Accurate logistics distribution method and system

Abstract

The invention discloses an accurate logistics distribution method and system, and relates to the technical field of logistics distribution, and the method comprises the following steps: collecting logistics user historical data, and based on a user historical express receiving mode, combining a workday and a dual-holiday, obtaining a logistics user historical express receiving mode; determining a preference type of the user and a corresponding express site preference receiving time period and a door-to-door distribution preference time period; historical delivery data of different logistics stations are collected, a real-time track prediction model is constructed in combination with the number of to-be-delivered parcels and delivery attributes of the current different stations, and predicted arrival time and predicted stay time of other delivery stations are output based on the current position of a courier; and collecting user logistics progress query times of different to-be-distributed packages on the express arrival day. According to the method, the delivery habit of the courier, the receiving preference of the user and the package emergency degree are combined, and accurate and personalized logistics delivery path planning is realized through big data analysis, so that the delivery efficiency and the user satisfaction are improved.

Description

Technical Field

[0001] This invention relates to the field of logistics and distribution technology, and more specifically, to a precise logistics and distribution method and system. Background Technology

[0002] With the booming development of the e-commerce industry, logistics and distribution, as a key link connecting goods and consumers, directly affect the quality of industry development and user consumption experience through their service quality and operational efficiency. Most existing logistics delivery route planning is based on fixed geographical distances or simple order aggregation rules to generate delivery routes, lacking consideration for the differentiated delivery habits of individual couriers. Couriers in different areas have developed their own unique delivery rhythms, familiar delivery area routes, and station operation habits over the long term. Traditional methods have not fully explored and utilized this historical data, resulting in the inability to achieve more accurate logistics delivery routes that better suit users and couriers. Furthermore, existing technologies do not pay enough attention to user needs, failing to accurately capture different users' preferred times for package collection, nor effectively differentiate the level of urgency users have for different packages, which can easily affect user satisfaction and result in low practicality. Summary of the Invention

[0003] In view of the problems in related technologies, the present invention proposes a precise logistics distribution method and system to overcome the aforementioned technical problems existing in the existing related technologies.

[0004] Therefore, the specific technical solution adopted by the present invention is as follows: A precise logistics delivery method and system, the method comprising the following steps: S1. Collect historical data of logistics users, and based on the user's historical express delivery collection methods, combined with weekdays and weekends, determine the user's preference type and the corresponding express delivery station's preferred collection time period and door-to-door delivery preferred time period; S2. Collect historical delivery data from different logistics stations, combine the number of parcels to be delivered at different stations and delivery attributes, build a real-time trajectory prediction model, and output the estimated arrival time and predicted dwell time of other delivery stations based on the current location of the courier. S3. Collect the number of times users query the logistics progress of different parcels to be delivered on the day the express delivery arrives, determine the urgency of different parcels to be delivered, and based on the urgency of different parcels to be delivered, the user's preferred collection time at the express delivery station and the preferred time for door-to-door delivery, construct a precise logistics delivery route model for the courier based on the real-time trajectory prediction model of the courier, and output the courier's delivery route for the day.

[0005] In a preferred embodiment, S1 includes the following steps: S11. Compile historical logistics data for different users on the current logistics platform, including basic user information and historical package receipt records. Combine the time tags of the historical data to classify the preference types and preference time periods of different users. The specific steps include: Edit user data into a collection Each record Including user ID Tracking number, payment method Collection time and collection date ,in S represents self-pickup at the courier point, and D represents door-to-door delivery. The user's preferred payment method is selected from the payment method m as the user's delivery attribute, and the payment date is used as the delivery date. Extracting the day of the week attribute 1 represents Monday and 7 represents Sunday; Create a collection of date types The corresponding day of the week The data below will be compiled into the corresponding group in the date type set W, and the time will be collected. Convert to minutes ,in , , These represent the time from which the collection begins. The extracted hours and minutes are based on user u, payment method m, and day of the week. Constructing time series ,in This represents the number of records in the current combination. These represent the various time points within the current combination; For each group, the kernel density estimation model is used. Define probability density function ,in: ; in, For adaptive bandwidth parameters, The standard Gaussian kernel function is used, based on the analytical probability density function. The first and second derivative properties are used to detect peak values, where the peak condition is... ,in This is a relative threshold parameter. Representing the first and second derivatives of the probability density function, respectively. Local maximum point Find the maximum value of the probability density function to obtain the set of candidate peak points. k represents the number of candidate peak points; Select the peak point with the highest density. And define the left and right boundaries, where: ; Left boundary To meet The smallest Value, right boundary To meet The largest Value, of which Using the boundary threshold parameter, we obtain the current user's preference time period under the current attribute. Convert preferred time periods to time format .

