Replenishment task generation method, configuration device, electronic equipment and storage medium

Abstract

The invention provides a replenishment task generation method and device, electronic equipment and a computer readable storage medium, and relates to the technical field of warehouse logistics. The replenishment task generation method comprises the steps of predicting a to-be-delivered quantity of articles in a target time period based on types and corresponding quantities of the articles in a historical delivery order; calculating a replenishment state value of the goods in the goods picking position, and determining a first goods in a replenishment state and a second group of goods in a near replenishment state based on a replenishment grading rule; determining to-be-associated articles entering a replenishment state in the target time period in the second group of articles based on the to-be-delivered number of the articles; identifying associated articles having an association relationship with the first group of articles in the to-be-associated articles based on an associated article identification model; and associating the first group of articles with the associated articles to generate a replenishment task. According to the technical scheme, the delivery efficiency of the articles can be improved, and the probability of order waiting when the articles are out of stock is reduced.

Description

Background Art

[0002] With the rapid development of e-commerce and modern logistics, online shopping has gradually become mainstream. After the order system receives an online shopping order, it will send the order-related item information to the warehouse management system for processing to facilitate the subsequent picking process.

[0003] However, current warehouse management systems have problems such as fixed replenishment intervals, time stagnation in the replenishment process, and order waiting when items are out of stock.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of the present disclosure is to provide a replenishment task generation method, a replenishment task generation device, an electronic device and a storage medium, which can at least to some extent improve the problem of low item replenishment efficiency in the related art.

[0006] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0007] According to one aspect of the present disclosure, a replenishment task generation method is provided, comprising: predicting the number of items to be shipped out in a target time period based on the types and corresponding quantities of items in historical shipping orders; calculating the replenishment status values ​​of the items in the picking position, and determining a first item in a replenishment state and a second group of items in an approaching replenishment state based on a replenishment grading rule; determining to-be-associated items in the second group of items that enter the replenishment state in the target time period based on the number of items to be shipped out; identifying associated items in the to-be-associated items that have an associated relationship with the first group of items based on an associated item identification model; and associating the first group of items with the associated items to generate a replenishment task.

[0008] In one embodiment, based on the types and corresponding quantities of items in historical outbound orders, the number of items to be shipped out in a target time period is predicted, including: statistically dividing the item outbound data of multiple time periods based on the types and corresponding quantities of the items; extracting a predetermined number of single-item outbound data from the item outbound data of a single time period; calculating the expected outbound data of a single time period based on the single-item outbound data; and predicting the corresponding number of items to be shipped out in the target time period based on the expected outbound data.

[0009] In one embodiment, the expected delivery data for a single time period is calculated based on the single product delivery data, including: for the single time period, determining the interval between the collection time of the single product delivery data and the current time; configuring the weight coefficient of each of the single product delivery data based on the negative correlation with the interval time; calculating a weighted average value based on the weight coefficient and each of the single product delivery data, and determining the weighted average value as the expected delivery data for the single time period.

[0010] In one embodiment, predicting the number of items to be shipped in the target time period based on the expected shipping data includes: constructing a Gaussian distribution of items to be shipped based on the number of items to be shipped in the target time period, so as to predict the number of items to be shipped in the target time period based on the Gaussian distribution.

[0011] In one embodiment, the replenishment status values ​​of the items at the picking location are calculated, and a first group of items at the replenishment status and a second group of items at a state close to replenishment are determined based on a replenishment grading rule, including: calculating the replenishment status values ​​of the items based on the number of items at the picking location; acquiring a first replenishment threshold and a second replenishment threshold configured based on the replenishment grading rule, wherein the first replenishment threshold is greater than the second replenishment threshold; determining the items whose replenishment status values ​​are greater than or equal to the first replenishment threshold as the first group of items at the replenishment status; and determining the items whose replenishment status values ​​are less than the first replenishment threshold and greater than the second replenishment threshold as the second group of items at the state close to replenishment.

[0012] In one embodiment, determining the items to be associated in the second group of items that enter the replenishment state in the target time period based on the number of the items to be shipped out includes: for a single type of item in the second group of items, obtaining a replenishment state update value of the single type of item based on the number of the single type of items located at the picking location and the number to be shipped out; if the replenishment state update value is greater than or equal to the first replenishment threshold, determining the single type of item as the item to be associated.

[0013] In one embodiment, before identifying the associated items in the to-be-associated items that have an associated relationship with the first group of items based on the associated item identification model, the method further includes:

[0014] An item association matrix is ​​constructed based on historical outbound orders; a model training operation is performed based on the Markov chain and the association matrix to obtain the associated item recognition model.

[0015] In one embodiment, the construction of an item association matrix based on historical outbound orders includes: detecting item association relationships based on the historical outbound orders; constructing an item sequence of multiple items based on the association relationships; respectively calculating the outbound rates of the multiple items; and constructing the item association matrix based on the item sequences and the corresponding outbound rates.

[0016] In one embodiment, constructing the item association matrix based on the item sequence and the corresponding out-of-stock rates includes: constructing a plurality of randomly initialized collaborative filtering matrices based on the item sequence; generating a priori probability matrix based on exponential processing of the out-of-stock rates of the plurality of items; and determining the plurality of collaborative filtering matrices and the priori probability matrix as the item association matrix.

[0017] In one embodiment, the performing of a model training operation based on the Markov chain and the association matrix to obtain the associated item recognition model includes: configuring a neural network for each collaborative filtering matrix, respectively, and initializing network parameters of each neural network; extracting a predicted subsequence of the item sequence with a maximum association probability based on the Markov chain, the collaborative filtering matrix, and the prior probability matrix; and performing a model training operation based on the relationship between the predicted subsequence and the actual subsequence to iteratively update the network parameters until it is detected that the extracted predicted subsequence satisfies a loss function, so as to obtain the associated item recognition model.

