Storage location distribution method and device, equipment, medium and product

Abstract

The invention discloses a storage location distribution method, device and equipment, a medium and a product. The method comprises the steps that a storage dynamic graph is acquired, and the storage dynamic graph comprises node feature vectors and edge feature vectors; the node feature vectors and the edge feature vectors of the storage dynamic graph are input into a pre-constructed graph neural network model, a storage location popularity score of each storage location is obtained, and the storage location popularity scores are used for representing the access probability of the storage locations; screening the initial storage location set based on the storage location popularity score of each storage location to obtain a candidate storage location set; and based on the candidate storage location set, solving a pre-constructed operational research multi-objective optimization model to obtain a storage location allocation result. According to the technical scheme of the invention, through graph neural network feature learning and operational research multi-objective optimization collaborative decision making, the accuracy and reliability of storage location allocation are improved, the time spent in the storage and taking process of goods is reduced, and the overall warehousing operation efficiency is improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of data processing, and particularly relate to a storage location allocation method, device, equipment, medium and product. BACKGROUND

[0002] In the field of warehouse management, traditional storage location allocation methods (such as genetic algorithm, rule engine, etc.) rely on pre-defined static warehouse models, cannot respond to real-time order fluctuations, and cannot effectively balance multi-objective conflicts, resulting in unreasonable storage location allocation, which makes goods need to spend more time in the storage and retrieval process, and further affects the overall warehouse operation efficiency. SUMMARY

[0003] Embodiments of the present application provide a storage location allocation method, device, equipment, medium and product to improve warehouse operation efficiency.

[0004] In a first aspect, embodiments of the present application provide a storage location allocation method, comprising:

[0005] obtaining a warehouse dynamic graph, wherein the warehouse dynamic graph comprises node feature vectors and edge feature vectors;

[0006] inputting the node feature vectors and the edge feature vectors of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent the probability of the storage location being accessed; and performing screening on an initial storage location set based on the storage location heat score of each storage location to obtain a candidate storage location set;

[0007] solving a pre-constructed operational research multi-objective optimization model based on the candidate storage location set to obtain a storage location allocation result.

[0008] In a second aspect, embodiments of the present application also provide a storage location allocation device, which comprises:

[0009] a warehouse dynamic graph obtaining module configured to obtain a warehouse dynamic graph, wherein the warehouse dynamic graph comprises node feature vectors and edge feature vectors;

[0010] a graph neural network feature learning module configured to input the node feature vectors and the edge feature vectors of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent the probability of the storage location being accessed; and perform screening on an initial storage location set based on the storage location heat score of each storage location to obtain a candidate storage location set;

[0011] an operational research multi-objective optimization module configured to solve a pre-constructed operational research multi-objective optimization model based on the candidate storage location set to obtain a storage location allocation result.

[0012] In a third aspect, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the storage location allocation method according to any of the embodiments of the present application when executing the program.

[0013] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and the program is executable on a processor to implement the storage location allocation method according to any of the embodiments of the present application.

[0014] In a fifth aspect, a computer program product is provided, which includes a computer program, and the computer program is executable on a processor to implement the storage location allocation method according to any of the embodiments of the present application.

[0015] In the embodiments of the present application, a warehouse dynamic graph is obtained, wherein the warehouse dynamic graph includes node feature vectors and edge feature vectors; the node feature vectors and the edge feature vectors of the warehouse dynamic graph are input into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent a probability of the storage location being accessed; based on the storage location heat score of each storage location, an initial storage location set is filtered to obtain a candidate storage location set; based on the candidate storage location set, a pre-constructed operational research multi-objective optimization model is solved to obtain a storage location allocation result. The above technical solution cooperatively decides through graph neural network feature learning and operational research multi-objective optimization, improves the accuracy and reliability of storage location allocation, reduces the time spent in the storage and retrieval process of goods, and thus improves the overall warehouse operation efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be considered as limiting the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0017] Figure 1 A flowchart of a storage location allocation method provided by the embodiments of the present application;

[0018] Figure 2 A flowchart of a storage location allocation method provided by the embodiments of the present application;

[0019] Figure 3 A flowchart of a storage location allocation method provided by the embodiments of the present application;

[0020] Figure 4A flowchart of a warehouse location allocation method provided for an embodiment of the present application;

[0021] Figure 5 A flowchart of a warehouse location multi-objective optimization allocation method provided for an embodiment of the present application, which combines a graph neural network with operations research.

