A method and system for monitoring overseas warehouse WMS inventory during high concurrency periods

Abstract

The present invention relates to the technical field of inventory monitoring, and invents an overseas warehouse WMS inventory monitoring method for high concurrency periods, comprising: extracting inventory time series features from historical inventory data, and performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analytical inventory data; performing inventory feature extraction and time series feature clustering on the analytical inventory data to obtain inventory feature sets, and generating inventory sharding rules based on the inventory feature sets; performing commodity barcode scanning and barcode recognition on warehouse commodities to obtain real-time inventory data; performing inventory data sharding on real-time inventory data according to inventory sharding rules to obtain sharded inventory messages, and performing subscription transmission on the sharded inventory messages to obtain updated inventory messages; and using updated inventory messages to update preset historical inventory monitoring data to obtain real-time inventory monitoring data. The present invention can improve the robustness of inventory monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of inventory monitoring, and in particular to a method and system for monitoring overseas warehouse WMS inventory during high concurrency periods. Background Art

[0002] Overseas warehouse management system (Warehouse Management System, referred to as WMS) is a software system used to manage and control the daily operations of overseas warehouses, helping warehouse managers to effectively track inventory, optimize warehouse layout, manage orders, and handle tasks such as shipping and receiving. Overseas warehouse management system can realize inventory monitoring, warehouse operation management, order processing, material scheduling and cargo tracking and other functions.

[0003] In order to optimize inventory management and improve enterprise operational efficiency, it is necessary to use a warehouse management system to monitor inventory. In actual applications, when the order volume surges or goods are put into storage, frequent warehouse operations will cause the warehouse management system to process a large number of read and write operations at the same time, which may lead to problems such as inaccurate inventory and processing delays, resulting in low robustness when performing inventory monitoring. Summary of the invention

[0004] The present invention provides a method and system for monitoring inventory of overseas warehouses (WMS) during high concurrency periods, the main purpose of which is to solve the problem of low robustness when performing inventory monitoring.

[0005] To achieve the above objectives, the present invention provides a method for monitoring overseas warehouse WMS inventory during high concurrency periods, comprising:

[0006] Extracting inventory time series features from pre-acquired historical inventory data, and performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analysis inventory data, wherein the extracting inventory time series features from the pre-acquired historical inventory data includes: extracting data timestamps from the historical inventory data, and performing time series sorting on the pre-acquired historical inventory data according to the data timestamps to obtain a historical data sequence; performing inventory feature extraction on the historical data sequence to obtain a historical inventory feature sequence; performing time series recursive convolution on the historical inventory feature sequence to obtain short-term inventory features; performing time series jump convolution on the historical inventory feature sequence to obtain long-term inventory features; performing attention encoding on the historical data sequence according to the attention mechanism to obtain attention inventory features; and performing feature fusion on the short-term inventory features, the long-term inventory features, and the attention inventory features using the following time series connection algorithm to obtain inventory time series features: in, The inventory timing characteristics are The eigenvalues ​​of the time steps, is the feature weight of the short-term inventory feature, The short-term inventory characteristics are The eigenvalues ​​of the time steps, , is the time step index, is the number of hidden units that the long-term inventory feature skips when performing temporal skip convolution, The long-term inventory characteristics are The feature weights at the time step, The long-term inventory characteristics are The eigenvalues ​​of the time steps, is the attention inventory feature in The eigenvalues ​​of the time steps, is the preset bias coefficient;

[0007] Performing inventory feature extraction and time series feature clustering on the analyzed inventory data to obtain an inventory feature set, and generating an inventory sharding rule according to the inventory feature set;

[0008] Scan and identify the barcodes of the preset warehouse products to obtain real-time inventory data;

[0009] Performing inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain sharded inventory messages, and subscribing and transmitting the sharded inventory messages to obtain updated inventory messages;

[0010] The preset historical inventory monitoring data is updated using the inventory update message to obtain real-time inventory monitoring data.

[0011] Optionally, performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analyzed inventory data includes:

[0012] Performing feature splitting on the inventory time series feature to obtain a split time series feature set;

[0013] Filtering a split feature node set from a preset split point set according to the split time series feature set;

[0014] Generate a time series feature decision tree set according to the split feature node set;

[0015] Performing inventory decision on the split feature node set according to the time series feature decision tree set to obtain a decision inventory data set;

[0016] A weighted regression is performed on the decision inventory data set to obtain analysis inventory data.

[0017] Optionally, the extracting inventory features and clustering time series features of the analyzed inventory data to obtain an inventory feature set includes:

[0018] Performing inventory data splitting on the analysis inventory data to obtain an analysis data set, and performing inventory feature extraction on the analysis data set to obtain a primary analysis inventory feature set;

[0019] Performing timestamp extraction and timestamp standardization operations on the analysis data set to obtain an analysis inventory time feature set;

[0020] Generate an analysis inventory feature set according to the analysis inventory time feature set and the primary analysis inventory feature set;

[0021] Splitting all analysis inventory features in the analysis inventory feature set into a plurality of primary analysis inventory feature groups, and randomly selecting a primary central inventory feature for each primary analysis inventory feature group;

[0022] The inventory feature distance between each primary center inventory feature and each analysis inventory feature is calculated using the following inventory feature distance algorithm: in, It refers to analyzing inventory characteristics and primary center inventory characteristics The inventory feature distance between is the exponential function symbol, is the inner product symbol, is the modulo symbol, refers to the analyzed inventory characteristics, refers to the primary center inventory characteristics, is the preset smoothness coefficient, is the preset threshold coefficient, is the transpose symbol, It refers to analyzing inventory characteristics and primary center inventory characteristics The covariance matrix of is the preset regularization coefficient, is the identity matrix;

[0023] Regrouping all analysis inventory features according to the inventory feature distance to obtain multiple secondary analysis inventory feature groups;

[0024] Calculating the secondary center inventory features corresponding to each secondary analysis inventory feature group, and calculating the center feature distance between each secondary center inventory feature and the corresponding primary center inventory feature;

[0025] Each secondary analysis inventory feature group is iteratively updated according to all center feature distances to obtain an inventory feature cluster.