[0018] In one embodiment, extracting the predicted subsequence of the item sequence with the maximum association probability based on the Markov chain, the collaborative filtering matrix and the prior probability matrix includes: calculating the transition probability of associated items in the item sequence based on the Markov chain and the prior probability matrix; and inputting the transition probability and the collaborative filtering matrix into the neural network to extract the predicted subsequence with the maximum association probability from the item sequence.

[0019] In one embodiment, the method further includes: constructing the loss function based on the Markov chain.

[0020] In one embodiment, the identifying, based on the associated item recognition model, associated items in the items to be associated that have an associated relationship with the first group of items comprises: constructing a sequence to be identified based on the first group of items and the items to be associated; and inputting the sequence to be identified into the associated item recognition model to output a target associated sequence including the first group of items and the associated items based on a maximum association probability.

[0021] According to another aspect of the present disclosure, a replenishment task generation device is provided, comprising: a prediction module, configured to predict the number of items to be shipped out in a target period based on the types and corresponding quantities of items in historical shipping orders; a calculation module, configured to calculate the replenishment status values ​​of the items in the picking position, and determine a first item in the replenishment status and a second group of items in an approaching replenishment status based on a replenishment grading rule; a determination module, configured to determine, based on the number of items to be shipped out, items to be associated in the second group of items that enter the replenishment status in the target period; an identification module, configured to identify associated items in the items to be associated that have an associated relationship with the first group of items based on an associated item identification model; and a generation module, configured to associate the first group of items with the associated items to generate a replenishment task.

[0022] According to another aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any one of the above-mentioned replenishment task generation methods by executing the executable instructions.

[0023] According to another aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, any one of the above replenishment task generation methods is implemented.

[0024] The replenishment task generation scheme provided by the embodiment of the present disclosure obtains historical outbound order data to predict the number of items to be shipped in a specified time period based on the item types and item quantities recorded in the historical outbound order data, so that it is possible to determine which items need to be replenished based on the number of items to be shipped. In addition, by calculating the replenishment status values ​​of the items in the picking position, the replenishment status of the items in the picking position is graded, and a first group of items in the replenishment state and a second group of items in the state of approaching replenishment are obtained. Combined with the predicted number of items to be shipped, the items in the second group of items that will become replenishment status in the specified time period can be determined as the items to be associated. The associated item identification model is further used to identify the associated items in the items to be associated that have a high degree of association with the current first group of items, that is, the associated items that can often form the same order, so as to generate replenishment tasks together. Therefore, based on the prediction of the replenishment operation of the associated items, the replenishment operation of more items can be executed based on one replenishment task quantity list, which is conducive to reducing the generation and issuance of the number of replenishment tasks, improving the efficiency of item shipping, and reducing the probability of order waiting when the items are out of stock.

[0025] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0027] Figure 1 A schematic diagram showing the structure of a replenishment task generation system in an embodiment of the present disclosure;

[0028] Figure 2 A schematic diagram showing a method for generating a replenishment task in an embodiment of the present disclosure is shown;

[0029] Figure 3 A schematic diagram showing another replenishment task generation method in an embodiment of the present disclosure;

[0030] Figure 4 A schematic diagram showing another replenishment task generation method in an embodiment of the present disclosure;

[0031] Figure 5 A schematic diagram showing another replenishment task generation method in an embodiment of the present disclosure;

[0032] Figure 6 A schematic diagram showing another replenishment task generation method in an embodiment of the present disclosure;

[0033] Figure 7 A schematic diagram showing another replenishment task generation method in an embodiment of the present disclosure;

[0034] Figure 8 A schematic diagram showing a replenishment task generating device in an embodiment of the present disclosure is shown;

[0035] Fig. 9 A schematic diagram showing a replenishment task generating device in an embodiment of the present disclosure is shown;

[0036] Fig.10 A schematic diagram of an electronic device in an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0037] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the disclosure will be more comprehensive and complete and to fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0038] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0039] The solution provided by the present application obtains historical outbound order data to predict the number of items to be shipped out in a specified period of time based on the item types and item quantities recorded in the historical outbound order data, so that it is possible to determine which items need to be replenished based on the number of items to be shipped out. In addition, by calculating the replenishment status values ​​of the items in the picking position, the replenishment status of the items in the picking position is graded, and a first group of items in a replenishment state and a second group of items in a state close to replenishment are obtained. Combined with the predicted number of items to be shipped out, it is possible to determine the items in the second group that will become replenishment status in a specified period of time as items to be associated, and further use the associated item recognition model to identify the associated items in the items to be associated that have a high degree of association with the current first group of items, that is, the associated items that can often form the same order, so as to generate replenishment tasks together. Therefore, based on the prediction of the replenishment operation of the associated items, it is possible to execute replenishment operations for more items based on one replenishment task quantity list, which is conducive to reducing the generation and issuance of replenishment tasks, improving the efficiency of item shipping, and reducing the probability of order waiting when items are out of stock.

[0040] To facilitate understanding, several terms involved in this application are first explained below.

[0041] In the warehouse, there are usually two types of storage locations: storage locations and picking locations. Storage locations are only responsible for storing items, while picking locations are used to locate and pick items during outbound operations.

[0042] In the process of warehouse positioning and picking, the picking location is often out of stock, and replenishment tasks need to be generated regularly or urgently to transfer items from the storage location to the picking location.

[0043] Figure 1 A structural schematic diagram of a replenishment task generation system in an embodiment of the present disclosure is shown, including multiple terminals 120 and a server cluster 140.