[0006] As a preferred implementation, in the process of classifying and statistically analyzing the preference types and time periods of different users, a hierarchical fallback mechanism is also set up for users with insufficient samples, specifically including the following steps: S111. Determine the sample situation based on the minimum sample threshold function, where the minimum sample threshold function is... for: ; in, This represents the minimum sample size threshold, where n is the input sample size, and the first level input is the original weekday dimension. ,in ,when If the result is True, calculate directly. Extracting preferred time periods, when If the condition is False, proceed to the second level of date type merging: For weekdays, ; Regarding Saturday, ; Regarding Sunday, ; The number of samples merged based on date type is determined using a minimum sample threshold function. If the threshold is True, the probability density function of the merged data is calculated and the preferred time period is extracted. If the threshold is False, the system enters the third-level global mode rollback to merge all data. The decision is made using a minimum sample threshold function. If the decision is True, the merged probability density function is calculated and the preferred time period is extracted. If the decision is False, it is marked as the default courier service. Represents the merged workday data. This represents the merged global data.

[0007] In a preferred embodiment, S2 includes the following steps: S21. Collect historical delivery data from different logistics stations, and based on the current average delivery speed of couriers, the average time a single courier stays at a station, and the average time a single courier is delivered to the door, combine the number of packages to be delivered by couriers at different stations and the delivery attributes to build a real-time trajectory prediction model. S22. Based on the current location of the courier, output the estimated arrival time and predicted dwell time of the remaining delivery stations through the real-time trajectory prediction model.

[0008] In a preferred embodiment, S21 includes the following steps: S211. Based on historical delivery data from different logistics stations and combined with the courier's historical trajectory, analyze the courier's service time at different stations. Perform the calculation: ; in, Represents the fixed operation time of site i. These represent the historical average door-to-door delivery time and average station dwell time for a single package at the current station i, respectively, in minutes per package. Represents the delivery attribute of the j-th express delivery, where =1 represents door-to-door delivery. =0 indicates self-pickup at the courier point; S212, Based on the historical average delivery speed of couriers Combining the actual road distances between different stations Calculate the travel time between different stations : ; Where i and e represent station i and station e, respectively. The unit is kilometers per hour. Let be the traffic impact factor at time t.

[0009] In a preferred embodiment, S22 includes the following steps: S221. Obtain the current location of the courier at the designated station via GPS signal. Based on the parcel delivery information entered by the courier at other stations, calculate the station service time for each of the remaining courier stations. As the predicted dwell time, where h represents the number of the other stations, the travel time from the central station to the other different stations is calculated based on the actual road distance from the other stations to the central station, taking the current station where the courier is located as the center. Accumulated predicted dwell time at central stations As the estimated arrival time.

[0010] In a preferred embodiment, step S3 includes the following steps: S31. Count the number of times users check the logistics status of each package to be delivered through the App on the expected delivery day, classify the packages as high-urgency packages based on the average number of times, and mark them in the corresponding package attributes. At the same time, record the user's preferred time period on the current date in the package attributes. S32. Based on the urgency and preferred time periods of the current parcels to be delivered, combined with the predicted dwell time and estimated arrival time of different express stations, comprehensively enumerate the logistics delivery routes, and comprehensively filter the logistics delivery routes based on the number of parcels that do not match the user's preferred time periods and the number of parcels with high urgency.

[0011] In a preferred embodiment, S31 includes the following steps: S311, Based on Time Window The number of times users checked the logistics status of each package awaiting delivery through the App within this time window was counted. These represent 00:00 and the current time, respectively. Based on the average number of times the logistics status of all packages awaiting delivery is queried, packages with a value greater than the average are classified as high-urgency packages. S312. Mark high-tension packages in the basic attributes of the corresponding packages to be delivered, and record the user's preferred time period for the current date of the packages to be delivered.

