Order group wave processing method and device

Abstract

The invention discloses an order group wave processing method and device, and relates to the field of warehouse logistics. A specific embodiment of the method comprises the steps of determining a current time period of current time, and obtaining related data of a first to-be-delivered order and order interception time of the current time period; obtaining related data of a second to-be-delivered order in a next time period through a Gaussian distribution model; obtaining a first ex-warehouse duration of a current time period, a second ex-warehouse duration of a next time period and an article type coincidence rate of the current time period and the next time period through a preset target model; and inputting the cut-off time, the first ex-warehouse duration, the second ex-warehouse duration and the article type coincidence rate into a wave combination mode determination model, and performing wave combination on to-be-ex-warehouse orders in the current time period and the next time period in response to the fact that an output result of the model is combinable, and issuing for picking. According to the embodiment, the current time period and the adjacent time period are combined into waves, the order number in one wave is increased, the order picking efficiency is improved, and the performance time is met.

Description

Technical Field

[0001] The present invention relates to the field of warehousing logistics, and in particular to an order group wave processing method and device. Background Art

[0002] With the rapid development of e-commerce and modern logistics, more and more users choose to shop online. After a wide range of items are ordered in the order system, the order system will send the order-related item information to the warehouse management system for positioning, pre-occupying inventory and generating waves, providing convenience for the subsequent picking process.

[0003] Currently, outbound operations are mainly carried out through order wave grouping. A batch of orders to be shipped that have recently entered the warehouse are manually aggregated according to certain standards or rules to generate wave tasks for subsequent picking tasks, review, delivery and other operations.

[0004] In the process of realizing the present invention, the inventors found that the prior art has at least the following problems: it is too dependent on manual operation, and the time interval between grouping waves is fixed, which makes it impossible to group more orders, affecting the efficiency of order picking and delivery. In addition, there is a required fulfillment time for delivery of orders. How to balance the fulfillment time and order grouping is also a major problem that needs to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides an order group wave processing method and device, which can at least solve the problem of low warehouse efficiency caused by reliance on manual order grouping, fixed group wave time intervals and unbalanced order fulfillment time in the prior art.

[0006] To achieve the above-mentioned purpose, according to one aspect of an embodiment of the present invention, a method for processing order group waves is provided, including: determining the current time period in which the current time is located, obtaining the first order-to-be-shipped related data of the current time period and the cut-off time in the order-to-be-shipped; processing the first order-to-be-shipped related data through a Gaussian distribution model to obtain the second order-to-be-shipped related data of the next time period; processing the first order-to-be-shipped related data and the second order-to-be-shipped related data through a preset target model to obtain the first shipment duration of the current time period, the second shipment duration of the next time period, and the overlap rate of item types in the current time period and the next time period; inputting the cut-off time, the first shipment duration, the second shipment duration, and the overlap rate of item types into a group wave determination model, and in response to the model output result being mergeable, grouping the orders to be shipped in the current time period and the next time period and sending them for picking.

[0007] Optionally, the method also includes: obtaining outbound order related data within a first time range; using a second time range to divide each day of the first time range to obtain multiple time periods per day; using the time periods to classify to obtain an outbound order related data set for each time period within the first time range, and dividing the outbound order related data set into a sample data set and a label data set; using the sample data set as input data and the label data set as output data to train the original model to obtain a trained target model.

[0008] Optionally, after obtaining the set of data related to outbound orders within the first time range for each time period, it also includes: extracting some time periods from multiple time periods; wherein the extraction is based on a first preset quantity or a first preset ratio; extracting some outbound order-related data from the set of data related to outbound orders for each time period; wherein the extraction is based on a second preset quantity or a second preset ratio.

[0009] Optionally, the outbound order related data include outbound order quantity, outbound time and warehouse related data, and the target model is an outbound time determination model; the outbound order related data set is divided into a sample data set and a label data set, including: combining the outbound order quantity set and warehouse related data set of each time period as a sample data set, and the outbound time set as a label data set.

[0010] Optionally, the method of combining the set of outbound order quantities and warehouse-related data for each time period as a sample data set and the set of outbound durations as a label data set also includes: for the set of outbound order quantities for the same time period, arranging each outbound order quantity in descending order from back to front in time, assigning a corresponding weight to each outbound order quantity according to a preset attenuation method, and obtaining the average outbound order quantity by taking a weighted averaging method; for the set of outbound durations for the same time period, arranging each outbound duration in descending order from back to front in time, assigning a corresponding weight to each outbound duration according to a preset attenuation method, and obtaining the average outbound duration by taking a weighted averaging method.

[0011] Optionally, the warehouse-related data includes: warehouse attribute data and warehouse operation data.

[0012] Optionally, the outbound order related data includes item type information and item attribute information in the outbound order, and the target model is a model for determining the item overlap rate in adjacent time periods; the dividing of the outbound order related data set into a sample data set and a label data set includes: determining a first item type information set in the outbound order in the first time period of each day, and a second item type information set in the outbound order in the second time period; wherein the first time period and the second time period are adjacent, and the first time period is located before the second time period; determining the number of common item types between the first item type information set and the second item type information set, and determining the number of item types in the second item type information set; calculating the ratio of the common item type number to the item type number, and using the ratio as the item type overlap rate between the second time period and the first time period; using the first item type information set and the item attribute information as sample data, and using the item type overlap rate between the second time period and the first time period as label data.

[0013] Optionally, after obtaining the set of data related to outbound orders within the first time range for each time period, the method further includes: calling a corresponding processing method to process the sample data according to the data type of the sample data to obtain a sample feature vector; wherein the data type is continuous data or enumerated data.

[0014] Optionally, the method of combining the sample data set as input data and the label data set as output data to train the original model to obtain a trained target model includes: inputting the sample data as input data into the original model to obtain an output result; performing loss calculation on the output result and the label data to iteratively optimize the parameters of the model by minimizing the model loss value until the loss value reaches a preset value, thereby obtaining the trained target model.

[0015] Optionally, the data related to the orders to be shipped include the number of orders to be shipped, and the method further includes: constructing a Gaussian distribution model based on Gaussian distribution; wherein the Gaussian distribution model is used to determine the number of orders to be shipped in adjacent time periods.

[0016] Optionally, the wave group determination model includes: fulfillment time restriction condition and wave group order quantity restriction condition; wherein, the fulfillment time restriction condition is: the difference between the cut-off time and the current time, the preset time period interval, the first delivery time, the second delivery time and the product of the item type overlap rate is greater than or equal to zero; the wave group order quantity restriction condition is: the sum of the number of orders to be shipped in the current time period and the number of orders to be shipped in the next time period is less than or equal to the upper limit threshold of the number of orders in a wave.

