Intelligent warehouse management method, device and system suitable for electronic commerce

Abstract

The invention provides an intelligent warehouse management method, device and system suitable for electronic commerce. The method comprises the steps of periodically obtaining goods order information from an e-commerce platform; according to the goods order information, the storage position of the goods to be delivered is adjusted through a pre-constructed storage digital twin model, and delivery strategy information is obtained; acquiring actual execution information corresponding to each link of the commodity shifting process of the sales order; acquiring predetermined execution information corresponding to each link of the sales order in the commodity position adjustment process; judging whether the completion degree of the sales order meets an early warning requirement or not according to the actual execution information and the predetermined execution information; and when the pre-warning requirement is met, stopping warehouse-out of the sales order, and generating pre-warning information and sending the pre-warning information to the user equipment. And the predetermined execution information is compared with the actual execution information, and the order whose sales order completion degree satisfies the early warning condition is alarmed, so that goods which do not meet requirements are prevented from being delivered out of a warehouse, and the overall delivery accuracy is improved.

Description

Technical Field

[0001] The present application relates to the fields of digital twin technology and e-commerce technology, and in particular to an intelligent warehousing management method, device and system suitable for e-commerce. Background Art

[0002] E-commerce refers to commercial transactions conducted electronically, using the Internet to achieve the sale of goods. Related technologies are also developing and improving day by day.

[0003] For example, taking the logistics warehousing management system disclosed in application number CN202211184691.6 as an example, the warehouse monitoring module monitors the remaining warehouse space, warehouse temperature and humidity, and warehouse cargo weight in real time to achieve effective monitoring of the warehouse space, the warehouse environment, and the weight of the cargo on the warehouse space, so as to avoid the situation where the warehouse manager cannot find out in time when there is a large change in the temperature and humidity environment of the warehouse space, resulting in damage to the cargo. At the same time, by displaying the cargo weight in real time, it plays a role in monitoring whether the cargo has been replaced, making the storage of cargo safer. However, its outbound cargo analysis module is only used to display the storage location of the outbound cargo by analyzing the reserved warehouse space corresponding to the required cargo in the inbound monitoring module, which can only facilitate the staff to quickly locate the storage location of the cargo, and cannot meet the warehousing management needs of e-commerce.

[0004] This is because the warehouse management of e-commerce platforms pays more attention to the logistics quality expected by customers of goods, and the logistics quality is reflected in the accuracy of outbound delivery in warehouse management. The warehouse management in related technologies, such as the logistics warehouse management system in the above example, pays more attention to the safe storage of goods and is biased towards storage, and cannot meet the circulation requirements of warehouse management goods in e-commerce. Therefore, this application aims to provide an intelligent warehouse management method, device and system suitable for e-commerce to solve the above problems. Summary of the invention

[0005] The purpose of this application is to provide an intelligent warehouse management method, device and system suitable for e-commerce, so as to meet the needs of customers who pay more attention to the logistics quality expected by goods in the warehouse management of e-commerce.

[0006] The purpose of this application is achieved by the following technical solutions:

[0007] The present application provides an intelligent warehouse management method applicable to e-commerce, the method comprising:

[0008] Step S100, periodically obtaining product order information from an e-commerce platform, wherein the product order information includes sales order data of multiple sales orders obtained by the e-commerce platform within the same time period; the sales order data includes order number, product ID, product quantity, customer information, and expected delivery time;

[0009] Step S200, adjusting the storage location of the goods to be shipped out through a pre-built warehouse digital twin model according to the goods order information, and obtaining shipping strategy information;

[0010] Step S300, for each sales order corresponding to the product order information, controlling the shifting device to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information, and obtaining the actual execution information corresponding to each link of the product shifting process of the sales order;

[0011] Step S400, obtaining the scheduled execution information corresponding to each link of the sales order in the commodity position adjustment process according to the outbound strategy information;

[0012] Step S500, judging whether the completion degree of the sales order meets the early warning requirement according to the actual execution information and the scheduled execution information; when the early warning requirement is met, stopping the delivery of the sales order, and generating early warning information to be sent to the user device.

[0013] Furthermore, the step S200 includes:

[0014] Step S210, putting the deferred sales order data into the product order information, and obtaining the priority sorting results of the corresponding multiple sales orders according to the product order information;

[0015] Step S220, obtaining the total delivery forecast duration of the sales orders corresponding to the product order information; when the total delivery forecast duration is greater than the preset duration, according to the priority sorting result, the sales order data corresponding to the lowest N sales orders are used as new deferred sales order data and removed from the product order information, so that the total delivery forecast duration is less than the preset duration;

[0016] Step S230, taking the multiple sales order data corresponding to the sales orders remaining after the removal and the priority sorting results as the intermediate order information;

[0017] Step S240: According to the intermediate order information, the storage location of the goods to be shipped out is adjusted through the pre-built warehouse digital twin model to obtain the shipping strategy information.

[0018] Furthermore, in step S210, obtaining the priority ranking results of the corresponding multiple sales orders according to the product order information includes:

[0019] Inputting the product order information into a sorting model, and outputting a plurality of sales order data arranged in sequence through the sorting model, wherein the plurality of sales order data arranged in sequence are sorted according to priority;

[0020] The training process of the sorting model is as follows:

[0021] Obtain a first training set, the first training set comprising a plurality of training data, each of the training data comprising a plurality of historical sales order data of a sample and labeled data of sequence information corresponding to the plurality of historical sales order data; each of the historical sales order data comprising an order number, a product ID, a product quantity, and an expected delivery time;

[0022] The preset deep learning model is trained using the first training set to obtain the sorting model.