[0022] Figure 6 A structural schematic diagram of a warehouse location allocation device provided for an embodiment of the present application.

[0023] Figure 7 A structural schematic diagram of an electronic device provided for an embodiment of the present application. DETAILED DESCRIPTION

[0024] The present application will be further described below in conjunction with the accompanying drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that, for the convenience of description, only the parts related to the present application are shown in the drawings, but not all the structures.

[0025] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance. The acquisition, storage, use, processing, etc. of data in the technical solution of the present application all comply with the relevant provisions of national laws and regulations.

[0026] It should be noted that, in the embodiments of the present application, some industry existing solutions, components, models, etc. may be mentioned, which should be considered as exemplary, and the purpose is only to illustrate the feasibility of the implementation of the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0027] Figure 1 A flowchart of a warehouse location allocation method provided for an embodiment of the present application, the present embodiment can be applicable to the automatic allocation of warehouse locations, the method can be executed by a warehouse location allocation device, which can be realized in the form of hardware and / or software, and the warehouse location allocation device can be configured in a terminal, a server, etc. electronic equipment. As shown in the figure, the method comprises: Figure 1

[0028] S110, acquiring a warehouse dynamic graph, wherein the warehouse dynamic graph comprises node feature vectors and edge feature vectors.

[0029] The warehouse dynamic graph is a graph structure updated in real time according to warehouse orders, transportation equipment and warehouse location states, which can include node feature vectors and edge feature vectors.​

[0030] Specifically, the node feature vector includes a storage location node, a cargo node, an order node, and a device node; the storage location node includes one or more of a storage location coordinate, a storage location bearing, a storage location temperature and humidity state (for cold chain), and a storage location occupancy state (free / occupied); the cargo node includes one or more of a commodity color and model code (SKU code), a cargo size, a cargo weight, a cargo category label, and a cargo turnover rate; the order node includes one or more of an order identifier, an order urgency label (such as "urgent"), a list of cargos included in the order, and a number of cargos included in the order; and the device node includes one or more of an automatic guided vehicle (AGV) number, an AGV real-time power, an AGV current position, and an AGV task queue state (running or non-running). The edge feature vector includes a weight value of a physical distance edge and a weight value of a co-picking probability edge; the weight value of the physical distance edge is the inverse of the Euclidean distance between the storage locations; and the weight value of the co-picking probability edge is calculated according to the formula:

[0031]

[0032] wherein the co-picking frequency represents the number of times that two cargos are ordered by the same order (such as cargos A and B are often ordered by the same order), and the time decay factor is a pre-set parameter value.

[0033] In some embodiments, the edge feature vector can further include a weight value of a device-storage location reachable edge. Specifically, the weight value of the device-storage location reachable edge can be dynamically adjusted according to the AGV real-time power, the path obstacle state, and the like, such as when the power is lower than 20%, the AGV cannot reach a long-distance storage location, and the weight value is 0. The edge feature vector can further include the type of the edge, and the type of the edge includes a physical type and a logical type.

[0034] On the basis of the above-mentioned embodiments, optionally, after the warehouse dynamic graph is obtained, the method further includes: in the case where a new order is detected to be generated, updating the weight value of the co-picking probability edge.

[0035] Illustratively, if the new order includes co-picking cargos, the co-picking frequency is increased by 1, and the total number of orders is increased by 1, and the weight value of the co-picking probability edge is recalculated based on the updated co-picking frequency and the total number of orders, so that the warehouse dynamic graph responds to the actual order fluctuation in real time.

[0036] S120, input the node feature vector and the edge feature vector of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent the probability of the storage location being accessed; and filter the initial storage location set based on the storage location heat score of each storage location to obtain a candidate storage location set.

[0037] The bin location heat score is a comprehensive score of the bin location, and can represent the probability of the bin location being accessed in the future. For example, the bin location heat score can be a value in (0, 1). The greater the value of the bin location heat score, the greater the probability of the bin location being accessed in the future.

[0038] Specifically, the node feature vector and the edge feature vector of the warehouse dynamic graph can be input into the graph neural network model as input features. The graph neural network model can perform feature extraction and prediction based on the warehouse dynamic graph, and output the bin location heat score. Then, the bin locations with the bin location heat score in a preset score range or greater than a preset score threshold can be selected as the candidate bin location set.