[0017] Optionally, the steps of grouping the orders to be shipped in the current time period and the next time period and sending them for picking include: determining the cut-off time for the next time period, repeating the steps of calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, so as to judge whether the orders to be shipped in the next time period and the next time period can be merged; in response to the judgment result that they can be merged, merging the orders to be shipped in the current time period, the next time period, and the next time period; repeating the steps of determining the cut-off time for the time period, calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, until the judgment result is that they cannot be merged, merging the orders to be shipped in multiple time periods with the judgment result of yes, to obtain the target wave and send them for picking.

[0018] To achieve the above-mentioned purpose, according to another aspect of an embodiment of the present invention, there is provided an order group wave processing device, including: an acquisition module, used to determine the current time period in which the current time is located, and obtain the first order-to-be-shipped related data of the current time period and the cut-off time in the order-to-be-shipped; a processing module, used to process the first order-to-be-shipped related data through a Gaussian distribution model to obtain the second order-to-be-shipped related data of the next time period; through a preset target model, the first order-to-be-shipped related data and the second order-to-be-shipped related data are processed to obtain the first shipment duration of the current time period, the second shipment duration of the next time period, and the overlap rate of the item types in the current time period and the next time period; a group wave module, used to input the cut-off time, the first shipment duration, the second shipment duration, and the overlap rate of the item types into a group wave mode determination model, and in response to the model output result being mergeable, group waves are performed on the orders to be shipped in the current time period and the next time period and send them for picking.

[0019] Optionally, the device also includes a model training module, which is used to: obtain outbound order related data within a first time range; use a second time range to divide each day of the first time range to obtain multiple time periods per day; use the time periods to classify, obtain an outbound order related data set for each time period within the first time range, and divide the outbound order related data set into a sample data set and a label data set; use the sample data set as input data and the label data set as output data to train the original model to obtain a trained target model.

[0020] Optionally, the model training module is also used to: extract part of the time periods from multiple time periods; wherein the extraction is based on a first preset quantity or a first preset ratio; extract part of the outbound order-related data from the outbound order-related data set of each time period; wherein the extraction is based on a second preset quantity or a second preset ratio.

[0021] Optionally, the outbound order related data includes the outbound order quantity, outbound time and warehouse related data, and the target model is the outbound time determination model; the model training module is used to: combine the outbound order quantity set and warehouse related data set for each time period as a sample data set, and the outbound time set as a label data set.

[0022] Optionally, the model training module is also used to: for a set of outbound order quantities in the same time period, arrange each outbound order quantity in descending order from the back to the front in time, assign a corresponding weight to each outbound order quantity according to a preset attenuation method, and adopt a weighted averaging method to obtain the average outbound order quantity; for a set of outbound durations in the same time period, arrange each outbound duration in descending order from the back to the front in time, assign a corresponding weight to each outbound duration according to a preset attenuation method, and adopt a weighted averaging method to obtain the average outbound duration.

[0023] Optionally, the warehouse-related data includes: warehouse attribute data and warehouse operation data.

[0024] Optionally, the outbound order related data includes item type information and item attribute information in the outbound order, and the target model is a model for determining the item overlap rate in adjacent time periods; the model training module is used to: determine a first item type information set in the outbound order in the first time period of each day, and a second item type information set in the outbound order in the second time period; wherein the first time period and the second time period are adjacent, and the first time period is located before the second time period; determine the number of common item types between the first item type information set and the second item type information set, and determine the number of item types in the second item type information set; calculate the ratio of the common item type number to the item type number, and use the ratio as the item type overlap rate between the second time period and the first time period; use the first item type information set and the item attribute information as sample data, and use the item type overlap rate between the second time period and the first time period as label data.

[0025] Optionally, the device further comprises a feature construction module, which is used to: according to the data type of the sample data, call a corresponding processing method to process the sample data to obtain a sample feature vector; wherein the data type is continuous data or enumerated data.

[0026] Optionally, the model training module is used to: input sample data as input data into the original model to obtain output results; perform loss calculation on the output results and label data to iteratively optimize the model parameters by minimizing the model loss value until the loss value reaches a preset value, thereby obtaining a trained target model.

[0027] Optionally, the data related to the orders to be shipped include the number of orders to be shipped, and the device also includes a Gaussian distribution model construction module, which is used to: construct a Gaussian distribution model based on Gaussian distribution; wherein the Gaussian distribution model is used to determine the number of orders to be shipped in adjacent time periods.

[0028] Optionally, the wave group determination model includes: fulfillment time restriction condition and wave group order quantity restriction condition; wherein, the fulfillment time restriction condition is: the difference between the cut-off time and the current time, the preset time period interval, the first delivery time, the second delivery time and the product of the item type overlap rate is greater than or equal to zero; the wave group order quantity restriction condition is: the sum of the number of orders to be shipped in the current time period and the number of orders to be shipped in the next time period is less than or equal to the upper limit threshold of the number of orders in a wave.

[0029] Optionally, the wave group module is also used to: determine the cut-off time for the next time period, repeat the above steps of calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, so as to judge whether the orders to be shipped in the next time period and the next time period can be merged; in response to the judgment result that they can be merged, merge the orders to be shipped in the current time period, the next time period, and the next time period; repeat the steps of determining the cut-off time for the time period, calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met until the judgment result is that they cannot be merged, merge the orders to be shipped in multiple time periods with the judgment result of yes, obtain the target wave and send them for picking.

[0030] To achieve the above objective, according to another aspect of an embodiment of the present invention, an electronic device for processing an order group wave is provided.

[0031] The electronic device of an embodiment of the present invention includes: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement any of the order group wave processing methods described above.

[0032] To achieve the above-mentioned purpose, according to another aspect of an embodiment of the present invention, a computer-readable medium is provided, on which a computer program is stored, and when the program is executed by a processor, any of the above-mentioned order group wave processing methods is implemented.

[0033] According to the scheme provided by the present invention, one embodiment of the above invention has the following advantages or beneficial effects: taking into full account the different distributions of the outbound time length of each time period of the warehouse and the repetition rate of goods in adjacent time periods, and constructing the number of orders to be outbound in adjacent time periods into a Gaussian distribution, better fitting the various distributions of the warehouse, and improving the accuracy of determining various data in different time periods. Then, considering the restrictions of the fulfillment time and the threshold of the number of group wave orders, iterative calculation is performed to determine whether adjacent time periods can be combined into a group wave, and finally a group wave scheme is obtained, the group wave is completed, and the purpose of dynamic group wave considering the fulfillment time is achieved.