[0023] Furthermore, the method for obtaining the preset duration in step S220 includes:

[0024] Obtain a duration matching relationship between duration and time period; using the duration matching relationship, calculate the duration corresponding to the time period for periodically obtaining product order information from the e-commerce platform and use it as the preset duration; the duration matching relationship is used to indicate the expected order processing capacity in different time periods.

[0025] Furthermore, the construction method of the warehouse digital twin model includes:

[0026] Obtaining the structural information of the warehouse and the information of the goods in the warehouse; the goods information includes the inventory quantity, inventory location and goods size of each product; the structural information is used to indicate the physical layout of the warehouse, including shelf data and channel data;

[0027] According to the structural information, a three-dimensional model of the warehouse is obtained using three-dimensional modeling software, the product information is associated with the three-dimensional model, and a digital twin model of the warehouse is constructed.

[0028] Further, the actual execution information includes actual execution time information and actual operation information of each link, and the scheduled execution information includes scheduled execution time information and scheduled operation information of each link. Step S500 includes:

[0029] Obtain completion scores based on actual execution time information and scheduled execution time information corresponding to two adjacent links;

[0030] When the completion score meets the preset first completion requirement, the completion scores of the next two adjacent links are compared;

[0031] When the completion score does not meet the first completion requirement, a first warning message is generated and the delivery of the current sales order is stopped; and the first warning message is sent to the user device;

[0032] When the completion score meets the first completion requirement but does not meet the second completion requirement, the corresponding link is used as the target link; the second completion requirement is higher than the first completion requirement;

[0033] When all links have completed position adjustment and the completion scores all meet the first completion requirement, actual operation information of the target link of the sales order is obtained;

[0034] If it is determined that the target link is subjected to repeated operations or the operation is omitted based on the actual operation information and the predetermined operation data in the predetermined operation information corresponding to the target link, a second warning information is generated; the second warning information includes the task information of repeated execution or omitted execution; and the second warning information is sent to the user device.

[0035] In a second aspect, the present application further provides an intelligent warehouse management device suitable for e-commerce, the device comprising:

[0036] An order acquisition module is used to periodically acquire product order information from an e-commerce platform, wherein the product order information includes sales order data of multiple sales orders acquired by the e-commerce platform within the same time period; the sales order data includes order number, product ID, product quantity, customer information, and expected delivery time;

[0037] A strategy acquisition module is used to adjust the storage location of the goods to be shipped out through a pre-built warehouse digital twin model according to the goods order information to obtain the shipping strategy information;

[0038] A position adjustment module, for controlling the shifting device to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information for each sales order corresponding to the goods order information, and obtaining the actual execution information corresponding to each link of the goods shifting process of the sales order;

[0039] A reservation execution information acquisition module, used to acquire the reservation execution information corresponding to each link of the sales order in the commodity position adjustment process according to the outbound strategy information;

[0040] The early warning judgment module is used to judge whether the completion degree of the sales order meets the early warning requirements based on the actual execution information and the scheduled execution information; when the early warning requirements are met, the delivery of the sales order is stopped, and an early warning message is generated and sent to the user device.

[0041] In a third aspect, the present application also provides an intelligent warehouse management system suitable for e-commerce, the system comprising the intelligent warehouse management device described in the second aspect; the system also comprises a user device, the user device is used to receive warning information and perform warning operations.

[0042] Beneficial effects of the present application: Different from the traditional warehouse management system, the warehouse operation of the e-commerce platform is monitored and predicted in real time through the digital twin model. The digital twin model can accurately simulate the actual warehouse environment and provide a more accurate outbound strategy, while the traditional system often lacks such advanced simulation and prediction capabilities. At the same time, compared with the warehouse management technology mentioned above that uses the twin model, the present application directly obtains data from the e-commerce platform and automatically updates the order information according to the preset time interval. It can be considered that the warehouse management system can process the orders within the corresponding time interval within a time period. Once the orders within the current time interval are processed, it will wait for the arrival of the next time interval and repeat the above process, thereby realizing the information processing cycle and ensuring that there will be no information congestion when the warehouse management method processes orders. Periodic processing ensures the continuous updating and processing of order information, so that the warehouse management system can operate stably, and allows the system to perform maintenance, updates or other optimization operations during non-processing time intervals. The time interval can be adjusted according to the processing capacity and business needs to optimize resource allocation, such as shortening the time interval during peak hours to quickly respond to order needs.

[0043] In summary, through periodic information processing cycles, the warehouse management system can achieve efficient and stable order processing while maintaining the ability to respond quickly to changes in e-commerce platform orders. The outbound strategy is obtained based on the product order information including the customer's expected delivery time, and the scheduled execution information is obtained at the same time as the outbound delivery. The scheduled execution information and the actual execution information are compared, and the sales order completion degree meets the early warning conditions. The orders are alarmed to avoid the delivery of goods that do not meet the requirements, thereby improving the overall outbound accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are only used to illustrate specific embodiments and are not considered to limit the present application. In the entire drawings, the same reference symbols represent the same components. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0045] Figure 1 It is a flow chart of an intelligent warehouse management method provided in an embodiment of the present application.

[0046] Figure 2 It is a flow chart of obtaining outbound delivery strategy information provided by an embodiment of the present application.

[0047] Figure 3 It is a flowchart of sending early warning information provided in an embodiment of the present application.

