screening the cargo flow list of the distribution configuration cargo flow list for the cargo information of a consignment company or a payment company, extracting the cargo weight field in the cargo list, comparing the cargo weight with a preset weight threshold, reserving the cargo data with the cargo weight greater than or equal to the preset weight threshold, counting the sum of the cargo weights of the cargo lists which are not marked as the undirected cargo amount information, counting the sum of the cargo weights of the cargo lists which are in accordance with the client type and the weight condition as the undirected cargo amount information, and presetting the preset weight threshold according to the straightening cost and efficiency parameters of the transportation network.

In some embodiments, the method comprises the steps of grouping to-be-straightened bill data with the same home allocation information, carrying out aggregation statistics on the transportation destination, the transportation route and the transportation aging requirement of each group of bill data, generating a bill flow record taking the home allocation center as a starting point and the transportation destination as an ending point according to an aggregation result corresponding to the aggregation statistics, wherein each flow record contains the bill quantity, the weight distribution and the configuration information of the common transportation route types of the flow, carrying out de-duplication and standardization processing on the flow records to form a distribution configuration bill flow table containing allocation center codes, destination codes, transportation route rules and cargo type adaptation conditions, and the distribution configuration bill flow table is used for guiding the sorting and transportation route planning of the allocation center.

In some embodiments, the generating analysis result information according to the to-be-straightened manifest data, the un-straightened cargo amount information and the to-be-straightened cargo amount information comprises establishing an analysis data model, associating the mail sending company information, the mail sending point information and the attribution allocation information in the to-be-straightened manifest data with the un-straightened cargo amount information and the to-be-straightened cargo amount information, calculating the proportion of the un-straightened cargo amount to the to-be-straightened cargo amount as a straightening rate index, analyzing the straightening rate difference of different mail sending companies, mail sending point parts or attribution allocation centers, combining the transportation cost data and allocation center processing efficiency data, generating the cargo amount straightening priority ordering under each dimension, and forming analysis result information comprising straightening potential evaluation, cost benefit analysis and operation suggestion, wherein the analysis result information is stored in a structured data table or a visual report form.

In some embodiments, the straightening of the customer manifest data is completed according to the analysis result information, and the method comprises the steps of generating a straightening operation instruction aiming at a high-priority consignment company or consignment point part based on the straightening priority ordering in the analysis result information, wherein the operation instruction comprises the steps of adjusting a cargo sorting rule of a distribution center to match an optimal line in a distribution configuration cargo flow table, sending a configuration request of a straightening transportation line to a transportation scheduling system, marking a straightened state of manifest data for executing the straightening operation, recording the execution time of the straightening operation, an operation main body and transportation line change information, periodically backtracking and verifying the straightened manifest data, and comparing actual transportation timeliness, transportation cost and expected indexes in the analysis result information to form a straightening effect evaluation report for optimizing a subsequent straightening strategy.

In a second aspect, the present application provides a customer manifest straightening apparatus, comprising:

The front-end filtering unit is used for carrying out general front-end filtering on the acquired customer bill data, removing the bill transported by the whole vehicle in a service mode, removing the bill transported by the whole vehicle in a transportation mode of direct flight and air-air, unifying the data format and field mapping of the customer bill data, and acquiring the bill data to be straightened according to the dimensions of a mail company and a mail point part of the customer bill data;

The flow direction generating unit is used for eliminating the waybill data with the allocation flow times less than or equal to 1 from the waybill data to be straightened, and obtaining the attributive allocation information corresponding to each piece of the waybill data to be straightened;

The information acquisition unit is used for extracting to-be-straightened bill data which is of a customer type of a consignment company or a payment company and has a weight greater than or equal to a preset weight from the distribution configuration cargo flow list, and acquiring corresponding un-straightened cargo quantity information and to-be-straightened cargo quantity information;

And the straightening completion unit is used for generating analysis result information according to the to-be-straightened freight volume information and the un-straightened freight volume information and the to-be-straightened freight volume information, and completing the straightening of the customer freight volume data according to the analysis result information.

In a third aspect, the present application also provides a computer device comprising:

a memory and a processor;

the memory is used for storing a computer program;

The processor is configured to execute the computer program and implement the steps of the customer manifest straightening method according to the first aspect described above when the computer program is executed.

In a fourth aspect, the present application also provides a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to implement the steps of the customer manifest straightening method according to the first aspect above.

The customer bill straightening method, device, computer equipment and storage medium provided by the embodiment of the application are used for removing invalid bills and unifying data formats through general pre-filtering, improving the accuracy and efficiency of subsequent analysis, avoiding invalid data interference, ensuring that a distribution configuration cargo flow table only contains bill data with actual distribution value based on distribution flow times screening and attribution distribution rule matching, providing scientific basis for sorting and line planning of a distribution center, counting the quantity of undirected cargo and the quantity of undirected cargo by combining a customer type and a weight threshold value, generating analysis result information through a data model, realizing priority sorting and effect evaluation of straightening operation, obviously improving the resource utilization efficiency of a transportation network, reducing transportation cost and shortening cargo turnover time.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application as claimed.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of steps of a method for straightening a customer manifest according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a customer bill straightening device according to an embodiment of the present application;

fig. 3 is a schematic block diagram of a computer device according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

The flow diagrams depicted in the figures are merely illustrative and not necessarily all of the elements and operations/steps are included or performed in the order described. For example, some operations/steps may be further divided, combined, or partially combined, so that the order of actual execution may be changed according to actual situations.

It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, the words "first", "second", etc. are used to distinguish identical items or similar items having substantially the same function and effect. It will be appreciated by those of skill in the art that the words "first," "second," and the like do not limit the amount and order of execution, and that the words "first," "second," and the like do not necessarily differ.

It is to be understood that the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should also be understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.

Some embodiments of the present application are described in detail below with reference to the accompanying drawings. The following embodiments and features of the embodiments may be combined with each other without conflict.