[0044] The terminal 120 may be a mobile terminal such as a mobile phone, a game console, a tablet computer, an e-book reader, smart glasses, an MP4 (Moving Picture Experts Group Audio Layer IV) player, a smart home device, an AR (Augmented Reality) device, a VR (Virtual Reality) device, or the terminal 120 may also be a personal computer (PC), such as a laptop computer and a desktop computer.

[0045] The terminal 120 may be installed with an application for generating replenishment tasks.

[0046] The terminal 120 is connected to the server cluster 140 via a communication network. Optionally, the communication network is a wired network or a wireless network.

[0047] The server cluster 140 is a server, or is composed of several servers, or is a virtualization platform, or is a cloud computing service center. The server cluster 140 is used to provide background services for providing replenishment task generation applications. Optionally, the server cluster 140 undertakes the main computing work, and the terminal 120 undertakes the secondary computing work; or, the server cluster 140 undertakes the secondary computing work, and the terminal 120 undertakes the main computing work; or, the terminal 120 and the server cluster 140 adopt a distributed computing architecture for collaborative computing.

[0048] In some optional embodiments, the server cluster 140 is used to store replenishment task generation models, etc.

[0049] Optionally, the client of the application installed in different terminals 120 is the same, or the client of the application installed on the two terminals 120 is the client of the same type of application of different control system platforms. Based on the different terminal platforms, the specific form of the client of the application may also be different, for example, the application client may be a mobile client, a PC client or a World Wide Web (Web) client.

[0050] Those skilled in the art will appreciate that the number of the above-mentioned terminals 120 may be more or less. For example, the above-mentioned terminal may be only one, or the above-mentioned terminals may be dozens or hundreds, or more. The embodiment of the present application does not limit the number and device type of the terminals.

[0051] Optionally, the system may also include a management device ( Figure 1(not shown), the management device is connected to the server cluster 140 via a communication network. Optionally, the communication network is a wired network or a wireless network.

[0052] Optionally, the above-mentioned wireless network or wired network uses standard communication technology and / or protocol. The network is usually the Internet, but it can also be any network, including but not limited to a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a dedicated network or any combination of a virtual private network). In some embodiments, the data exchanged through the network is represented by technologies and / or formats including Hyper Text Mark-up Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec) can also be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.

[0053] Below, the replenishment task generation method, mobile terminal, writing device and processing method in this example implementation will be described in more detail with reference to the accompanying drawings and embodiments.

[0054] like Figure 2 As shown, a replenishment task generation method according to an embodiment of the present disclosure includes:

[0055] Step S202: predict the quantity of items to be shipped out during a target period based on the types and corresponding quantities of items in historical shipping orders.

[0056] Among them, the number of historical outbound orders can be the number of historical outbound orders for multiple items within a specified number of days from the current moment. Furthermore, multiple time periods are divided into days. For the target time period, the number of orders for each target time period within the specified number of days can be used to predict the number of items to be shipped in the target time period.

[0057] Step S204, calculating the replenishment status values ​​of the items at the picking location, and determining a first group of items in a replenishment status and a second group of items in an upcoming replenishment status based on a replenishment grading rule.

[0058] Among them, the replenishment classification rules are rules for detecting whether a certain type of items is in need of replenishment, is about to be replenished, or does not need replenishment. Through the classification rules, the current items can be divided into a state of needing replenishment, a state of being about to be replenished, or a state of not needing replenishment. The first replenishment state is a state of needing replenishment, and the second replenishment state is a state of being about to be replenished.

[0059] Step S206: determining the to-be-associated items in the second group of items that enter the replenishment state during the target period based on the to-be-shipped quantity of the items.

[0060] Step S208: identifying associated items that have an associated relationship with the first group of items among the items to be associated based on the associated item identification model.

[0061] Among them, by training the associated item recognition model, based on the associated item recognition model, the items to be associated in the second group of items that have an associated relationship with the first group of items, that is, the associated items, can be identified. The associated items can be understood as items that have an associated relationship in sale or use.

[0062] Step S210: Associating the first group of items with the associated items to generate a replenishment task.

[0063] The generating of the replenishment task based on the first group of items and the associated items specifically includes: completing the replenishment operation of multiple items including the first group of items and the associated items on one replenishment task.

[0064] In this embodiment, by acquiring historical outbound order data, the number of items to be shipped in a specified period of time is predicted based on the item types and item quantities recorded in the historical outbound order data, so that it is possible to determine which items need to be replenished based on the number of items to be shipped. In addition, by calculating the replenishment status values ​​of the items in the picking position, the replenishment status of the items in the picking position is graded, and a first group of items in a replenishment state and a second group of items in a state close to replenishment are obtained. Combined with the predicted number of items to be shipped, it is possible to determine the items in the second group that will become replenishment status in a specified period of time as the items to be associated, and further use the associated item recognition model to identify the associated items in the items to be associated that have a high degree of association with the current first group of items, that is, the associated items that can often form the same order, so as to generate replenishment tasks together. Therefore, based on the prediction of the replenishment operation of the associated items, it is possible to execute replenishment operations for more items based on one replenishment task quantity list, which is conducive to reducing the generation and issuance of replenishment tasks, improving the efficiency of item shipping, and reducing the probability of order waiting when the items are out of stock.

[0065] like Figure 3 As shown, in one embodiment, based on the types and corresponding quantities of items in historical outbound orders, predicting the quantity of items to be shipped out in a target period includes:

[0066] Step S302: Count the item outbound data of multiple time periods based on the types of items and the corresponding quantities.

[0067] Among them, a day is divided into multiple time periods, and the number of items shipped out in each time period is constructed. Combined with storage attributes, operation attributes, etc., the number of items to be shipped out in the same time period of the next day can be predicted.

