Method, device and medium for predicting express material demand

Abstract

The application provides a method and device for predicting express material demand, computer equipment and a storage medium. The method comprises the following steps: obtaining courier feature information, historical piece quantity information and a to-be-predicted day feature information of a to-be-predicted courier, determining a corresponding type of piece quantity prediction model according to the historical piece quantity information, inputting the courier feature information and the to-be-predicted day feature information into the piece quantity prediction model, obtaining a predicted piece quantity corresponding to the to-be-predicted courier according to an output result of the piece quantity prediction model, and determining a material predicted consumption quantity of the to-be-predicted courier according to a material consumption rate of the to-be-predicted courier. The above scheme avoids high consumption rate caused by material distribution based on a network point, realizes prediction of the material consumption quantity based on the courier dimension, and improves the accuracy of the prediction of the express material demand.

Description

Technical Field

[0001] This application relates to the field of logistics technology, and in particular to a method, apparatus, computer equipment, and storage medium for predicting the demand for express delivery materials. Background Technology

[0002] With the continuous development of e-commerce and the rapid growth of express delivery services, the usage of express delivery materials has also gradually increased. For production materials such as document envelopes, packaging bags, and transparent tape, the usage is large and the inventory cost is high. Express delivery companies need to predict the volume of express deliveries in order to allocate express delivery materials.

[0003] Currently, express delivery companies typically predict and distribute parcels based on their network of delivery points. Delivery personnel collect the necessary materials from these points, and they often collect a large amount of materials, resulting in a high material consumption rate. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, computer equipment, and storage medium for predicting the demand for express delivery materials, addressing the technical problem of high consumption rate of express delivery materials in current technologies.

[0005] A method for predicting the demand for express delivery materials, the method comprising:

[0006] Obtain the courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted;

[0007] Based on the historical quantity information, determine the corresponding type of quantity prediction model;

[0008] The courier feature information and the day to be predicted feature information are input into the parcel volume prediction model, and the predicted parcel volume corresponding to the courier to be predicted is obtained according to the output of the parcel volume prediction model.

[0009] Obtain the predicted material consumption rate for the courier to be predicted; the predicted material consumption rate is obtained based on the material requisition records of the courier to be predicted.

[0010] Based on the predicted quantity of items and the predicted material consumption rate, the predicted material requisition amount for the delivery personnel to be predicted is determined.

[0011] In one embodiment, determining the corresponding type of quantity prediction model based on the historical quantity information includes:

[0012] If the time range corresponding to the historical parcel volume information of the courier to be predicted is greater than or equal to a preset time threshold, the parcel volume prediction model is determined to be a machine learning model; the machine learning model is trained based on courier feature information, historical parcel volume information and date feature information corresponding to the historical parcel volume information; the machine learning model includes at least two types of decision tree models fused through an ensemble algorithm;

[0013] If the time range corresponding to the historical parcel volume information of the courier to be predicted is less than a preset time threshold, the parcel volume prediction model is determined to be a time series model; the time series model is predicted based on the courier characteristic information, historical parcel volume information and the date characteristic information corresponding to the historical parcel volume information.

[0014] In one embodiment, obtaining the predicted material consumption rate of the courier to be predicted includes:

[0015] Obtain the material requisition records of the delivery personnel to be predicted, and determine the material consumption rate model corresponding to the delivery personnel to be predicted based on the requisition cycle corresponding to the material requisition records.

[0016] The material requisition records and the historical quantity information are input into the material consumption rate model, and the predicted material consumption rate is obtained based on the output of the material consumption rate model.

[0017] In one embodiment, determining the predicted material requisition amount for the courier to be predicted based on the predicted quantity of items and the predicted material consumption rate includes:

[0018] The predicted material requisition amount for the courier to be predicted is obtained by multiplying the predicted quantity of items and the predicted material consumption rate.

[0019] In one embodiment, after determining the predicted material requisition amount for the courier to be predicted, the method further includes:

[0020] Obtain the minimum packaging quantity corresponding to the material; the minimum packaging quantity represents the quantity of the material that can be configured in the smallest unit of packaging;

[0021] Based on the predicted material usage of the couriers to be predicted and the minimum packaging quantity, determine multiple delivery batches for the couriers to be predicted and the delivery quantity corresponding to each delivery batch.

[0022] In one embodiment, it further includes:

[0023] Obtain the material delivery weights corresponding to the multiple delivery batches;

[0024] If the total delivery weight of the materials in the multiple delivery batches is greater than the first preset threshold, and there is a delivery batch whose corresponding material delivery weight is greater than the second preset threshold, the material delivery weight corresponding to the delivery batch will be allocated to other delivery batches.