[0034] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0036] Figure 1 It is a main flow chart of an order group wave processing method according to an embodiment of the present invention;

[0037] Figure 2 is a flow chart of an optional order group wave processing method according to an embodiment of the present invention;

[0038] Figure 3 is a flow chart of another optional order group wave processing method according to an embodiment of the present invention;

[0039] Figure 4 is a schematic diagram of the process of training a model for determining the outbound delivery time according to an embodiment of the present invention;

[0040] Figure 5 is a schematic diagram of a flow chart of a model for determining the overlap rate of items in adjacent time periods for training according to an embodiment of the present invention;

[0041] Figure 6 is a flow chart of another optional order group wave processing method according to an embodiment of the present invention;

[0042] Figure 7 is a flowchart of another optional order group wave processing method according to an embodiment of the present invention;

[0043] Figure 8 is a schematic diagram of main modules of an order group wave processing device according to an embodiment of the present invention;

[0044] Fig. 9 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0045] Fig.10It is a schematic diagram of the structure of a computer system of a mobile device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0047] It should be pointed out that the collection, collection, updating, analysis, processing, use, transmission, storage and other aspects of the user personal information involved in the technical solution of this disclosure are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Take necessary measures for user personal information, such as de-identification or anonymization, to protect user privacy, prevent illegal access to user personal information data, and maintain user personal information security, network security and national security. Once these user personal information data are no longer needed, the risks should be minimized by restricting or even prohibiting data collection and / or deleting data.

[0048] The problems of the prior art are described in detail here:

[0049] 1. The traditional order wave aggregation method mainly relies on the warehouse management end to pre-set the system wave management conditions according to the on-site operation situation, compile the orders that meet the same screening conditions into a picking list and print them, and then pick the goods according to the picking list. The screening conditions of the order aggregation are not comprehensive enough, resulting in unreasonable distribution of picking lists, which in turn affects the efficiency of outbound delivery. During promotional activities, it is easy to have a warehouse explosion.

[0050] 2. Currently, the time interval for order grouping is a fixed value set according to the warehouse management system, such as 15 minutes. In warehouse delivery scenarios such as different warehouse times and different order quantities, a fixed time interval grouping method is used, which lacks flexibility and makes it impossible to group more orders, thus affecting the efficiency of picking and subsequent delivery.

[0051] 3. There is a required fulfillment time for warehouse orders to be shipped out. If the fulfillment time is exceeded, there will be customer complaints and compensation. How to balance the order group and the fulfillment time is also a major problem that warehousing needs to solve.

[0052] See also Figure 1 , which shows a main flow chart of an order group wave processing method provided by an embodiment of the present invention, including the following steps:

[0053] S101: Determine the current time period in which the current time is located, and obtain the first order to be shipped in the current time period and the cut-off time of the order to be shipped;

[0054] S102: Processing the first order-to-be-shipped related data by using a Gaussian distribution model to obtain second order-to-be-shipped related data for the next time period;

[0055] S103: Processing the first order-to-be-shipped related data and the second order-to-be-shipped related data through a preset target model to obtain a first shipment duration in a current time period, a second shipment duration in a next time period, and an item type overlap rate between the current time period and the next time period;

[0056] S104: Input the cut-off time, the first delivery time, the second delivery time, and the item type overlap rate into a wave grouping method determination model, and in response to the model output result being mergeable, group the orders to be delivered in the current time period and the next time period and send them for picking.

[0057] In the above implementation, for steps S101 and S102, the time period in which the current time is located is determined, and the time period is taken as the current time period, and the first order to be shipped data in the current time period is obtained, specifically, the number of orders to be shipped, current warehouse attributes (such as automated warehouse, manual warehouse, etc.) and current warehouse operation data (such as busy procedures, personnel configuration, etc.), item type information, and item attribute information. Since orders are continuous, and the orders in the next time period are uncertain, it is necessary to predict the second order to be shipped data in the next time period through the first order to be shipped data in the current time period.

[0058] This solution pre-builds a Gaussian distribution model, or a model for determining the number of orders to be shipped in adjacent time periods. Assume that the number of orders to be shipped in the current time period is N batch_now , the number of orders to be shipped in the next time period is N batch_next , N batch_next Obey N batch_now is the mean, σ 2 is the Gaussian distribution of variance. Therefore, through the Gaussian distribution model, based on N batch_now and σ 2 Construct a Gaussian distribution (or normal distribution). Gaussian distribution follows the laws of nature. Randomly extract data samples from the Gaussian distribution as the number of orders to be shipped in the next time period N batch_next .

[0059] For warehouses, their attributes and operation data in different time periods are usually pre-set, so it is also possible to easily obtain warehouse attributes and warehouse operation data for the next time period, combined with the estimated number of orders to be shipped N batch_next, get the relevant data of the second order to be shipped in the next time period.

[0060] Among them, warehouse attribute data includes but is not limited to: warehouse scale, personnel attributes, storage location attributes, item attributes, etc., such as the number of personnel and proficiency of each position, the degree of automation of the warehouse, the number of storage locations, storage location types, item types, etc. Warehouse operation data includes but is not limited to: warehouse outbound process configuration data, such as the warehouse outbound operation process is "initialization-positioning-picking-review-packaging-shipping" or "initialization-positioning-picking-shipping", etc. The dynamic attributes of warehouse operation data can be selected from the average daily outbound volume of the warehouse in the previous month and the volume of goods to be outbound in the current time period as dynamic attributes, or the busyness of each operation process in the current time period as dynamic attributes.

[0061] For step S103, after obtaining the relevant data of the first order to be shipped in the current time period and the relevant data of the second order to be shipped in the next time period, the first shipping duration in the current time period, the second shipping duration in the next time period, and the overlap rate of item types in the current time period and the next time period can be estimated. Therefore, the preset target model here can estimate both the shipping duration and the overlap rate of item types.

[0062] Specifically, the target model includes a model for determining the delivery time and a model for determining the overlap rate of items in adjacent time periods. The delivery time determination model is used to estimate the first delivery time T of the current time period based on the data related to the first order to be delivered. current , based on the relevant data of the second order to be shipped, estimate the second shipping time T in the next time period next The model for determining the overlap rate of items in adjacent time periods is used to estimate the overlap rate p of item types in the current time period and the next time period based on the first order-to-be-shipped data and the second order-to-be-shipped data.

[0063] Furthermore, the data related to the orders to be shipped can be digitized to obtain feature vectors, such as the shipping time feature vector and the item overlap rate feature vector. Since the order item information for the next time period is unknown, only the data related to the first order to be shipped includes the set of types of items to be shipped and item attribute data.

[0064] For step S104, this solution constructs a group wave method determination model based on the fulfillment time restriction condition and the group wave order quantity restriction condition in advance, wherein:

[0065] Time constraints for fulfillment: t ddl -t now -T gap -T current -T next *p≥0, where t ddlis the cut-off time for orders to be shipped out in the current time period, t now is the current time, T gap is the time interval, usually preset, such as 15 minutes, T current is the first outbound delivery time of the current time period, T next is the second outbound time of the next time period, and p is the overlap rate of item types between the current time period and the next time period.