[0048] Figure 4 It is a device schematic diagram of an intelligent warehouse management device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] Below, in conjunction with the accompanying drawings and specific implementation methods, the present application is further described. It should be noted that, under the premise of no conflict, the various embodiments described below or the various technical features can be arbitrarily combined to form a new embodiment. The implementation procedures of the present application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific implementation procedures, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for illustrating the present application, not for limiting the scope of protection of the present application.

[0050] Below, some terms in this application are explained to facilitate understanding by those skilled in the art.

[0051] Machine Learning (ML) is a multi-disciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. A computer program can learn from experience E given a certain type of task T and performance metric P. If its performance in task T can be measured by P, it will improve with experience E. Machine learning specifically studies how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are in all areas of artificial intelligence.

[0052] Deep learning is a special type of machine learning that achieves tremendous power and flexibility by learning to use nested hierarchies of concepts to represent, where each concept is defined in relation to simpler concepts, and more abstract representations are computed in less abstract ways. Machine learning and deep learning typically include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and teaching learning.

[0053] Digital twin technology refers to the establishment of a virtual, corresponding digital model of a physical entity or system in the real world through digital means, so as to realize the monitoring, simulation, prediction and optimization of the physical entity or system. Digital twin consists of two parts: physical entity and digital model. Physical entity is something that exists, and digital model is modeled based on the data and information of physical entity, and performs simulation, prediction, optimization and other operations on physical entity.

[0054] Digital twin technology is developing very fast. In warehouse management, digital twin technology can build a three-dimensional digital model scene of the entire warehouse based on data collection, carry out warehouse layout planning and design, and access the data of various application systems of the industrial Internet to display key information such as warehousing process and storage equipment status in real time; it can use data to drive the digital twin model, monitor the goods and equipment status in the warehouse in real time, optimize the storage and sorting system configuration, and realize the intelligent management and control of warehousing. However, it has not been well promoted in the warehouse management of e-commerce. This is because traditional warehouse management believes that only if the goods are stored reasonably, the safety of goods storage and the accuracy of goods and order matching will inevitably be improved, that is, the optimization of goods storage takes up too much computing power. In specific applications, cloud servers are used to sort orders according to the order of placing orders, and the servers at the local end of the warehouse process the sorted orders in sequence. Although digital twin technology can simulate different storage and sorting strategies, it still depends on the optimization of goods storage, and the attention and optimization of goods out of the warehouse are limited. In the technical solution of this application, the accuracy of goods out of the warehouse corresponding to e-commerce refers to the accuracy of the outbound time and the accuracy of the outbound quality.

[0055] Taking the logistics warehousing management method and system based on digital twins disclosed in application number CN202411389829.5 as an example, the logistics data and structural layout data of the warehouse are obtained to build a warehouse digital twin model; the order data of goods in the warehouse is obtained, the importance of the goods order is analyzed by the projection tracking algorithm, and the goods order is prioritized by the directed graph analysis method; based on the priority sorting results, the storage location of the goods is optimized through the constructed warehouse digital twin model to determine the optimal storage location of each product; according to the optimal storage location of each product, the placement of goods in the warehouse is dynamically adjusted. By integrating the priority sorting results into the warehouse digital twin model and simulating the storage locations of different goods, it can accurately predict and evaluate the warehousing efficiency under different storage strategies, and take minimizing the goods picking path as the objective function, ensuring that while meeting the priority requirements, the storage location of the goods can minimize the picking time and distance, thereby significantly improving the efficiency and response speed of warehousing operations. In other words, its core idea is still to optimize the storage location of goods, and there is limited attention and optimization on the accuracy of goods leaving the warehouse.

[0056] Based on this, the present application provides an intelligent warehouse management method, device and system applicable to e-commerce, which takes into account the need for more attention to the logistics quality expected by customers of goods in warehouse management of e-commerce, obtains the outbound strategy according to the goods order information including the expected delivery time of the customer, and obtains the scheduled execution information at the same time of outbound delivery, compares the scheduled execution information with the actual execution information, and alarms the orders whose sales order completion meets the early warning conditions, so as to avoid the outbound delivery of goods that do not meet the requirements (of the ordering user of the sales order), thereby improving the overall outbound delivery accuracy. The following will first describe the intelligent warehouse management method applicable to e-commerce, and then describe the system, etc.

[0057] Method embodiment.

[0058] See also Figure 1 , Figure 1 It is a flow chart of an intelligent warehouse management method provided in an embodiment of the present application.

[0059] This embodiment provides an intelligent warehouse management method suitable for e-commerce. The intelligent warehouse management method can be run on the edge computing device of the warehouse. The edge computing device is close to the data source and can process data in real time, reduce delays, and improve response speed. The e-commerce platform is used to provide an API for external systems to access and exchange data. The edge computing device interacts with the e-commerce platform through the API to obtain order information. The communicating parties follow data formats and communication protocols such as JSON, XML, REST, and SOAP, and use Kafka message queues to process asynchronous communications. The method includes:

[0060] Step S100, periodically obtaining product order information from an e-commerce platform, wherein the product order information includes sales order data of multiple sales orders obtained by the e-commerce platform within the same time period;

[0061] Periodically obtain product order information from the e-commerce platform. Product order information includes all sales order data generated within a specific time period. Sales order data includes, for example, order number, product ID, product quantity, customer information, and expected delivery time, providing basic data for subsequent warehouse management.