[0012] In a preferred embodiment, S32 includes the following steps: S321, Based on the current station where the courier is located The collection of delivery stations is as follows Where M is the number of remaining stations, the delivery route is defined as starting from the current station. Depart and visit all remaining sites. A one-and-only permutation, where the route is represented as... , =0 indicates the current site. , yes An arrangement; S322, Completely enumerate all delivery routes, for the index set Generate all possible permutations, each permutation corresponding to a delivery route, and for each route... By substituting the information of the packages to be delivered and the historical data of the couriers into the prediction model in S22, the estimated arrival time and station service time of each station on different routes are obtained and converted into a time format. S323. Accumulate the estimated arrival time and service time of each station to obtain the predicted delivery time for the corresponding station. Compare this time with the preferred time period in the basic attributes of the packages to be delivered at each station, and count the number of packages whose predicted delivery time differs from the preferred time period. For the preferred self-pickup time at the courier point, packages that are different from the preferred time are those whose predicted delivery time is later than the preferred time. For the preferred door-to-door delivery time, packages that are different from the preferred time are those whose predicted delivery time is not within the preferred time. S324, targeting The number of packages with the basic attribute of "high stress" was counted. , cumulative and Get Judgment Score For the judgment score Sort the routes in ascending order and select the top-ranked route as the current delivery path for the courier.

[0013] A precision logistics and distribution system includes a data collection module, a predictive model building module, and a delivery route generation module; The data collection module collects historical data of logistics users, and based on the user's historical express delivery collection methods, combined with weekdays and weekends, determines the user's preference type and the corresponding express delivery station preferred collection time period and door-to-door delivery preferred time period; The prediction model building module collects historical delivery data from different logistics stations, combines the number of parcels to be delivered at different stations and delivery attributes, and builds a real-time trajectory prediction model. Based on the courier's current location, it outputs the estimated arrival time and predicted dwell time at other delivery stations. The delivery route generation module collects the number of times users query the logistics progress of different packages to be delivered on the day the express delivery arrives, determines the urgency of different packages to be delivered, and constructs a precise logistics delivery route model for couriers based on the urgency of different packages to be delivered, the user's preferred collection time at the express delivery station, and the preferred time for door-to-door delivery, based on the real-time trajectory prediction model of the courier, and outputs the courier's delivery route for the day.

[0014] The beneficial effects of this invention are as follows: 1. This invention analyzes the delivery habits of couriers in different areas based on historical big data. It then outputs the logistics delivery routes for the day's packages based on the delivery data of different couriers, users' preferred time slots for package collection, and users' urgency levels for different packages. This prioritizes the preferred time slots for most users and provides special protection for high-urgency packages, thereby enhancing the accuracy of logistics delivery. 2. This invention analyzes historical average delivery speed, station operation time and other habitual data of different couriers to build a personalized real-time trajectory prediction model for each courier, so that the output path conforms to the actual situation of different areas, realizes global collaborative optimization of multi-dimensional resources, and improves user satisfaction under comprehensive conditions. 3. This invention plans the most efficient driving route for each courier and aligns delivery times with users' actual needs as much as possible. Through real-time collection and big data analysis of massive user behavior and courier operations, the logistics delivery route is made more in line with the actual delivery situation, thus enhancing the practicality of route planning. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a precise logistics distribution method according to an embodiment of the present invention; Figure 2 This is a block diagram of a precision logistics distribution system according to an embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0018] According to embodiments of the present invention, a precise logistics distribution method and system are provided.