[0026] Optionally, generating inventory sharding rules according to the inventory feature set includes:

[0027] Performing feature number statistics on each inventory feature class in the inventory feature class set to obtain an inventory frequency set;

[0028] Performing cluster center screening on the inventory feature set to obtain a standard center inventory feature set;

[0029] Generate an inventory standard feature set according to the inventory frequency set and the standard center inventory feature set;

[0030] Performing database mapping on the inventory standard feature set to obtain a shard database set;

[0031] An inventory sharding rule is generated according to the standard central inventory feature set and the sharding database set.

[0032] Optionally, the barcode scanning and barcode recognition of the preset warehouse commodities to obtain real-time inventory data includes:

[0033] Scan the barcodes of the preset warehouse products to obtain the inventory barcode images;

[0034] Performing image denoising and image equalization operations on the inventory barcode image to obtain a balanced inventory barcode image;

[0035] Performing grayscale threshold segmentation on the balanced inventory barcode image to obtain a segmented barcode image;

[0036] Performing a barcode width normalization operation on the segmented barcode image to obtain a standard barcode image;

[0037] Recognize the width information of the standard barcode image to obtain the product barcode;

[0038] Performing code verification on the commodity barcode to obtain verification information;

[0039] The commodity barcode is decoded according to the verification information to obtain real-time inventory data.

[0040] Optionally, performing grayscale threshold segmentation on the balanced inventory barcode image to obtain a segmented barcode image includes:

[0041] Calculating the grayscale of the balanced inventory barcode image to obtain a barcode grayscale group;

[0042] Selecting barcode gray levels in the barcode gray level group one by one as target gray levels, performing image segmentation on the balanced inventory barcode image according to the target gray levels, and obtaining a target inventory barcode image and a target background barcode image;

[0043] The image grayscale distance between the target inventory barcode image and the target background barcode image is calculated using the following image grayscale distance algorithm: in, refers to the image grayscale distance, refers to the pixel length of the balanced inventory barcode image, refers to the pixel width of the balanced inventory barcode image, Refers to the total number of pixels of the target inventory barcode image, Refers to the total number of pixels of the target background barcode image, , is the pixel index, is the gray value symbol;

[0044] The image grayscale distances of all target grayscale levels in the barcode grayscale group are collected into an image grayscale distance group, and the image grayscale distance with the largest value in the image grayscale distance group is used as the segmentation grayscale distance;

[0045] The target inventory barcode image corresponding to the segmentation grayscale distance is used as the segmentation barcode image.

[0046] Optionally, performing inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain sharded inventory messages includes:

[0047] Extracting inventory features from the real-time inventory data to obtain real-time inventory features;

[0048] Extracting a standard center inventory feature set from the inventory sharding rule, and calculating a feature distance between the real-time inventory feature and each standard center inventory feature in the standard center inventory feature set to obtain a sharding distance set;

[0049] Selecting the shard distance with the smallest value from the shard distance set as the target shard distance;

[0050] The standard central inventory feature corresponding to the target shard distance in the standard central inventory feature set is used as the shard inventory feature;

[0051] Filtering a target shard database from the inventory sharding rules according to the shard inventory characteristics;

[0052] The real-time inventory data is polled and packaged according to the target shard database to obtain shard inventory information.

[0053] Optionally, polling and packaging the real-time inventory data according to the target shard database to obtain shard inventory information includes:

[0054] Enqueue the real-time inventory data according to the target shard database to obtain an inventory information queue;

[0055] Performing polling scheduling scanning on the inventory information queue to obtain scheduling information, and extracting a trigger time from the scheduling information;

[0056] Filtering the inventory information queue according to the trigger time to obtain triggered inventory information;

[0057] The triggered inventory information is packaged according to the target shard database to obtain a shard inventory message.

[0058] Optionally, the subscribing to and transmitting the shard inventory message to obtain an updated inventory message includes:

[0059] Enqueue the shard inventory message to obtain a subscription inventory message queue;

[0060] Performing a message persistence operation on the subscription inventory message queue to obtain a persistent inventory message queue;

[0061] Subscribing to the persistent inventory message queue to obtain a subscribed inventory message;

[0062] The subscribed inventory message is asynchronously transmitted to obtain an updated inventory message.

[0063] In order to solve the above problems, the present invention also provides an overseas warehouse WMS inventory monitoring system for high concurrency periods, the system comprising:

[0064] The inventory analysis module is used to extract inventory time series features from pre-acquired historical inventory data, and perform feature splitting and feature weighted regression operations on the inventory time series features to obtain analysis inventory data, wherein the extraction of inventory time series features from the pre-acquired historical inventory data includes: extracting data timestamps from the historical inventory data, and performing time series sorting on the pre-acquired historical inventory data according to the data timestamp to obtain a historical data sequence; extracting inventory features from the historical data sequence to obtain a historical inventory feature sequence; performing time series recursive convolution on the historical inventory feature sequence to obtain short-term inventory features; performing time series jump convolution on the historical inventory feature sequence to obtain long-term inventory features; performing attention encoding on the historical data sequence according to the attention mechanism to obtain attention inventory features; and performing feature fusion on the short-term inventory features, the long-term inventory features, and the attention inventory features using the following time series connection algorithm to obtain inventory time series features: in, The inventory timing characteristics are The eigenvalues ​​of the time steps, is the feature weight of the short-term inventory feature, The short-term inventory characteristics are The eigenvalues ​​of the time steps, , is the time step index, is the number of hidden units that the long-term inventory feature skips when performing temporal skip convolution, The long-term inventory characteristics are The feature weights at the time step, The long-term inventory characteristics are The eigenvalues ​​of the time steps, is the attention inventory feature in The eigenvalues ​​of the time steps, is the preset bias coefficient;

[0065] A clustering and sharding module is used to extract inventory features and cluster time series features of the analyzed inventory data to obtain inventory feature sets, and generate inventory sharding rules based on the inventory feature sets;

[0066] The data entry module is used to scan and identify the barcodes of the preset warehouse commodities to obtain real-time inventory data;

[0067] A shard transmission module, used to perform inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain shard inventory messages, and to subscribe to and transmit the shard inventory messages to obtain updated inventory messages;

[0068] The data updating module is used to update the preset historical inventory monitoring data using the inventory update message to obtain real-time inventory monitoring data.