[0066] The number of group wave orders is limited to: N batch_now +N batch_next ≤N batch_limit , where N batch_now is the number of orders to be shipped in the current time period, N batch_next is the number of orders to be shipped in the next time period, N batch_limit It is the upper threshold of the order quantity in a wave.

[0067] Only when the above two restrictions are met can the orders to be shipped in the current time period and the next time period be combined into one wave for picking, so that the number of orders in one wave can be as large as possible, thereby improving the order picking efficiency and meeting the fulfillment time.

[0068] The method provided in the above embodiment predicts the relevant data of the orders to be shipped in the next time period based on the relevant data of the orders to be shipped in the current time period, and then determines whether adjacent time periods can be grouped according to the fulfillment time restriction conditions and the group order quantity restriction conditions, so that the number of orders in one wave is as large as possible, thereby improving the order picking efficiency and meeting the fulfillment time.

[0069] See also Figure 2 , which shows a main flow chart of an optional order group wave processing method provided by an embodiment of the present invention, comprising the following steps:

[0070] S201: Determine the current time period in which the current time is located, and obtain the first order to be shipped in the current time period and the cut-off time of the order to be shipped;

[0071] S202: Processing the first order-to-be-shipped related data through a Gaussian distribution model to obtain second order-to-be-shipped related data for the next time period;

[0072] S203: Processing the first order-to-be-shipped related data and the second order-to-be-shipped related data through a preset target model to obtain a first shipment duration in a current time period, a second shipment duration in a next time period, and an item type overlap rate between the current time period and the next time period;

[0073] S204: inputting the order cut-off time, the first delivery time, the second delivery time, and the overlap rate of the item types into a wave grouping mode determination model, and in response to the model output result being mergeable, merging the orders to be delivered in the current time period and the next time period; wherein the wave grouping mode determination model includes: a fulfillment time restriction condition and a wave grouping order quantity restriction condition;

[0074] S205: Determine the cut-off time for the next time period, repeat the above steps of calculating the data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, so as to judge whether the orders to be shipped in the next time period and the next time period can be merged;

[0075] S206: In response to the judgment result that the orders can be merged, the orders to be shipped in the current time period, the next time period, and the next next time period are merged;

[0076] S207: Repeat the steps of determining the cut-off time of the time period, calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met until the result is that they cannot be merged. Merge the orders to be shipped in multiple time periods with a judgment result of yes to obtain the target wave and send them for picking.

[0077] In the above embodiment, for steps S201 to S204, see Figure 1 The above description will not be repeated here.

[0078] For steps S205 to S207, assume that there are multiple time periods, 10:15-10:30, 10:30-10:45, 10:45-11:00, 11:00-11:15, ..., through Figure 1 In the steps shown, it is determined that the orders to be shipped in the two time periods of 10:15-10:30 and 10:30-10:45 can be merged, then the calculation continues to determine whether the orders to be shipped in the two time periods of 10:30-10:45 and 10:45-11:00 can be merged. If the result is yes, then the orders to be shipped in the two time periods of 10:45-11:00 and 11:00-11:15 can be merged, and the calculation is continued to be iterative until the calculation result is no.

[0079] Assuming that the calculation results are the four time periods of 10:15-10:30, 10:30-10:45, 10:45-11:00, and 11:00-11:15, the orders to be shipped in these four time periods will be merged into one wave for picking, so as to complete more order picking and shipping in one wave. The iterative calculation shows that the time interval of the group wave is 1 hour. Compared with the current 15 minutes, the number of orders in the group wave is greatly increased.

[0080] It should be noted that the cut-off time can be determined directly based on the time period, which is basically fixed in the warehouse. For example, orders placed after 20:00 p.m. are required to be shipped out before 12:00 the next day, and orders placed on the same day are required to be shipped out at 22:00. Therefore, for the above-mentioned 10:15-10:30, 10:30-10:45, 10:45-11:00, and 11:00-11:15, the cut-off time is 22:00.

[0081] The method provided in the above embodiment determines whether adjacent time periods can be grouped through iterative calculation, and finally obtains a group-grouping scheme. At this time, the time interval of the group-grouping is greatly improved compared with the existing fixed time interval, so the number of orders in the group-group also increases significantly, while meeting the fulfillment time.

[0082] See also Figure 3 , which shows a main flow chart of another optional order group wave processing method provided by an embodiment of the present invention, comprising the following steps:

[0083] S301: Obtaining data related to outbound orders within a first time range;

[0084] S302: Use the second time range to divide each day of the first time range into multiple time periods per day;

[0085] S303: Classify using time periods to obtain a data set related to outbound orders within a first time range for each time period, and divide the data set related to outbound orders into a sample data set and a label data set;

[0086] S304: The sample dataset is used as input data and the label dataset is used as output data to train the original model and obtain a trained target model.

[0087] In the above implementation, for steps S301 to S303, this implementation mainly describes the training process of the target model. First, determine a certain period of time in history (the first time range), such as the last month or three months (the value is only an example and is actually adjustable). Obtain the outbound order related data within the first time range, including one or more of the following: order identification, order quantity, order item composition, item identification, item quantity, outbound time, warehouse-related data, item type information, and item attribute information.

[0088] The first time range is divided into several time periods each day according to a certain time interval (second time range). The second time range can be the interval time of the waves in the warehouse, such as 15 minutes. One day is divided into 96 time periods according to the above 15 minutes, and time period 1 (0:00-0:15), time period 2 (0:15-0:30), ... time period 96 (23:45-24:00) are obtained.

[0089] According to the time period, the outbound order related data within the first time range are divided to obtain the classification results of multiple time periods. For example, time period 1 includes 30 outbound order related data from April 1 to 30, 2023, time period 2 includes 30 outbound order related data from April 1 to 30, 2023, time period 3 includes 30 outbound order related data from April 1 to 30, 2023…, a total of 96 time periods of outbound order related data sets.

[0090] Furthermore, considering that there are a large number of time periods and a large number of outbound order-related data in each time period, the subsequent calculation workload can be reduced. Some time periods can be extracted from the 96 time periods, and the extraction basis is the first preset number N or the first preset ratio, such as 50%, 80%, 100%. Or, from the outbound order-related data set of each time period, some outbound order-related data can be extracted, and the extraction basis is the second preset number M or the second preset ratio, such as 50%, 80%. Or, a combination of the two, first extract the time period, and then extract some outbound order-related data from the outbound order-related data set of the extracted time period. The amount of calculation can be greatly reduced through the extraction method. In addition, from a probability perspective, random extraction can better reflect the randomness of the prediction.

[0091] Through the above steps, whether it is the classification results of the original multiple time periods, or the classification results are extracted therefrom, or the outbound order related data are extracted therefrom, or both the classification results and the outbound order related data are extracted, the purpose is to construct sample data. This solution preferably constructs it by "extracting both the classification results and the outbound order related data". And considering that the current time period (such as 12:00-12:30) must have similar rules with the outbound order related data of the same time period in history (such as the past 1 month). And the more recent the time, the more similar or close the data is, so here we start from the perspective of time period.