[0025] In one embodiment, it further includes:

[0026] Determine the employee number of the courier to be predicted, and sort it among the employee numbers of all couriers corresponding to the courier outlet to which the courier to be predicted belongs.

[0027] Based on the sorting and the multiple delivery batches of the courier to be predicted, the material collection order of the courier to be predicted at the express delivery outlet is determined, and the material collection order is sent to the courier to be predicted.

[0028] A device for predicting the demand for express delivery materials, the device comprising:

[0029] The information acquisition module is used to acquire the courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted;

[0030] The quantity model module is used to determine the corresponding type of quantity prediction model based on the historical quantity information.

[0031] The predicted shipment volume acquisition module is used to input the courier feature information and the feature information of the day to be predicted into the shipment volume prediction model, and obtain the predicted shipment volume corresponding to the courier to be predicted based on the output of the shipment volume prediction model.

[0032] The predicted consumption rate acquisition module is used to acquire the predicted material consumption rate of the courier to be predicted; the predicted material consumption rate is obtained based on the material requisition records of the courier to be predicted.

[0033] The material forecast quantity acquisition module is used to determine the material forecast requisition quantity of the courier to be forecasted based on the forecast quantity of the items and the material forecast consumption rate.

[0034] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method for predicting express material demand in any of the above embodiments.

[0035] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for predicting the demand for express delivery materials in any of the above embodiments.

[0036] The aforementioned method, apparatus, computer equipment, and storage medium for predicting express delivery material demand acquire courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted. Based on the historical parcel volume information, a corresponding parcel volume prediction model is determined. The courier characteristic information and daily characteristic information are input into the parcel volume prediction model. Based on the output of the model, the predicted parcel volume for the courier is obtained. Finally, based on the courier's material consumption rate, the predicted material requisition quantity for that courier is determined. This solution avoids the high consumption rate caused by material allocation based on network points, achieves material requisition prediction based on the courier dimension, and improves the accuracy of express delivery material demand prediction. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a method for predicting the demand for express delivery materials in one embodiment.

[0038] Figure 2 This is a flowchart illustrating a method for predicting the demand for express delivery materials in another embodiment.

[0039] Figure 3 This is a flowchart illustrating a method for predicting the demand for express delivery materials in another embodiment.

[0040] Figure 4 This is a structural block diagram of a device for predicting the demand for express delivery materials in one embodiment;

[0041] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0043] It should be noted that the terms "first" and "second" used in the embodiments of the present invention are merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first" and "second" can be interchanged in a specific order or sequence where permissible. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein.

[0044] In one embodiment, such as Figure 1As shown, a method for predicting the demand for express delivery materials is provided. This embodiment illustrates the application of this method to a server. It is understood that this method can also be applied to a terminal, or to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0045] Step S101: Obtain the courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted.

[0046] In this disclosure, the delivery personnel to be predicted can be a specific delivery personnel or a category of delivery personnel with common characteristics. The server can perform material demand forecasting for a specific delivery personnel or for a category of delivery personnel.

[0047] In this disclosure, the courier characteristic information refers to information such as the age, length of service, number of days on the job, attendance rate, and branch information of the courier to be predicted. The server can pre-classify couriers based on the courier characteristic information, classifying a courier into a certain category.

[0048] Each courier can be associated with a courier outlet. The volume of packages at each outlet varies, and the server can distinguish the outlet where a courier is located through category codes and other means.

[0049] In this disclosure, historical parcel volume information refers to the number of parcels delivered by the courier to be predicted during the corresponding time period in the past, as well as the statistical value of the parcel volume. For example, this could include the courier's sales volume over the previous n days, the average sales volume over the previous n days, the minimum sales volume over the previous n days, the maximum sales volume over the previous n days, the median sales volume over the previous n days, the standard deviation of sales volume over the previous n days, and the skewness value of sales volume over the previous n days. Alternatively, it could be statistical information on the number of parcels delivered to the courier category to which the courier to be predicted belongs, such as the average number of parcels delivered by all couriers in that category during a certain time period, and the statistical value of the parcel volume.

[0050] In this disclosure, the feature information of the day to be predicted refers to the time dimension features corresponding to the day to be predicted, including seasonal characteristics, holiday characteristics, and promotional day characteristics. Each feature information of the day to be predicted can be identified by different class codes. For example, January, February, and March can be designated as spring, April, May, and June as summer, July, August, and September as autumn, and October, November, and December as winter, with corresponding seasonal category codes configured for each season; holidays such as New Year's Day and Spring Festival, as well as the n days before them, can be marked and their corresponding holiday category codes configured; promotional days, as well as the n days before and after them, can be marked and their corresponding promotional day category codes configured.