[0062] Step S200: According to the product order information, the storage location of the products to be shipped is adjusted through the pre-built warehouse digital twin model to obtain the shipping strategy information; the pre-built warehouse digital twin model is used to adjust the storage location of the products to be shipped according to the obtained order information to optimize the shipping strategy. The digital twin model is a virtual copy of the warehouse system that can simulate the storage and shipping process of the products, thereby predicting and optimizing the shipping strategy without affecting the actual inventory.

[0063] Step S300, for each sales order corresponding to the product order information, control the shifting device to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information, and obtain the actual execution information corresponding to each link of the goods shifting process of the sales order; according to the outbound strategy information, control the shifting device (such as an automated guided vehicle, a robot, etc.) to adjust the position of the goods in the intelligent warehouse. The outbound strategy information takes into account the outbound priority, the storage location of the goods, and the customer order requirements to achieve automatic adjustment of the goods position and ensure efficient outbound.

[0064] Step S400, according to the outbound strategy information, obtain the scheduled execution information corresponding to each link of the sales order in the commodity position adjustment process; in the commodity position adjustment process, obtain the scheduled execution information of each sales order in each link, the scheduled execution information includes, for example, the scheduled execution start time and the scheduled execution end time, which is used to monitor and evaluate the actual execution situation. In this case, the actual execution information includes the actual execution start time and the actual execution end time.

[0065] Step S500, judging whether the completion degree of the sales order meets the early warning requirement according to the actual execution information and the scheduled execution information; when the early warning requirement is met, stopping the delivery of the sales order, and generating early warning information to be sent to the user device.

[0066] That is, it is possible to determine whether the completion of the sales order meets the early warning requirements based on the actual execution information and the scheduled execution information. If the completion does not meet the preset early warning requirements, the relevant sales order will be automatically stopped from being shipped out, and an early warning message will be generated and sent to the user device so that timely measures can be taken. The user device is, for example, a laptop, desktop computer or tablet computer, and the early warning message is displayed on the user device in the form of a pop-up window, text message or voice prompt.

[0067] The technical solution provided by this application is different from the traditional warehouse management system. It uses a digital twin model to conduct real-time monitoring and predictive analysis of the warehouse operations of the e-commerce platform. The digital twin model can accurately simulate the actual warehouse environment and provide a more accurate outbound strategy, while the traditional system often lacks such advanced simulation and prediction capabilities. At the same time, compared with the related warehouse management technology using the twin model mentioned above, the technical solution of this application directly obtains data from the e-commerce platform and automatically updates the order information according to the preset time interval. It can be considered that the warehouse management system can process the orders within the corresponding time interval within a time period. Once the orders within the current time interval are processed, it will wait for the arrival of the next time interval and repeat the above process, thereby realizing the cycle of information processing and ensuring that there will be no information congestion when the warehouse management method processes orders. Periodic processing ensures the continuous updating and processing of order information, so that the warehouse management system can operate stably, and allows the system to perform maintenance, updates or other optimization operations during non-processing time intervals. The time interval can be adjusted according to the processing capacity and business needs to optimize resource allocation, such as shortening the time interval during peak hours to quickly respond to order needs.

[0068] In summary, through this periodic information processing cycle, the warehouse management system can achieve efficient and stable order processing while maintaining the ability to respond quickly to changes in orders on the e-commerce platform. Taking into account the need for e-commerce warehouse management to pay more attention to the logistics quality expected by customers of goods, the outbound strategy is obtained based on the goods order information including the customer's expected delivery time, and the scheduled execution information is obtained at the same time as the outbound delivery. The scheduled execution information and the actual execution information are compared, and the sales order completion degree meets the warning conditions. The orders are alarmed to avoid the outbound delivery of goods that do not meet the requirements, thereby improving the overall outbound accuracy.

[0069] See also Figure 2 , Figure 2 It is a flow chart of obtaining outbound delivery strategy information provided by an embodiment of the present application.

[0070] In an exemplary embodiment, step S200 includes:

[0071] Step S210, putting the deferred sales order data into the product order information, and obtaining the priority sorting results of the corresponding multiple sales orders according to the product order information;

[0072] Step S220, obtaining the total delivery forecast duration of the sales order corresponding to the product order information; when the total delivery forecast duration is greater than the preset duration, according to the priority sorting result, the sales order data corresponding to the lowest N sales orders are used as new deferred sales order data and removed from the product order information, so that the total delivery forecast duration is less than the preset duration; N is a positive integer not less than 1;

[0073] Step S230, taking the multiple sales order data corresponding to the sales orders remaining after the removal and the priority sorting results as the intermediate order information;

[0074] Step S240: According to the intermediate order information, the storage location of the goods to be shipped out is adjusted through the pre-built warehouse digital twin model to obtain the shipping strategy information.

[0075] First, the deferred sales order data that was not processed in the previous time period is re-added to the product order information of the current time period. Then, all sales orders are sorted according to the preset priority rules (for example, based on the urgency of the order, the importance of the customer, the size of the order, etc.) to determine the order of processing. The total outbound forecast time of all sales orders in the current product order information is calculated, that is, the total time required to complete all orders. If the total time exceeds the preset time threshold, measures will be taken to optimize the outbound process. If the total outbound forecast time exceeds the preset time, the N sales orders with the lowest ranking will be selected according to the priority sorting results, marked as new deferred sales orders, and removed from the currently processed product order information, which can ensure that the total outbound forecast time of the remaining orders is controlled within the preset time. For the remaining sales orders, the pre-built warehouse digital twin model is used to simulate and optimize the storage location of the goods to obtain the best outbound strategy information. The digital twin model can simulate the actual warehouse environment and help predict the impact of different outbound strategies on outbound efficiency.

