CN-122022863-A - Goods-code customer portrait generation method based on transportation path optimization

CN122022863ACN 122022863 ACN122022863 ACN 122022863ACN-122022863-A

Abstract

The invention discloses a cargo-agent customer portrait generation method based on transportation path optimization, which relates to the technical field of cargo-agent customer portrait generation and comprises the following steps of extracting a path deviation direction consistency rate, a path reproduction frequency and a path stay inertia span based on a path adoption deviation sequence under the condition that a customer continuously adopts a non-recommended path for a long time, combining to generate a path deviation inertia parameter set for determining the stability degree of the customer path behavior, mapping the path deviation inertia parameter set to a freezing reliability interval, judging whether a label freezing signal is generated, and identifying that the customer path behavior has entered a stable preference state and correspondingly triggering the updating stop operation of a portrait label. The invention solves the problem that the prior goods-substituted customer portrait cannot identify the stability of the route behavior when the customer stably adopts the non-recommended transportation route for a long time, so as to cause the false update of the portrait label, and realizes the accurate freezing and dynamic release control of the customer route preference label.

Inventors

ZHU JINTANG
LUO LI

Assignees

上海必晟物流科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260107

Claims (8)

1. The goods-code customer portrait generating method based on the transportation path optimization is characterized by comprising the following steps: S1, carrying out complete path mapping on a recommended path and an actual path in each transportation task of a goods-agent customer, generating a path deviation label based on the path node coincidence degree, the path deviation direction trend and the road section selection rejection degree, and forming a path adoption deviation sequence according to the task completion time sequence to judge whether the customer continuously adopts a non-recommended path for a long time; S2, under the condition that a client continuously adopts a non-recommended path for a long time, extracting a path deviation direction consistency rate, a path reproduction frequency and a path stay inertia span based on a path adoption deviation sequence, and combining to generate a path deviation inertia parameter set for determining the stability of the path behavior of the client; s3, mapping the path offset inertia parameter set to a freezing reliability interval, judging whether a label freezing signal is generated or not, and identifying that the client path behavior has entered a stable preference state and correspondingly triggering the updating stop operation of the portrait label; s4, after the tag freezing signal is generated, a behavior monitoring mark is established, a path adoption offset sequence generated by a subsequent transportation task is subjected to extended recording, the path behavior trend is continuously tracked, and the offset degree with the historical parameters is calculated; S5, judging whether a freeze release condition is met or not based on the path offset inertia parameter set updated in the behavior monitoring mark, releasing the label frozen state under the condition that the freeze release condition is met, and recovering the portrait label updating flow so as to realize dynamic regulation and control of the portrait label of the client path.
2. The shipping path optimization-based customer representation generation method of claim 1, wherein S1 specifically comprises the steps of: S101, comparing a path node sequence corresponding to a recommended path in each transportation task of a commodity-code customer with a path node sequence corresponding to an actual path node by node, establishing a mapping relation between the recommended path node and the actual path node based on a spatial position corresponding relation of the path nodes, and forming a complete path mapping covering the whole course of the recommended path and the actual path on the basis of the mapping relation; S102, calculating the path node overlap ratio between a recommended path and an actual path based on complete path mapping, extracting the overall path deviation direction trend of the actual path relative to the recommended path, determining the road section selection rejection ratio based on the avoidance condition of the actual path to a specific road section in multiple transportation tasks, and generating a path deviation label corresponding to each transportation task by combining the path node overlap ratio, the path deviation direction trend and the road section selection rejection ratio; S103, arranging the path offset labels according to the time sequence of completing the transportation task to form a path adoption offset sequence, dividing paragraphs based on whether adjacent path offset labels in the path adoption offset sequence are continuously non-recommended path types, and judging that a client continuously adopts a non-recommended path for a long time when the number of the non-recommended path offset labels in the continuous paragraphs exceeds a preset number threshold and the path offset direction trend does not have reverse transition in the continuous paragraphs.
3. The shipping path optimization-based customer representation generation method according to claim 2, wherein S102 is specifically: Extracting a path node sequence corresponding to a recommended path and a path node sequence corresponding to an actual path based on complete path mapping, and calculating the path node coincidence ratio between the recommended path and the actual path by counting the ratio of the number of nodes spatially matched in the two groups of path node sequences to the total number of path nodes; Calculating the overall offset direction change track of the actual path relative to the recommended path based on the spatial arrangement sequence of each path node in the complete path mapping, and extracting and reflecting the overall path offset direction trend of the actual path relative to the recommended path by carrying out consistency analysis on the offset directions of the continuous path nodes; Based on the complete path mapping result of the multiple transportation tasks, counting the repeated detouring or non-adopted road segment set in the actual path, calculating the occurrence frequency of each road segment avoided in the multiple transportation tasks to determine the road segment selection rejection degree, and carrying out combined mapping on the path node coincidence degree, the path deviation direction trend and the road segment selection rejection degree to generate the path deviation label corresponding to each transportation task.