[0069] The present invention can capture short-term fluctuations, long-term trends and global correlations by extracting inventory time series features, making time series modeling more accurate. By performing feature splitting and feature weighted regression operations, inventory data can be analyzed, and the accuracy of inventory data analysis in future time periods can be improved through a gradient optimization algorithm. By performing inventory feature extraction and time series feature clustering, inventory goods can be clustered according to the characteristics of the goods and characteristics such as inventory turnover rate, thereby facilitating inventory sharding. By generating inventory sharding rules according to the inventory feature set, inventory goods data can be sharded according to the read and write frequency and product characteristics, thereby improving the flexibility of the system, reducing the waiting time of the system, and improving the stability of the inventory monitoring system.

[0070] By scanning and identifying the barcodes of goods, the efficiency of identifying goods entering or leaving the warehouse can be improved, thereby improving the efficiency of inventory monitoring. By sharding and polling the inventory data, the real-time inventory data can be cached at high speed, and the data can be temporarily stored in the memory to reduce the pressure on the system. By subscribing and transmitting, the stability of the inventory data when it is updated can be guaranteed, the fault tolerance of the system can be improved, and the robustness of inventory monitoring can be improved. Therefore, the overseas warehouse WMS inventory monitoring method and system proposed by the present invention for high concurrency periods can solve the problem of low robustness when performing inventory monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 A flowchart of a method for monitoring overseas warehouse WMS inventory during a high concurrency period provided by an embodiment of the present invention;

[0072] Figure 2 A schematic diagram of a flow chart of a method for generating inventory sharding rules provided by an embodiment of the present invention;

[0073] Figure 3 A schematic diagram of a flow chart of a method for generating a shard inventory message provided by an embodiment of the present invention;

[0074] Figure 4 A functional module diagram of an overseas warehouse WMS inventory monitoring system for high concurrency periods provided by an embodiment of the present invention;

[0075] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0076] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0077] The embodiment of the present application provides a WMS inventory monitoring method for overseas warehouses during high concurrency periods. The execution subject of the WMS inventory monitoring method for overseas warehouses during high concurrency periods includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the WMS inventory monitoring method for overseas warehouses during high concurrency periods can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0078] Reference Figure 1 FIG. 1 is a flow chart of a method for monitoring overseas warehouse WMS inventory during a high concurrency period provided by an embodiment of the present invention. In this embodiment, the method for monitoring overseas warehouse WMS inventory during a high concurrency period includes:

[0079] S1. Extracting inventory time series features from pre-acquired historical inventory data, and performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analytical inventory data.

[0080] In detail, the historical inventory data refers to the inventory data of various types of goods in overseas warehouses entered in the past time period. The historical inventory data can be obtained through the overseas warehouse management system (Warehouse Management System, referred to as WMS). The overseas warehouse management system is a software system used to manage and control the daily operations of overseas warehouses, helping warehouse managers to effectively track inventory, optimize warehouse layout, manage orders, and handle tasks such as shipping and receiving. The historical inventory data includes data such as the location, name, price and size of various commodities.

[0081] In the embodiment of the present invention, extracting inventory time series features from pre-acquired historical inventory data includes:

[0082] Extracting a data timestamp from the historical inventory data, and performing time-series sorting on the pre-acquired historical inventory data according to the data timestamp to obtain a historical data sequence;

[0083] Extracting inventory features from the historical data sequence to obtain a historical inventory feature sequence;

[0084] Performing time series recursive convolution on the historical inventory feature sequence to obtain short-term inventory features;

[0085] Performing time-series jump convolution on the historical inventory feature sequence to obtain long-term inventory features;

[0086] Attention encoding is performed on the historical data sequence according to the attention mechanism to obtain an attention inventory feature;

[0087] The short-term inventory feature, the long-term inventory feature and the attention inventory feature are fused to obtain an inventory time series feature.

[0088] In detail, the data timestamp refers to the timestamp corresponding to the entry of each data in the historical inventory data. The format of each historical data in the historical data sequence is, for example, "Men's clothing, European region, outbound, M size". The inventory feature extraction refers to data standardization and data vectorization operations on each historical data in the historical data sequence.

[0089] Specifically, the following time series connection algorithm is used to perform feature fusion to obtain inventory time series features: in, The inventory timing characteristics are The eigenvalues ​​of the time steps, is the feature weight of the short-term inventory feature, The short-term inventory characteristics are The eigenvalues ​​of the time steps, , is the time step index, is the number of hidden units that the long-term inventory feature skips when performing temporal skip convolution, The long-term inventory characteristics are The feature weights at the time step, The long-term inventory characteristics are The eigenvalues ​​of the time steps, is the attention inventory feature in The eigenvalues ​​of the time steps, is the preset bias coefficient.

[0090] Specifically, the gate structure of the pre-trained long-short time series neural network can be used to perform the time series recursive convolution and the time series jump convolution. The time series connection algorithm can simultaneously capture short-term fluctuations, long-term trends and global correlations, making time series modeling more accurate, and can dynamically adjust the contribution of different features in the final result according to the characteristics of the data, thereby achieving more flexible time series modeling. The analyzed inventory data refers to the inventory data that needs to be entered into the warehouse management system in the future time period analyzed based on the historical inventory data.

[0091] In detail, performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analyzed inventory data includes:

[0092] Performing feature splitting on the inventory time series feature to obtain a split time series feature set;

[0093] Filtering a split feature node set from a preset split point set according to the split time series feature set;

[0094] Generate a time series feature decision tree set according to the split feature node set;

[0095] Performing inventory decision on the split feature node set according to the time series feature decision tree set to obtain a decision inventory data set;

[0096] A weighted regression is performed on the decision inventory data set to obtain analysis inventory data.