[0051] In practice, the server can obtain the courier characteristic information, historical shipment volume information, and daily characteristic information of the courier to be predicted from the storage module.

[0052] In some embodiments, the server can obtain the courier characteristic information corresponding to the courier, as well as the courier's historical shipment volume information and the courier's daily characteristic information, from the storage module based on the unique identifier information such as the courier's employee number.

[0053] In some embodiments, the server can obtain the courier characteristic information corresponding to the courier of the category to which the courier to be predicted belongs, as well as the statistical value of the historical parcel volume information of the courier of that category, and obtain the characteristic information of the day to be predicted.

[0054] Step S102: Determine the corresponding type of part quantity prediction model based on historical part quantity information.

[0055] In this disclosure, the parcel volume prediction model can predict future parcel volumes for couriers based on historical information. Different parcel volume prediction models can be used for different historical parcel volumes to improve the accuracy of the predictions. The parcel volume prediction model can be constructed based on courier characteristic information, historical parcel volume information, and corresponding date characteristic information of multiple couriers.

[0056] The server can pre-establish a correspondence between historical parcel volume information and parcel volume prediction models. This correspondence is determined based on the time span of the historical parcel volume information corresponding to each courier. For example, if the historical parcel volume information of a courier spans a long period, the factors affecting the parcel volume prediction value of that courier become more complex, and a more complex neural network model can be used for prediction.

[0057] In practice, the server can determine the shipment prediction model for the shipment prediction of the shipment to be predicted based on the historical shipment volume information of the shipment to be predicted, from the pre-established correspondence between historical shipment volume information and shipment volume prediction model.

[0058] Step S103: Input the courier characteristic information and the day characteristic information to be predicted into the parcel volume prediction model, and obtain the predicted parcel volume corresponding to the courier to be predicted based on the output of the parcel volume prediction model.

[0059] In this disclosure, the predicted parcel volume can be the predicted parcel volume of the courier on the predicted date. The server can obtain the predicted parcel volume of the courier within a certain time period, such as one month or one quarter, based on the predicted parcel volumes of multiple predicted dates.

[0060] In some embodiments, the shipment volume prediction model can also be configured to predict the shipment volume for a certain time period. In this case, the server can obtain the courier characteristic information and the characteristic information of the day to be predicted corresponding to that time period for prediction.

[0061] For example, the parcel volume prediction model can be configured with the monthly historical parcel volume information and corresponding date feature information of the courier. The server can obtain the courier feature information, monthly historical parcel volume information and the feature information of the date to be predicted for the month to be predicted for the courier.

[0062] Step S104: Obtain the predicted material consumption rate for the delivery personnel to be predicted.

[0063] In this disclosure, the predicted material consumption rate can be obtained from the material requisition records of the couriers to be predicted. The material requisition records may include the quantity and time of material requisition for the couriers to be predicted in historical periods, and the predicted material consumption rate can be expressed as the ratio of the quantity of materials requisitioned to the number of express parcels.

[0064] The predicted material consumption rate for each delivery person is linked to their respective branch office and cannot exceed that branch office's consumption rate. The branch office's consumption rate can be adjusted based on historical consumption rates.

[0065] Among them, the working period of the couriers will affect the judgment of the future predicted consumption rate. The server can provide different calculation models for the predicted consumption rate of materials for couriers with different working periods.

[0066] In some cases, couriers may have temporary express delivery needs during historical periods, leading to an abnormal increase in the amount of materials used. The server can adjust the predicted material consumption rate for the courier to meet the express delivery needs.

[0067] In practice, the server can determine the corresponding predicted material consumption rate based on the material requisition records of the couriers to be predicted.

[0068] Step S105: Determine the predicted material requisition quantity for the delivery personnel to be predicted based on the predicted quantity of items and the predicted material consumption rate.

[0069] In this disclosure, the predicted material usage refers to the predicted amount of materials required for a courier to complete the predicted number of packages over a certain period. The server can issue materials to the courier based on this predicted material usage.

[0070] The aforementioned method for predicting express delivery material demand involves acquiring the courier's characteristic information, historical parcel volume information, and daily characteristic information for the courier to be predicted. Based on the historical parcel volume information, a corresponding parcel volume prediction model is determined. The courier's characteristic information and the daily characteristic information are then input into this model. Based on the model's output, the predicted parcel volume for the courier is obtained. Finally, based on the courier's material consumption rate, the predicted material requisition quantity for that courier is determined. This approach avoids the high consumption rate resulting from point-based material allocation and achieves courier-based material requisition prediction, thus improving the accuracy of express delivery material demand forecasting.