[0068] Exemplarily, a certain historical time period (first time range) may be the number of historical outbound orders in the last month or three months; the time interval (second time range) may be the interval time of waves in the warehouse, such as 15 minutes. The number of outbound orders in one day is divided into 96 time periods according to the above 15 minutes. For simple expression, we mark the above 96 time periods with time period 1 (0:00-0:15), time period 2 (0:15-0:30), ... time period 96 (23:45-24:00).

[0069] Furthermore, the sample characteristic data may include statistical characteristic data, wherein the statistical characteristic 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.

[0070] Step S304: extract a predetermined number of single item outbound data from the item outbound data in a single period.

[0071] In one embodiment, extracting a predetermined amount of single item outbound data from the item outbound data in a single period includes:

[0072] A predetermined number of single product outbound data are extracted based on a random operation; or a predetermined number of single product outbound data are extracted from the item outbound data in a single period based on a preset ratio.

[0073] For example, assuming that the historical outbound orders are historical orders within the past 30 days, the outbound quantities of items within 10 days are randomly selected from the outbound data of items within the target period within the past 30 days.

[0074] Step S306, calculating the expected delivery data for a single period based on the delivery data of the single product.

[0075] Among them, the expected shipment data is obtained by averaging the shipment data of a predetermined number of items.

[0076] Step S308: predicting the quantity to be shipped in the corresponding target period based on the expected shipping data.

[0077] In this embodiment, the expected outbound data of a single period is calculated to obtain the quantity to be outbound based on the expected outbound quantity, so that it is possible to determine which items are about to enter the second replenishment state from the first replenishment state based on the quantity to be outbound.

[0078] In one embodiment, the expected delivery data for a single period is calculated based on the delivery data of a single product, including:

[0079] For a single time period, determine the interval between the collection time of the single product outbound data and the current time.

[0080] The weight coefficient of each product's outbound data is configured based on the negative correlation with the interval length.

[0081] Among them, the weight coefficient can be determined based on the inverse of the time interval from the current moment, or arranged in descending order according to the time interval from the current moment, and each single product shipment data is assigned a different weight coefficient according to exponential decay or linear decay.

[0082] A weighted average is calculated based on the weight coefficient and the shipment data of each single product, and the weighted average is determined as the expected shipment data for a single period.

[0083] like Figure 4 As shown, according to an embodiment of the present disclosure, a method for determining the quantity to be shipped out includes:

[0084] Step 402, obtaining the product shipment data for the same period of multiple days.

[0085] Step 404: randomly select N single product shipment data from the multiple single product shipment data or select N single product shipment data at a preset ratio.

[0086] Step 406 , sort the data in descending order from new to old, and configure different weights for the N single product delivery data.

[0087] Step 408, obtaining the expected delivery data for the period according to different weights and corresponding single product delivery data.

[0088] Step 410, taking the expected outbound data as the mean of the Gaussian distribution, constructing a Gaussian distribution model of the outbound quantity in the period, and obtaining the quantity to be outbound based on the random output data of the model.

[0089] In one embodiment, predicting the number of items to be shipped in the corresponding target period based on the expected shipping data includes: constructing a Gaussian distribution of items to be shipped based on the number of items to be shipped in the target period, so as to predict the number of items to be shipped in the target period based on the Gaussian distribution.

[0090] In this embodiment, the quantity to be shipped out in the next time period is determined according to the Gaussian distribution model to ensure the reliability of the prediction of the quantity to be shipped out.

[0091] In one embodiment, the replenishment status values ​​of the items at the picking location are calculated, and a first group of items at the replenishment status and a second group of items at a state close to replenishment are determined based on a replenishment grading rule, including: calculating the replenishment status values ​​of the items based on the number of items at the picking location; obtaining a first replenishment threshold and a second replenishment threshold configured based on the replenishment grading rule, wherein the first replenishment threshold is greater than the second replenishment threshold; determining the items having a replenishment status value greater than or equal to the first replenishment threshold as the first group of items at the replenishment status; and determining the items having a replenishment status value less than the first replenishment threshold and greater than the second replenishment threshold as the second group of items at a state close to replenishment.

[0092] Among them, when calculating the replenishment status value, the number of items in the current picking position is used as the calculation object. The number of items in the picking position and the replenishment threshold set by the system are mapped from continuous values ​​to enumeration values ​​to determine the replenishment status value of each item, which can be replenishment, upcoming replenishment, or no replenishment required.

[0093] According to the number of items in the picking position at the current moment, the replenishment status value of each item is determined, which can be a value between 0 and 1. According to the predicted outbound quantity in the next time period obtained by 1, the replenishment status value of each item is obtained.

[0094] Exemplarily, specifically, the replenishment status classification can be divided into:

[0095] Replenishment required state (first replenishment state): that is, the out-of-stock level reaches the threshold set by the system, marked as Status-1.

[0096] Near replenishment status (second replenishment status): the out-of-stock level reaches 80%-100% of the threshold set by the system, marked as Status-2.

[0097] No need to replenish: The out-of-stock level is below 80% of the system-set threshold, marked as Status-3.

[0098] In one embodiment, determining the associated items in the second group of items that enter the replenishment state in the target period based on the quantity of items to be shipped out includes:

[0099] For a single type of item in the second group of items, a replenishment status update value of the single type of item is obtained based on the quantity of the single type of item located at the picking location and the quantity to be shipped out.

[0100] If the replenishment status update value is greater than or equal to the first replenishment threshold, the single-category item is determined as an item to be associated.

[0101] In this embodiment, for the second group of items that are about to be replenished, the predicted number of items to be shipped out can be used to detect whether the items in the second group will switch from the state to be replenished to the state to be replenished in the next target period. For items that will switch from the state to be replenished to the state to be replenished, they can be determined as items to be associated, so as to further obtain the associated replenishment tasks based on the association operation.