[0039] In S130, the pre-constructed operations research multi-objective optimization model is solved based on the candidate bin location set, and a bin location allocation result is obtained.

[0040] The operations research multi-objective optimization model can realize multi-objective optimization of path, space, and load through operations research methods, so as to obtain an optimal bin location allocation result. Specifically, the operations research multi-objective optimization model can include multiple objective functions and one or more constraint conditions.

[0041] Specifically, the candidate bin location set can be input into the operations research multi-objective optimization model, and the operations research multi-objective optimization model is solved to obtain the bin location allocation result.

[0042] In the embodiments of the present application, a warehouse dynamic graph is obtained, wherein the warehouse dynamic graph includes a node feature vector and an edge feature vector. The node feature vector and the edge feature vector of the warehouse dynamic graph are input into a pre-constructed graph neural network model to obtain a bin location heat score of each bin location, wherein the bin location heat score represents the probability of the bin location being accessed. Based on the bin location heat score of each bin location, an initial bin location set is screened to obtain a candidate bin location set. Based on the candidate bin location set, a pre-constructed operations research multi-objective optimization model is solved to obtain a bin location allocation result. The technical solution of the present application improves the accuracy and reliability of bin location allocation through collaborative decision-making of graph neural network feature learning and operations research multi-objective optimization, reduces the time spent in the storage and retrieval process of goods, and thus improves the overall warehouse operation efficiency.

[0043] Figure 2A flowchart of a warehouse location allocation method provided by an embodiment of the present application, the method of the present embodiment can be combined with the various optional schemes of the warehouse location allocation method provided in the above embodiments. The warehouse location allocation method provided by the present embodiment is further optimized. Optionally, the inputting of the node feature vector and the edge feature vector of the warehouse dynamic graph into the pre-constructed graph neural network model to obtain the warehouse location heat score of each warehouse location comprises: inputting the node feature vector and the edge feature vector of the warehouse dynamic graph into a spatio-temporal graph convolution network of the graph neural network model to obtain the spatio-temporal feature of the warehouse location; inputting the spatio-temporal feature of the warehouse location into a graph attention network of the graph neural network model to obtain the spatio-temporal feature of the warehouse location after attention weighting; and inputting the spatio-temporal feature of the warehouse location after attention weighting into a fully connected layer to obtain the warehouse location heat score of each warehouse location.

[0044] As shown in Figure 2 , the method comprises:

[0045] S210, acquiring a warehouse dynamic graph, wherein the warehouse dynamic graph comprises a node feature vector and an edge feature vector.

[0046] S220, inputting the node feature vector and the edge feature vector of the warehouse dynamic graph into a spatio-temporal graph convolution network of the graph neural network model to obtain a spatio-temporal feature of the warehouse location.

[0047] The spatio-temporal graph convolution network (ST-GCN) is used to fuse the spatial graph convolution and the time convolution of the node feature vector and the edge feature vector, capture the time sequence change of the warehouse location state, and obtain the spatio-temporal feature of the warehouse location.

[0048] S230, inputting the spatio-temporal feature of the warehouse location into a graph attention network of the graph neural network model to obtain the spatio-temporal feature of the warehouse location after attention weighting.

[0049] The graph attention network (GAT) can focus on the warehouse location where the high-frequency turnover goods are located, i.e., obtain the spatio-temporal feature of the warehouse location after attention weighting.

[0050] S240, inputting the spatio-temporal feature of the warehouse location after attention weighting into a fully connected layer to obtain a warehouse location heat score of each warehouse location, wherein the warehouse location heat score is used to represent the probability of the warehouse location being accessed.

[0051] S250, screening an initial warehouse location set based on the warehouse location heat score of each warehouse location to obtain a candidate warehouse location set.

[0052] S260, solving a pre-constructed operational research multi-objective optimization model based on the candidate warehouse location set to obtain a warehouse location allocation result.

[0053] In some embodiments, the graph neural network model can also be compressed by knowledge distillation and INT8 quantization technology to reduce the volume of the graph neural network model, so as to realize the lightweight deployment of edge computing.

[0054] In the embodiment of the application, the node feature vector and the edge feature vector of the warehouse dynamic graph are input into the spatio-temporal graph convolution network of the graph neural network model to obtain the spatio-temporal features of the storage locations; the spatio-temporal features of the storage locations are input into the graph attention network of the graph neural network model to obtain the spatio-temporal features of the storage locations after attention weighting; and the spatio-temporal features of the storage locations after attention weighting are input into a full connection layer to obtain the storage location heat scores of each storage location, thereby realizing the accurate prediction of the storage location heat scores.