[0097] In detail, the feature splitting refers to dividing the feature space into multiple regions according to the size of the feature value of the inventory time series feature, and making the split feature node set corresponding to each feature region have the smallest mean square error. The split point set can be a split point set of a pre-trained XGBoost model, and the XGBoost model is a model trained with a large amount of input being time series features and outputting training set data for analyzing inventory data.

[0098] Specifically, each time series feature decision tree in the time series feature decision tree set is obtained by splitting each split feature node in the split feature node set according to the optimal split result, and the time series feature decision tree set can be generated by using the XGBoost model.

[0099] In detail, the inventory decision refers to taking the predicted value of the leaf node corresponding to each split feature node in the split feature node set in the time series feature decision tree set as the decision inventory data, and the weighted regression refers to weighted accumulation of each decision inventory data in the decision inventory data set, and performing regression analysis on the accumulated results to obtain analysis inventory data.

[0100] In the embodiment of the present invention, by extracting inventory time series features, short-term fluctuations, long-term trends and global correlations can be captured simultaneously, making time series modeling more accurate. By performing feature splitting and feature weighted regression operations, analyzed inventory data is obtained, and the accuracy of inventory data analysis in future time periods can be improved through a gradient optimization algorithm.

[0101] S2. Extract inventory features and cluster time series features of the analyzed inventory data to obtain inventory feature sets, and generate inventory segmentation rules based on the inventory feature sets.

[0102] In detail, each inventory feature class in the inventory feature class set corresponds to the inventory feature of a group of commodity inventory data in the analyzed inventory data.

[0103] In the embodiment of the present invention, the inventory feature extraction and time series feature clustering of the analyzed inventory data to obtain an inventory feature set includes:

[0104] Performing inventory data splitting on the analysis inventory data to obtain an analysis data set, and performing inventory feature extraction on the analysis data set to obtain a primary analysis inventory feature set;

[0105] Performing timestamp extraction and timestamp standardization operations on the analysis data set to obtain an analysis inventory time feature set;

[0106] Generate an analysis inventory feature set according to the analysis inventory time feature set and the primary analysis inventory feature set;

[0107] Splitting all analysis inventory features in the analysis inventory feature set into a plurality of primary analysis inventory feature groups, and randomly selecting a primary central inventory feature for each primary analysis inventory feature group;

[0108] Calculate the inventory feature distance between each primary center inventory feature and each analysis inventory feature;

[0109] Regrouping all analysis inventory features according to the inventory feature distance to obtain multiple secondary analysis inventory feature groups;

[0110] Calculating the secondary center inventory features corresponding to each secondary analysis inventory feature group, and calculating the center feature distance between each secondary center inventory feature and the corresponding primary center inventory feature;

[0111] Each secondary analysis inventory feature group is iteratively updated according to all center feature distances to obtain an inventory feature cluster.

[0112] In detail, the inventory data splitting refers to splitting the analysis inventory data into an analysis data set consisting of multiple inventory data, wherein each analysis data in the analysis data set corresponds to the inventory data, and the method of extracting inventory features is consistent with the method in the above step S1, which will not be repeated here.

[0113] Specifically, the timestamp standardization refers to updating all timestamps corresponding to the analysis data set to a form with a mean of 0 and a standard deviation of 1, and generating the analysis inventory feature set according to the analysis inventory time feature set and the primary analysis inventory feature set refers to splicing each analysis inventory time feature of the analysis inventory time feature set as a one-dimensional feature of the primary analysis inventory feature set onto each primary analysis inventory feature of the primary analysis inventory feature set to obtain the analysis inventory feature set.

[0114] In detail, the inventory feature distance is calculated using the following inventory feature distance algorithm: in, It refers to analyzing inventory characteristics and primary center inventory characteristics The inventory feature distance between is the exponential function symbol, is the inner product symbol, is the modulo symbol, refers to the analyzed inventory characteristics, refers to the primary center inventory characteristics, is the preset smoothness coefficient, is the preset threshold coefficient, is the transpose symbol, It refers to analyzing inventory characteristics and primary center inventory characteristics The covariance matrix of is the preset regularization coefficient, is the identity matrix.

[0115] Specifically, by using the inventory feature distance algorithm to calculate the inventory feature distance between each primary center inventory feature and each analysis inventory feature, the distance of the features can be determined while retaining the correlation between the features, thereby improving the efficiency of feature clustering.

[0116] In detail, regrouping all analysis inventory features according to the inventory feature distance to obtain multiple secondary analysis inventory feature groups means allocating the analysis inventory features to the analysis inventory feature group where the primary center inventory feature with the closest inventory feature distance is located to obtain secondary analysis inventory feature groups.

[0117] Specifically, calculating the secondary center inventory features corresponding to each secondary analysis inventory feature group refers to calculating the mean features of each secondary analysis inventory feature of the secondary analysis inventory feature group, and the center feature distance refers to the inventory feature distance between the secondary center inventory feature and the corresponding primary center inventory feature.

[0118] In an embodiment of the present invention, the iterative updating of each secondary analysis inventory feature group according to all center feature distances to obtain an inventory feature set includes: taking the average value of all center feature distances as the standard center feature distance, and judging whether the standard center feature distance is less than a preset distance threshold; if so, taking the secondary center inventory feature as the primary center inventory feature, and returning to the step of calculating the inventory feature distance between each primary center inventory feature and each analysis inventory feature using the following inventory feature distance algorithm; if not, taking the secondary analysis inventory feature groups one by one as inventory feature classes to obtain an inventory feature set.

[0119] Specifically, the inventory sharding rule refers to a rule for sharding transmission and storage of inventory data, for example, allocating inventory data corresponding to inventory feature class A in the inventory feature class set to database node No. 1.

[0120] For details, refer to Figure 2 As shown, the generating of inventory sharding rules according to the inventory feature set includes:

[0121] S21, performing feature number statistics on each inventory feature class in the inventory feature class set to obtain an inventory frequency set;

[0122] S22, performing cluster center screening on the inventory feature set to obtain a standard center inventory feature set;

[0123] S23, generating an inventory standard feature set according to the inventory frequency set and the standard center inventory feature set;

[0124] S24, performing database mapping on the inventory standard feature set to obtain a shard database set;

[0125] S25. Generate inventory sharding rules according to the standard center inventory feature set and the sharding database set.