[0102] In one embodiment, before identifying the associated items that have an associated relationship with the first group of items among the items to be associated based on the associated item identification model, the method further includes:

[0103] Build an item association matrix based on historical outbound orders.

[0104] A model training operation is performed based on the Markov chain and the association matrix to obtain an associated item recognition model.

[0105] Among them, the Markov chain is a random process of transition from one state to another in the state space. The Markov chain is a random process of transition from one state to another in the state space. Based on the principle that the probability of state transition at a certain moment of the Markov chain only depends on its previous state, the dependency of state transition is applied to the dependency solving process of associativity.

[0106] In this embodiment, the conditional probability of multiple items appearing at the same time is modeled by combining historical outbound orders and Markov chains, and the maximum association probability is determined based on the maximum probability solution. Based on the above processing procedure, a model training operation is performed to obtain an associated item recognition model to ensure the accuracy of associated item recognition.

[0107] In one embodiment, an item association matrix is ​​constructed based on historical outbound orders, including: detecting item association relationships based on historical outbound orders; constructing item sequences of multiple items based on the association relationships; calculating outbound rates of multiple items respectively; and constructing an item association matrix based on the item sequences and corresponding outbound rates.

[0108] Among them, the probability of each item leaving the warehouse is calculated by the ratio of the number of items leaving the warehouse in a certain period of history (which can be 1 month, 3 months, etc.) to the total number of items leaving the warehouse, that is, the prior probability P(S i ), as shown in formula (1).

[0109]

[0110] There are many (tens of thousands) types of items in the warehouse, and the prior probability of each item is very small. In order to make the probability of each item more distinguishable. as the final prior probability.

[0111] In one embodiment, an item association matrix is ​​constructed based on an item sequence and a corresponding out-of-stock rate, including: constructing a plurality of randomly initialized collaborative filtering matrices based on the item sequence; generating a priori probability matrix based on exponential processing of out-of-stock rates of a plurality of items; and determining the plurality of collaborative filtering matrices and the priori probability matrix as an item association matrix.

[0112] In one embodiment, a model training operation is performed based on the Markov chain and the association matrix to obtain an associated item recognition model, including:

[0113] A neural network is configured for each collaborative filtering matrix respectively, and the network parameters of each neural network are initialized.

[0114] like Figure 5 As shown, according to the collaborative filtering matrix obtained by the above random initialization, the neural network to be trained is configured accordingly, and the network parameters of the neural network are randomly initialized.

[0115] After initializing the network parameters of each neural network, the predicted subsequence of the item sequence with the maximum association probability is extracted based on the Markov chain, collaborative filtering matrix and prior probability matrix.

[0116] After obtaining the predicted subsequence, a model training operation is performed based on the relationship between the predicted subsequence and the actual subsequence to iteratively update the network parameters until it is detected that the extracted predicted subsequence satisfies the loss function to obtain an associated item recognition model.

[0117] like Figure 6 As shown, according to an embodiment of the present disclosure, a method for generating an associated item identification model includes:

[0118] Step S602, obtaining the prior probability of each item in the warehouse.

[0119] Step S604, constructing an item association matrix based on the prior probability and the item sequence, assigning a neural network to each item association matrix, and initializing the network parameters of the neural network.

[0120] Step S606, using the association matrix as the input of the neural network, and outputting the predicted item sequence with the maximum association probability.

[0121] Step S608, calculating the loss function based on the output item sequence and the actual physical sequence, reversely iterating the association matrix based on the calculation result, and training the neural network.

[0122] Step S610: It is detected that the extracted item sequence satisfies the maximization of the loss function, and a trained associated item recognition model is obtained.

[0123] In one embodiment, extracting a predicted subsequence of an item sequence with a maximum association probability based on a Markov chain, a collaborative filtering matrix, and a prior probability matrix includes: calculating transition probabilities of associated items in the item sequence based on the Markov chain and the prior probability matrix; and inputting the transition probabilities and the collaborative filtering matrix into a neural network to extract a predicted subsequence with a maximum association probability from the item sequence.

[0124] In one embodiment, the method further includes: constructing a loss function based on a Markov chain.

[0125] In one embodiment, identifying associated items in the to-be-associated items that have an associated relationship with the first group of items based on an associated item identification model includes: constructing a to-be-associated sequence based on the first group of items and the to-be-associated items; and inputting the to-be-associated sequence into the associated item identification model to output a target associated sequence including the first group of items and the associated items based on a maximum association probability.

[0126] like Figure 7 As shown, according to another embodiment of the present disclosure, a replenishment task generation method includes:

[0127] Step S702, predicting the quantity of different types of items to be shipped out.

[0128] Among them, the number of outbound orders for multiple items within a certain historical period (first time range) is obtained, and each day of the above time is divided according to a certain time interval (second time range) to obtain classification results for multiple time periods.

[0129] A certain historical time period (the first time range) can be the number of historical outbound orders in the last month or three months. The time interval (the second time range) can be the interval time of waves in the warehouse, such as 15 minutes. The number of outbound orders in one day is divided into 96 time periods according to the above 15 minutes. For simple expression, the above 96 time periods are marked with time period 1 (0:00-0:15), time period 2 (0:15-0:30), ... time period 96 (23:45-24:00).

[0130] Furthermore, the sample characteristic data may include statistical characteristic data, which includes one or more of the following: mean, first quantile value, median, standard deviation and maximum value of historical data, wherein the first quantile value can be flexibly set, such as 80% quantile value or 95% quantile value.

[0131] Among them, N items are randomly selected from the classification results, or the data of the number of items shipped out of the warehouse in each time period are selected according to a certain probability (such as 50%, 100%). For the sake of simplicity, the above data is called the set data of the number of items shipped out of the warehouse.