4. The shipping path optimization-based commodity customer representation generating method according to claim 1, wherein S2 specifically comprises the steps of: S201, under the condition that a client continuously adopts a non-recommended path for a long time, based on non-recommended path offset labels continuously appearing in a path adoption offset sequence, counting the path offset directions corresponding to the offset labels, calculating the proportion of the occurrence times of the same offset direction in the continuous offset labels to the total times of the continuous offset labels, extracting the consistency rate of the path offset directions, and based on the occurrence times of the path offset labels corresponding to the same actual path in the path adoption offset sequence, counting the path reproduction times; S202, determining the continuous length of a customer under the condition of keeping the same actual path unchanged based on the continuous distribution condition of path offset labels corresponding to the same actual path in a path adoption offset sequence in the time dimension, extracting path stay inertia spans according to the continuous length, and combining the path offset direction consistency rate, the path reproduction times and the path stay inertia spans after unified dimension processing to generate a path offset inertia parameter set; S203, constructing parameter judgment sections for distinguishing different stability levels based on the value distribution condition of each parameter in the path adoption offset sequence in the path offset inertial parameter set, and determining the stability degree of the client path behavior when the whole path offset inertial parameter set falls into the corresponding stability section.
5. The shipping path optimization-based customer representation generation method according to claim 4, wherein S203 is specifically: Based on the path adoption offset sequence, respectively extracting historical value distribution intervals of the path deviation direction consistency rate, the path recurrence times and the path stay inertia span in a plurality of customer transportation tasks, and determining a single-parameter value threshold section for representing stable behavior characteristics based on the variation range and the concentrated trend of each parameter; after the stability judgment sections are respectively set for the path deviation direction consistency rate, the path reproduction times and the path stay inertia span, comparing the current value of each parameter in the path deviation inertia parameter set with the corresponding stability judgment section, and identifying whether each parameter falls into the corresponding stability judgment section or not, so as to construct the stability hit condition of the current path deviation inertia parameter set; And counting the number of parameters falling into a stability judging section based on the stability hit condition of the path offset inertia parameter set, and determining that the stability degree of the client path behavior reaches a stable state when the current values of all the parameters fall into the corresponding stability judging sections at the same time, wherein the stability degree is used for supporting the state control operation of the follow-up portrait tags.
6. The shipping path optimization-based customer representation generation method of claim 1, wherein S3 is specifically: Respectively carrying out normalization processing on a path deviation direction consistency rate, a path reproduction frequency and a path stay inertia span in a path deviation inertia parameter set to construct a joint evaluation vector, and mapping the path deviation inertia parameter set to a freezing credibility interval based on stability distribution of the joint evaluation vector in historical task data, wherein the reliability grade is reached by deviation inertia characteristics for reflecting the path behavior of a client; setting a freezing critical range in a freezing credibility interval, correspondingly comparing a path offset inertial parameter set mapping result with the freezing critical range, and generating a tag freezing signal when the path offset inertial parameter set falls into the freezing critical range as a whole so as to represent that the client path behavior has entered a stable preference state; After the label freezing signal is generated, the label freezing signal is related to the path adoption class label in the customer image, and the updating and stopping operation of the path adoption class label is executed by taking the label freezing signal as a control triggering basis, so that the path adoption class label is not changed in a follow-up path adoption offset sequence.
7. The shipping path optimization-based customer representation generation method of claim 1, wherein S4 is specifically: After the tag freezing signal is generated, a behavior monitoring mark is established, a path adoption offset sequence and a path offset inertial parameter set at the moment of generating the tag freezing signal are written into the behavior monitoring mark, and a client identifier is bound for continuous comparison of subsequent path behavior trend changes; After each subsequent transportation task is completed, adding the newly generated path deviation label into a path adoption deviation sequence according to the transportation task completion time sequence to form an extended record containing the pre-freezing and post-freezing path behaviors, and writing the extended path adoption deviation sequence into a behavior monitoring annotation; And re-extracting the path deviation direction consistency rate, the path reproduction times and the path stay inertia span based on the path adoption deviation sequence after the record expansion, calculating a current path deviation inertia parameter set, and comparing the current path deviation inertia parameter set with a historical path deviation inertia parameter set recorded in the behavior monitoring annotation to obtain the deviation degree between the current and the history, so as to continuously track the path behavior trend.
8. The shipping path optimization-based customer representation generation method of claim 1, wherein S5 is specifically: Based on the updated path offset inertia parameter set in the behavior monitoring annotation, respectively performing difference calculation on the current path offset direction consistency rate, the path recurrence times and the path stay inertia span and corresponding parameters in the historical path offset inertia parameter set recorded when the tag freezing signal is generated to form three groups of parameter offset amplitude data, comparing each parameter offset amplitude with a corresponding release judgment threshold value, and judging that the freezing release condition is met when the three groups of parameter offset amplitudes exceed the corresponding release judgment threshold values; When judging that the freezing release condition is met, setting a tag freezing signal associated with a client identifier to be in an invalid state, releasing the tag freezing state, and recording the current state switching time and a state change result in a behavior monitoring mark; and after the label freezing state is released, re-incorporating the extended path adoption offset sequence into the portrait label updating flow, and continuously calculating the path adoption class label based on the path adoption offset sequence so as to recover the dynamic regulation and control of the portrait label of the client path.