[0055] Figure 3 A flowchart of a storage location allocation method provided in the embodiment of the application can be combined with each optional scheme in the storage location allocation method provided in the above-mentioned embodiments. The storage location allocation method provided in the embodiment is further optimized. Optionally, the initial storage location set is filtered based on the storage location heat scores of each storage location to obtain a candidate storage location set, which includes: sorting the storage locations in the initial storage location set from large to small based on the storage location heat scores of each storage location; and taking the first K storage locations in the sorting result as the candidate storage location set, wherein K = the total number of storage locations in the initial storage location set x M%, and M is a value between 0 and 100.

[0056] As shown in Figure 3 the method includes:

[0057] S310, a warehouse dynamic graph is acquired, wherein the warehouse dynamic graph includes a node feature vector and an edge feature vector.

[0058] S320, the node feature vector and the edge feature vector of the warehouse dynamic graph are input into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent the probability of the storage location being accessed.

[0059] S330, the storage locations in an initial storage location set are sorted from large to small based on the storage location heat scores of each storage location; and the first K storage locations in the sorting result are taken as a candidate storage location set, wherein K = the total number of storage locations in the initial storage location set x M%, and M is a value between 0 and 100.

[0060] For example, the first 10% of the storage locations in the initial storage location set with the highest storage location heat scores can be selected as the candidate storage location set as the input of the subsequent operational research multi-objective optimization model, which can reduce the search space by 90%.

[0061] S340, based on the candidate set of positions, solve the pre-constructed operational research multi-objective optimization model to obtain a position allocation result.

[0062] In the embodiment of the application, the positions in the initial position set are sorted from large to small based on the position heat scores of each position; the first K positions in the sorting result are taken as the candidate position set, thereby realizing pre-selection of high-quality positions.

[0063] Figure 4 A flowchart of a position allocation method provided in the embodiment of the application can be combined with each optional scheme in the position allocation method provided in the above-mentioned embodiments. The position allocation method provided in the embodiment is further optimized. Optionally, based on the candidate position set, the pre-constructed operational research multi-objective optimization model is solved by using a mixed integer programming or a multi-objective genetic algorithm to obtain a position allocation result.

[0064] As shown in Figure 4 , the method comprises:

[0065] S410, obtaining a warehouse dynamic graph, wherein the warehouse dynamic graph comprises a node feature vector and an edge feature vector.

[0066] S420, inputting the node feature vector and the edge feature vector of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a position heat score of each position, wherein the position heat score is used to represent the probability of the position being accessed; based on the position heat score of each position, an initial position set is screened to obtain a candidate position set.

[0067] S430, based on the candidate position set, using a mixed integer programming or a multi-objective genetic algorithm to solve a pre-constructed operational research multi-objective optimization model to obtain a position allocation result.

[0068] Specifically, for the mixed integer programming, Gurobi / CPLEX can be used for solving. For the multi-objective genetic algorithm, interactive selection of a decision maker can be supported, a Pareto frontier solution is generated, and the position allocation result is obtained.

[0069] The objective function of the operational research multi-objective optimization model comprises:

[0070] Minimize the total picking path function: ∑ i,j,k d jk ·x ijk ;

[0071] Maximize the space utilization function:

[0072] Balancing device load function:

[0073] wherein d jk represents the distance from device k to the storage location j; x ijk is a binary variable, indicating whether the goods i are allocated to the storage location j and executed by the device k; used volume j represents the volume already used by the storage location j, and total capacity j represents the total volume of the storage location j;

[0074] The constraint conditions of the operational research multi-objective optimization model include:

[0075] ∑ i volume i x ijk ≤ capacity j;

[0076] ∑ j d jk x ijk ≤ remaining endurance k;

[0077] wherein volume i represents the volume of the goods i, capacity j represents the current available volume of the storage location j, and remaining endurance k represents the remaining endurance of the device k.

[0078] In some embodiments, for a warehouse system containing a cold storage, the objective function of the operational research multi-objective optimization model further includes:

[0079]

[0080] wherein the number of door openings j represents the number of door openings of the storage location j, and the temperature control stability coefficient is a pre-set custom parameter.