[0071] In one embodiment, step S102, which involves determining the corresponding type of quantity prediction model based on historical quantity information, includes:

[0072] If the time range corresponding to the historical parcel volume information of the courier to be predicted is greater than or equal to a preset time threshold, the parcel volume prediction model is determined to be a machine learning model. If the time range corresponding to the historical parcel volume information of the courier to be predicted is less than the preset time threshold, the parcel volume prediction model is determined to be a time series model.

[0073] In this disclosure, there is a correspondence between historical parcel volume information and parcel volume prediction models. This correspondence can be related to the time range of the historical parcel volume information of the couriers to be predicted. The longer the time span of the courier's historical parcel volume information, the more complex the influencing factors; the shorter the time span, the stronger the linear correlation.

[0074] The preset time threshold can be 3 months or other time span values.

[0075] For example, if the historical parcel volume information of the courier to be predicted includes more than or equal to 3 months of historical parcel receipt data, a machine learning model that integrates at least two decision tree models using an ensemble algorithm can be used to predict the parcel volume. If the historical parcel receipt data is less than 3 months, a time series model can be used to predict the parcel volume.

[0076] In this disclosure, the training data used for training the machine learning model and the data used for building the time series model can have the same data source, including the characteristic information of each courier, historical parcel volume information, and historical parcel volume information.

[0077] The machine learning model can be at least two decision tree models fused together using an ensemble algorithm. These at least two decision tree models can be XGBoost or LightGBM. XGBoost is an optimized distributed gradient boosting library designed for high efficiency, flexibility, and portability, enabling rapid and accurate solutions to many data science problems. LightGBM (LightGradient Boosting Machine) is an open-source framework from Microsoft that implements the GBDT algorithm and supports highly efficient parallel training.

[0078] The time series model can be established based on time series data obtained from system observations, through curve fitting and parameter estimation.

[0079] The solution in the above embodiment obtains the historical parcel volume information of the couriers to be predicted, and determines the corresponding parcel volume prediction model based on the time range corresponding to the historical parcel volume information. This fully considers the influencing factors of parcel volume and improves the accuracy of parcel volume prediction.

[0080] In one embodiment, step S103, which involves determining the predicted material consumption rate for the delivery person to be predicted, includes:

[0081] Obtain the material requisition records of the couriers to be predicted. Based on the requisition cycle corresponding to the material requisition records, determine the material consumption rate model corresponding to the couriers to be predicted. Input the material requisition records and historical quantity information into the material consumption rate model. Obtain the predicted material consumption rate based on the output of the material consumption rate model.

[0082] In this disclosure, the material consumption rate model refers to a consumption rate model derived from the material requisition records and historical replenishment information of the couriers to be predicted. The material consumption rate model is related to the requisition period corresponding to the material requisition records of the couriers, and the server can suggest the correspondence between the requisition period and the material consumption rate model.

[0083] The server can divide the requisition period into intervals and configure a corresponding material consumption rate model for each interval. The couriers falling within that interval can use this material consumption rate model to determine the predicted material consumption rate.

[0084] In some embodiments, the requisition period can be divided into three intervals: interval 1: requisition records for more than 3 months; interval 2: requisition records for more than 1 month but no more than 3 months; and interval 3: no requisition records.

[0085] For couriers falling into interval 1, the corresponding predicted material consumption rate model is: Predicted material consumption rate = Sum of material requisitions in the previous three months / Sum of item quantities in the previous three months.

[0086] For couriers falling into interval 2, the corresponding predicted material consumption rate model is: Predicted material consumption rate = Sum of material consumption in the previous month / Sum of item quantity in the previous month.

[0087] For couriers falling into interval 3, the corresponding predicted consumption rate model can be clustered based on historical requisition records and item quantity data to obtain the material consumption rate of each type of courier, which is then used as the predicted material consumption rate for that courier.

[0088] In some embodiments, the server can obtain the material requisition records, requisition cycles, and historical quantity information of the couriers as training datasets, and use the actual material consumption rate corresponding to each courier as a validation set to train a neural network model corresponding to each requisition cycle, which serves as the material consumption rate model corresponding to each requisition cycle.

[0089] In some embodiments, if a courier has a temporary delivery need, the server can increase its predicted material consumption rate by a certain percentage. If no temporary need occurs, the predicted material consumption rate can be appropriately decreased.