[0019] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments: Example 1: like Figure 1 As shown, according to an embodiment of the present invention, a precise logistics delivery method includes the following steps: S1. Collect historical data of logistics users, and based on the user's historical express delivery collection methods, combined with weekdays and weekends, determine the user's preference type and the corresponding express delivery station's preferred collection time period and door-to-door delivery preferred time period; S11. Compile historical logistics data for different users on the current logistics platform, including basic user information and historical package receipt records. Combine the time tags of the historical data to classify the preference types and preference time periods of different users. The specific steps include: Edit user data into a collection Each record Including user ID Tracking number, payment method Collection time and collection date ,in S represents self-pickup at the courier point, and D represents door-to-door delivery. The user's preferred payment method is selected from the payment method m as the user's delivery attribute, and the payment date is used as the delivery date. Extracting the day of the week attribute 1 represents Monday and 7 represents Sunday; It should be noted that the basic information includes the user ID and tracking number. Historical package collection records include the collection method and collection time. Based on the timestamp of the tracking number, the date type of the package collection is obtained, where the date type is Monday to Sunday, the collection method includes self-pickup at the courier point and door-to-door delivery, and the collection time is the time when the package was signed for by the user, derived from the timestamp. Extracting the day of the week attribute Where 1 represents Monday, 7 represents Sunday, and so on for 2, 3, 4, 5, 6, etc. The user's payment method m needs to be determined based on the user's historical payment history or logistics information notes. That is, the user's preferred payment method is the one in which the number of self-pickup at the express delivery point and door-to-door delivery accounts for the majority of the current user's historical express delivery records. This will be used as the current user's delivery attribute.

[0020] Create a collection of date types The corresponding day of the week The data below will be compiled into the corresponding group in the date type set W, and the time will be collected. Convert to minutes ,in , , These represent the time from which the collection begins. The extracted hours and minutes are based on user u, payment method m, and day of the week. Constructing time series ,in This represents the number of records in the current combination. These represent the various time points within the current combination; For each group, the kernel density estimation model is used. Define probability density function ,in: ; in, For adaptive bandwidth parameters, The standard Gaussian kernel function is used, based on the analytical probability density function. The first and second derivative properties are used to detect peak values, where the peak condition is... ,in This is a relative threshold parameter. Representing the first and second derivatives of the probability density function, respectively. Local maximum point Find the maximum value of the probability density function to obtain the set of candidate peak points. k represents the number of candidate peak points; It should be noted that peak values ​​are filtered using first and second derivatives and significance criteria. The relative threshold parameter is typically set to 0.3, but this value can be adjusted according to the specific circumstances. Standard Gaussian kernel function The adaptive bandwidth parameter is: ; Where 0.9 is the empirical coefficient, and 1.34 is a constant factor based on the normal distribution. The interquartile range represents the sample. This is the skewness adjustment factor. ,in This is the sample skewness coefficient. -1 / 5 is the index of the sample size. This represents the sample standard deviation.

[0021] Select the peak point with the highest density. And define the left and right boundaries, where: ; Left boundary To meet The smallest Value, right boundary To meet The largest Value, of which Using the boundary threshold parameter, we obtain the current user's preference time period under the current attribute. Convert preferred time periods to time format .

[0022] It should be noted that, among them The boundary threshold parameter is usually set between 0.6 and 0.8. The specific setting needs to be determined by consulting experts in the relevant field based on the actual situation.

[0023] In the process of classifying and statistically analyzing the preferences of different users across different time periods, a tiered fallback mechanism is also implemented for users with insufficient samples. This mechanism includes the following steps: S111. Determine the sample situation based on the minimum sample threshold function, where the minimum sample threshold function is... for: ; in, This represents the minimum sample size threshold, where n is the input sample size, and the first level input is the original weekday dimension. ,in ,when If the result is True, calculate directly. Extracting preferred time periods, when If the condition is False, proceed to the second level of date type merging: For weekdays, ; Regarding Saturday, ; Regarding Sunday, ; The number of samples merged based on date type is determined using a minimum sample threshold function. If the threshold is True, the probability density function of the merged data is calculated and the preferred time period is extracted. If the threshold is False, the system enters the third-level global mode rollback to merge all data. The decision is made using a minimum sample threshold function. If the decision is True, the merged probability density function is calculated and the preferred time period is extracted. If the decision is False, it is marked as the default courier service. Represents the merged workday data. This represents the merged global data.

[0024] It should be noted that using the minimum sample threshold function can prevent data distortion caused by an insufficient number of user samples. This represents the minimum sample size threshold, which is usually set to 5. However, the minimum sample size threshold can be adjusted according to the actual situation. For logistics items marked as default delivery, the default delivery will be based on the courier's own situation during the actual delivery.