[0081] It should be noted that by minimizing the number of door openings through storage location allocation, strict temperature control of the cold storage is achieved, which reduces temperature exceeding events by 90% and reduces energy consumption by 15% in actual application.

[0082] In the embodiments of the present application, the pre-constructed operational research multi-objective optimization model is solved by using mixed integer programming or multi-objective genetic algorithm based on the candidate storage location set, multi-objective balancing is achieved, and accurate storage location allocation results are obtained.

[0083] Figure 5 A flowchart of a warehouse storage location multi-objective optimization allocation method combining a graph neural network with operational research provided by the embodiments of the present application is shown in FIG. Figure 5 As shown in the figure, the method includes:

[0084] 1. Collecting warehouse data through a data real-time collection module, which can include storage location coordinates, goods turnover rate, order identification, etc.

[0085] 2. The warehouse dynamic graph is constructed by the warehouse data through the warehouse dynamic graph construction module.

[0086] 3. The warehouse dynamic graph is input into a pre-constructed graph neural network model through the graph neural network feature learning module, and a bin heat score of each bin is obtained. The initial bin set is filtered based on the bin heat score of each bin to obtain a candidate bin set.

[0087] 4. The warehouse dynamic graph is input into a pre-constructed graph neural network model through the graph neural network feature learning module, and a bin heat score of each bin is obtained. The initial bin set is filtered based on the bin heat score of each bin to obtain a candidate bin set.

[0088] 5. The warehouse dynamic graph is input into a pre-constructed graph neural network model through the graph neural network feature learning module, and a bin heat score of each bin is obtained. The initial bin set is filtered based on the bin heat score of each bin to obtain a candidate bin set.

[0089] Specifically, the bin allocation result can be sent to the AGV controller through the 5G network facility in the warehouse hardware of the warehouse system, and the AGV controller controls the AGV to plan the path and avoid obstacles in real time by using the A* algorithm. After obtaining the bin allocation result, the bin-goods binding information can be written into the bin through the RFID in the warehouse hardware, and the information is fed back to the data real-time acquisition module.

[0090] In some embodiments, the actual picking time, path deviation, device utilization rate, bin occupation conflict, device failure and order cancellation can also be fed back to the data real-time acquisition module, thereby improving the reliability of the acquired data.

[0091] Figure 6 A structure diagram of a bin allocation device provided by an embodiment of the application is shown in FIG. 1. Figure 6 As shown in the figure, the device comprises:

[0092] The warehouse dynamic graph acquisition module 510 is configured to acquire a warehouse dynamic graph, wherein the warehouse dynamic graph comprises a node feature vector and an edge feature vector.

[0093] The graph neural network feature learning module 520 is configured to input the node feature vector and the edge feature vector of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a bin heat score of each bin, wherein the bin heat score is used to represent the probability of the bin being accessed. The initial bin set is filtered based on the bin heat score of each bin to obtain a candidate bin set.

[0094] The operations research multi-objective optimization module 530 is configured to solve a pre-constructed operations research multi-objective optimization model based on the candidate bin set to obtain a bin allocation result.

[0095] In the embodiment of the present application, a warehouse dynamic graph is obtained, wherein the warehouse dynamic graph includes a node feature vector and an edge feature vector; the node feature vector and the edge feature vector of the warehouse dynamic graph are input into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent a probability of the storage location being accessed; based on the storage location heat score of each storage location, an initial storage location set is screened to obtain a candidate storage location set; based on the candidate storage location set, a pre-constructed operational research multi-objective optimization model is solved to obtain a storage location allocation result. The technical scheme of the present application cooperates decision-making of graph neural network feature learning and operational research multi-objective optimization, improves the accuracy and reliability of storage location allocation, reduces the time spent in the storage and retrieval process of goods, and thus improves the overall warehouse operation efficiency.

[0096] In some optional embodiments, the node feature vector includes a storage location node, a goods node, an order node, and a device node;

[0097] The storage location node includes one or more of a storage location coordinate, a storage location load-bearing capacity, a storage location temperature and humidity state, and a storage location occupancy state;

[0098] The goods node includes one or more of a commodity color and model code, a goods size, a goods weight, a goods category tag, and a goods turnover rate;

[0099] The order node includes one or more of an order identifier, an order urgency label, a list of goods included in the order, and a number of goods included in the order;

[0100] The device node includes one or more of an automatic guided vehicle number, an automatic guided vehicle real-time power, an automatic guided vehicle current position, and an automatic guided vehicle task queue state.