The traditional method does not accurately filter specific transportation modes (such as whole vehicle transportation and direct flight/air-to-air transportation), so that subsequent analysis contains a large amount of invalid data, and the traditional method lacks unified data format and field mapping standard and has poor data compatibility;

The allocation flow direction configuration is blind, namely, the screening rule of the number of times of the allocation flow of the freight bill is not clear, the low-value allocation data cannot be effectively removed, the determination of the attribution allocation information depends on manual experience, and a systematic attribution rule based on an actual circulation log is not formed, so that the scientificity of the allocation flow direction table is insufficient;

the straightening strategy lacks quantitative basis, wherein the existing scheme does not combine the customer type (such as a consignment company and a payment company) with the cargo weight threshold value to analyze the cargo amount, and cannot accurately evaluate the difference between the un-straightened cargo amount and the cargo amount to be straightened, so that the priority ordering and effect evaluation of the straightening operation lack data support, and the optimal allocation of transportation resources is difficult to realize.

Although partial waybill data processing or line optimization methods exist in the prior art, a complete systematic method is not formed by multi-dimensional waybill filtering, allocation flow times screening, attribution allocation information matching, cargo flow table generation and cargo quantity difference analysis, and particularly, precise filtering for a specific transportation mode, attribution analysis based on actual allocation node circulation logs and quantitative evaluation combining a customer type and a weight threshold value are lacking. Therefore, the prior art does not give a technical suggestion for realizing efficient straightening of the customer freight bill through the multi-step collaboration, and an improved method capable of improving the data processing precision, scientifically configuring the allocation flow direction and quantifying the straightening strategy is needed.

In order to solve the above-mentioned problems, referring to fig. 1, fig. 1 is a schematic flow chart of a customer bill straightening method according to an embodiment of the present application. The customer manifest straightening method may be implemented by a computer device, which may be deployed on a single server or a cluster of servers. The method can also be deployed in a handheld terminal, a notebook computer, a wearable device or a robot, etc.

It should be noted that, any information related to the provided method is obtained in accordance with the relevant regulations and is carried out on the premise of user consent, so that privacy of the user is not violated and relevant legal regulations are not violated.

Specifically, as shown in fig. 1, the provided customer manifest straightening method includes steps S101 to S104 as follows:

S101, carrying out general pre-filtering on the acquired customer bill data, removing the bill with a service mode of whole vehicle transportation, removing the bill with a transportation mode of direct flight and empty-to-empty, unifying the data format and field mapping of the customer bill data, and acquiring the bill data to be straightened according to the dimensions of a mail sending company and a mail sending point part of the customer bill data.

Specifically, for the natural 'direct' attribute of a specific transportation mode (whole transportation, direct flight/air-to-air transportation) in logistics transportation, the data of the freight bill needing to be straightened and optimized is precisely screened out through data cleaning and format unification, and the objects to be processed are divided according to service dimensions (mail company, mail point part), so that a high-quality data set is provided for subsequent allocation analysis.

The transportation mode precise filtering comprises the steps of defining a filtering rule, removing the whole transportation bill through a service mode label (such as whole transportation and part transportation) in a transportation bill field, and removing the direct flight and air-to-air transportation bill through a transportation mode label (such as direct flight and air-to-air transit transportation), wherein the transportation has no allocation optimizing space. A preliminary valid data set is generated by batch screening of raw manifest data using regular expressions or database query statements (e.g., WHERE conditions of SQL).

Data format and field mapping unification maps fields of different data sources (e.g., ERP systems, third party platforms) by establishing a standardized data model by defining unified field specifications (e.g., consignment address, consignee address, cargo weight, allocation node, etc.) (e.g., mapping "shipper" to "consignee" and "origin" to "consignee point"). Data heterogeneity is eliminated by field cleaning, type conversion (e.g., converting string weight to numerical value), such as by an ETL (extract-convert-load) tool or programming script (Python/PySpark).

The dimension extraction of the to-be-straightened waybill comprises dimension division, namely taking a 'consignment company' (customer main body) and a 'consignment point part' (delivery point) as grouping dimensions, and ensuring the follow-up analysis to focus on batch waybill optimization of the same customer or point. Grouping and aggregating the cleaned data sets according to the parts of the consignment company and the consignment points, extracting the waybill data with the indirect (to be allocated) transportation mode in each group, and generating a to-be-straightened waybill data set

S102, eliminating the waybill data with the allocation flow times less than or equal to 1 from the to-be-straightened waybill data, acquiring the attribution allocation information corresponding to each piece of to-be-straightened waybill data, and generating a distribution configuration cargo flow table according to the attribution allocation information.

Specifically, low-value allocation data (the number of times of the flow is too small and the optimization space is small) is eliminated through quantitative screening of the allocation flow times, a systematic attribution rule is established based on the actual allocation node circulation log, the attribution allocation nodes of the freight list are automatically determined, and a scientific cargo flow table is finally generated to provide data support for allocation configuration.

The screening of the number of times of allocation flow is realized by defining a screening rule, and the waybill with the number of times of flow less than or equal to 1 (because the waybill is close to direct, the straightening space is limited) is removed by extracting the number of allocation nodes (namely the number of times of allocation centers through which the waybill passes from sending to signing) in the waybill circulation log. And associating the waybill data with the circulation log (through the waybill number main key), calculating the count of allocation nodes (such as the number of lines of allocation center operation record in the log) of each waybill, and filtering the data by using WHERE allocation flow times > 1.

The automatic matching of the attribution allocation information comprises the steps of establishing an attribution rule model, and defining a judgment logic of an attribution allocation center (for example, an originating point part of a waybill belongs to a certain allocation center A, and more than 80% of first allocation nodes of the waybill with the same website are A by default attribution A) through machine learning (such as a decision tree) or a rule engine based on a historical circulation log. And for each to-be-straightened waybill, automatically filling a field of a home allocation center according to data such as a mail point part, a history allocation record, a geofence (an allocation center service range) and the like through database association inquiry or algorithm matching, and replacing manual experience judgment.

Generating a goods flow list of distribution configuration comprises the step of data integration, namely, carrying out structural organization on the screened freight flow list data (containing attribution distribution information) according to a flow chain of a mail point part, an attribution distribution center, a target distribution center and a delivery address, and generating a goods flow list containing distribution node levels and transport paths. And constructing an allocation node relation graph by using a pivot table or a graph database (such as Neo4 j), visualizing the cargo flow direction, and providing visual basis for the follow-up line optimization.