[0025] Example 2: S2. Collect historical delivery data from different logistics stations, combine the number of parcels to be delivered at different stations and delivery attributes, build a real-time trajectory prediction model, and output the estimated arrival time and predicted dwell time of other delivery stations based on the current location of the courier. S21. Collect historical delivery data from different logistics stations, and based on the current average delivery speed of couriers, the average time a single courier stays at a station, and the average time a single courier is delivered to the door, combine the number of packages to be delivered by couriers at different stations and the delivery attributes to build a real-time trajectory prediction model. S211. Based on historical delivery data from different logistics stations and combined with the courier's historical trajectory, analyze the courier's service time at different stations. Perform the calculation: ; in, Represents the fixed operation time of site i. These represent the historical average door-to-door delivery time and average station dwell time for a single package at the current station i, respectively, in minutes per package. Represents the delivery attribute of the j-th express delivery, where =1 represents door-to-door delivery. =0 indicates self-pickup at the courier point; It should be noted that the average door-to-door delivery time and average station dwell time of a single courier at different courier stations can be obtained by averaging the historical data of the current courier at the current station. The station fixed operation time is the average fixed time spent by the current courier at the current station, such as parking and handover. By analyzing historical big data, we can analyze the courier transportation habits of different couriers at different stations, which will facilitate subsequent precise logistics and distribution planning.

[0026] S212, Based on the historical average delivery speed of couriers Combining the actual road distances between different stations Calculate the travel time between different stations : ; Where i and e represent station i and station e, respectively. The unit is kilometers per hour. Let be the traffic impact factor at time t.

[0027] It should be noted that the actual road distance can be obtained from the urban road conditions recorded in the GIS system. This is a traffic impact factor at time t, used to measure road congestion during peak hours in different cities, to allow for travel time leeway between different stations. It needs to be set according to the traffic conditions of different roads at different time stages in the actual application environment. =1 represents normal traffic, greater than 1 represents congestion, and less than 1 represents smooth traffic.

[0028] S22. Based on the current location of the courier, output the estimated arrival time and predicted dwell time of the remaining delivery stations through the real-time trajectory prediction model; S221. Obtain the current location of the courier at the designated station via GPS signal. Based on the parcel delivery information entered by the courier at other stations, calculate the station service time for each of the remaining courier stations. As the predicted dwell time, where h represents the number of the other stations, the travel time from the central station to the other different stations is calculated based on the actual road distance from the other stations to the central station, taking the current station where the courier is located as the center. Accumulated predicted dwell time at central stations As the estimated arrival time.

[0029] It should be noted that by combining the current location of the courier with the status of packages to be delivered at each different station, and using the predictive model, the predicted dwell time and estimated arrival time at other stations can be output. This allows for subsequent adjustments to delivery times based on different users' preferred delivery times.

[0030] S3. Collect the number of times users query the logistics progress of different parcels to be delivered on the day the express delivery arrives, determine the urgency of different parcels to be delivered, and based on the urgency of different parcels to be delivered, the user's preferred collection time at the express delivery station and the preferred time for door-to-door delivery, construct a precise logistics delivery route model for the courier based on the real-time trajectory prediction model of the courier, and output the courier's delivery route for the day. S31. Count the number of times users check the logistics status of each package to be delivered through the App on the expected delivery day, classify the packages as high-urgency packages based on the average number of times, and mark them in the corresponding package attributes. At the same time, record the user's preferred time period on the current date in the package attributes. S311, Based on Time Window The number of times users checked the logistics status of each package awaiting delivery through the App within this time window was counted. These represent 00:00 and the current time, respectively. Based on the average number of times the logistics status of all packages awaiting delivery is queried, packages with a value greater than the average are classified as high-urgency packages. It should be noted that as the logistics and delivery progresses, the number of user queries for packages awaiting delivery needs to be updated in real time in order to re-identify and update high-urgency packages.

[0031] S312. Mark high-tension packages in the basic attributes of the corresponding packages to be delivered, and record the user's preferred time period for the current date of the packages to be delivered.

[0032] It should be noted that the preferred time period needs to be determined based on the user's delivery attributes of the current package to be delivered, so as to determine whether the preferred time period of the package is the self-pickup time period at the courier point or the door-to-door delivery time period.