[0101] The edge feature vector includes a weight value of a physical distance edge and a weight value of a common picking probability edge;

[0102] The weight value of the physical distance edge is the inverse of the Euclidean distance between the storage locations;

[0103] The calculation formula of the weight value of the common picking probability edge is:

[0104]

[0105] Wherein, the common picking frequency represents the number of times that two goods are ordered by the same order, and the time decay factor is a pre-set parameter value.

[0106] In some optional embodiments, the storage location allocation device further comprises:

[0107] The weight value updating module of the co-picking probability edge is configured to update the weight value of the co-picking probability edge when a new order is detected.

[0108] In some optional embodiments, the graph neural network feature learning module 520 can be specifically configured to:

[0109] input the node feature vector and the edge feature vector of the warehouse dynamic graph into a spatio-temporal graph convolution network of the graph neural network model to obtain the spatio-temporal feature of the storage location;

[0110] input the spatio-temporal feature of the storage location into a graph attention network of the graph neural network model to obtain the spatio-temporal feature of the storage location after attention weighting;

[0111] input the spatio-temporal feature of the storage location after attention weighting into a full connection layer to obtain the storage location heat score of each storage location.

[0112] In some optional embodiments, the graph neural network feature learning module 520 can be specifically configured to:

[0113] sort the storage locations in the initial storage location set from large to small based on the storage location heat score of each storage location;

[0114] take the first K storage locations in the sorting result as a candidate storage location set, where K = total number of storage locations in the initial storage location set × M%, and M is a value between 0 and 100.

[0115] In some optional embodiments, the operations research multi-objective optimization module 530 can be specifically configured to:

[0116] based on the candidate storage location set, use a mixed integer programming or a multi-objective genetic algorithm to solve a pre-constructed operations research multi-objective optimization model to obtain a storage location allocation result.

[0117] In some optional embodiments, the objective function of the operations research multi-objective optimization model includes:

[0118] minimize the total picking path function: ∑ i,j,k d jk ·x ijk ;

[0119] maximize the space utilization function:

[0120] balance the device load function:

[0121] where d jk represents the distance from device k to storage location j; x ijkis a binary variable, indicating whether the goods i are allocated to the storage location j and executed by the device k; used volume j represents the volume already used by the storage location j, and total capacity j represents the total volume of the storage location j;

[0122] The constraint condition of the operation optimization model includes:

[0123] ∑ i volume i x ijk ≤ capacity j;

[0124] ∑ j d jk x ijk ≤ remaining endurance k;

[0125] wherein volume i represents the volume of the goods i, capacity j represents the current available volume of the storage location j, and remaining endurance k represents the remaining endurance of the device k.

[0126] The storage location allocation device provided in the embodiments of the present application can execute the storage location allocation method provided in any of the embodiments of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0127] Figure 7 Fig. 1 is a structural schematic diagram of an electronic device provided by an embodiment of the present disclosure. Hereinafter, the Figure 7 , which shows a structural schematic diagram of an electronic device (for example, a terminal device or a server) 500 suitable for implementing an embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure can include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (Personal Digital Assistant), a PAD (Tablet Personal Computer), a PMP (Portable Multimedia Player), a vehicle-mounted terminal (for example, a vehicle-mounted navigation terminal), and the like, and a fixed terminal such as a digital TV, a desktop computer, and the like. Figure 7 The electronic device shown is merely an example, and should not bring any limitation to the functions and use range of the embodiments of the present disclosure. Figure 7

[0128] As shown in Figure 7 , the electronic device 500 can include a processing device (for example, a central processor, a graphics processor, and the like) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An edit / output (I / O) interface 505 is also connected to the bus 504.

[0129] ​In general, the following devices can be connected to the I / O interface 505: input devices 506 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, and the like; output devices 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, and the like; storage devices 508 including, for example, a magnetic tape, a hard disk, and the like; and communication devices 509. The communication devices 509 can allow the electronic device 500 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 7 The electronic device 500 is shown with various devices, but it is understood that all of the shown devices are not required to be implemented or present. More or fewer devices can alternatively be implemented or present.

[0130] In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network through the communication devices 509, or installed from the storage devices 508, or installed from the ROM 502. When the computer program is executed by the processing devices 501, the above-described functions defined in the methods of the embodiments of the present disclosure are performed.