[0132] First, the average outbound quantity of item s is calculated, preferably the expected outbound data N is calculated for item s. avg , as shown in formula (2).

[0133]

[0134] Among them, considering the weighted average of the sales volume in the past T1 days, {w t ≥0} is a decreasing sequence, t is the current time point, w t is the reciprocal of the distance from the current time, that is, the closer to the current time w t The larger the value.

[0135] Specifically, another implementation form is: the above-obtained item outbound quantity set can be arranged in descending order of time, 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 item outbound quantity of each item in each time period. The specific weight exponential decay formula is shown in formula (3).

[0136]

[0137]

[0138] Finally, the expected outbound quantity N for the obtained time period is avg As the expected value μ of the Gaussian distribution, the variance is σ 2 , construct N(μ,σ 2 ) is a Gaussian distribution outbound quantity. Randomly select values ​​from the above Gaussian distribution to obtain the number of goods to be outbound in the specified time period N ot .

[0139] Step S704, by calculating the replenishment status value of each type of item, the different types of items are divided into a replenishment status, a near replenishment status, and a no-replenishment status.

[0140] Build a replenishment status hierarchy for warehouse items.

[0141] Step 706, obtaining a second group of items that are currently in a state of near replenishment, and obtaining a set of items to be associated that are in a state of replenishment at the next moment according to the predicted quantity to be shipped out at the next moment.

[0142] Step S708, obtaining a first group of items currently in a replenishment state, and identifying associated items with the maximum association probability in the set of items to be associated based on the associated item identification model.

[0143] Step S710: Generate and issue a replenishment task based on the first group of items and related items.

[0144] According to the Markov chain, the correlation matrix and multiple neural networks, the items with high correlation that make up the order are obtained.

[0145] The conditional probability of multiple items appearing at the same time is modeled in the form of Markov transition probability.

[0146] For example, in the warehouse outbound orders, due to the consumer's purchasing habits and the relationship between items, each item will depend on other items. For example, when buying a mobile phone or tablet, you often need to buy accessories such as mobile phone cases and mobile phone films at the same time, and consumers usually buy paper towel products at the same time when they buy milk powder. Therefore, the items that appear in the same order actually include the relationship between SKUs. That is, the appearance of item A is more likely to be associated with the appearance of item B.

[0147] Specifically, suppose S = [s1, s2, ..., s n ] is the sorting sequence corresponding to the item (the meaning of sorting is s i The most relevant item is s i+1 ), the corresponding association probability of the sequence is shown in formula (4).

[0148]

[0149] Among them, p(s i+1 |s i ) is the probability that item si and item si+1 appear at the same time.

[0150] Then for a known sorted sequence S = [s1, s2, ..., s n ], in the given set of items {x1, x2, x3, ..., xn}, select the item sequence that maximizes the above probability Ps, as shown in formula (5).

[0151]

[0152] In order to simplify the complexity of the above model, we use the fact that the ownership of some items has a certain Markov chain dependency, such as mobile phone->screen protector, etc. Dynamic programming can be used to solve the above problem.

[0153] because:

[0154]

[0155] Then the sequence that maximizes Ps must contain

[0156]

[0157] The subsequence with the highest probability, P sn-1is the probability of the subsequence.

[0158] The above Ps can be solved using the Viterbi algorithm.

[0159] Finally, the most likely subsequent sequence of the currently known item sequence S = [s1, s2, ..., sn] can be obtained. For example, the current items are sku-1, sku-2, and the given item combination is {sku-11, sku-22, sku-99, sku-88}. After the above solution, the most likely item sequence is [sku-1, sku-2, sku-99, sku-22].

[0160] The probability of each item leaving the warehouse (i.e., prior probability) is calculated by the ratio of the number of items leaving the warehouse in a certain period of history (which can be 1 month, 3 months, etc.) to the total number of items leaving the warehouse, as shown in formula (1).

[0161]

[0162] There are many (tens of thousands) types of items in the warehouse, and the prior probability of each item is very small. In order to make the probability of each item more distinguishable, As the final prior probability, we get Table 1.

[0163] Table 1

[0164] thing sku-1 sku-2 sku-3 sku-4 … sku-M <![CDATA[P(S i )]]> 0.38 0.23 0.43 0.13 … 0.78

[0165] In addition, the randomly initialized collaborative filtering matrix is ​​shown in Table 2.

[0166] Table 2

[0167] <![CDATA[S j / S i ]]> sku-1 sku-2 sku-3 sku-4 sku-1 1 0.4 0.6 0.5 sku-2 0.8 1 0.55 0.35 sku-3 0.57 0.5 1 0.75 sku-4 0.1 0.6 0.45 1

[0168] According to the association matrix obtained by the random initialization above, the parameters of N neural training networks are randomly initialized (N is a finite number, which can be an item to be replenished, N items with high association) and the model structure is as follows: Figure 5 shown.

[0169] Specifically, in the case of p(S j |S i ) Before learning, you can use the association between items, such as the collaborative filtering matrix between warehouse items, to do probability normalization as the initialization matrix, which will help you learn the transfer matrix faster.

[0170] Use multiple neural network task loss functions, use sigmoid / Relu functions as activation functions, and use stochastic gradient descent method (specifically Adam, SGD, etc.) to optimize the neural network.

[0171] Select a set of items from multiple historical orders that have been combined to obtain an item sequence [S1, S2, ..., Sn], and use the following maximum likelihood function as the loss function, as shown in formula (8).

[0172]

[0173] The neural network is trained to obtain network parameter values ​​of N neural networks.