And S103, extracting to-be-straightened freight bill data which is of a customer type of a consignment company or a payment company and is greater than or equal to a preset weight from the distribution configuration freight flow list, and acquiring corresponding information of the quantity of the un-straightened freight and information of the quantity of the straightened freight.

Specifically, by combining the customer attribute (consignment company/payment company) and the cargo weight threshold, the high-value optimization object is precisely positioned, and the priority ranking basis is provided for the straightening strategy by quantifying the difference between the 'un-straightened cargo quantity' (actually allocated cargo quantity) and the 'straightened cargo quantity' (theoretically directly-available cargo quantity).

Customer type and weight threshold screening is accomplished by extracting the manifest for which the customer type is "consignment" (actual shipper) or "payment" (fee carrier), as such customers are typically more sensitive to shipping costs and timeliness. And screening out heavy freight bills needing important optimization according to the load limit (such as the maximum load of a trunk truck and the limit of an air freight single piece) of the transport means and the cost benefit analysis, and presetting a weight threshold (such as more than or equal to 300 kg). Data filtering is performed through WHERE customer type IN ('sender's ',' payment's') AND cargo weight is greater than or equal to a preset threshold.

The information of the undirected freight volume comprises the actual allocation freight volume (such as total weight and number) of the freight list after screening, namely the total freight volume which needs to be allocated for a plurality of times in the current flow. The information of the amount of goods to be straightened includes theoretical amounts of goods (consistent with the amount of goods not to be straightened, different in the number of times of allocation) which are assumed that the manifest can be directly transported by straightening (reducing the allocation links) for comparative analysis of the optimizing potential. And (3) performing aggregation calculation (SUM (weight of cargoes) and COUNT (number of cargoes)) on the selected freight list data to generate indexes such as 'total weight not straightened', 'total weight to be straightened', and the like.

And S104, generating analysis result information according to the to-be-straightened bill data, the un-straightened cargo quantity information and the cargo quantity information to be straightened, and completing straightening of the customer bill data according to the analysis result information.

And generating an analysis result comprising priority ordering and expected effects by integrating to-be-straightened freight bill data and freight volume difference indexes, guiding the adjustment of a transportation line, and realizing simplified allocation links and optimal resource allocation.

The analysis result information generation comprises the steps that core indexes comprise quantity of shipping sheets to be straightened, quantity difference of not straightened goods/quantity of goods to be straightened, estimated time shortening time (calculated according to reduction of distribution times) and cost saving estimation (based on a distribution link cost model). The priority order generates a priority list according to the 'goods volume difference multiplied by the client value coefficient' (the client value can be calculated through weighting of historical order volume, profit contribution degree and the like), and the manifest of the high-goods volume and high-value clients is explicitly optimized first. The structured analysis report is generated by a data reporting tool (e.g., tableau) or a custom algorithm, supporting visual presentation. The bill straightening execution comprises line optimization, namely, designing a direct line (such as skipping an intermediate redundant allocation center) for the bill with high priority according to a cargo flow chart, and adjusting an allocation routing rule (such as configuring a certain allocation center to be forbidden to be accessed in a WMS system). By recording the actual distribution times, transportation timeliness and cost data after straightening, and comparing with an analysis result, closed loop optimization (such as updating a preset weight threshold value and attribution distribution rules regularly) is formed. And synchronizing the optimized routing instruction to a Transportation Management System (TMS) through an API interface, and triggering line scheduling update.

By analyzing the service mode and the transportation mode identification in the waybill field, the transportation type (whole vehicle, direct flight/air-to-air) which is not required to be straightened is accurately filtered, the field standardization processing is carried out on the residual data, the problem of data compatibility is solved, and a unified data base is provided for subsequent analysis.

The field analysis and identification matching extracts field values (such as 'whole vehicle transportation', 'part transportation', 'direct flight', 'air-to-air') by analyzing a service mode field (such as 'transportation type', 'service type') and a transportation mode field (such as 'transportation mode', 'transportation path') in the freight list data.

The establishment of the preset identification dictionary comprises a whole vehicle transportation identification comprising keywords such as whole vehicle, whole vehicle transportation, package vehicle and the like, a direct flight transportation mode identification comprising keywords such as direct flight, direct air transportation, no transfer and the like, and an air-to-air transportation mode identification comprising keywords such as air-to-air, aviation direct airport direct transportation and the like. And comparing the field values by using a character string matching algorithm (such as regular expression and fuzzy matching), and identifying the waybill meeting the filtering condition.

The data filtering process is performed by performing a "logical delete" or "tag filter status" on the manifest matching the filter identity (marked as "no need to straighten" status, rather than physically deleted, facilitating auditing) (adding a "filter reason" tag, such as "whole car transportation" or "direct flight mode").

The field standardization process includes field mapping, namely, establishing a cross-system field comparison table (such as unified mapping of 'shipping enterprises' of A system and 'shipper names' of B system into standard fields 'consignment companies'), and replacing field names in batches through ETL tools or code scripts. Structured conversion ambiguous addresses are resolved using Natural Language Processing (NLP) by splitting unstructured data (e.g., address text) into structured fields (province/city/district/street). Missing value processing is followed by manual verification by filling the manifest with default values (e.g., "unknown") or marked as "data anomalies" for missing key fields (e.g., weight, mailpiece points).

The distribution node circulation log is used for counting the distribution times of the actual flow of the waybill, eliminating the low-value waybill (the flow times are less than or equal to 1), determining the attribution distribution center based on the last operation node and the preset rule, replacing manual experience judgment, and improving the accuracy of distribution information.

The distribution flow times statistics comprises the steps of associating the waybill data with distribution node circulation logs (through a waybill number main key), and extracting all records with the operation type of distribution center processing in the logs. The number of allocation center nodes of each bill (namely the occurrence times of different allocation center codes in the log) is counted and used as the allocation flow times.

The waybill screening comprises the steps of reserving the waybill with the allocation flow times of more than 1 (redundant allocation links are arranged and straightening optimizing space exists), and eliminating the waybill with the frequency less than or equal to 1 (approaching direct arrival and low optimizing value).

The home allocation information determines the "last actually operated allocation center node" (i.e., the allocation center closest to the destination address) in the allocation node flow log by extracting the reserved manifest.