[0092] For step S304, according to the training purpose, the outbound order related data set is divided into a sample data set and a label data set. For example, if the training is to estimate the outbound time, the sample data set includes the outbound order quantity and warehouse related data, and the label data includes the outbound time. If the training is to determine the overlap rate of item types in adjacent time periods, the sample data set includes item type information and item attribute information in the outbound order, and the label data includes the overlap rate of item types.

[0093] The sample dataset is taken as input and the label dataset is taken as output. The original model is trained to obtain the target model. The obtained target model can estimate the outbound delivery time and the overlap rate of item types in adjacent time periods.

[0094] Furthermore, the sample data may also include statistical feature data, wherein the statistical feature data includes one or more of the following: such as the mean, first quantile value, median, standard deviation and maximum value of historical data. The first quantile value may be flexibly set, such as the 80% quantile value or the 95% quantile value.

[0095] The method provided in the above embodiment divides each day of a historical period into multiple time periods, and constructs a data set related to the outbound orders in the same time period in the historical time period to train the model, thereby obtaining a target model that can estimate the outbound time and the overlap rate of item types in adjacent time periods.

[0096] See also Figure 4 , which shows a main flow chart of a training outbound delivery time determination model provided by an embodiment of the present invention, including the following steps:

[0097] S401: Acquire outbound order related data within a first time range; wherein the outbound order related data includes outbound order quantity, outbound time and warehouse related data, and the warehouse related data includes: warehouse attribute data and warehouse operation data;

[0098] S402: Use the second time range to divide each day of the first time range into multiple time periods per day;

[0099] S403: Classify using time periods to obtain a data set related to outbound orders within a first time range for each time period;

[0100] S404: The outbound order quantity set and warehouse-related data set in each time period are combined as a sample data set, and the outbound time set is combined as a label data set;

[0101] S405: The sample data set is used as input data and the label data set is used as output data to train the original model, thereby obtaining a trained model for determining the delivery time.

[0102] In the above-mentioned implementation mode, this implementation mode is mainly used to obtain and construct sample data and label data of the warehouse delivery time determination model, and obtain the warehouse delivery time determination model based on this training.

[0103] For steps S401 to S405, first obtain the outbound order related data within the first time range, then divide it according to the second time range to obtain the classification results of multiple time periods, and randomly or according to a certain number or probability extract multiple outbound order related data in the same time period to construct the outbound time period sample data. For example, the number of outbound orders and warehouse attributes, as well as the required outbound time length, in the same time period of 10:15-10:30 every day in the history of one month. In addition to being related to the number of outbound orders, the outbound time length is also related to the warehouse operation attributes. Therefore, this solution also needs to construct warehouse attribute sample data and warehouse operation sample data.

[0104] Furthermore, for the outbound order quantity set and the outbound time set in the outbound order related data set, they are arranged in chronological order from back to front, and different weights are assigned to the elements in the set according to exponential decay or linear decay, and the sum of the element weights is 1, so as to obtain the outbound time period sample data and the corresponding outbound time label, so the outbound time period sample data only corresponds to the outbound order quantity set.

[0105] The specific weight exponential decay formula is as follows, where w i is the weight of the i-th element, with a total of n elements:

[0106]

[0107]

[0108] For example, the current time is the 618 promotion. Before 618, the order sales data of 617 were close to those of 618, followed by 616 and 615, so the weight should be gradually reduced. For example, the weight of 617 is set to 0.6. After weighting, the weighted average method is used to calculate the average number of outbound orders and the average outbound time. Assuming that the original data is the number of 30 outbound orders in the same time period of 10:15-10:30 every day for a month in history, after this operation, only one data is left. The same is true for the outbound time, so as to further reduce the amount of calculation.

[0109] The method provided in the above embodiment, for the model for determining the outbound delivery time, after obtaining the sample data and the label data, in addition to using the extraction method, can also use the weighted averaging method, and the weighted method decreases one by one in order from the back to the front, so that the later the time, the higher the weight, thereby increasing its influence on the current time period, making the estimated outbound delivery time more accurate.

[0110] See also Figure 5 , which shows a main flow chart of a model for determining the overlap rate of items in adjacent time periods provided by an embodiment of the present invention, including the following steps:

[0111] S501: Acquire relevant data of outbound orders within a first time range; wherein the relevant data of outbound orders include item type information and item attribute information in the outbound orders;

[0112] S502: Use the second time range to divide each day of the first time range into multiple time periods per day;

[0113] S503: Classify using time periods to obtain a data set related to outbound orders within a first time range for each time period;

[0114] S504: Determine a first item category information set in a first time period outbound order and a second item category information set in a second time period outbound order every day; wherein the first time period and the second time period are adjacent, and the first time period is before the second time period;

[0115] S505: Determine the number of common item types between the first item type information set and the second item type information set, and determine the number of item types in the second item type information set;

[0116] S506: Calculate the ratio of the number of the common item types to the number of the item types, and use the ratio as the overlap rate of item types between the second time period and the first time period;

[0117] S507: taking the first item category information set and the item attribute information as sample data, and taking the item category overlap rate between the second time period and the first time period as label data;

[0118] S508: The sample data set is combined as input data and the label data set is combined as output data to train the original model, and obtain a model for determining the overlap rate of items in adjacent time periods after training.

[0119] In the above-mentioned implementation mode, this implementation mode is used to obtain and construct sample data and label data of a model for determining the overlap rate of items in adjacent time periods.

[0120] For steps S501 to S503, the classification results here are different from the classification results of calculating the outbound time, which are a set of multiple outbound order item types in the same time period of each day (for example, a set of multiple item types from 10:15 to 10:30 in the past month), and the overlap rate of item types between the next time period and the current time period. As described above, this solution needs to randomly extract N items from the classification results, or extract them according to a certain probability (for example, 50%, 100%) to reduce the amount of calculation.

[0121] In addition to the order itself, the overlap rate also needs to consider the attributes of the items, such as the type of item and whether it is a hot item. For example, in summer, the overlap rate of watermelon, cold drinks, and air conditioners is relatively high. Therefore, the outbound order-related data in this embodiment includes the item type information and item attribute information in the outbound order.

[0122] For steps S504 to S506, the overlap rate refers to the overlap rate of item types between two adjacent time periods. Take Table 1 as an example:

[0123] Table 1

[0124]

[0125] The item type set for the time period 15:15-15:30 is {skuA, skuB, skuC, skuE, skuF}, and the item type set for the time period 15:30-15:45 is {skuA, skuB, skuC, skuF, skuX, skuY}. Then the overlap rate of item types in the above adjacent time periods is 4 / 6=66.7%. The calculation formula is:

[0126] p=N cross / N all

[0127] Among them, p is the overlap rate of item types between the second time period and the first time period, N cross N is the number of common item types in the first time period and the second time period. all It is the sum of all item types in the second time period. The first time period and the second time period are adjacent, and the first time period is before the second time period.