[0174] For example, according to the above prior probability matrix, sku-1: 0.38, sku-2: 0.23, sku-3: 0.43, sku-4: 0.13; combined with the initial correlation matrix, p(sku-2|sku-1) = 0.4, p(sku-3|sku-2) = 0.55, p(sku-4|sku-3) = 0.75, p(sku-3|sku-1) = 0.6. The maximum probability subsequence can be calculated:

[0175] sku-1->sku-2->sku-3:0.38*(0.4*0.23)*(0.55*0.43)=0.00826804

[0176] sku-1->sku-3->sku-4:0.38*(0.6*0.43)*(0.75*0.13)=0.0095589

[0177] Since the subsequence with the highest probability is: sku-1->sku-3->sku-4, the result during prediction is: the sku sequence of sku-1->sku-3->sku-4. Compare it with the real sku sequence (that is, the sequence of items that make up the order in history), reversely iterate the conditional probabilities between the skus, that is, update the collaborative filtering matrix (association matrix) of the sku, and continuously iterate to obtain the final association matrix and associated item recognition model.

[0178] The collaborative filtering matrix after neural network training is shown in Table 3:

[0179] Table 3

[0180] <![CDATA[S j / S i ]]> sku-1 sku-2 sku-3 sku-4 sku-1 1 0.46 0.91 0.05 sku-2 0.65 1 0.75 0.56 sku-3 0.41 0.76 1 0.36 sku-4 0.04 0.67 0.45 1

[0181] Get one of the items (skui) in the replenishment state (i.e., Status-1), and get the set of items that are currently in the replenishment state (i.e., Status-2), set S = [s1, s2, ..., s n ].

[0182] According to the obtained quantity N to be shipped ot , calculate S nThe set of items that are converted from Status-2 to Status-1 at the next moment is denoted by S n ′.

[0183] Furthermore, according to the associated item recognition model, skui is used as the first element of the sequence, and the n ′ forms the sequence with the highest probability (N elements), recorded as S, and the N items in S are combined into a replenishment task, and the replenishment is issued.

[0184] like Figure 8 As shown, according to an embodiment of the present disclosure, a replenishment task generating device 800 includes:

[0185] Setting module 802, setting basic configuration information;

[0186] Data acquisition module 804: used to collect historical data of the warehouse management system.

[0187] Replenishment task construction module 806: sample data processing, picking status classification, and association matrix construction.

[0188] Replenishment module 808: Calculate various indicators and determine the replenishment method based on the constraints.

[0189] It should be noted that the above figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0190] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0191] Refer to the following Fig. 9 The replenishment task generating device 900 according to this embodiment of the present invention is described. Fig. 9 The replenishment task generating device 900 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.

[0192] The replenishment task generation device 900 is expressed in the form of a hardware module. The components of the replenishment task generation device 900 may include but are not limited to: a prediction module 902, which is used to predict the number of items to be shipped out in the target period based on the types and corresponding quantities of items in the historical shipping orders; a calculation module 904, which calculates the replenishment status value of the items in the picking position, and determines the first item in the replenishment state and the second group of items in the near replenishment state based on the replenishment classification rule; a determination module 906, which is used to determine the items to be associated in the second group of items that enter the replenishment state in the target period based on the number of items to be shipped out; an identification module 908, which is used to identify the associated items in the items to be associated that have an associated relationship with the first group of items based on the associated item identification model; a generation module 910, which is used to associate the first group of items with the associated items to generate a replenishment task.

[0193] Refer to the following Fig.10 The electronic device 1000 according to this embodiment of the present invention will be described. Fig.10 The electronic device 1000 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.

[0194] like Fig.10 As shown, the electronic device 1000 is in the form of a general computing device. The components of the electronic device 1000 may include but are not limited to: the at least one processing unit 1010, the at least one storage unit 1020, and a bus 1030 connecting different system components (including the storage unit 1020 and the processing unit 1010).

[0195] The storage unit stores program codes, which can be executed by the processing unit 1010, so that the processing unit 1010 performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 1010 can perform the following steps: Figure 2 Steps S202 to S208 shown in , and other steps defined in the replenishment task generation method of the present disclosure.

[0196] The storage unit 1020 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 10201 and / or a cache storage unit 10202 , and may further include a read-only storage unit (ROM) 10203 .

[0197] The storage unit 1020 may also include a program / utility 10204 having a set (at least one) of program modules 10205, such program modules 10205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0198] Bus 1030 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0199] The electronic device 1000 may also communicate with one or more external devices 1060 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device, and / or communicate with any device that enables the electronic device 1000 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1050. Furthermore, the electronic device 1000 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 1050. As shown, the network adapter 1050 communicates with other modules of the electronic device 1000 via a bus 1030. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0200] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0201] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present invention may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of the present specification.

[0202] The program product for implementing the above method according to an embodiment of the present invention may adopt a portable compact disk read-only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus, or device.

[0203] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0204] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0205] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0206] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0207] In addition, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0208] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0209] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A replenishment task generation method, characterized in that: include: Based on the types and corresponding quantities of items in historical outbound orders, predict the quantity of the items to be shipped out during the target period; Calculate the replenishment status values ​​of the items at the picking location, and determine a first group of items in the replenishment status and a second group of items in the adjacent replenishment status based on the replenishment classification rule; Determine, based on the quantity of the items to be shipped out, the items in the second group of items to be associated that enter the replenishment state during the target period; Identifying, based on an associated item identification model, associated items among the items to be associated that have an associated relationship with the first group of items; The first group of items is associated with the associated items to generate a replenishment task.