The application of the preset attribution rule to match the attribution allocation center comprises a geographical area rule, an operation main body rule, a logistic network attribution matching the bill according to a company (such as a self-owned allocation center and an allied allocation center) to which the allocation center belongs, and a network level rule, wherein the company is used for preferentially matching a trunk allocation center (instead of an end distribution center) as an attribution node to ensure coverage of a core allocation link.

And screening high-value optimization objects by combining the client type (consignment/payment company) and the cargo weight threshold, and quantifying the difference between the 'un-straightened cargo quantity' (current allocation cargo quantity) and the 'straightened cargo quantity' (theoretical directly-available cargo quantity) by counting the sum of the freight bill weights meeting the conditions so as to provide data support for strategy formulation.

Customer type and weight screening is performed by screening the shipping notes from the distribution profile flow chart for "consignment" or "payment" for the type of customer field (such customers are more sensitive to cost and age). And (3) extracting the freight weight field of the freight bill, and comparing the freight bill with a preset weight threshold (such as 300kg, which is configured based on parameters such as the lower load limit of the trunk transport vehicle and the air freight cost critical point), wherein the freight bill with the weight more than or equal to the threshold is reserved (the heavy freight allocation cost is high, and the straightening benefit is more remarkable).

The calculation of the cargo volume information includes un-straightened cargo volume information, which is to count the sum of the cargo weights of the manifest that is not marked as a "straightened" state (i.e., the total amount of heavy cargo that still needs to be distributed multiple times at present). The cargo volume information should be straightened out by counting the sum of the cargo weights of all the manifest that meet the customer type and weight conditions (whether or not straightened out, representing a theoretically optimizable potential cargo volume). The filtered waybill data is weight summed by an SQL aggregation function (SUM) or data framework (e.g., pandas) to distinguish "straightened" status markers (e.g., filtered by the Boolean field "is_ STRAIGHTENED").

By grouping the same waybills of the home allocation center, information such as a transportation destination, a line, timeliness and the like is aggregated, a standardized cargo flow table is generated, the optimal line configuration from the allocation center to the destination is defined, and rule guidance is provided for sorting and scheduling.

Grouping and aggregation statistics are carried out according to the reserved to-be-straightened waybill data grouping (each group corresponds to an outflow waybill of the allocation center). For each group of bills, information such as the destination (the province/city/district level of the destination address, or the custom area code), the common transportation route (such as "distribution center A→trunk transportation→destination distribution center B"), the aging requirement (the latest delivery time specified by the customer, the transportation aging history statistics), the number of bills, and the weight distribution of the cargoes (such as the statistical duty ratio by weight interval) is aggregated.

The flow record generation is to generate a flow record for each combination of the home allocation center and the destination, wherein the flow record generation comprises basic information including allocation center codes and destination codes, flow information including quantity of the waybills, average weight and weight distribution interval, and line information including common transportation line types (highway trunk lines, railways and air transportation), line cost and ageing parameters.

The method comprises the steps of carrying out de-duplication and standardization by de-duplication on the flow records according to the distribution center code and the destination code, merging statistical data (such as average timeliness and total traffic) of the same flow direction, and finally forming a structured cargo flow table containing the distribution center code, the destination code, the transportation line rule and cargo type adaptation conditions (such as applicable highway trunk with the weight less than or equal to 500 kg) by a standardized field format (such as unified line type named as 'highway and regional distribution'), and storing the structured cargo flow table in a database or a configuration file.

And establishing an analysis model by associating multidimensional data, calculating a straightening rate index, analyzing straightening potentials of different dimensions (a mail company, a point part and an allocation center), and generating priority orders by combining cost and efficiency data to form an analysis report capable of guiding operation.

The data model is established by constructing a multidimensional analysis table, and the data comprises basic information of a shipping list to be straightened, a shipping company, a shipping point part and a home allocation center, goods volume difference indexes, an additional attribute, a customer history order volume, an allocation center processing cost (yuan/kg) and an average transportation timeliness (hour), wherein the goods volume difference indexes comprise a non-straightened goods volume, a goods volume to be straightened (the non-straightened goods volume/the goods volume to be straightened is multiplied by 100 percent).

And (3) analyzing the straightening rate and calculating the priority, grouping according to dimensions such as a mail company, a mail point part, a home allocation center and the like, calculating the straightening rate of each group, and identifying the high-potential group with low straightening rate (namely high non-straightening cargo amount). And introducing a weight factor to calculate priority, wherein the priority is = (the amount of goods to be straightened is multiplied by a customer value coefficient) + (1-straightening rate) multiplied by an allocation cost saving coefficient, and the customer value coefficient is assigned according to the amount of historical orders, the cooperative age and the like, and the allocation cost saving coefficient is calculated according to the number of times of allocation reduced after straightening multiplied by the single allocation cost.

The analysis result output comprises the steps of generating a structured form which comprises fields such as groups, straightening rates, amounts of goods to be straightened, expected cost saving, priority levels and the like, visually reporting, wherein the visual report comprises the steps of displaying the difference of the straightening rates of different groups through a histogram, displaying the straightening potential distribution of a distribution center through a thermodynamic diagram, supporting interactive screening and drill-down analysis, and storing the result in a data warehouse or generating a PDF report for reference by a business department.

And generating a straightening operation instruction based on the analysis result, adjusting an allocation rule and a transportation line, marking an operation record, and forming a strategy optimization closed loop through backtracking verification and evaluation effects.

The straightening operation instruction generation comprises the steps of sorting according to priority, generating operation instructions for high-priority groups (such as mail company with straightening rate <30% and straightening cargo amount >10 tons), wherein allocation rule adjustment is carried out by configuring sorting rules in WMS (warehouse management system), such as ' a ticket of mail point part X, a home allocation center Y and a destination Z ', entering an intermediate allocation center M ' is forbidden, and a line configuration request sends line optimization instructions to TMS (transportation management system) through an API, and a direct line (such as a trunk transportation line of ' allocation center Y-destination Z ') is specified.