[0131] The names of the messages or information exchanged between the plurality of devices in the embodiments of the present disclosure are only for illustrative purposes, and are not intended to limit the scope of the messages or information.

[0132] The electronic device provided by the embodiments of the present disclosure and the storage location allocation method provided by the above embodiments belong to the same inventive concept, and the technical details not described in detail in the present embodiment can be referred to the above embodiments, and the present embodiment has the same beneficial effects as the above embodiments.

[0133] The embodiments of the present disclosure provide a computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the storage location allocation method provided by the above embodiments.

[0134] It should be noted that the computer-readable medium described above can be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable storage medium, for example, can be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or any suitable combination of the foregoing. More specific examples of the computer-readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program used by or in connection with an instruction execution system, apparatus or device. In the disclosure, the computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, in which the computer-readable program code is contained. Such a propagated data signal can take any of a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination of the foregoing. The computer-readable signal medium can also be any computer-readable medium that is not a storage medium and that can communicate, propagate or transport a program for use by or in connection with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including, but not limited to, wire, cable, RF (radio frequency), etc., or any suitable combination of the foregoing.

[0135] In some embodiments, the client, server, or both can communicate using any current known or future developed network protocol, such as HTTP (HyperText Transfer Protocol), and can be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks ("LAN"), wide area networks ("WAN"), the Internet, and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any current known or future developed networks.

[0136] The computer-readable medium described above can be included in the electronic device; or exist separately from the electronic device, and not be assembled into the electronic device.

[0137] The computer-readable medium described above carries one or more programs, when the one or more programs are executed by the electronic device, cause the electronic device to:

[0138] An in-warehouse dynamic graph is acquired, wherein the in-warehouse dynamic graph comprises node feature vectors and edge feature vectors;

[0139] The node feature vectors and the edge feature vectors of the in-warehouse dynamic graph are input into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent a probability of the storage location being accessed; and the initial storage location set is filtered based on the storage location heat score of each storage location to obtain a candidate storage location set.

[0140] The pre-constructed operational research multi-objective optimization model is solved based on the candidate storage location set to obtain a storage location allocation result.

[0141] Computer program code for carrying out operations of the present disclosure can be written in one or more programming languages or combinations of languages including object oriented programming languages such as Java, Smalltalk, C++ as well as conventional procedural programming languages such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0142] The flow and block diagrams in the drawings show architectural, functional, and operational representations of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flow and block diagrams can represent a module, a segment, or a portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or in the reverse order, depending on the functionality involved. Also, it is noted that each block of the block diagrams and / or flowchart illustrations, and combinations thereof, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or combinations of hardware and software.

[0143] The units described in the embodiments of the present disclosure can be implemented by means of software, or by hardware. In some cases, the name of the unit does not constitute a limitation on the unit itself.

[0144] The functions described above in the present document can be performed, at least in part, by one or more hardware logic components. For example, and without limitation, illustrative types of hardware logic components that can be used include Field-programmable Gate Arrays (FPGAs), Application-specific Integrated Circuits (ASICs), Application-specific Standard Products (ASSPs), System-on-a-chip systems (SOCs), Complex Programmable Logic Devices (CPLDs), etc.

[0145] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0146] The embodiments of the present application also provide a computer program product, comprising a computer program which, when executed by a processor, implements the warehouse location allocation method provided by any of the embodiments of the present application.

[0147] The computer program product, in its implementation, can be written in one or more programming languages or combinations of the same to implement the computer program code for performing the operations of the present application, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" language or similar programming languages. The program code can be executed entirely on a user computer, partially on a user computer, as a separate software package, partially on a user computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0148] Note that the above merely describes preferred embodiments of the application and the principles of the technology applied. Those skilled in the art will understand that the application is not limited to the specific embodiments described herein, and that various obvious changes, modifications and substitutions can be made to the application without departing from the scope of the application. Therefore, although the application has been described in detail by the above embodiments, the application is not limited to the above embodiments, and can include more other equivalent embodiments without departing from the concept of the application, and the scope of the application is determined by the appended claims.

Claims

1. A storage location allocation method characterized by, The method comprises: acquiring a warehouse dynamic graph, wherein the warehouse dynamic graph comprises node feature vectors and edge feature vectors; inputting the node feature vectors and the edge feature vectors of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a storage location heat score of each storage location, wherein the storage location heat score is used to represent a probability of the storage location being accessed; and screening an initial storage location set based on the storage location heat score of each storage location to obtain a candidate storage location set; solving a pre-constructed operational research multi-objective optimization model based on the candidate storage location set to obtain a storage location allocation result.