[0090] In some embodiments, the server can set a material consumption rate for each branch and adjust it according to the actual material consumption of each branch. The predicted material consumption rate for each courier can be configured not to exceed this material consumption rate to improve the efficiency and accuracy of material management.

[0091] The above embodiment obtains material requisition records, determines the material consumption rate model, and inputs the material requisition records and historical quantity information into the material consumption rate model to obtain the predicted material consumption rate of the deliveryman to be predicted. This fully measures the historical records of individual deliverymen and improves the accuracy of obtaining the predicted material consumption rate.

[0092] In one embodiment, step S105, which involves determining the predicted material requisition quantity for the courier based on the predicted quantity and the predicted material consumption rate, includes:

[0093] The predicted material requisition quantity for the courier is obtained by multiplying the predicted quantity of items by the predicted material consumption rate.

[0094] In this embodiment, the server can obtain the predicted material requisition quantity of the delivery person to be predicted based on the product of the predicted quantity of items and the predicted material consumption rate.

[0095] Specifically, if the predicted quantity corresponds to the quantity of items on a certain day, then the predicted material requisition quantity is the predicted requisition quantity for that day. If the predicted quantity corresponds to the quantity of items on a certain month, then the predicted material requisition quantity is the predicted requisition quantity for that month, thus obtaining the predicted material requisition quantity for the courier to be predicted based on the predicted quantity of items and the predicted material consumption rate.

[0096] In one embodiment, the steps following step S105, which determine the predicted material requisition quantity for the courier to be predicted, include:

[0097] Obtain the minimum packaging quantity corresponding to the material; based on the predicted material usage and minimum packaging quantity of the couriers to be predicted, determine the multiple delivery batches of the couriers to be predicted and the delivery quantity corresponding to each delivery batch.

[0098] In this disclosure, the minimum packaging quantity represents the quantity of the material that can be configured in the smallest unit of packaging, referring to the quantity of the material that can be contained in each minimum packaging unit when the material is delivered from the warehouse to each courier. For example, if the material is document envelopes, the minimum packaging unit for packaging document envelopes when the warehouse delivers the material to the courier can contain 50 document envelopes.

[0099] In this disclosure, a delivery batch refers to the predicted material usage of a courier, delivered from the warehouse to the courier in multiple batches. The delivery quantity refers to the quantity of material corresponding to each batch.

[0100] For example, if the predicted material usage of a delivery person in a certain month is 'a' and the minimum packaging quantity is 'b', the delivery batch can be determined according to the splitting rules in the table below.

[0101] Table 1. Requirements Breakdown Rules:

[0102]

[0103] The above embodiment improves the material delivery optimization based on couriers and increases the efficiency of material delivery by obtaining the minimum packaging quantity corresponding to the material.

[0104] In one embodiment, the above method further includes:

[0105] Obtain the material delivery weight corresponding to each of the multiple delivery batches; if the total delivery weight of the materials in the multiple delivery batches is greater than the first preset threshold, and there is a delivery batch whose material delivery weight is greater than the second preset threshold, allocate the material delivery weight corresponding to the delivery batch to other delivery batches.

[0106] In this disclosure, the material delivery weight can be obtained based on the delivery quantity and material weight parameters corresponding to each delivery batch. For a particular courier, if the delivery weight of a certain delivery batch is too large, or if the delivery weight corresponding to the overall material forecast demand is too large, it will affect the overall delivery efficiency of the warehouse. The server can be configured to dynamically adjust the delivery weight for couriers or delivery batches with excessive delivery weight.

[0107] The first preset threshold can be the maximum total material delivery weight of multiple delivery batches of the courier to be predicted, and the second preset threshold can be the maximum material delivery weight of each delivery batch of the courier to be predicted.

[0108] For example, the first preset threshold can be 80kg, and the second preset threshold can be 20kg. If the total material delivery weight of multiple delivery batches for the courier to be predicted is less than 80kg, and the material delivery weight of a certain delivery batch is greater than 20kg, the server can add 25% of the material weight of that delivery batch (greater than 20kg) to other delivery batches.

[0109] The above-described embodiment optimizes each delivery batch for the courier by adjusting the material delivery weight, thereby improving the efficiency of material delivery to the courier.

[0110] In one embodiment, the above method further includes:

[0111] Determine the employee ID of the courier to be predicted and sort it among the employee IDs of all couriers at the courier outlet to which the courier belongs; based on the sorting and the multiple delivery batches of the courier to be predicted, determine the material retrieval order of the courier at the outlet and send the material retrieval order to the courier to be predicted.