The operation record and the state mark record key information, namely operation time, an operation main body (automatic system trigger or manual confirmation), line change details (an original allocation path and a new direct path) and expected indexes in a correlation analysis result (such as expected ageing is shortened by 2 hours, and cost is saved by 5 percent) through the "straightened" state of the waybill mark for executing the straightening operation. Backtracking verification and effect evaluation, namely extracting actual transportation data (distribution times, timeliness and cost) of a straightened freight bill periodically (such as weekly/month), comparing the actual transportation time and predicted shortening time with expected indexes in analysis results, carrying out timeliness evaluation, carrying out cost evaluation, namely actual transportation cost and predicted saving cost (calculated based on reduction amount of distribution times), generating an effect evaluation report, summarizing optimization results, identifying abnormal cases (such as that timeliness does not reach standard after a certain line is straightened, analyzing whether the timeliness is caused by bottleneck of processing efficiency of a distribution center), adjusting preset weight threshold, attribution rule and other parameters according to the actual transportation data, and optimizing subsequent strategies.

In some embodiments, natural Language Processing (NLP) technology is introduced to solve the limitation of fuzzy or nonstandard fields in traditional rule matching, so that semantic understanding is performed on unstructured text fields such as service modes, transportation modes and the like, robustness of filtering rules is improved, and meanwhile missing key information (such as an implicit transportation mode) is automatically complemented.

The domain corpus is constructed by collecting service modes and transportation mode description texts (such as whole car package, aviation direct flight without transfer and air-to-air transfer) in historical waybill data, and marking standard categories (whole car/non-whole car and direct flight/non-direct flight), so that a logistics domain exclusive corpus is constructed. The pretightening model such as BERT is used for fine tuning, a transportation mode classification model is trained, and multi-label classification (such as synchronous identification of 'air-to-air' and 'direct flight' labels) is supported.

Intelligent parsing and complementation includes NLP parsing of free text fields (e.g., "remarks", "special requirements") of the original manifest, extracting implicit transportation mode information (e.g., "direct flight mode" inferred from "do not transit", through to destination "). And deducing the completion (for example, combining the distance between the mail address and the destination address and the weight of goods to judge whether the transportation is finished automobile) of the freight list with the missing key field (such as not explicitly marking the service mode) through the upper part and the lower part Wen Yuyi.

The dynamic rule optimization comprises the steps of carrying out cross verification on a classification result output by the model and a traditional rule matching result, automatically labeling a sample with the classification accuracy lower than a threshold value (such as 90%), reversely optimizing a corpus, and forming a closed loop of data labeling, model training and rule updating.

In some embodiments, the complex association among the allocation center, the mail point part and the geographic area is modeled by using a Graph Neural Network (GNN) by constructing an allocation node relation graph, the limitation of the traditional rule engine is broken through, the dynamic attribution rule is automatically learned, and the network structure change (such as starting of a new allocation center and regional service range adjustment) is adapted.

The allocation network map construction comprises node definition, namely an allocation center (attribute: coordinates, service range and processing capacity), a mail point part (attribute: coordinates and historical allocation preference) and a geographic area (province/city/district). The edge defines the service relation (history attribution record) of the allocation center and the mail point part, the trunk line transportation connection between the allocation centers and the coverage relation between the geographic area and the allocation centers.

The GNN model training comprises the steps of inputting the coordinates of a mail point part and a historical allocation node circulation log (forming a path sequence). And outputting, namely predicting the optimal attribution allocation center (considering the current allocation center load, the transportation distance and the time constraint) of the consignment point part corresponding to the freight bill. Training data, namely taking a true attribution allocation center of a historical freight bill as a label, and learning a dependency relationship among nodes by adopting a graph rolling network (GCN) or a graph annotation force network (GAT).

The real-time attribution matching comprises the steps of inputting the information of the consignment point part for a new waybill, outputting an attribution allocation center in real time through a trained GNN model, replacing the static matching of the traditional rule, and being particularly suitable for dynamic adjustment of a new opening area or an abnormal allocation path.

In some embodiments, by modeling the straightening strategy formulation as a sequence decision problem, the priority ordering is dynamically optimized by using Reinforcement Learning (RL), and real-time optimal allocation of transportation resources is realized by comprehensively considering real-time transportation resource states (such as allocation center congestion degree, trunk vehicle empty rate) and customer dynamic demands (such as time-lapse emergency order proportion).

The state space definition comprises the environmental state, namely the processing load (current backlog amount/processing capacity) of an allocation center, the real-time freight rate of a trunk transportation line, the historical straightening response rate of a client (such as the complaint rate of a certain client after straightening), and the available transport capacity of a current period.

The action space definition includes the optional act of performing a straightening operation (grouped by customer type/weight threshold) on a group of waybills of different priorities, adjusting the straightening strategy parameters (e.g., temporarily increasing the weight threshold of an area to match the vehicle load).

The reward function design comprises positive rewards, negative rewards, line conflict costs (such as default gold generated by temporary adjustment of lines) caused by straightening, and reputation losses caused by customer complaints, wherein the positive rewards are the actual cost saving amount after straightening, the ageing shortening time period and the improvement range of the processing efficiency of the distribution center.

The strategy optimization is realized by training a model by using a deep reinforcement learning algorithm (such as DQN and PPO), inputting a real-time state, outputting an optimal priority ordering strategy, and updating the weight of the reward function according to historical straightening effect data (such as backtracking verification result) periodically to adapt to the change of a business scene (such as aging weight is higher than cost during the period of e-commerce promotion).

In some embodiments, by introducing a space-time sequence model (such as an LSTM+graph network), and combining historical straightening data (allocation times, time efficiency and cost) and geospatial features (allocation center position and transportation line distance), the expected effect of different straightening strategies is predicted, and the hysteresis problem that the traditional method depends on historical average data is solved.

The space-time characteristic engineering comprises time characteristics, namely, the creation time (hour/workday/quarter) of a waybill, historical synchronous straightening and ageing data, and space characteristics, namely, the distance between a mail point part and a home allocation center and the distribution of trunk transportation duration of the allocation center and a destination.

The prediction model construction comprises the steps of inputting the customer type, the cargo weight and the current allocation path (node sequence) of a to-be-straightened freight bill, outputting the predicted allocation times reduction amount, the time-lapse shortening duration and the cost saving amount after straightening, adopting LSTM (least squares) to process time sequence dependence, embedding (Graph Embedding) a coding allocation network structure in combination with a diagram, and constructing an end-to-end space-time prediction model.