2. The replenishment task generation method according to claim 1, characterized in that: Based on the types and corresponding quantities of items in historical outbound orders, the quantity of the items to be shipped out during the target period is predicted, including: Counting the item outbound data of multiple time periods based on the types and corresponding quantities of the items; Extracting a predetermined number of single item outbound data from the item outbound data in a single period; Calculate the expected delivery data for a single period based on the delivery data of the single product; The quantity to be shipped out during the target time period is predicted based on the expected shipping data.

3. The replenishment task generation method according to claim 2, characterized in that: Calculating expected delivery data for a single period based on the delivery data of the single product includes: For the single time period, determine the interval between the collection time of the single product delivery data and the current time; Configuring a weight coefficient of each of the single product delivery data based on a negative correlation with the interval time; A weighted average value is calculated based on the weight coefficient and each of the single product delivery data, and the weighted average value is determined as the expected delivery data for the single period.

4. The replenishment task generation method according to claim 2, characterized in that: Predicting the quantity to be shipped in the target period corresponding to the expected shipping data includes: A Gaussian distribution of the goods to be shipped is constructed based on the quantity of goods to be shipped in the target time period, so as to predict the quantity of goods to be shipped in the target time period based on the Gaussian distribution.

5. The replenishment task generation method according to claim 1, characterized in that: Calculating the replenishment status values ​​of the items at the picking location, and determining a first item in the replenishment status and a second group of items in the adjacent replenishment status based on the replenishment classification rule, including: Calculate the replenishment status value of the item based on the number of items located at the picking location; Acquire a first replenishment threshold and a second replenishment threshold configured based on the replenishment grading rule, wherein the first replenishment threshold is greater than the second replenishment threshold; determining the items whose replenishment status values ​​are greater than or equal to the first replenishment threshold as the first group of items in the replenishment status; The items whose replenishment status values ​​are less than the first replenishment threshold and greater than the second replenishment threshold are determined as the second group of items in the impending replenishment status.

6. The replenishment task generation method according to claim 5, characterized in that: Determining the to-be-associated items in the second group of items that enter the replenishment state in the target time period based on the to-be-out quantity of the items includes: For a single type of item in the second group of items, obtaining a replenishment status update value of the single type of item based on the quantity of the single type of item located at the picking location and the quantity to be shipped out; If the replenishment status update value is greater than or equal to the first replenishment threshold, the single-category item is determined as the item to be associated.

7. The replenishment task generation method according to any one of claims 1 to 6, characterized in that: Before identifying the associated items in the to-be-associated items that have an associated relationship with the first group of items based on the associated item identification model, the method further includes: Build an item association matrix based on historical outbound orders; A model training operation is performed based on the Markov chain and the association matrix to obtain the associated item recognition model.

8. The replenishment task generation method according to claim 7, characterized in that: Build an item association matrix based on historical outbound orders, including: Detecting item association relationships based on the historical outbound orders; constructing an item sequence of multiple items based on the association relationship; Calculate the outbound rates of the multiple items respectively; The item association matrix is ​​constructed based on the item sequence and the corresponding out-of-stock rate.

9. The replenishment task generation method according to claim 8, characterized in that: Constructing the item association matrix based on the item sequence and the corresponding out-of-stock rate includes: Constructing a plurality of randomly initialized collaborative filtering matrices based on the item sequence; Based on the indexation processing of the outbound rates of the plurality of items, a priori probability matrix is ​​generated; The multiple collaborative filtering matrices and the prior probability matrix are determined as the item association matrix.

10. The replenishment task generation method according to claim 9, characterized in that: Performing a model training operation based on the Markov chain and the association matrix to obtain the associated item recognition model includes: respectively configuring a neural network for each of the collaborative filtering matrices, and initializing network parameters of each of the neural networks; Extracting a predicted subsequence of the item sequence having a maximum association probability based on the Markov chain, the collaborative filtering matrix, and the prior probability matrix; Based on the relationship between the predicted subsequence and the actual subsequence, a model training operation is performed to iteratively update the network parameters until it is detected that the extracted predicted subsequence satisfies the loss function, so as to obtain the associated item recognition model.

11. The replenishment task generation method according to claim 10, characterized in that: Extracting a predicted subsequence of the item sequence having a maximum association probability based on the Markov chain, the collaborative filtering matrix, and the prior probability matrix includes: Calculating the transition probability of associated items in the item sequence based on the Markov chain and the prior probability matrix; The transition probability and the collaborative filtering matrix are input into the neural network to extract the predicted subsequence with the maximum association probability from the item sequence.

12. The replenishment task generation method according to claim 10, characterized in that: Also includes: The loss function is constructed based on the Markov chain.

13. The method for generating replenishment tasks according to any one of claims 1 to 12, characterized in that: The step of identifying the associated items among the items to be associated that have an associated relationship with the first group of items based on the associated item identification model includes: constructing a sequence to be identified based on the first group of items and the items to be associated; The sequence to be identified is input into the associated item identification model to output a target associated sequence including the first group of items and the associated items based on the maximum association probability.

14. A replenishment task generation device, characterized in that: include: A prediction module, used to predict the quantity of the items to be shipped out in a target period based on the types and corresponding quantities of the items in the historical shipping orders; A calculation module, which calculates the replenishment status values ​​of the items at the picking location, and determines a first group of items in the replenishment status and a second group of items in the near replenishment status based on a replenishment classification rule; A determination module, configured to determine, based on the quantity of the items to be shipped out, the items in the second group of items to be associated that enter the replenishment state in the target time period; an identification module, configured to identify, based on an associated item identification model, associated items among the items to be associated that have an associated relationship with the first group of items; A generating module is used to associate the first group of items with the associated items to generate a replenishment task.

15. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; The processor is configured to execute the replenishment task generation method according to any one of claims 1 to 13 by executing the executable instructions.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the replenishment task generation method according to any one of claims 1 to 13 is implemented.

Images