[0128] For steps S507 to S508, the item type information set and item attribute information are used as sample data, the corresponding item type overlap rate is used as label data, the sample data set is used as input data, and the label data set is used as output data to train the model and obtain a model for determining the item overlap rate in adjacent time periods.

[0129] The method provided in the above embodiment, for determining the model for the overlap rate of items in adjacent time periods, in addition to considering the item type information of the outbound orders, also needs to consider the item attribute information, so as to first calculate the overlap rate of items in adjacent time periods and use this to train the model. Even if the subsequent model does not obtain the item types of the outbound orders in the next time period, it can also estimate the overlap rate of item types based on the relevant data of the outbound orders in the current time period.

[0130] See also Figure 6 , which shows a main flow chart of another order group wave processing method provided by an embodiment of the present invention, comprising the following steps:

[0131] S601: Obtaining data related to outbound orders within a first time range;

[0132] S602: Use the second time range to divide each day of the first time range into multiple time periods per day;

[0133] S603: Classify using time periods to obtain a data set related to outbound orders within a first time range for each time period, and divide the data set related to outbound orders into a sample data set and a label data set;

[0134] S604: according to the data type of the sample data, calling a corresponding processing method to process the sample data to obtain a sample feature vector; wherein the data type is continuous data or enumeration data;

[0135] S605: The sample dataset is used as input data and the label dataset is used as output data to train the original model and obtain a trained target model.

[0136] In the above implementation, for steps S601 to S603 and S605, see Figure 3 to Figure 5 The description shown is not repeated here.

[0137] For step S604, the sample data is processed according to the data type to obtain a sample feature vector. For the model for determining the time of shipment, the obtained feature vector is the feature vector of the time of shipment, and for the model for determining the overlap rate of items in adjacent time periods, the obtained feature vector is the feature vector of the overlap rate of items.

[0138] In this solution, the data type includes one of continuous data and / or enumerated data; wherein, continuous data may include time period, number of outbound orders, outbound time, and item overlap rate, etc.; enumerated data includes outbound process type and storage location type, etc., such as warehouse attribute data and warehouse operation data.

[0139] For the processing of continuous data, the sample data can be transformed mathematically to make the original data dimensionless, that is, the values ​​of each indicator are at the same level, avoiding the impact of data dimension on model training. For example, the number of items can be converted into a compressed value between [0,1] through min-max standardization / z-score standardization. For the processing of enumerated data, such as order types such as sales outbound, return outbound, 2B orders, etc., the vectors corresponding to various enumerations are generated through one-hot encoding.

[0140] The method provided in the above embodiment greatly reduces the computational workload of subsequent training models by performing feature processing on sample data, and digitizes the data to process a large range of sample data and realize standardized data processing, thereby further improving the model training speed.

[0141] See also Figure 7 , which shows a main flow chart of another order group wave processing method provided by an embodiment of the present invention, comprising the following steps:

[0142] S701: Obtaining data related to outbound orders within a first time range;

[0143] S702: Use the second time range to divide each day of the first time range into multiple time periods per day;

[0144] S703: Classify using time periods to obtain a data set related to outbound orders within a first time range for each time period, and divide the data set related to outbound orders into a sample data set and a label data set;

[0145] S704: input the sample data as input data into the original model to obtain output results;

[0146] S705: Perform loss calculation on the output results and label data to iteratively optimize the model parameters by minimizing the model loss value until the loss value reaches a preset value, thereby obtaining a trained target model.

[0147] In the above implementation, for steps S701 to S703, see Figure 3 to Figure 6 The description shown is not repeated here.

[0148] For steps S704 to S705, the model is trained based on the sample data and the label data, and a neural network model, an integrated learning algorithm, etc. can be used for training. This solution mainly uses the training process of the neural network model as an illustration. The sample feature vector is used as the input data of the model, and the label data is used as the output data of the model. The model is trained to learn the relationship between the sample feature data and the sample label data.

[0149] Furthermore, multiple sample data of multiple dimensions are input into the multi-layer perceptron for training. The dimensionless sample data processed above is used as input. The relevant parameters of the multi-layer perceptron model generated by the library transfer are constructed, including the number of layers of the neural network, the random initialization weight value, the learning rate, etc.

[0150] The above sample data and label data are used as training data for the multi-layer neural network model. Gradient descent or Adam algorithms are used to optimize the solution. Sigmod / ReLu is used as the activation function to obtain the output result yi. The loss function is constructed based on the mean square error between the output result and the label data. By minimizing the model loss value, the parameters of the model are continuously iterated and optimized until the set accuracy value is reached. The loss function is as follows:

[0151]

[0152] Among them, K is the total sample size, Y i is the model output result, y i The model is supervised training. Take the model for determining the outbound delivery time as an example. The sample data is the outbound delivery quantity and warehouse-related data. The model output is an outbound delivery time. The outbound delivery time and the label data (that is, the actual outbound delivery time) are used for loss calculation. The model parameters are adjusted according to the loss results to minimize the loss until the loss value reaches the preset value, such as 0.1.

[0153] The above training process is the same for the model for determining the outbound delivery time and the model for determining the overlap rate of items in adjacent time periods. Furthermore, the outbound delivery time feature vector or the item repetition rate feature vector can be used as the input data of the model, and the label data can be used as the comparison object for training.

[0154] The method provided in the above embodiment trains the model based on the loss function to minimize the model error and further improve the accuracy of subsequent predictions of the model.

[0155] Compared with the prior art, the method provided by the embodiment of the present invention has at least the following beneficial effects:

[0156] 1. Based on the warehousing operation background, obtain the data related to the outbound orders within the first time range, divide the above time every day according to the second time range, and obtain the classification results of multiple time periods: sample data of outbound time, sample data of commodity type repetition rate in adjacent time periods, and obtain the outbound time feature vector and commodity repetition rate feature vector after processing. Combined with warehousing attributes, warehousing operation attributes, commodity attributes, etc., train machine learning algorithms such as neural network algorithms and integrated algorithms to respectively build outbound time determination models and adjacent time period commodity repetition rate determination models.

[0157] 2. Construct a Gaussian distribution model to calculate the number of orders to be shipped in adjacent time periods.

[0158] 3. Based on the fulfillment time constraints and the group order quantity constraints, a group order method determination model is constructed.

[0159] By constructing the first three models, we fully consider the different distributions of the warehouse delivery time in each time period and the repetition rate of goods in adjacent time periods, and construct the number of orders to be shipped in adjacent time periods as a Gaussian distribution, which better fits the various distributions of the warehouse and improves the accuracy of the data determined in different time periods. Then, considering the restrictions of fulfillment time and the threshold of the number of group wave orders, we iteratively calculate whether adjacent time periods can be combined into a group wave, and finally obtain the group wave solution, complete the group wave, and realize the dynamic determination of the group wave interval considering the fulfillment time.