[0126] Specifically, each inventory frequency in the inventory frequency set corresponds to the number of features in each inventory feature class in the inventory feature class set, and the inventory frequency set corresponds to the number of times of warehousing in and out of each inventory feature class in the inventory feature class set.

[0127] In detail, the inventory standard feature set is obtained by fusing the inventory frequency feature and the standard center inventory feature set, and the database mapping refers to mapping each inventory standard feature in the inventory standard feature set and a database node according to a preset mapping rule, for example, mapping features with higher inventory frequency in the inventory standard features to high-performance database nodes, and mapping features with lower inventory frequency in the inventory standard features to low-performance database nodes.

[0128] Specifically, generating inventory sharding rules according to the standard center inventory feature set and the sharding database set refers to establishing a mapping relationship between each standard center inventory feature in the standard center inventory feature set and each sharding database in the sharding database set.

[0129] In the embodiment of the present invention, by performing inventory feature extraction and time series feature clustering, inventory items can be clustered according to the characteristics of the items and characteristics such as inventory turnover rate, thereby facilitating inventory sharding. By generating inventory sharding rules based on the inventory feature set, inventory item data can be sharded according to the read and write frequency and item characteristics, thereby improving the flexibility of the system, reducing the waiting time of the system, and improving the stability of the inventory monitoring system.

[0130] S3. Scan and identify the barcodes of the preset warehouse products to obtain real-time inventory data.

[0131] In detail, the warehouse goods refer to the inventory goods in the warehouse that need to be monitored. The surface of the inventory goods is covered with corresponding product barcodes, and each product barcode corresponds to an inventory data. The real-time inventory data is the inventory data corresponding to the warehouse goods entered in real time.

[0132] In the embodiment of the present invention, the barcode scanning and barcode recognition of the preset warehouse commodities to obtain real-time inventory data includes:

[0133] Scan the barcodes of the preset warehouse products to obtain the inventory barcode images;

[0134] Performing image denoising and image equalization operations on the inventory barcode image to obtain a balanced inventory barcode image;

[0135] Performing grayscale threshold segmentation on the balanced inventory barcode image to obtain a segmented barcode image;

[0136] Performing a barcode width normalization operation on the segmented barcode image to obtain a standard barcode image;

[0137] Recognize the width information of the standard barcode image to obtain the product barcode;

[0138] Performing code verification on the commodity barcode to obtain verification information;

[0139] The commodity barcode is decoded according to the verification information to obtain real-time inventory data.

[0140] Specifically, a barcode scanner or a smart phone may be used to scan a product barcode, a bilateral filtering algorithm or a median filtering algorithm may be used to perform image denoising, and a histogram equalization algorithm or an adaptive histogram equalization algorithm may be used to perform image equalization.

[0141] Specifically, the barcode width normalization operation can be performed using linear interpolation scaling or frequency domain normalization methods. The width information identification refers to encoding the width information of each barcode in the standard barcode image, for example, encoding a wider barcode as 1 and encoding a narrower barcode as 0.

[0142] In detail, the code verification can be performed using a modulo verification algorithm, and the verification information includes verification success and verification failure. The data decoding of the product barcode code according to the verification information to obtain real-time inventory data means that the data is decoded when the verification information is successfully verified. The data decoding can be performed using algorithms such as EAN-13, UPC-A or Code128.

[0143] In detail, performing grayscale threshold segmentation on the balanced inventory barcode image to obtain a segmented barcode image includes:

[0144] Calculating the grayscale of the balanced inventory barcode image to obtain a barcode grayscale group;

[0145] Selecting barcode gray levels in the barcode gray level group one by one as target gray levels, performing image segmentation on the balanced inventory barcode image according to the target gray levels, and obtaining a target inventory barcode image and a target background barcode image;

[0146] Calculating the image grayscale distance between the target inventory barcode image and the target background barcode image;

[0147] The image grayscale distances of all target grayscale levels in the barcode grayscale group are collected into an image grayscale distance group, and the image grayscale distance with the largest value in the image grayscale distance group is used as the segmentation grayscale distance;

[0148] The target inventory barcode image corresponding to the segmentation grayscale distance is used as the segmentation barcode image.

[0149] In detail, the grayscale calculation can be performed using a weighted average method or an average method, the target inventory barcode image is an image area composed of some pixels whose pixel grayscale is greater than the target grayscale set, and the target background barcode image is an image area composed of some pixels whose pixel grayscale is less than or equal to the target grayscale set.

[0150] Specifically, the image grayscale distance is calculated using the following image grayscale distance algorithm: in, refers to the image grayscale distance, refers to the pixel length of the balanced inventory barcode image, refers to the pixel width of the balanced inventory barcode image, Refers to the total number of pixels of the target inventory barcode image, Refers to the total number of pixels of the target background barcode image, , is the pixel index, is the gray value symbol.

[0151] In detail, the image grayscale distance algorithm may help find the best segmentation threshold or area by maximizing the variance of different regions, thereby improving the accuracy of barcode recognition segmentation.

[0152] In the embodiment of the present invention, by performing commodity barcode scanning and barcode recognition, the recognition efficiency of commodity entry or exit can be improved, thereby improving the efficiency of inventory monitoring.

[0153] S4. Perform inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain sharded inventory messages, and subscribe to and transmit the sharded inventory messages to obtain updated inventory messages.

[0154] In detail, the fragmented inventory message refers to a message used for updating inventory data after being fragmented and packaged. In a high concurrency period, in order to prevent information congestion, the updated inventory data needs to be fragmented and then transmitted.

[0155] In the embodiment of the present invention, the step of performing inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain sharded inventory messages includes:

[0156] Extracting inventory features from the real-time inventory data to obtain real-time inventory features;

[0157] Extracting a standard center inventory feature set from the inventory sharding rule, and calculating a feature distance between the real-time inventory feature and each standard center inventory feature in the standard center inventory feature set to obtain a sharding distance set;

[0158] Selecting the shard distance with the smallest value from the shard distance set as the target shard distance;

[0159] The standard central inventory feature corresponding to the target shard distance in the standard central inventory feature set is used as the shard inventory feature;

[0160] Filtering a target shard database from the inventory sharding rules according to the shard inventory characteristics;

[0161] The real-time inventory data is polled and packaged according to the target shard database to obtain shard inventory information.