[0076] Therefore, by prioritizing and intelligently adjusting the order processing sequence, the most important and urgent orders can be processed first, thereby improving the overall outbound efficiency. By evaluating the total outbound forecast time and making adjustments, resources (such as manpower and equipment) can be reasonably allocated. The order processing strategy can be dynamically adjusted according to the actual situation, and orders that exceed the processing capacity can be delayed, increasing the flexibility to deal with emergencies. Using digital twin models for simulation and prediction can achieve more intelligent decision support, reduce human errors and improve decision quality.

[0077] In summary, this technical solution not only improves delivery efficiency and customer satisfaction through intelligent order processing and delivery strategy optimization, but also reduces operating costs, enhances the intelligence and adaptability of the system, and is suitable for the fast-paced e-commerce environment.

[0078] In an exemplary embodiment, the step of obtaining the total delivery forecast duration of the sales order corresponding to the product order information includes:

[0079] Through the pre-built warehouse digital twin model, the warehouse operation of the sales order corresponding to the product order information is simulated to obtain the total delivery forecast time.

[0080] Specifically, the process can be:

[0081] Simulate order processing and input sales order data (including order number, product ID, product quantity, customer information, and expected delivery time) into the digital twin model. Simulate the process of picking workers picking goods in the warehouse according to order information, and record the picking path and time. Simulate the packaging process and estimate the packaging time based on the type of goods and packaging requirements. Simulate the process of loading the packaged goods into the outbound area and registering the outbound inspection, and record the loading and inspection time. Add up the simulation time of each order in each link to get the total outbound time for a single order. Add up the total outbound time of all orders to get the total outbound forecast time.

[0082] In an exemplary embodiment, when the total delivery forecast duration in step S220 is not greater than the preset duration, in step S230, multiple sales order data corresponding to the sales orders that have not been removed and their priority sorting results are used as intermediate order information.

[0083] In an exemplary embodiment, in step S210, obtaining the priority ranking results of the corresponding multiple sales orders according to the product order information includes:

[0084] Inputting the product order information into a sorting model, and outputting a plurality of sales order data arranged in sequence through the sorting model, wherein the plurality of sales order data arranged in sequence are sorted according to priority;

[0085] The training process of the sorting model is as follows:

[0086] Obtain a first training set, the first training set comprising a plurality of training data, each of the training data comprising a plurality of historical sales order data of a sample and labeled data of sequence information corresponding to the plurality of historical sales order data; each of the historical sales order data comprising an order number, a product ID, a product quantity, and an expected delivery time;

[0087] The preset deep learning model is trained using the first training set to obtain the sorting model.

[0088] In this embodiment, the received product order information is input into a pre-trained sorting model. The sorting model can prioritize the orders according to the built-in logic. The sorting model processes the input data and outputs a plurality of sales order data arranged in sequence.

[0089] In a specific application, in order to train the sorting model, a training set containing multiple training data is obtained, each of which contains not only historical sales order data, but also labeled data of sequence information corresponding to these orders, and the labeled data indicates the priority of the order. Use the preset deep learning model and train with the first training set. During the training process, the model learns how to predict the priority of the order based on the input features and labeled data. After training, the deep learning model can capture the patterns in the data and form a sorting model that can accurately prioritize new order data.

[0090] Therefore, the orders can be prioritized quickly and accurately, ensuring that non-urgent orders are processed later when the current order processing capacity cannot meet the order processing requirements. For e-commerce application scenarios, if non-urgent orders are responded to and shipped out faster, it will not bring better feedback to customers. At the same time, since non-urgent orders can be processed later, urgent orders are guaranteed to be processed first, which can bring better feedback to customers.

[0091] In an exemplary embodiment, the method for obtaining the preset duration in step S220 includes:

[0092] Obtain a duration matching relationship between duration and time period; using the duration matching relationship, calculate the duration corresponding to the time period for periodically obtaining product order information from the e-commerce platform and use it as the preset duration; the duration matching relationship is used to indicate the expected order processing capacity in different time periods.

[0093] By analyzing historical data, a time matching relationship can be established, which describes the expected order processing capacity in different time periods. According to the order acquisition cycle of the e-commerce platform, the expected order processing capacity in each cycle is calculated using the time matching relationship, and this is used as the preset time. At the end of each cycle, order information is automatically obtained from the e-commerce platform, and whether the order processing meets expectations is determined based on the preset time.

[0094] Therefore, by accurately matching the preset duration, resource allocation can be optimized, order processing efficiency can be improved, and unnecessary delays can be reduced. The duration matching relationship provides an accurate prediction of order processing capacity in different time periods, which helps to plan and adjust operations in advance.

[0095] As an example, the step of obtaining the duration matching relationship may include:

[0096] Data collection step: collect historical order data. Historical order data refers to the order data for each time period in the past. The order data includes order volume, time period, date data, whether it is a holiday, processing capacity data, promotion activity data, seasonal change data, etc.

[0097] Data processing step: clean the collected historical order data, remove outliers and incomplete records, and extract useful features from the raw data, such as time period, date, holiday labels, etc.

[0098] In the model building step, descriptive statistical analysis is used to identify the distribution and change trend of order processing capacity in different time periods. Time series analysis techniques such as moving average and exponential smoothing are applied to identify the change pattern of order processing capacity over time. ARIMA and seasonal decomposition are combined to predict order processing capacity in different time periods, and a polynomial regression model is used to establish a corresponding relationship for time matching. The polynomial regression model can make predictions based on the linear or nonlinear relationship between order processing capacity and time period, date, holiday and other characteristics.