[0112] In this disclosure, "network point" refers to the assigned area of ​​the courier to be predicted, which can be distinguished by network point category code. Different network points have different parcel volumes. Each network point corresponds to multiple couriers. When the server is distributing materials, it can optimize the distribution process to ensure that each network point has material orders sent to the warehouse.

[0113] In one embodiment, the total number of material delivery requests for a branch within a certain period can be the sum of delivery batches from all couriers at the branch. The server can determine the number of material deliveries required by the branch per day based on the total number of material delivery requests and plan warehouse shipments accordingly.

[0114] In some embodiments, couriers can be arranged according to their employee ID numbers. The server can determine the material collection order of a courier at a branch based on the employee ID number and the courier's delivery batch, and then deliver the materials in that order.

[0115] The material collection order can include the material collection plan of the courier within a certain period. It can be configured to ensure that the materials of each courier can be delivered. There are no specific restrictions here.

[0116] For example, courier #2 has a predicted material requisition quantity of 200, a minimum package quantity of 50, and four delivery batches, with each batch requiring 50 items. The server can assign the courier to collect materials on the 2nd, 10th, 15th, and 25th of the month and notify the courier accordingly.

[0117] The above-described embodiment uses employee ID and delivery batch number to determine the order in which the delivery personnel will pick up materials at the branch, thereby improving the efficiency of material delivery and retrieval.

[0118] In one embodiment, such as Figure 2 As shown, a method for predicting the demand for express delivery materials is provided. The method includes:

[0119] Step S201: Obtain the courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted.

[0120] Step S202: If the time range corresponding to the historical parcel volume information of the courier to be predicted is greater than or equal to a preset time threshold, the parcel volume prediction model is determined to be a machine learning model; the machine learning model is trained based on the courier feature information, historical parcel volume information, and the date feature information corresponding to the historical parcel volume information; the machine learning model includes at least two types of decision tree models fused through an ensemble algorithm; if the time range corresponding to the historical parcel volume information of the courier to be predicted is less than the preset time threshold, the parcel volume prediction model is determined to be a time series model; the time series model is predicted based on the courier feature information, historical parcel volume information, and the date feature information corresponding to the historical parcel volume information.

[0121] Step S203: Input the courier characteristic information and the day characteristic information to be predicted into the parcel volume prediction model, and obtain the predicted parcel volume corresponding to the courier to be predicted based on the output of the parcel volume prediction model.

[0122] Step S204: Obtain the material requisition records of the couriers to be predicted; determine the material consumption rate model corresponding to the couriers to be predicted based on the requisition cycle corresponding to the material requisition records; input the material requisition records and historical quantity information into the material consumption rate model; and obtain the predicted material consumption rate based on the output of the material consumption rate model.

[0123] Step S205: Based on the product of the predicted quantity of items and the predicted material consumption rate, the predicted material requisition quantity of the delivery personnel to be predicted is obtained.

[0124] Step S206: Obtain the minimum packaging quantity corresponding to the material; the minimum packaging quantity represents the quantity of materials that can be configured in the smallest unit of packaging; based on the predicted material usage of the courier to be predicted and the minimum packaging quantity, determine the multiple delivery batches of the courier to be predicted and the delivery quantity corresponding to each delivery batch.

[0125] Step S207: Obtain the material delivery weights corresponding to multiple delivery batches; if the total delivery weight of materials in multiple delivery batches is greater than the first preset threshold, and there is a delivery batch whose material delivery weight is greater than the second preset threshold, allocate the material delivery weights corresponding to the delivery batches to other delivery batches.

[0126] The above embodiments acquire courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted. Based on the historical parcel volume information, a corresponding parcel volume prediction model is determined. The courier characteristic information and daily characteristic information are input into the parcel volume prediction model. Based on the output results, the predicted parcel volume for the courier is obtained. Furthermore, based on the courier's material consumption rate, the predicted material requisition quantity for the courier is determined. Material distribution planning is then performed for the courier based on this predicted material requisition quantity. This solution avoids the high consumption rate caused by material allocation based on network points, achieves prediction of material requisition quantity based on the courier dimension, and material distribution planning based on the courier dimension, thus improving the accuracy of express delivery material demand prediction.

[0127] It should be understood that, although Figure 1-2 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1-2 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0128] In one embodiment, the method for predicting the demand for express delivery materials according to the above embodiments of this application may be as follows: Figure 3 As shown.