The application of the prediction result comprises the steps of replacing the traditional estimation based on the historical mean value with the predicted cost saving in the step S103, providing a more accurate quantification basis for priority ranking, comparing the prediction effect with the actual effect in the backtracking verification in the step S104, and continuously optimizing the prediction precision through model updating (transfer learning) to form an intelligent closed loop of prediction-execution-feedback.

In some embodiments, the clustering algorithm (such as DBSCAN, K-means) is utilized to carry out multidimensional clustering on the transportation destination, the cargo weight and the aging requirement according to the limitation of the traditional grouping depending on manual preset dimension, an implicit allocation flow mode is automatically found, and a dynamic and fine-grained cargo flow table is generated.

The multidimensional feature clustering comprises feature selection including destination longitude and latitude (space distance), cargo weight interval, delivery aging window required by clients and historical transportation line preference (such as whether air transport is preferred or not). The clustering algorithm identifies waybill groups (e.g., "close-range heavy-duty groups" and "long-range age-sensitive groups") with similar transportation characteristics by performing unsupervised clustering on the reserved waybill data to be straightened.

Flow pattern mining works by analyzing the optimal allocation path for each cluster group, e.g., close range heavy groups are adapted to bypass regional allocation centers, directly through the trunk, and far range age sensitive groups are adapted to preferentially match air transit nodes.

Generating the dynamic flow direction rule comprises automatically adjusting cargo distribution configuration (such as rule of adding aging sensitive cargoes to distribute direct flight lines preferentially) according to the clustering result to replace the traditional fixed line configuration.

The dynamic updating of the flow chart is realized by periodically (such as weekly) reclustering based on the latest waybill data, detecting the change of the flow pattern (such as the weight distribution change caused by seasonal cargoes), automatically updating the cargo flow chart of the distribution configuration, and adapting to business fluctuation.

In some embodiments, by utilizing anomaly detection algorithms such as isolated forests, one-Class SVMs and the like, inefficient or invalid straightening operations are identified from retrospective verification data, the problem of low efficiency of traditional manual troubleshooting is solved, and policy vulnerabilities (such as load imbalance of an allocation center caused by excessive straightening) are accurately located.

The abnormal index definition comprises a core index that the number of distribution times after straightening is not reduced (abnormal type: invalid straightening), the ageing is increased beyond a threshold (such as +2 hours), and the cost is increased instead (abnormal type: negative benefit straightening).

The anomaly detection model deployment includes training data including actual effect data (number of assignments, time-varying, cost-varying rate) of historical straightening operations, marking known anomaly samples (e.g., manually confirmed inefficiency cases). Model selection combines supervised (e.g., random forest classification) and unsupervised (e.g., isolated forest) algorithms to detect abnormal patterns that never occur (e.g., a certain dispatch center is subject to explosion due to straightening).

The closed loop optimization triggering automatically marks the detected abnormal operation, associates information such as a distribution path, a customer type and the like, generates an optimization suggestion (such as recovering the original distribution path for the customer or adjusting a weight threshold value), periodically gathers abnormal cases, and reversely optimizes input parameters of a straightening strategy (such as correcting geographical region division errors in a home distribution rule).

In some embodiments, an extensible customer straightening recognition algorithm model is built, and the method can cover three core business scenes (a receiving customer is not straightened, a home allocation is not straightened, a background planning is not straightened), automatically judge whether the un-straightened goods quantity in different distance intervals reaches a business threshold or not, check the consistency of a waybill path and a planning path based on a path matching mechanism of multi-data source fusion, and output periodic statistical results according to the sending time and the sending point dimension. The universal pre-filtering module eliminates the waybill of 'service mode=whole vehicle' transportation, eliminates the waybill of transportation mode=direct flight and air-to-air, unifies data format and field mapping, and ensures that the subsequent modules judge on the premise of the same data. According to the dimension of 'mail company+mail point part', calculating the freight bill of the same receiving customer in a week for more than or equal to 3 days, wherein the distance between the mail point part and the receiving point part is less than 400km and the amount of the un-straightened freight is more than or equal to 6T, the distance between the mail point part and the receiving point part is E [400km, 1000km ] and the amount of the un-straightened freight is more than or equal to 9T, and the distance between the mail point part and the receiving point part is more than or equal to 1000km and the amount of the un-straightened freight is more than or equal to 13T. And the consistency of the receiving clients is judged that ① receiving addresses are identical, ② or any field of the receiving mobile phone/signing mobile phone/receiving company is identical, and the distance between the receiving addresses is less than or equal to 1km (the grid codes are identical). And judging the undirected cargo quantity, namely judging that the heavy accumulated value of the [ field cargo ] unloading record exists.

After the waybill with the allocation flow times less than or equal to 1 is removed, comparing whether the unloading allocation of the waybill is consistent with the port entering allocation or the route attribution allocation of the [ cargo allocation information ], identifying an allocation node to be attributed through the route binding relation of the [ route background ] trunk line type = point part branch line/secondary branch line, selecting the allocation with the most binding 'return point taking part' as attribution if a plurality of return point allocations exist, and setting the undirected cargo quantity threshold (6T, 9T and 13T) according to the distance interval for judgment. The method comprises the steps of extracting a straightening freight bill meeting the condition (customer type = consignment company or payment company, weight is more than or equal to 4T), calculating straightening freight volume, namely matching the flow direction to final origin and the associated final origin, wherein the final origin is organized consistently, weight is more than or equal to 2T, judging that the backstage planning is not straightened if the undirected freight volume/the straightening freight volume is more than or equal to 30% in any 3 days of 7 consecutive days, and defining that the first unloading place is not equal to the final origin and not equal to the allocation organization.

The statistics time comprises the time of sending the mail, the organization dimension comprises a sending point part, the period comprises a period of sliding a time window for one week, and the output fields comprise the types of problems, the number of days without straightening, the average weight of the days, the quantity of goods without straightening, the information of analysts and the like, and support and analysis record are associated to form closed loop management.