[0099] In the model evaluation step, the mean square error (MSE) or mean absolute error (MAE) between the predicted value and the actual value is calculated. The training set and the test set are obtained. The training set is used to build and adjust the model, and the test set is used to evaluate the prediction performance of the model. Use the training set data to train the model and adjust the model parameters. Use the trained model to make predictions on the test set. Calculate the evaluation index based on the prediction results of the model and the actual values ​​on the test set. Analyze the results of the evaluation index to determine the prediction accuracy and generalization ability of the model. If the evaluation passes, a polynomial regression model of the duration matching relationship is obtained.

[0100] Through the above steps, the duration matching relationship is obtained, so as to better predict the fluctuation of order volume in different time periods and obtain a reasonable total outbound forecast duration. The duration matching relationship can reflect the pattern and trend of order volume changes over time, and the output preset duration indicates the expected order processing capacity in different time periods.

[0101] In an exemplary embodiment, the warehouse digital twin model is constructed in the following manner:

[0102] Obtaining the structural information of the warehouse and the information of the goods in the warehouse; the goods information includes the inventory quantity, inventory location and goods size of each product; the structural information is used to indicate the physical layout of the warehouse, including shelf data and channel data;

[0103] According to the structural information, a three-dimensional model of the warehouse is obtained using three-dimensional modeling software, the product information is associated with the three-dimensional model, and a digital twin model of the warehouse is constructed.

[0104] For example, use Blender as a 3D modeling software to draw a digital 3D model. Blender has built-in import and export functions for glTF and glb formats, so that the model can be directly used in the digital twin platform. Export the constructed model as a glb file. As a binary format, glb has a relatively small file size. Use BIMFACE to lightweight the model to simplify the model. Import the lightweight model into the digital twin visualization platform. By visually displaying various types of data in the form of columns, lines, and pie charts, users can see important data such as inventory resources, commodity sales rate, and inventory turnover rate at a glance. At the same time, in order to increase the model loading speed and reduce the loading time, you can use the method of separating textures and models, parallel loading, and use tools such as Draco and gltfpack to compress models and textures.

[0105] In this embodiment, it is first necessary to collect the structural information and product information of the warehouse. The structural information includes the physical layout of the warehouse, such as the location, size and aisle layout of the shelves. The product information includes the inventory, specific storage location and size of each commodity. Using 3D modeling software, a 3D model of the warehouse is created based on the collected structural information. The 3D model reflects the physical structure and spatial layout of the warehouse. The product information is associated with the 3D model so that each product has a corresponding position and attributes in the model, that is, data such as inventory, inventory location and product size are embedded into the 3D model. By integrating the structural information and product information, a complete warehouse digital twin model is constructed. The warehouse digital twin model can be considered as a virtual and accurate copy of the warehouse, which can simulate the real-world warehousing operations in the digital world.

[0106] In specific applications, the warehouse digital twin model can be updated in real time, such as the addition of new goods and changes in inventory, to reflect the latest changes in the warehouse. At the same time, the digital twin model can be used to simulate operations, such as the picking, storage and transportation of goods, to optimize warehouse operation processes and improve efficiency.

[0107] As a result, the digital twin model allows simulation and testing without affecting actual warehouse operations, thereby improving the efficiency of warehouse operations. Through simulation and optimization, unnecessary resource waste can be reduced and operating costs can be reduced. The digital twin model provides an experimental platform that enables managers to test different strategies and layouts before implementation, enhancing the scientificity and accuracy of decision-making.

[0108] See also Figure 3 , Figure 3 It is a flowchart of sending early warning information provided in an embodiment of the present application.

[0109] In an exemplary embodiment, the actual execution information includes actual execution time information and actual operation information of each link, and the scheduled execution information includes scheduled execution time information and scheduled operation information of each link. Step S500 includes:

[0110] Step S510, obtaining a completion score according to the actual execution time information and the scheduled execution time information corresponding to two adjacent links;

[0111] Step S520, when the completion score meets the preset first completion requirement, the completion scores of the next two adjacent links are compared;

[0112] Step S530: when the completion score does not meet the first completion requirement, generate a first warning message and stop the delivery of the current sales order; send the first warning message to the user device;

[0113] Step S540: when the completion score meets the first completion requirement but does not meet the second completion requirement, the corresponding link is used as the target link; the second completion requirement is higher than the first completion requirement;

[0114] Step S550, when all links have completed position adjustment and the completion scores all meet the first completion requirement, obtaining actual operation information of the target link of the sales order;

[0115] Step S560: If it is determined that the target link is subjected to repeated operations or the operation is omitted based on the actual operation information and the predetermined operation data in the predetermined operation information corresponding to the target link, a second warning message is generated; the second warning message is sent to the user device; the second warning message includes the task information of repeated execution or omitted execution.

[0116] In this technical solution, the actual execution time information and actual operation information of each link, as well as the corresponding scheduled execution time information and scheduled operation information, are collected. For two adjacent links, the actual execution time is compared with the scheduled execution time to calculate the completion score. If the completion score meets the preset first completion requirement, the completion of the next two adjacent links will continue to be monitored. If the score does not meet the first completion requirement, the first warning information is directly generated and the outbound process of the current sales order is stopped, and the warning information is sent to the user device. At the same time, this technical solution also subdivides the completion requirements. If the completion score meets the first completion requirement but does not meet the higher second completion requirement, the link is marked as a target link for more detailed analysis and monitoring. Once all links have completed the position adjustment and the completion score meets the first completion requirement, the actual operation information of the target link will be obtained. It can be considered that if the target link is empty, the process of subdividing the completion is skipped. In the process of subdividing the completion, the actual operation information of the target link is compared with the scheduled operation information to determine whether there is a repeated operation or an omitted operation. If a repeated or omitted operation is detected, a second warning information is generated, which includes specific task information, and then the warning information is sent to the user device.