[0129] The server can preprocess data such as historical parcel volume data, parcel characteristic information, and date characteristic information corresponding to historical parcel volume data. After extracting feature engineering for each data point, the data is divided into training and testing sets. Parameter optimization and algorithm modeling are then performed to generate a parcel volume prediction model. The server can also extract data from the source data using ETL technology. ETL (Extract-Transform-Load) describes the process of extracting, transforming, and loading data from the source to the destination. While ETL is commonly used in data warehousing, its application is not limited to data warehousing.

[0130] The server can input the courier's characteristics and historical parcel volume information into the parcel volume prediction model to obtain the corresponding predicted parcel volume.

[0131] The server can retrieve the material requisition records corresponding to the delivery personnel to be predicted from the source data. These material requisition records can be direct requisition records from the delivery personnel at the express delivery outlet. The server can then obtain the predicted consumption rate for the delivery personnel to be predicted based on these direct requisition records. Furthermore, the server uses the predicted quantity of items and the predicted consumption rate for the delivery personnel to be predicted to obtain the predicted material requisition quantity for that delivery personnel.

[0132] In one embodiment, such as Figure 4 As shown, a forecasting device for express delivery material demand is provided. The device 400 includes:

[0133] The information acquisition module 401 is used to acquire the courier characteristic information, historical parcel volume information and the characteristic information of the day to be predicted of the courier to be predicted;

[0134] The part quantity model module 402 is used to determine the corresponding type of part quantity prediction model based on historical part quantity information.

[0135] The predicted shipment volume acquisition module 403 is used to input the courier feature information and the feature information of the day to be predicted into the shipment volume prediction model, and obtain the predicted shipment volume corresponding to the courier to be predicted based on the output of the shipment volume prediction model.

[0136] The predicted consumption rate acquisition module 404 is used to acquire the predicted material consumption rate of the courier to be predicted; the predicted material consumption rate is obtained based on the material requisition records of the courier to be predicted.

[0137] The material forecast quantity acquisition module 405 is used to determine the material forecast requisition quantity of the courier to be forecasted based on the forecast quantity of parts and the material forecast consumption rate.

[0138] In one embodiment, the shipment volume model module 402 includes: a shipment volume model unit, configured to determine that the shipment volume prediction model is a machine learning model if the time range corresponding to the historical shipment volume information of the courier to be predicted is greater than or equal to a preset time threshold; the machine learning model is trained based on courier feature information, historical shipment volume information, and date feature information corresponding to the historical shipment volume information; the machine learning model includes at least two types of decision tree models fused by an ensemble algorithm; if the time range corresponding to the historical shipment volume information of the courier to be predicted is less than the preset time threshold, the shipment volume prediction model is determined to be a time series model; the time series model is predicted based on courier feature information, historical shipment volume information, and date feature information corresponding to the historical shipment volume information.

[0139] In one embodiment, the predicted consumption rate acquisition module 404 includes: a predicted consumption rate acquisition unit, used to acquire the material requisition records of the delivery personnel to be predicted, determine the material consumption rate model corresponding to the delivery personnel to be predicted based on the requisition cycle corresponding to the material requisition records, input the material requisition records and historical quantity information into the material consumption rate model, and obtain the predicted material consumption rate based on the output of the material consumption rate model.

[0140] In one embodiment, the material forecast quantity acquisition module 405 includes: a material forecast quantity acquisition unit, used to obtain the forecast material requisition quantity of the delivery person to be predicted based on the product of the forecast quantity and the forecast material consumption rate.

[0141] In one embodiment, the apparatus 400 further includes: a delivery batch unit, configured to obtain the minimum packaging quantity corresponding to the material; the minimum packaging quantity represents the quantity of material that can be configured in the smallest unit of packaging; and to determine multiple delivery batches of the courier to be predicted and the delivery quantity corresponding to each delivery batch based on the predicted material usage of the courier to be predicted and the minimum packaging quantity.

[0142] In one embodiment, the delivery batch unit is further configured to obtain the material delivery weight corresponding to multiple delivery batches; if the total delivery weight of the materials in multiple delivery batches is greater than a first preset threshold, and there is a delivery batch whose material delivery weight is greater than a second preset threshold, the material delivery weight corresponding to the delivery batch is allocated to other delivery batches.

[0143] In one embodiment, the delivery batch unit is further configured to determine the employee number of the courier to be predicted, and sort it among the employee numbers of each courier at the courier outlet to which the courier to be predicted belongs; based on the sorting and multiple delivery batches of the courier to be predicted, determine the material collection order of the courier to be predicted at the courier outlet, and send the material collection order to the courier to be predicted.