[0033] S32. Based on the urgency and preferred time period of the current parcels to be delivered, combined with the predicted dwell time and estimated arrival time of different express stations, comprehensively enumerate the logistics delivery routes, and comprehensively filter the logistics delivery routes based on the number of parcels that do not match the user's preferred time period and the number of parcels with high urgency. S321, Based on the current station where the courier is located The collection of delivery stations is as follows Where M is the number of remaining stations, the delivery route is defined as starting from the current station. Depart and visit all remaining sites. A one-and-only permutation, where the route is represented as... , =0 indicates the current site. , yes An arrangement; S322, Completely enumerate all delivery routes, for the index set Generate all possible permutations, each permutation corresponding to a delivery route, and for each route... By substituting the information of the packages to be delivered and the historical data of the couriers into the prediction model in S22, the estimated arrival time and station service time of each station on different routes are obtained and converted into a time format. It should be noted that during the calculation of the estimated arrival time, the location of the central station needs to be continuously changed based on the changes in the location of the courier's station during the prediction process, so as to output the predicted arrival time of the remaining stations.

[0034] S323. Accumulate the estimated arrival time and service time of each station to obtain the predicted delivery time for the corresponding station. Compare this time with the preferred time period in the basic attributes of the packages to be delivered at each station, and count the number of packages whose predicted delivery time differs from the preferred time period. For the preferred self-pickup time at the courier point, packages that are different from the preferred time are those whose predicted delivery time is later than the preferred time. For the preferred door-to-door delivery time, packages that are different from the preferred time are those whose predicted delivery time is not within the preferred time. S324, targeting The number of packages with the basic attribute of "high stress" was counted. , cumulative and Get Judgment Score For the judgment score Sort the routes in ascending order and select the top-ranked route as the current delivery path for the courier.

[0035] It should be noted that at each courier station, steps S321 to S324 need to be repeated to determine and generate the latest logistics delivery route, taking into account the high-urgency packages updated in real time, so as to ensure that the route always meets the latest user needs.

[0036] Example 3: like Figure 2 As shown, a precision logistics and distribution system includes a data collection module, a predictive model building module, and a delivery route generation module; The data collection module collects historical data of logistics users. Based on the user's historical express delivery collection methods, combined with weekdays and weekends, it determines the user's preference type and the corresponding preferred collection time period and preferred door-to-door delivery time period of the express delivery station. The prediction model building module collects historical delivery data from different logistics stations, combines the number of parcels to be delivered at different stations and delivery attributes, and builds a real-time trajectory prediction model. Based on the courier's current location, it outputs the estimated arrival time and predicted dwell time at other delivery stations. The delivery route generation module collects the number of times users query the logistics progress of different packages awaiting delivery on the day the package arrives, determines the urgency of different packages, and, based on the urgency of different packages, the user's preferred pickup time at the courier station, and the preferred time for door-to-door delivery, constructs a precise logistics delivery route model for couriers based on a real-time trajectory prediction model for couriers, outputting the courier's delivery route for the day. In summary, this invention, by statistically analyzing historical big data on the delivery habits of couriers in different areas, and based on delivery data from different couriers, users' preferred collection times, and users' urgency levels for different packages, comprehensively outputs the logistics delivery routes for the day. It prioritizes fulfilling the preferred time slots for the majority of users and provides special support for high-urgency packages, enhancing the accuracy of logistics delivery. By analyzing historical average delivery speeds, station operation times, and other habitual data of different couriers, it constructs personalized real-time trajectory prediction models for each courier, ensuring that the output routes conform to the actual conditions of different areas. This achieves global collaborative optimization of multi-dimensional resources, improving user satisfaction under overall conditions. By planning the most efficient driving routes for each courier and aligning delivery times with users' actual needs as much as possible, and through real-time collection and big data analysis of massive user behavior and courier operations, the logistics delivery routes are made more consistent with actual delivery conditions.

[0037] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A precise logistics distribution method, characterized in that, The method includes the following steps: S1. Collect historical data of logistics users, and based on the user's historical express delivery collection methods, combined with weekdays and weekends, determine the user's preference type and the corresponding express delivery station's preferred collection time period and door-to-door delivery preferred time period; S2. Collect historical delivery data from different logistics stations, combine the number of parcels to be delivered at different stations and delivery attributes, build a real-time trajectory prediction model, and output the estimated arrival time and predicted dwell time of other delivery stations based on the current location of the courier. S3. Collect the number of times users query the logistics progress of different parcels to be delivered on the day the express delivery arrives, determine the urgency of different parcels to be delivered, and based on the urgency of different parcels to be delivered, the user's preferred collection time at the express delivery station and the preferred time for door-to-door delivery, construct a precise logistics delivery route model for the courier based on the real-time trajectory prediction model of the courier, and output the courier's delivery route for the day.