[0117] Therefore, through real-time monitoring and scoring, the efficiency and accuracy of order processing can be ensured, and delays and errors can be reduced. Managers can understand the order processing status in real time and improve the transparency of operations. The early warning system enables managers to respond to potential problems in a timely manner and reduce losses. Fast problem response and accurate order processing can improve customer satisfaction. At the same time, it reduces the additional costs caused by delays and errors and reduces operating costs. Through real-time monitoring, scoring and early warning mechanisms, the efficiency and accuracy of e-commerce intelligent warehousing management are improved, the problem response capability is enhanced, and operating costs are reduced. The segmented completion requirements make the early warning more refined, and early warnings can be issued for problems of different degrees. Through real-time monitoring and early warning mechanisms, the accuracy of outbound time and outbound quality is avoided from being affected, which affects the accuracy of goods outbound. Although digital twin technology can simulate different storage and sorting strategies, this solution further optimizes the strategy through real-time monitoring and early warning mechanisms, reduces excessive reliance on goods storage optimization, and improves the accuracy and efficiency of goods outbound.

[0118] In specific applications, sensors and monitoring equipment, such as barcode scanners, RFID readers, cameras, weight sensors, etc., are deployed in various links of warehousing to capture data during the operation in real time. The data from sensors and monitoring equipment are automatically collected through the data acquisition system. The data acquisition system can monitor the operations in the warehouse in real time and record the actual operation status of each link. At the same time, after completing the tasks in each link, the operator enters confirmation information through a mobile device (such as a handheld terminal, a tablet computer, etc.), and the data acquisition system updates the operation status accordingly. The above collected data can be used as actual operation information.

[0119] In an exemplary embodiment, the intelligent warehouse management method can be run on the edge computing device of the warehouse; the links in the storage location adjustment process include: picking and checking link, packaging link, loading link and outbound inspection and registration link.

[0120] It can be considered that the operations in different links are different, and the possibility of repeated operations or omissions is also different. For example, the picking and checking link includes: repeated picking and repeated checking. Repeated picking means that the same order of goods is picked multiple times due to misunderstanding or operational errors. Repeated checking means that when checking the picked goods, the same batch of goods is repeatedly checked without a clear check completion mark. The packaging link includes: repeated packaging and repeated labeling. Repeated packaging means that a product is mistakenly packaged multiple times, such as repacking after mistakenly thinking that the packaging has failed. Repeated labeling means that the logistics label of the product is mistakenly affixed multiple times. The loading link includes repeated loading and repeated distribution. Repeated loading means that the same batch of goods is mistakenly loaded on the transport vehicle multiple times due to recording errors or communication errors. Repeated distribution refers to the situation where the goods are mistakenly distributed to the wrong transport vehicle or transportation route, and then need to be redistributed and loaded. The outbound inspection registration link includes repeated registration, which means that the registration information of the warehouse goods is mistakenly entered multiple times.

[0121] In an exemplary embodiment, step S510 includes:

[0122] The actual execution time information and the scheduled execution time information corresponding to two adjacent links are input into a similarity model to obtain the similarity between the actual execution time information and the scheduled execution time information corresponding to the two links and use it as a completion score, wherein the similarity model is obtained by training using a preset deep learning neural network.

[0123] The training process of the similarity model is as follows:

[0124] A second training set is obtained, where the second training set includes a plurality of training data, and the training data includes first time information, second time information, and a labeling similarity between the first time information and the second time information.

[0125] Inputting the first time information and the second time information in the second training set into the preset deep learning neural network to obtain the predicted similarity between the first time information and the second time information;

[0126] Based on the predicted similarity and the labeled similarity of the first time information and the second time information, a predicted loss value is calculated; and based on the predicted loss value, the parameters of the preset deep learning neural network are updated to obtain the similarity model.

[0127] As an example, the actual execution time information and the scheduled execution time information corresponding to two adjacent links are input into the similarity model, and the similarity between the actual execution time information and the scheduled execution time information corresponding to the two links is 0.93, which is used as the completion score. The first completion requirement is not less than 0.95, and the second completion requirement is not less than 0.99.

[0128] It can be considered that each sales order has actual execution time information and scheduled execution time information in each link of the commodity position adjustment process (such as picking and checking, packaging, loading and outbound inspection registration, etc.), which is used as input data. A similarity model is used to evaluate the similarity between the actual execution time and the scheduled execution time. The similarity model is built based on a deep learning neural network and can process and compare time series data. The similarity model is trained using a preset deep learning neural network, which can learn the complex relationship between the actual and scheduled time information and predict the similarity between them. The similarity value obtained by the similarity model is used as a completion score, which reflects the degree of match between the actual execution and the scheduled plan.

[0129] Therefore, the similarity model obtained through deep learning neural network training can more accurately evaluate the matching degree between actual execution time and scheduled time, thereby improving the accuracy of warehouse management. It can monitor the execution of each link in real time, and timely detect deviations through completion scores, triggering early warning mechanisms to avoid potential errors. The similarity model can quickly process large amounts of data, quickly respond to different orders, and improve overall work efficiency. The deep learning model has good generalization ability, can adapt to different warehouse environments and order processing processes, and has better adaptability and flexibility.