Compared with the existing path analysis mode, the method has the remarkable advantages that multiple scenes are covered, the unified model frame can simultaneously identify three main 'unbent' scenes, and maintenance cost caused by scattered business rules is reduced. And (3) fine threshold control, namely setting cargo quantity thresholds according to different distance intervals, wherein the judgment standard is scientific and extensible. And the multi-data source fusion is to automatically integrate multi-system data such as a waybill, a route, a site, a distribution configuration and the like, so as to realize cross-system path consistency verification. And the periodic statistics and attribution are carried out according to the consignment time and the consignment point part dimension uniformly, so that the performance assessment and operation analysis are facilitated. And the expandability and maintainability are that the modularized design can flexibly add/adjust business rules and meet the requirements of different business stages.

In conclusion, the customer straightening algorithm model provided by the invention can greatly improve the path abnormality recognition efficiency and accuracy, provide powerful data support for logistics enterprises to optimize transportation paths, reduce operation cost and improve customer satisfaction, and has higher industrial application value and popularization prospect.

Referring to fig. 2, fig. 2 is a schematic structural diagram of a customer bill straightening device 200 according to an embodiment of the application. The customer order straightening device 200 is used for performing the steps of the customer order straightening method described in the above embodiments. The customer order straightening device 200 may be a single server or a cluster of servers, or the customer order straightening device 200 may be a terminal, which may be a handheld terminal, a notebook, a wearable device, a robot, or the like.

As shown in fig. 2, the customer order straightening device 200 includes:

the prepositive filtering unit 201 is used for carrying out general prepositive filtering on the acquired customer waybill data, removing the waybill transported by the whole car in a service mode, removing the waybill transported by the whole car in a transport mode of direct flight and air-air, unifying the data format and field mapping of the customer waybill data, and acquiring the to-be-straightened waybill data according to the dimensions of a mail sending company and a mail sending point part of the customer waybill data;

The flow direction generating unit 202 is configured to exclude the waybill data with the allocation flow times less than or equal to 1 from the waybill data to be straightened, and obtain the attributive allocation information corresponding to each piece of the waybill data to be straightened;

An information obtaining unit 203, configured to extract to-be-straightened bill data, which is of a customer type of a consignment company or a payment company and is greater than or equal to a preset weight, from the distribution configuration cargo flow table, and obtain corresponding un-straightened cargo amount information and to-be-straightened cargo amount information;

and the straightening completion unit 204 is configured to generate analysis result information according to the to-be-straightened waybill data, the un-straightened freight volume information and the to-be-straightened freight volume information, and complete straightening of the customer waybill data according to the analysis result information.

It should be noted that, for convenience and brevity of description, specific working processes of the customer bill straightening device and each module described above may refer to corresponding processes in the customer bill straightening method embodiments described in the above embodiments, which are not repeated herein.

The customer order straightening method described above may be implemented in the form of a computer program that can be run on the apparatus as shown in fig. 2.

Referring to fig. 3, fig. 3 is a schematic block diagram of a computer device according to an embodiment of the present application. The computer device includes a processor, a memory, and a network interface connected by a device bus, where the memory may include storage media and internal memory.

The storage medium may store an operating device and a computer program. The computer program includes program instructions that, when executed, cause the processor to perform any of a number of customer order straightening methods.

The processor is used to provide computing and control capabilities to support the operation of the entire computer device.

The internal memory provides an environment for the execution of a computer program in a non-volatile storage medium that, when executed by a processor, causes the processor to perform any of a number of customer order straightening methods.

The network interface is used for network communication such as transmitting assigned tasks and the like. It will be appreciated by those skilled in the art that the architecture shown in fig. 3 is merely a block diagram of some of the architecture associated with the present inventive arrangements and is not limiting of the terminal to which the present inventive arrangements are applicable, and that a particular computer device may include more or less components than those shown, or may combine some of the components, or have a different arrangement of components.

It should be appreciated that the Processor may be a central processing unit (Central Processing Unit, CPU), it may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

Wherein in one embodiment the processor is configured to run a computer program stored in the memory to implement the steps of:

An embodiment of the present application further provides a computer readable storage medium, where the computer readable storage medium stores a computer program, where the computer program includes program instructions, and the processor executes the program instructions to implement the steps of the customer manifest straightening method provided in the foregoing embodiments of the present application.

The computer readable storage medium may be an internal storage unit of the computer device according to the foregoing embodiment, for example, a hard disk or a memory of the computer device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SMART MEDIA CARD, SMC), a Secure Digital (SD) card, a flash memory card (FLASH CARD), or the like, which are provided on the computer device.

While the application has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the application. Therefore, the protection scope of the application is subject to the protection scope of the claims.

Claims

1. A method for straightening customer waybills, characterized in that it includes:

The acquired customer waybill data is subjected to general pre-filtering to exclude waybills with full truckload transportation service mode and waybills with direct flight and air-to-air transportation mode, and to unify the data format and field mapping of customer waybill data; the customer waybill data is then processed by sending company and sending point to obtain the waybill data to be straightened out.

In the data of waybills awaiting direct shipment, waybill data with a distribution flow count less than or equal to 1 is excluded. The distribution information corresponding to each waybill is then obtained, including: for waybill data filtered by sending company and sending point, the distribution node flow log of the waybill is extracted, and the number of distribution center nodes actually passed through by each waybill in the transportation process is counted as the distribution flow count; waybill data with a distribution flow count greater than 1 is retained, and waybill data with a distribution flow count less than or equal to 1 is removed; for the retained waybill data, based on the last actual operation of the distribution center node in the distribution node flow log, combined with a preset distribution center attribution rule, the corresponding distribution center is determined as the attribution information for that waybill. The attribution rule for the distribution center includes factors such as the geographical region of the distribution node, The system performs attribution matching at the operational entity or network level; it generates a cargo flow table based on multiple attribution and distribution information, including: grouping the order data to be straightened with the same attribution and distribution information into groups, and aggregating and statistically analyzing the transportation destination, transportation route, and transportation timeliness requirements of each group of order data; generating cargo flow records with the attribution and distribution center as the starting point and the transportation destination as the ending point based on the aggregation results of the aggregation statistics, with each flow record containing the number of order data, cargo weight distribution, and configuration information of commonly used transportation route types for that flow; and performing deduplication and standardization processing on the flow records to form a cargo flow table containing distribution center code, destination code, transportation route rules, and cargo type adaptation conditions, which is used to guide cargo sorting and transportation route planning at the distribution center.