2. The precise logistics distribution method according to claim 1, characterized in that, S1 includes the following steps: S11. Compile historical logistics data for different users on the current logistics platform, including basic user information and historical package receipt records. Combine the time tags of the historical data to classify the preference types and preference time periods of different users. The specific steps include: Edit user data into a collection Each record Including user ID Tracking number, payment method Collection time and collection date ,in S represents self-pickup at the courier point, and D represents door-to-door delivery. The user's preferred payment method is selected from the payment method m as the user's delivery attribute, and the payment date is used as the delivery date. Extracting the day of the week attribute 1 represents Monday and 7 represents Sunday; Create a collection of date types The corresponding day of the week The data below will be compiled into the corresponding group in the date type set W, and the time will be collected. Convert to minutes ,in , , These represent the time from which the collection begins. The extracted hours and minutes are based on user u, payment method m, and day of the week. Constructing time series ,in This represents the number of records in the current combination. These represent the various time points within the current combination; For each group, the kernel density estimation model is used. Define probability density function ,in: ； in, For adaptive bandwidth parameters, The standard Gaussian kernel function is used, based on the analytical probability density function. The first and second derivative properties are used to detect peak values, where the peak condition is... ,in This is a relative threshold parameter. Representing the first and second derivatives of the probability density function, respectively. Local maximum point Find the maximum value of the probability density function to obtain the set of candidate peak points. k represents the number of candidate peak points; Select the peak point with the highest density. And define the left and right boundaries, where: ； Left boundary To meet The smallest Value, right boundary To meet The largest Value, of which Using the boundary threshold parameter, we obtain the current user's preference time period under the current attribute. Convert preferred time periods to time format .

3. The precise logistics distribution method according to claim 2, characterized in that, In the process of classifying and statistically analyzing the preference types and time periods of different users, a tiered fallback mechanism is also set up for users with insufficient samples, which specifically includes the following steps: S111. Determine the sample situation based on the minimum sample threshold function, where the minimum sample threshold function is... for: ； in, This represents the minimum sample size threshold, where n is the input sample size, and the first level input is the original weekday dimension. ,in ,when If the result is True, calculate directly. Extracting preferred time periods, when If the condition is False, proceed to the second level of date type merging: For weekdays, ; Regarding Saturday, ; Regarding Sunday, ; The number of samples merged based on date type is determined using a minimum sample threshold function. If the threshold is True, the probability density function of the merged data is calculated and the preferred time period is extracted. If the threshold is False, the system enters the third-level global mode rollback to merge all data. The decision is made using a minimum sample threshold function. If the decision is True, the merged probability density function is calculated and the preferred time period is extracted. If the decision is False, it is marked as the default courier service. Represents the merged workday data. This represents the merged global data.

4. The precise logistics distribution method according to claim 2, characterized in that, S2 includes the following steps: S21. Collect historical delivery data from different logistics stations, and based on the current average delivery speed of couriers, the average time a single courier stays at a station, and the average time a single courier is delivered to the door, combine the number of packages to be delivered by couriers at different stations and the delivery attributes to build a real-time trajectory prediction model. S22. Based on the current location of the courier, output the estimated arrival time and predicted dwell time of the remaining delivery stations through the real-time trajectory prediction model.

5. The precise logistics distribution method according to claim 4, characterized in that, S21 includes the following steps: S211. Based on historical delivery data from different logistics stations and combined with the courier's historical trajectory, analyze the courier's service time at different stations. Perform the calculation: ； in, Represents the fixed operation time of site i. These represent the historical average door-to-door delivery time and average station dwell time for a single package at the current station i, respectively, in minutes per package. Represents the delivery attribute of the j-th express delivery, where =1 represents door-to-door delivery. =0 indicates self-pickup at the courier point; S212, Based on the historical average delivery speed of couriers Combining the actual road distances between different stations Calculate the travel time between different stations : ； Where i and e represent station i and station e, respectively. The unit is kilometers per hour. Let be the traffic impact factor at time t.