[0160] See also Figure 8 , showing a schematic diagram of the main modules of an order group wave processing device 800 provided by an embodiment of the present invention, including: an acquisition module 801, used to determine the current time period in which the current time is located, and obtain the first order-to-be-shipped related data of the current time period and the cut-off time in the order-to-be-shipped; a processing module 802, used to process the first order-to-be-shipped related data through a Gaussian distribution model to obtain the second order-to-be-shipped related data of the next time period; through a preset target model, the first order-to-be-shipped related data and the second order-to-be-shipped related data are processed to obtain the first shipment duration of the current time period, the second shipment duration of the next time period, and the overlap rate of the item types in the current time period and the next time period; a group wave module 803, used to input the cut-off time, the first shipment duration, the second shipment duration, and the overlap rate of the item types into a group wave determination model, and in response to the model output result being mergeable, group waves are performed on the orders to be shipped in the current time period and the next time period and send them for picking.

[0161] The implementation device of the present invention also includes a model training module, which is used to: obtain outbound order related data within a first time range; use a second time range to divide each day of the first time range to obtain multiple time periods per day; use the time periods to classify, obtain an outbound order related data set for each time period within the first time range, and divide the outbound order related data set into a sample data set and a label data set; use the sample data set as input data and the label data set as output data to train the original model to obtain a trained target model.

[0162] In the implementation device of the present invention, the model training module is also used to: extract part of the time period from multiple time periods; wherein the extraction is based on a first preset quantity or a first preset ratio; extract part of the outbound order-related data from the outbound order-related data set of each time period; wherein the extraction is based on a second preset quantity or a second preset ratio.

[0163] In the implementation device of the present invention, the outbound order related data includes the outbound order quantity, outbound time and warehouse related data, and the target model is the outbound time determination model; the model training module is used to: combine the outbound order quantity set and warehouse related data of each time period as a sample data set, and the outbound time set as a label data set.

[0164] In the implementation device of the present invention, the model training module is also used for: for a set of outbound order quantities in the same time period, arranging each outbound order quantity in descending order according to the order of time from back to front, assigning a corresponding weight to each outbound order quantity according to a preset attenuation method, and obtaining the average outbound order quantity by weighted averaging; for a set of outbound durations in the same time period, arranging each outbound duration in descending order according to the order of time from back to front, and assigning a corresponding weight to each outbound duration according to a preset attenuation method, and obtaining the average outbound duration by weighted averaging.

[0165] In the implementation device of the present invention, the warehouse-related data includes: warehouse attribute data and warehouse operation data.

[0166] In the implementation device of the present invention, the outbound order related data includes item type information and item attribute information in the outbound order, and the target model is an item overlap rate determination model for adjacent time periods; the model training module is used to: determine a first item type information set in the outbound order in the first time period of each day and a second item type information set in the outbound order in the second time period; wherein the first time period and the second time period are adjacent, and the first time period is located before the second time period; determine the number of common item types between the first item type information set and the second item type information set, and determine the number of item types in the second item type information set; calculate the ratio of the number of common item types to the number of item types, and use the ratio as the item type overlap rate between the second time period and the first time period; use the first item type information set and the item attribute information as sample data, and use the item type overlap rate between the second time period and the first time period as label data.

[0167] The implementation device of the present invention also includes a feature construction module, which is used to: according to the data type of the sample data, call a corresponding processing method to process the sample data to obtain a sample feature vector; wherein the data type is continuous data or enumeration data.

[0168] In the implementation device of the present invention, the model training module is used to: input sample data as input data into the original model to obtain output results; perform loss calculation on the output results and label data to iteratively optimize the parameters of the model by minimizing the model loss value until the loss value reaches a preset value, thereby obtaining a trained target model.

[0169] In the implementation device of the present invention, the data related to the orders to be shipped include the number of orders to be shipped, and the device also includes a Gaussian distribution model construction module, which is used to: construct a Gaussian distribution model based on Gaussian distribution; wherein the Gaussian distribution model is used to determine the number of orders to be shipped in adjacent time periods.

[0170] In the implementation device of the present invention, the group wave mode determination model includes: fulfillment time restriction conditions and group wave order quantity restriction conditions; wherein, the fulfillment time restriction conditions are: the difference between the cut-off time and the current time, the preset time period interval, the first outbound delivery time, the second outbound delivery time and the product of the item type overlap rate is greater than or equal to zero; the group wave order quantity restriction conditions are: the sum of the number of orders to be shipped in the current time period and the number of orders to be shipped in the next time period is less than or equal to the upper limit threshold of the number of orders in a wave.

[0171] In the implementation device of the present invention, the wave group module 803 is also used to: determine the cut-off time of the next time period, repeat the above steps of calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, so as to judge whether the orders to be shipped in the next time period and the next time period can be merged; in response to the judgment result that they can be merged, merge the orders to be shipped in the current time period, the next time period, and the next time period; repeat the steps of determining the cut-off time of the time period, calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, until the judgment result is that they cannot be merged, merge the orders to be shipped in multiple time periods with the judgment result of yes, obtain the target wave and send them for picking.

[0172] In addition, the specific implementation content of the device described in the embodiment of the present invention has been described in detail in the method described above, so the repeated content will not be described again here.

[0173] Fig. 9 An exemplary system architecture 900 to which embodiments of the present invention may be applied is shown, including terminal devices 901 , 902 , 903 , a network 904 and a server 905 (only an example).

[0174] Terminal devices 901, 902, 903 can be various electronic devices with display screens and supporting web browsing, and various communication client applications are installed. Users can use terminal devices 901, 902, 903 to interact with server 905 through network 904 to receive or send messages, etc.

[0175] The network 904 is used to provide a medium for communication links between the terminal devices 901, 902, 903 and the server 905. The network 904 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0176] The server 905 may be a server that provides various services. It should be noted that the method provided in the embodiment of the present invention is generally executed by the server 905 , and accordingly, the device is generally set in the server 905 .

[0177] It should be understood that Fig. 9 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0178] Reference below Fig.10 , which shows a schematic diagram of the structure of a computer system 1000 of a terminal device suitable for implementing an embodiment of the present invention. Fig.10 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0179] like Fig.10 As shown, the computer system 1000 includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage part 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the system 1000 are also stored. The CPU 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0180] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage section 1008 as needed.

[0181] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the central processing unit (CPU) 1001, the above-mentioned functions defined in the system of the present invention are executed.

[0182] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0183] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0184] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor. For example, they may be described as: a processor includes an acquisition module, a processing module, and a wave group module. The names of these modules do not constitute limitations on the modules themselves in some cases. For example, the processing module may also be described as an "order data processing module."