[0162] In detail, the method for extracting inventory features is consistent with the method in the above step S2, and the method for calculating feature distance can be a cosine feature distance algorithm or a Euclidean distance algorithm.

[0163] Specifically, refer to Figure 3 As shown, the real-time inventory data is polled and packaged according to the target shard database to obtain shard inventory information, including:

[0164] S31, according to the target shard database, the real-time inventory data is queued to obtain an inventory information queue;

[0165] S32, performing a polling scheduling scan on the inventory information queue to obtain scheduling information, and extracting a trigger time from the scheduling information;

[0166] S33, filtering the inventory information queue according to the trigger time to obtain triggered inventory information;

[0167] S34. Packaging the triggered inventory information according to the target shard database to obtain shard inventory information.

[0168] In detail, the information enqueue refers to storing the real-time inventory data in the information queue corresponding to the target shard database. The inventory information queue can be a Redis queue. The polling scheduled scanning refers to scanning each message in the inventory information queue at a preset time interval. The scheduling information refers to information such as the trigger time assigned to each message.

[0169] Specifically, a polling scheduling scan can be performed through a timer, and a trigger event can be extracted through a regular matching method. The information screening refers to screening out the information whose processing time in the trigger inventory information is at the trigger time as the trigger inventory information. The message packaging refers to packaging the trigger inventory information into a message corresponding to the target shard database to facilitate subsequent asynchronous subscription operations.

[0170] In detail, the subscribing and transmitting the shard inventory message to obtain the updated inventory message includes:

[0171] Enqueue the shard inventory message to obtain a subscription inventory message queue;

[0172] Performing a message persistence operation on the subscription inventory message queue to obtain a persistent inventory message queue;

[0173] Subscribing to the persistent inventory message queue to obtain a subscribed inventory message;

[0174] The subscribed inventory message is asynchronously transmitted to obtain an updated inventory message.

[0175] In detail, the message enqueuing refers to storing the shard inventory message into a preset message queue, the subscription inventory message queue can be a RocketMQ queue, and the message persistence refers to storing each message in the subscription inventory message queue according to the queue topic of the subscription inventory message queue.

[0176] Specifically, the queue subscription refers to subscribing to the persistent inventory message queue according to the target shard database, and the asynchronous transmission refers to transmitting the subscribed inventory message to the target shard database according to the consumer of the target shard database.

[0177] In the embodiment of the present invention, by performing inventory data sharding and polling packaging, real-time inventory data can be cached at high speed, and the data can be temporarily stored in memory to reduce system pressure. By performing subscription transmission, the stability of inventory data updates can be guaranteed, the system's fault tolerance rate can be improved, and the robustness of inventory monitoring can be improved.

[0178] S5. Using the inventory update message, the preset historical inventory monitoring data is updated to obtain real-time inventory monitoring data.

[0179] In detail, the historical inventory monitoring data refers to the inventory data stored by the overseas warehouse management system in the past period of time.

[0180] Specifically, the using the updated inventory message to update the preset historical inventory monitoring data to obtain real-time inventory monitoring data includes: unpacking the updated inventory message to obtain real-time inventory data, and using the real-time inventory data to update the preset historical inventory monitoring data to obtain real-time inventory monitoring data.

[0181] In the embodiment of the present invention, by updating the preset historical inventory monitoring data, the stability of data updating of the inventory monitoring data during a high concurrency period can be ensured, thereby improving the robustness of inventory monitoring.

[0182] The present invention can capture short-term fluctuations, long-term trends and global correlations by extracting inventory time series features, making time series modeling more accurate. By performing feature splitting and feature weighted regression operations, inventory data can be analyzed, and the accuracy of inventory data analysis in future time periods can be improved through a gradient optimization algorithm. By performing inventory feature extraction and time series feature clustering, inventory goods can be clustered according to the characteristics of the goods and characteristics such as inventory turnover rate, thereby facilitating inventory sharding. By generating inventory sharding rules according to the inventory feature set, inventory goods data can be sharded according to the read and write frequency and product characteristics, thereby improving the flexibility of the system, reducing the waiting time of the system, and improving the stability of the inventory monitoring system.

[0183] By scanning and identifying the barcodes of goods, the efficiency of identifying goods entering or leaving the warehouse can be improved, thereby improving the efficiency of inventory monitoring. By sharding and polling the inventory data, the real-time inventory data can be cached at high speed, and the data can be temporarily stored in the memory to reduce the pressure on the system. By subscribing and transmitting, the stability of the inventory data when it is updated can be guaranteed, the fault tolerance of the system can be improved, and the robustness of inventory monitoring can be improved. Therefore, the overseas warehouse WMS inventory monitoring method for high concurrency periods proposed by the present invention can solve the problem of low robustness when performing inventory monitoring.

[0184] like Figure 4 , which is a functional module diagram of an overseas warehouse WMS inventory monitoring system for high concurrency periods provided by an embodiment of the present invention.