2. The method of claim 1, wherein, The node feature vectors comprise storage location nodes, cargo nodes, order nodes and equipment nodes. The storage location nodes comprise one or more of storage location coordinates, storage location bearing capacity, storage location temperature and humidity status and storage location occupancy status. The cargo nodes comprise one or more of commodity color and model code, cargo size, cargo weight, cargo category label and cargo turnover rate. The order nodes comprise one or more of order identification, order urgency label, a list of cargos contained in the order and the number of cargos contained in the order. The equipment nodes comprise one or more of automatic guided vehicle number, automatic guided vehicle real-time power, automatic guided vehicle current position and automatic guided vehicle task queue status. The edge feature vectors comprise weight values of physical distance edges and weight values of common picking probability edges. The weight value of the physical distance edge is the inverse of the Euclidean distance between storage locations. The calculation formula of the weight value of the common picking probability edge is: wherein the common picking frequency represents the number of times that two cargos are ordered by the same order, and the time decay factor is a pre-set parameter value.

3. The method of claim 2, wherein, After the warehouse dynamic graph is acquired, the method further comprises: updating the weight value of the common picking probability edge when a new order is detected.

4. The method of claim 1, wherein, The method of inputting the node feature vectors and the edge feature vectors of the warehouse dynamic graph into the pre-constructed graph neural network model to obtain the storage location heat score of each storage location comprises: inputting the node feature vectors and the edge feature vectors of the warehouse dynamic graph into a spatio-temporal graph convolution network of the graph neural network model to obtain the spatio-temporal features of the storage location; inputting the spatio-temporal features of the storage location into a graph attention network of the graph neural network model to obtain the spatio-temporal features of the storage location after attention weighting; inputting the spatio-temporal features of the storage location after attention weighting into a fully connected layer to obtain the storage location heat score of each storage location.

5. The method of claim 1, wherein, The method of screening the initial storage location set based on the storage location heat score of each storage location to obtain the candidate storage location set comprises: sorting the storage locations in the initial storage location set from large to small based on the storage location heat score of each storage location; taking the first K storage locations in the sorting result as the candidate storage location set, wherein K = the total number of storage locations in the initial storage location set × M%, and M is a value between 0 and 100.

6. The method of claim 1, wherein, The method of solving the pre-constructed operational research multi-objective optimization model based on the candidate storage location set to obtain the storage location allocation result comprises: Based on the candidate set of locations, a pre-constructed operational research multi-objective optimization model is solved using mixed integer programming or multi-objective genetic algorithm to obtain a location allocation result.

7. The method of claim 6, wherein, The objective function of the operational research multi-objective optimization model includes: Minimize total picking path function:∑ i,j,k d jk ·x ijk ; Maximize the space utilization function: Balancing device load function: where d jk represents the distance from device k to bin j; x ijk is a binary variable indicating whether item i is assigned to bin j and executed by device k; used volume j represents the volume already used by bin j, and total capacity j represents the total volume of bin j; The constraint condition of the operational research multi-objective optimization model includes: ∑ i Volume i · x ijk ≤ Capacity j; ∑ j d jk ·x ijk ≤ remaining range k; Wherein, the volume i represents the volume of the goods i, the capacity j represents the current available volume of the location j, and the remaining endurance k represents the remaining endurance of the device k.

8. A storage location allocation apparatus characterized by comprising: It includes: A warehouse dynamic graph acquisition module is configured to acquire a warehouse dynamic graph, wherein the warehouse dynamic graph includes node feature vectors and edge feature vectors. A graph neural network feature learning module is configured to input the node feature vectors and the edge feature vectors of the warehouse dynamic graph into a pre-constructed graph neural network model to obtain a location heat score of each location, wherein the location heat score is used to represent the probability of the location being accessed; and the initial location set is filtered based on the location heat score of each location to obtain a candidate location set. An operational research multi-objective optimization module is configured to solve a pre-constructed operational research multi-objective optimization model based on the candidate location set to obtain a location allocation result.

9. An electronic device, comprising: The program is executed by the processor to implement the location allocation method of any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the location allocation method of any one of claims 1-7.

11. A computer program product comprising a computer program, characterized in that, The program is executed by the processor to implement the location allocation method of any one of claims 1-7.

Images