Extract the data of waybills to be straightened from the cargo flow table of the cargo distribution configuration. The customer type is either the sending company or the paying company and the weight is greater than or equal to the preset weight. Obtain the corresponding information on the unstraightened cargo volume and the cargo volume to be straightened.

Based on the data of waybills to be straightened, the information of unstraightened cargo volume, and the information of cargo volume to be straightened, an analysis result is generated, and the straightening of the customer waybill data is completed based on the analysis result.

2. The method according to claim 1, characterized in that, the step of performing general pre-filtering on the acquired customer waybill data to exclude waybills with full truckload transportation as the service method, and to exclude waybills with direct flight and air-to-air transportation modes, and to unify the data format and field mapping of customer waybill data, includes:

The service method field and transportation mode field in the customer waybill data are parsed. The value of the service method field is matched with the preset full truckload transportation service identifier, and the value of the transportation mode field is matched with the preset direct flight transportation mode identifier and air-to-air transportation mode identifier.

Logically delete or mark as filtered waybill data that match the service method with the full truckload transportation identifier or the transportation mode with the direct flight or air-to-air identifier.

Perform field standardization processing on unfiltered waybill data, including mapping data items with the same meaning but different field names from different customer systems to a unified standard field, converting unstructured data into structured data format, and filling missing fields with default values or marking abnormal data.

3. The method according to claim 1, characterized in that, obtaining the corresponding information on unstraightened goods and information on goods to be straightened includes:

Filter waybill data from the cargo flow table where the customer type field is either the sending company or the paying company, and extract the cargo weight field from the waybill.

The cargo weight is compared with a preset weight threshold, and waybill data with cargo weight greater than or equal to the preset weight threshold are retained. For the retained waybill data, the total cargo weight of waybills not marked as straightened is calculated as the unstraightened cargo volume information, and the total cargo weight of all waybills that meet the customer type and weight conditions is calculated as the straightened cargo volume information. The preset weight threshold is pre-configured based on the straightening cost and efficiency parameters of the transportation network.

4. The method according to claim 1, characterized in that, generating analysis result information based on the data of the waybills to be straightened, the information of the quantity of goods not yet straightened, and the information of the quantity of goods to be straightened includes:

Establish an analytical data model to associate the information of the shipping company, shipping point, and distribution center in the data of waybills to be straightened with the information of the volume of goods that have not been straightened and the volume of goods that should be straightened.

The proportion of unstraightened cargo to straightened cargo is calculated as the straightening rate indicator. The differences in straightening rates among different shipping companies, shipping points, or affiliated distribution centers are analyzed. Combined with transportation cost data and distribution center processing efficiency data, a cargo straightening priority ranking is generated under each dimension, forming an analysis result information that includes straightening potential assessment, cost-benefit analysis, and operational suggestions. The analysis result information is stored in the form of structured data tables or visual reports.

5. The method according to claim 4, characterized in that, the step of straightening the customer waybill data based on the analysis result information includes:

Based on the priority ranking in the analysis results, a straightening operation instruction is generated for high-priority parcel delivery companies or parcel delivery points. The operation instruction includes adjusting the cargo sorting rules of the distribution center to match the optimal route in the cargo flow table of the cargo distribution configuration and sending a configuration request for straightening the transportation route to the transportation scheduling system.

The waybill data that has undergone straightening operation is marked as straightened, and the execution time, operation subject, and transportation route change information of the straightening operation are recorded;

Regularly backtest and verify the data of straightened waybills, compare the actual transportation time and cost with the expected indicators in the analysis results, and generate a straightening effect evaluation report to optimize subsequent straightening strategies.

6. A customer waybill straightening device, characterized in that it comprises:

The pre-filtering unit is used to perform general pre-filtering on the acquired customer waybill data, excluding waybills with full truckload transportation as the service method, and excluding waybills with direct flight and air-to-air transportation modes, and unifying the data format and field mapping of customer waybill data; and to obtain the waybill data to be straightened out by the dimensions of the sending company and the sending point of the customer waybill data.

The flow generation unit is used to exclude waybill data with a distribution flow count less than or equal to 1 from the data of waybills to be straightened, and to obtain the corresponding distribution information for each waybill to be straightened, including: for the waybill data to be straightened filtered by the dimensions of sending company and sending point, extracting the distribution node flow log of the waybill, and counting the number of distribution center nodes actually passed by each waybill in the transportation process as the distribution flow count; retaining waybill data with a distribution flow count greater than 1, and removing waybill data with a distribution flow count less than or equal to 1; for the retained waybill data, determining the corresponding distribution center as the distribution information based on the last actual operation of the distribution center node in the distribution node flow log, combined with the preset distribution center attribution rules, the distribution center attribution rules include the distribution node's geographical area, operating entity, or network. The process involves: hierarchical attribution matching; generating a cargo flow configuration table based on multiple attribution and distribution information, including: grouping the order data to be straightened with the same attribution and distribution information into groups; aggregating and statistically analyzing the transportation destination, transportation route, and transportation timeliness requirements of each group of order data; generating cargo flow records with the attribution and distribution center as the starting point and the transportation destination as the ending point based on the aggregation results; each flow record containing the number of orders, cargo weight distribution, and configuration information of commonly used transportation route types for that flow; deduplicating and standardizing the flow records to form a cargo flow configuration table containing distribution center codes, destination codes, transportation route rules, and cargo type adaptation conditions, which is used to guide cargo sorting and transportation route planning at the distribution center; and generating the cargo flow configuration table based on multiple attribution and distribution information.

The information acquisition unit is used to extract the data of the waybills to be straightened that are customer type sender company or payer company and have a weight greater than or equal to the preset weight from the cargo flow table of the cargo distribution configuration, and to obtain the corresponding information on the quantity of goods not straightened and the quantity of goods to be straightened.

The straightening completion unit is used to generate analysis result information based on the straightening data of the waybill to be straightened, the unstraightened cargo quantity information, and the cargo quantity to be straightened information, and to complete the straightening of the customer waybill data based on the analysis result information.

7. A computer device, characterized in that the computer device includes a memory and a processor;

The memory is used to store computer programs;

The processor is configured to execute the computer program and, in executing the computer program, implement the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, causes the processor to implement the method as described in any one of claims 1 to 5.