[0185] The overseas warehouse WMS inventory monitoring system 100 for high concurrency periods of the present invention can be installed in an electronic device. According to the functions implemented, the overseas warehouse WMS inventory monitoring system 100 for high concurrency periods can include an inventory analysis module 101, a clustering and sharding module 102, a data entry module 103, a sharding transmission module 104 and a data update module 105. The module of the present invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0186] In this embodiment, the functions of each module / unit are as follows:

[0187] The inventory analysis module 101 is used to extract inventory time series features from pre-acquired historical inventory data, and perform feature splitting and feature weighted regression operations on the inventory time series features to obtain analysis inventory data, wherein the extraction of inventory time series features from the pre-acquired historical inventory data includes: extracting data timestamps from the historical inventory data, and performing time series sorting on the pre-acquired historical inventory data according to the data timestamp to obtain a historical data sequence; extracting inventory features from the historical data sequence to obtain a historical inventory feature sequence; performing time series recursive convolution on the historical inventory feature sequence to obtain short-term inventory features; performing time series jump convolution on the historical inventory feature sequence to obtain long-term inventory features; performing attention encoding on the historical data sequence according to the attention mechanism to obtain attention inventory features; and performing feature fusion on the short-term inventory features, the long-term inventory features, and the attention inventory features using the following time series connection algorithm to obtain inventory time series features: in, The inventory timing characteristics are The eigenvalues ​​of the time steps, is the feature weight of the short-term inventory feature, The short-term inventory characteristics are The eigenvalues ​​of the time steps, , is the time step index, is the number of hidden units that the long-term inventory feature skips when performing temporal skip convolution, The long-term inventory characteristics are The feature weights at the time step, The long-term inventory characteristics are The eigenvalues ​​of the time steps, is the attention inventory feature in The eigenvalues ​​of the time steps, is the preset bias coefficient;

[0188] The clustering and sharding module 102 is used to extract inventory features and cluster time series features of the analyzed inventory data to obtain inventory feature sets, and generate inventory sharding rules according to the inventory feature sets;

[0189] The data entry module 103 is used to scan and identify the barcodes of the preset warehouse commodities to obtain real-time inventory data;

[0190] The shard transmission module 104 is used to perform inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain shard inventory messages, and to subscribe to and transmit the shard inventory messages to obtain updated inventory messages;

[0191] The data updating module 105 is used to update the preset historical inventory monitoring data using the inventory update message to obtain real-time inventory monitoring data.

[0192] In detail, each module described in the overseas warehouse WMS inventory monitoring system 100 for high concurrency periods in the embodiment of the present invention is used in the same manner as described above. Figure 1 The same technical means are used as the overseas warehouse WMS inventory monitoring method for high concurrency periods described in, and can produce the same technical effects, so I will not go into details here.

[0193] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0194] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0195] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0196] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0197] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0198] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0199] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or systems stated in the system embodiment can also be implemented by one unit or system through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any specific order.

[0200] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A WMS inventory monitoring method for overseas warehouses during high concurrency periods, characterized in that: The method comprises: Extracting inventory time series features from pre-acquired historical inventory data, and performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analysis inventory data, wherein the extracting inventory time series features from the pre-acquired historical inventory data includes: extracting data timestamps from the historical inventory data, and performing time series sorting on the pre-acquired historical inventory data according to the data timestamps to obtain a historical data sequence; performing inventory feature extraction on the historical data sequence to obtain a historical inventory feature sequence; performing time series recursive convolution on the historical inventory feature sequence to obtain short-term inventory features; performing time series jump convolution on the historical inventory feature sequence to obtain long-term inventory features; performing attention encoding on the historical data sequence according to the attention mechanism to obtain attention inventory features; and performing feature fusion on the short-term inventory features, the long-term inventory features, and the attention inventory features using the following time series connection algorithm to obtain inventory time series features: in, The inventory timing characteristics are The eigenvalues ​​of the time steps, is the feature weight of the short-term inventory feature The short-term inventory characteristics are The eigenvalues ​​of the time steps, , is the time step index, is the number of hidden units that the long-term inventory feature skips when performing temporal skip convolution, The long-term inventory characteristics are The feature weights at the time step, The long-term inventory characteristics are The eigenvalues ​​of the time steps, is the attention inventory feature in The eigenvalues ​​of the time steps, is the preset bias coefficient; Performing inventory feature extraction and time series feature clustering on the analyzed inventory data to obtain an inventory feature set, and generating an inventory sharding rule according to the inventory feature set; Scan and identify the barcodes of the preset warehouse products to obtain real-time inventory data; Performing inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain sharded inventory messages, and subscribing and transmitting the sharded inventory messages to obtain updated inventory messages; The preset historical inventory monitoring data is updated using the inventory update message to obtain real-time inventory monitoring data.

2. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 1 is characterized in that: The performing feature splitting and feature weighted regression operations on the inventory time series features to obtain analyzed inventory data includes: Performing feature splitting on the inventory time series feature to obtain a split time series feature set; Filtering a split feature node set from a preset split point set according to the split time series feature set; Generate a time series feature decision tree set according to the split feature node set; Performing inventory decision on the split feature node set according to the time series feature decision tree set to obtain a decision inventory data set; A weighted regression is performed on the decision inventory data set to obtain analysis inventory data.

3. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 1 is characterized in that: The inventory feature extraction and time series feature clustering are performed on the analyzed inventory data to obtain Inventory feature set, including: Performing inventory data splitting on the analysis inventory data to obtain an analysis data set, and performing inventory feature extraction on the analysis data set to obtain a primary analysis inventory feature set; Performing timestamp extraction and timestamp standardization operations on the analysis data set to obtain an analysis inventory time feature set; Generate an analysis inventory feature set according to the analysis inventory time feature set and the primary analysis inventory feature set; Splitting all analysis inventory features in the analysis inventory feature set into a plurality of primary analysis inventory feature groups, and randomly selecting a primary central inventory feature for each primary analysis inventory feature group; The inventory feature distance between each primary center inventory feature and each analysis inventory feature is calculated using the following inventory feature distance algorithm: in, It refers to analyzing inventory characteristics and primary center inventory characteristics The inventory feature distance between is the exponential function symbol, is the inner product symbol, is the modulo symbol, refers to the analyzed inventory characteristics, refers to the primary center inventory characteristics, is the preset smoothness coefficient, is the preset threshold coefficient, is the transpose symbol, It refers to analyzing inventory characteristics and primary center inventory characteristics The covariance matrix of is the preset regularization coefficient, is the identity matrix; Regrouping all analysis inventory features according to the inventory feature distance to obtain multiple secondary analysis inventory feature groups; Calculating the secondary center inventory features corresponding to each secondary analysis inventory feature group, and calculating the center feature distance between each secondary center inventory feature and the corresponding primary center inventory feature; Each secondary analysis inventory feature group is iteratively updated according to all center feature distances to obtain an inventory feature cluster.

4. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 1 is characterized in that: Generating inventory sharding rules according to the inventory feature set includes: Performing feature number statistics on each inventory feature class in the inventory feature class set to obtain an inventory frequency set; Performing cluster center screening on the inventory feature set to obtain a standard center inventory feature set; Generate an inventory standard feature set according to the inventory frequency set and the standard center inventory feature set; Performing database mapping on the inventory standard feature set to obtain a shard database set; An inventory sharding rule is generated according to the standard central inventory feature set and the sharding database set.

5. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 1 is characterized in that: The barcode scanning and barcode recognition of the preset warehouse goods are performed to obtain real-time inventory data, including: Scan the barcodes of the preset warehouse products to obtain the inventory barcode images; Performing image denoising and image equalization operations on the inventory barcode image to obtain a balanced inventory barcode image; Performing grayscale threshold segmentation on the balanced inventory barcode image to obtain a segmented barcode image; Performing a barcode width normalization operation on the segmented barcode image to obtain a standard barcode image; Recognize the width information of the standard barcode image to obtain the product barcode; Performing code verification on the commodity barcode to obtain verification information; The commodity barcode is decoded according to the verification information to obtain real-time inventory data.

6. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 5 is characterized in that: The step of performing grayscale threshold segmentation on the balanced inventory barcode image to obtain a segmented barcode image includes: Calculating the grayscale of the balanced inventory barcode image to obtain a barcode grayscale group; Selecting barcode gray levels in the barcode gray level group one by one as target gray levels, performing image segmentation on the balanced inventory barcode image according to the target gray levels, and obtaining a target inventory barcode image and a target background barcode image; The image grayscale distance between the target inventory barcode image and the target background barcode image is calculated using the following image grayscale distance algorithm: in, refers to the image grayscale distance, refers to the pixel length of the balanced inventory barcode image, refers to the pixel width of the balanced inventory barcode image, Refers to the total number of pixels of the target inventory barcode image, Refers to the total number of pixels of the target background barcode image, , is the pixel index, is the gray value symbol; The image grayscale distances of all target grayscale levels in the barcode grayscale group are collected into an image grayscale distance group, and the image grayscale distance with the largest value in the image grayscale distance group is used as the segmentation grayscale distance; The target inventory barcode image corresponding to the segmentation grayscale distance is used as the segmentation barcode image.

7. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 1 is characterized in that: The step of performing inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain shard inventory messages includes: Extracting inventory features from the real-time inventory data to obtain real-time inventory features; Extracting a standard center inventory feature set from the inventory sharding rule, and calculating a feature distance between the real-time inventory feature and each standard center inventory feature in the standard center inventory feature set to obtain a sharding distance set; Selecting the shard distance with the smallest value from the shard distance set as the target shard distance; The standard central inventory feature corresponding to the target shard distance in the standard central inventory feature set is used as the shard inventory feature; Filtering a target shard database from the inventory sharding rules according to the shard inventory characteristics; The real-time inventory data is polled and packaged according to the target shard database to obtain shard inventory information.

8. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 7, characterized in that: The step of polling and packaging the real-time inventory data according to the target shard database to obtain shard inventory information includes: Enqueue the real-time inventory data according to the target shard database to obtain an inventory information queue; Performing polling scheduling scanning on the inventory information queue to obtain scheduling information, and extracting a trigger time from the scheduling information; Filtering the inventory information queue according to the trigger time to obtain triggered inventory information; The triggered inventory information is packaged according to the target shard database to obtain a shard inventory message.

9. The overseas warehouse WMS inventory monitoring method for high concurrency period as claimed in claim 1, characterized in that: The subscribing and transmitting the shard inventory message to obtain the updated inventory message includes: Enqueue the shard inventory message to obtain a subscription inventory message queue; Performing a message persistence operation on the subscription inventory message queue to obtain a persistent inventory message queue; Subscribing to the persistent inventory message queue to obtain a subscribed inventory message; The subscribed inventory message is asynchronously transmitted to obtain an updated inventory message.

10. A WMS inventory monitoring system for overseas warehouses during high concurrency periods, characterized in that: The system comprises: The inventory analysis module is used to extract inventory time series features from pre-acquired historical inventory data, and perform feature splitting and feature weighted regression operations on the inventory time series features to obtain analysis inventory data, wherein the extraction of inventory time series features from the pre-acquired historical inventory data includes: extracting data timestamps from the historical inventory data, and performing time series sorting on the pre-acquired historical inventory data according to the data timestamp to obtain a historical data sequence; extracting inventory features from the historical data sequence to obtain a historical inventory feature sequence; performing time series recursive convolution on the historical inventory feature sequence to obtain short-term inventory features; performing time series jump convolution on the historical inventory feature sequence to obtain long-term inventory features; performing attention encoding on the historical data sequence according to the attention mechanism to obtain attention inventory features; and performing feature fusion on the short-term inventory features, the long-term inventory features, and the attention inventory features using the following time series connection algorithm to obtain inventory time series features: in, The inventory timing characteristics are The eigenvalues ​​of the time steps, is the feature weight of the short-term inventory feature, The short-term inventory characteristics are The eigenvalues ​​of the time steps, , is the time step index, is the number of hidden units that the long-term inventory feature skips when performing temporal skip convolution, The long-term inventory characteristics are The feature weights at the time step, The long-term inventory characteristics are The eigenvalues ​​of the time steps, is the attention inventory feature in The eigenvalues ​​of the time steps, is the preset bias coefficient; A clustering and sharding module is used to extract inventory features and cluster time series features of the analyzed inventory data to obtain inventory feature sets, and generate inventory sharding rules based on the inventory feature sets; The data entry module is used to scan and identify the barcodes of the preset warehouse commodities to obtain real-time inventory data; A shard transmission module, used to perform inventory data sharding and polling packaging on the real-time inventory data according to the inventory sharding rule to obtain shard inventory messages, and to subscribe to and transmit the shard inventory messages to obtain updated inventory messages; The data updating module is used to update the preset historical inventory monitoring data using the inventory update message to obtain real-time inventory monitoring data.

Images