[0185] As another aspect, the present invention further provides a computer-readable medium, which may be included in the device described in the above embodiment; or may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device executes any of the above order group wave processing methods.

[0186] The computer program product of the present invention comprises a computer program, and when the computer program is executed by a processor, the order group wave processing method in the embodiment of the present invention is implemented.

[0187] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for processing order group waves, characterized in that: include: Determine the current time period in which the current time is located, obtain the first order to be shipped in the current time period and the cut-off time of the order to be shipped; Processing the first order-to-be-shipped related data through a Gaussian distribution model to obtain second order-to-be-shipped related data for the next time period; By using a preset target model, the first data related to the order to be shipped and the second data related to the order to be shipped are processed to obtain the first shipping duration in the current time period, the second shipping duration in the next time period, and the overlap rate of item types in the current time period and the next time period; The cut-off time, the first delivery time, the second delivery time, and the item type overlap rate are input into a wave grouping method determination model. In response to the model output result that the orders to be delivered in the current time period and the next time period are combined, the orders are grouped and sent for picking.

2. The method according to claim 1, characterized in that The method further comprises: Obtain the outbound order related data within the first time range; Use the second time range to divide each day of the first time range to obtain multiple time periods per day; Use time periods for classification to obtain a data set related to outbound orders within the first time range for each time period, and divide the data set related to outbound orders into a sample data set and a label data set; The sample dataset is used as input data and the label dataset is used as output data to train the original model and obtain the trained target model.

3. The method according to claim 2, characterized in that After obtaining the data set related to the outbound orders in each time period within the first time range, the method further includes: Extracting a portion of time periods from a plurality of time periods; wherein the extraction is based on a first preset number or a first preset ratio; Extract part of the outbound order related data from the outbound order related data set of each time period; wherein the extraction is based on the second preset quantity or the second preset ratio.

4. The method according to claim 2 or 3, characterized in that: The outbound order related data includes outbound order quantity, outbound time and warehouse related data, and the target model is an outbound time determination model; The step of dividing the outbound order related data set into a sample data set and a label data set includes: The outbound order quantity set and warehouse-related data set in each time period are combined as the sample data set, and the outbound time set is taken as the label data set.

5. The method according to claim 4, characterized in that The method combines the outbound order quantity set and warehouse-related data set in each time period as a sample data set, and the outbound time set as a label data set, and also includes: For the set of outbound order quantities in the same time period, sort the outbound order quantities in descending order from the last time to the first time, assign corresponding weights to each outbound order quantity according to the preset attenuation method, and use the weighted average method to obtain the average outbound order quantity; For the set of outbound time durations in the same time period, each outbound time duration is arranged in descending order from the back to the front, and a corresponding weight is assigned to each outbound time duration according to the preset attenuation method, so as to obtain the average outbound time duration by adopting the weighted averaging method.

6. The method according to claim 4, characterized in that The warehouse related data includes: warehouse attribute data and warehouse operation data.

7. The method according to claim 2 or 3, characterized in that: The outbound order related data includes item type information and item attribute information in the outbound order, and the target model is a model for determining the overlap rate of items in adjacent time periods; The step of dividing the outbound order related data set into a sample data set and a label data set includes: Determine a first item category information set in a first time period of each day's outbound orders and a second item category information set in a second time period of each day's outbound orders; wherein the first time period and the second time period are adjacent, and the first time period is before the second time period; Determining the number of common item types between the first item type information set and the second item type information set, and determining the number of item types in the second item type information set; Calculating the ratio of the number of the common item types to the number of the item types, and using the ratio as the overlap rate of the item types between the second time period and the first time period; The first item category information set and item attribute information are used as sample data, and the item category overlap rate between the second time period and the first time period is used as label data.

8. The method according to claim 2, characterized in that: After obtaining the data set related to the outbound orders in each time period within the first time range, the method further includes: According to the data type of the sample data, a corresponding processing method is called to process the sample data to obtain a sample feature vector; wherein the data type is continuous data or enumeration data.

9. The method according to claim 2, characterized in that: The method uses the sample dataset as input data and the label dataset as output data to train the original model to obtain a trained target model, including: The sample data is used as input data, input into the original model, and the output result is obtained; The loss is calculated for the output results and label data to minimize the model loss value and iteratively optimize the model parameters until the loss value reaches the preset value to obtain the trained target model.

10. The method according to claim 1, characterized in that The data related to the orders to be shipped include the quantity of the orders to be shipped, and the method further includes: Based on Gaussian distribution, a Gaussian distribution model is constructed; wherein the Gaussian distribution model is used to determine the number of orders to be shipped in adjacent time periods.

11. The method according to claim 1, characterized in that: The wave grouping method determination model includes: fulfillment time constraint conditions and wave grouping order quantity constraint conditions; The fulfillment time restriction condition is: the difference between the cut-off time and the product of the current time, the preset time interval, the first delivery time, the second delivery time and the overlap rate of the item types is greater than or equal to zero; The limit condition for the number of orders in a wave is: the sum of the number of orders to be shipped in the current time period and the number of orders to be shipped in the next time period is less than or equal to the upper threshold of the number of orders in a wave.

12. The method according to claim 1, characterized in that The step of grouping the outbound orders in the current time period and the next time period and sending them for picking includes: Determine the cut-off time for the next time period, repeat the above steps of calculating the data of the orders to be shipped, estimating the shipping time and the overlap rate of the item types, and judging whether the restriction conditions are met, so as to determine whether the orders to be shipped in the next time period and the next time period can be merged; In response to the judgment result that the orders can be merged, the orders to be shipped in the current time period, the next time period, and the next next time period are merged; Repeat the steps of determining the cut-off time for the time period, calculating the relevant data of the orders to be shipped, estimating the shipping time and the overlap rate of item types, and judging whether the restriction conditions are met until the result is that they cannot be merged. Merge the orders to be shipped in multiple time periods with a judgment result of yes to obtain the target wave and send them for picking.

13. An order group wave processing device, characterized in that: include: An acquisition module is used to determine the current time period in which the current time is located, and to obtain the first order to be shipped in the current time period and the cut-off time in the order to be shipped; A processing module, configured to process the first order-to-be-shipped related data through a Gaussian distribution model to obtain the second order-to-be-shipped related data for the next time period; By using a preset target model, the first data related to the order to be shipped and the second data related to the order to be shipped are processed to obtain the first shipping duration in the current time period, the second shipping duration in the next time period, and the overlap rate of item types in the current time period and the next time period; The wave grouping module is used to input the cut-off time, the first outbound delivery time, the second outbound delivery time, and the item type overlap rate into the wave grouping method determination model. In response to the model output result that the orders to be shipped in the current time period and the next time period are mergeable, the orders are grouped and sent for picking.

14. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 12.

15. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 12 is implemented.

Images