6. The precise logistics distribution method according to claim 5, characterized in that, S22 includes the following steps: S221. Obtain the current location of the courier at the designated station via GPS signal. Based on the parcel delivery information entered by the courier at other stations, calculate the station service time for each of the remaining courier stations. As the predicted dwell time, where h represents the number of the other stations, the travel time from the central station to the other different stations is calculated based on the actual road distance from the other stations to the central station, taking the current station where the courier is located as the center. Accumulated predicted dwell time at central stations As the estimated arrival time.

7. The precise logistics distribution method according to claim 6, characterized in that, S3 includes the following steps: S31. Count the number of times users check the logistics status of each package to be delivered through the App on the expected delivery day, classify the packages as high-urgency packages based on the average number of times, and mark them in the corresponding package attributes. At the same time, record the user's preferred time period on the current date in the package attributes. S32. Based on the urgency and preferred time periods of the current parcels to be delivered, combined with the predicted dwell time and estimated arrival time of different express stations, comprehensively enumerate the logistics delivery routes, and comprehensively filter the logistics delivery routes based on the number of parcels that do not match the user's preferred time periods and the number of parcels with high urgency.

8. The precise logistics distribution method according to claim 7, characterized in that, S31 includes the following steps: S311, Based on Time Window The number of times users checked the logistics status of each package awaiting delivery through the App within this time window was counted. These represent 00:00 and the current time, respectively. Based on the average number of times the logistics status of all packages awaiting delivery is queried, packages with a value greater than the average are classified as high-urgency packages. S312. Mark high-tension packages in the basic attributes of the corresponding packages to be delivered, and record the user's preferred time period for the current date of the packages to be delivered.

9. A precise logistics distribution method according to claim 8, characterized in that, S32 includes the following steps: S321, Based on the current station where the courier is located The collection of delivery stations is as follows Where M is the number of remaining stations, the delivery route is defined as starting from the current station. Depart and visit all remaining sites. A one-and-only permutation, where the route is represented as... , =0 indicates the current site. , yes An arrangement; S322, Completely enumerate all delivery routes, for the index set Generate all possible permutations, each permutation corresponding to a delivery route, and for each route... By substituting the information of the packages to be delivered and the historical data of the couriers into the prediction model in S22, the estimated arrival time and station service time of each station on different routes are obtained and converted into a time format. S323. Accumulate the estimated arrival time and service time of each station to obtain the predicted delivery time for the corresponding station. Compare this time with the preferred time period in the basic attributes of the packages to be delivered at each station, and count the number of packages whose predicted delivery time differs from the preferred time period. For the preferred self-pickup time at the courier point, packages that are different from the preferred time are those whose predicted delivery time is later than the preferred time. For the preferred door-to-door delivery time, packages that are different from the preferred time are those whose predicted delivery time is not within the preferred time. S324, targeting The number of packages with the basic attribute of "high stress" was counted. , cumulative and Get Judgment Score For the judgment score Sort the routes in ascending order and select the top-ranked route as the current delivery path for the courier.

10. A precision logistics distribution system, characterized in that, The system employs a precise logistics delivery method as described in any one of claims 1-9, comprising a data collection module, a prediction model construction module, and a delivery route generation module; The data collection module collects historical data of logistics users, and based on the user's historical express delivery collection methods, combined with weekdays and weekends, determines the user's preference type and the corresponding express delivery station preferred collection time period and door-to-door delivery preferred time period; The prediction model building module collects historical delivery data from different logistics stations, combines the number of parcels to be delivered at different stations and delivery attributes, and builds a real-time trajectory prediction model. Based on the courier's current location, it outputs the estimated arrival time and predicted dwell time at other delivery stations. The delivery route generation module collects the number of times users query the logistics progress of different packages to be delivered on the day the express delivery arrives, determines the urgency of different packages to be delivered, and constructs a precise logistics delivery route model for couriers based on the urgency of different packages to be delivered, the user's preferred collection time at the express delivery station, and the preferred time for door-to-door delivery, based on the real-time trajectory prediction model of the courier, and outputs the courier's delivery route for the day.

Images