Description

Goods-code customer portrait generation method based on transportation path optimization Technical Field The invention relates to the technical field of commodity customer portrait generation, in particular to a commodity customer portrait generation method based on transportation path optimization. Background The generation of the freight generation customer portrait based on the optimization of the transportation path refers to the mining of the behavior characteristics of customers in the logistics transportation process through the collection and analysis of a large amount of transportation path data in freight generation business, thereby constructing the customer portrait capable of reflecting the multidimensional characteristics of the customer such as transportation preference, business habit, service requirement and the like. In the prior art, the transportation path data associated with the customer is collected in the manners of track tracking, order recording, scheduling logs and the like, and the transportation path is analyzed and evaluated in an optimized manner by utilizing a path optimization algorithm (such as a shortest path algorithm, time window optimization, cost function modeling and the like), so that the behavior characteristics of the customer in the dimensions of transportation time sensitivity, destination distribution, line selection preference, frequency law and the like are extracted from the transportation path data. The process generally comprises four key links, namely multisource acquisition and normalization processing of transportation path data, path optimization analysis, construction of an optimal transportation model of individual clients under multiple scenes, structured extraction of client behavior characteristics, including path preference, service level, cost trend and the like, and portrait generation and classification modeling, wherein visual expression and differentiated service support of client groups are realized through a label system or vector space. Through the flow, the commodity enterprise can realize accurate customer identification and personalized service optimization based on the real transportation behaviors, so that the transportation efficiency and customer satisfaction are improved. The prior art has the following defects: In the commodity circulation business, when part of high-frequency business clients adopt the actual transportation path inconsistent with the transportation path optimization result for a long time and stably due to objective reasons such as contracted carrier constraint, enterprise compliance route requirement or transportation confidentiality requirement, the situation that the actual transportation behavior of the clients deviates from the recommended path of the system for a long time can be formed. Because the actual transport paths of such customers remain highly consistent over a longer period, they reflect that their transport path selections have explicit and stable business decision logic, rather than sporadic deviations. However, in the conventional system, in order to improve the image updating efficiency in the process of realizing the generation of the commodity generation customer image based on the optimization of the transportation path, the path adoption class label in the customer image is continuously and dynamically updated usually based on the consistency degree of the recommended path and the actual path in a fixed time window. In the above case, since the system does not distinguish whether the client path deviation is caused by an occasional behavior or a long-term stability policy, the portrait update is continuously performed with the recommended path as a reference basis, and thus a wrong change in the label is caused. The conventional goods-code customer portrait generation technology based on transport path optimization cannot judge whether the portrait labels should stop dynamic updating according to the stability of the path behaviors of customers under the condition that the customers continuously adopt non-recommended paths for a long time, so that the labels reflecting the path preferences in the customer portrait drift, the depiction accuracy of the customer portrait on the real transport behaviors is reduced, and the matching effect of transport scheduling decisions and service strategies based on the portrait is further influenced. The above information disclosed in the background section is only for enhancement of understanding of the background of the disclosure and therefore it may include information that does not form the prior art that is already known to a person of ordinary skill in the art. Disclosure of Invention The invention aims to provide a cargo generation customer portrait generation method based on transportation path optimization so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical sc