[0144] Specific limitations regarding the forecasting device for express delivery material demand can be found in the limitations of the forecasting method for express delivery material demand described above, and will not be repeated here. Each module in the aforementioned forecasting device for express delivery material demand can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0145] The method for predicting the demand for express delivery materials provided in this application can be applied to computer equipment, which can be a server, and its internal structure diagram can be as follows: Figure 5 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data such as shipment volume prediction models, historical shipment volume information, and courier data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a method for predicting express delivery material demand.

[0146] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0147] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0148] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0149] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0150] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0151] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for predicting the demand for express delivery materials, characterized in that, The method includes: Obtain the courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted; Based on the historical quantity information, determine the corresponding type of quantity prediction model; The courier feature information and the day to be predicted feature information are input into the parcel volume prediction model, and the predicted parcel volume corresponding to the courier to be predicted is obtained according to the output of the parcel volume prediction model. Obtain the predicted material consumption rate for the courier to be predicted; the predicted material consumption rate is obtained based on the material requisition records of the courier to be predicted. Based on the predicted quantity of items and the predicted material consumption rate, the predicted material requisition amount for the delivery personnel to be predicted is determined.

2. The method according to claim 1, characterized in that, The step of determining the corresponding type of quantity prediction model based on the historical quantity information includes: If the time range corresponding to the historical parcel volume information of the courier to be predicted is greater than or equal to a preset time threshold, the parcel volume prediction model is determined to be a machine learning model; the machine learning model is trained based on courier feature information, historical parcel volume information and date feature information corresponding to the historical parcel volume information; the machine learning model includes at least two types of decision tree models fused through an ensemble algorithm; If the time range corresponding to the historical parcel volume information of the courier to be predicted is less than a preset time threshold, the parcel volume prediction model is determined to be a time series model; the time series model is predicted based on the courier characteristic information, historical parcel volume information and the date characteristic information corresponding to the historical parcel volume information.

3. The method according to claim 1, characterized in that, The step of obtaining the predicted material consumption rate for the courier to be predicted includes: Obtain the material requisition records of the delivery personnel to be predicted, and determine the material consumption rate model corresponding to the delivery personnel to be predicted based on the requisition cycle corresponding to the material requisition records. The material requisition records and the historical quantity information are input into the material consumption rate model, and the predicted material consumption rate is obtained based on the output of the material consumption rate model.

4. The method according to claim 1, characterized in that, The step of determining the predicted material requisition quantity for the courier to be predicted based on the predicted quantity of items and the predicted material consumption rate includes: The predicted material requisition amount for the courier to be predicted is obtained by multiplying the predicted quantity of items and the predicted material consumption rate.

5. The method according to claim 1, characterized in that, After determining the predicted material requisition quantity for the courier to be predicted, the method further includes: Obtain the minimum packaging quantity corresponding to the material; the minimum packaging quantity represents the quantity of the material that can be configured in the smallest unit of packaging; Based on the predicted material usage of the couriers to be predicted and the minimum packaging quantity, determine multiple delivery batches for the couriers to be predicted and the delivery quantity corresponding to each delivery batch.

6. The method according to claim 5, characterized in that, Also includes: Obtain the material delivery weights corresponding to the multiple delivery batches; If the total delivery weight of the materials in the multiple delivery batches is greater than the first preset threshold, and there is a delivery batch whose corresponding material delivery weight is greater than the second preset threshold, the material delivery weight corresponding to the delivery batch will be allocated to other delivery batches.

7. The method according to claim 5, characterized in that, Also includes: Determine the employee number of the courier to be predicted, and sort it among the employee numbers of all couriers corresponding to the courier outlet to which the courier to be predicted belongs. Based on the sorting and the multiple delivery batches of the courier to be predicted, the material collection order of the courier to be predicted at the express delivery outlet is determined, and the material collection order is sent to the courier to be predicted.

8. A device for predicting the demand for express delivery materials, characterized in that, The device includes: The information acquisition module is used to acquire the courier characteristic information, historical parcel volume information, and daily characteristic information of the courier to be predicted; The part quantity model module is used to determine the corresponding type of part quantity prediction model based on the historical part quantity information. The predicted shipment volume acquisition module is used to input the courier feature information and the feature information of the day to be predicted into the shipment volume prediction model, and obtain the predicted shipment volume corresponding to the courier to be predicted based on the output of the shipment volume prediction model. The predicted consumption rate acquisition module is used to acquire the predicted material consumption rate of the courier to be predicted; the predicted material consumption rate is obtained based on the material requisition records of the courier to be predicted. The material forecast quantity acquisition module is used to determine the material forecast requisition quantity of the courier to be forecasted based on the forecast quantity of the items and the material forecast consumption rate.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images