[0130] In a specific application, an intelligent warehouse management method applicable to e-commerce is provided, wherein the intelligent warehouse management method runs on an edge computing device of a warehouse, and the method includes:

[0131] Periodically obtain product order information from an e-commerce platform, the product order information including sales order data of multiple sales orders obtained by the e-commerce platform within the same time period; the sales order data includes order number, product ID, product quantity, customer information and expected delivery time; the sales order data includes: order number SO001, product ID is P1001, product quantity is 2, customer information is Zhang San, and expected delivery time is 10:00 on December 5, 2000; order number SO002, product ID is P1003, product quantity is 3, customer information is Li Si, and expected delivery time is 14:00 on December 6, 2000;

[0132] Putting the deferred sales order data into the product order information, inputting the product order information into a sorting model, and outputting a plurality of sales order data arranged in sequence through the sorting model, wherein the plurality of sales order data arranged in sequence are sorted according to priority;

[0133] Obtain a duration matching relationship between duration and time period; use the duration matching relationship to calculate the duration corresponding to the time period for periodically obtaining product order information from the e-commerce platform and use it as a preset duration; the duration matching relationship is used to indicate the expected order processing capacity in different time periods; when the total delivery forecast duration is greater than the preset duration, based on the priority sorting result, the sales order data corresponding to the lowest-ranked N sales orders are used as new deferred sales order data and removed from the product order information, so that the total delivery forecast duration is less than the preset duration;

[0134] The multiple sales order data corresponding to the remaining sales orders after the removal and the priority sorting results thereof are used as the intermediate order information;

[0135] According to the intermediate order information, the storage location of the goods to be shipped out is adjusted through the pre-built warehouse digital twin model to obtain the shipping strategy information.

[0136] For each sales order corresponding to the product order information, the shifting device is controlled to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information, and the actual execution information corresponding to each link of the product shifting process of the sales order is obtained; the links in the storage position adjustment process include: picking and checking link, packaging link, loading link and outbound inspection and registration link;

[0137] Acquire the scheduled execution information corresponding to each link of the product location adjustment process of the sales order according to the outbound strategy information;

[0138] The actual execution information includes the actual execution time information and the actual operation information of each link, and the scheduled execution information includes the scheduled execution time information and the scheduled operation information of each link; the actual execution time information and the scheduled execution time information corresponding to two adjacent links are input into a similarity model to obtain the similarity between the actual execution time information and the scheduled execution time information corresponding to the two links and use it as a completion score, wherein the similarity model is obtained by training using a preset deep learning neural network;

[0139] When the completion score meets the preset first completion requirement, the completion scores of the next two adjacent links are compared;

[0140] When the completion score does not meet the first completion requirement, a first warning message is generated and the delivery of the current sales order is stopped; and the first warning message is sent to the user device;

[0141] When the completion score meets the first completion requirement but does not meet the second completion requirement, the corresponding link is used as the target link; the second completion requirement is higher than the first completion requirement;

[0142] When all links have completed position adjustment and the completion scores all meet the first completion requirement, actual operation information of the target link of the sales order is obtained;

[0143] If it is determined that the target link is subjected to repeated operations or the operation is omitted based on the actual operation information and the predetermined operation data in the predetermined operation information corresponding to the target link, a second warning information is generated; the second warning information includes the task information of repeated execution or omitted execution; and the second warning information is sent to the user device.

[0144] Device Embodiments.

[0145] See also Figure 4 , Figure 4 It is a device schematic diagram of an intelligent warehouse management device provided in an embodiment of the present application.

[0146] This embodiment provides an intelligent warehouse management device suitable for e-commerce, and its specific implementation method is consistent with the implementation method and technical effects recorded in the embodiment of the above method, and some contents are not repeated here. The device includes:

[0147] An order acquisition module is used to periodically acquire product order information from an e-commerce platform, wherein the product order information includes sales order data of multiple sales orders acquired by the e-commerce platform within the same time period; the sales order data includes order number, product ID, product quantity, customer information, and expected delivery time;

[0148] A strategy acquisition module is used to adjust the storage location of the goods to be shipped out through a pre-built warehouse digital twin model according to the goods order information to obtain the shipping strategy information;

[0149] A position adjustment module, for controlling the shifting device to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information for each sales order corresponding to the goods order information, and obtaining the actual execution information corresponding to each link of the goods shifting process of the sales order;

[0150] A reservation execution information acquisition module, used to acquire the reservation execution information corresponding to each link of the sales order in the commodity position adjustment process according to the outbound strategy information;

[0151] The early warning judgment module is used to judge whether the completion degree of the sales order meets the early warning requirements based on the actual execution information and the scheduled execution information; when the early warning requirements are met, the delivery of the sales order is stopped, and an early warning message is generated and sent to the user device.

[0152] In an exemplary embodiment, the policy acquisition module includes:

[0153] A sorting unit, used to put the deferred sales order data into the product order information, and obtain the priority sorting results of the corresponding multiple sales orders according to the product order information;

[0154] A duration acquisition unit is used to acquire the total delivery forecast duration of the sales order corresponding to the product order information; when the total delivery forecast duration is greater than a preset duration, the sales order data corresponding to the lowest-ranked N sales orders are used as new deferred sales order data and removed from the product order information according to the priority sorting result, so that the total delivery forecast duration is less than the preset duration;

[0155] An intermediate order acquisition unit, used to take multiple sales order data corresponding to the sales orders remaining after the removal and the priority sorting results thereof as intermediate order information;

[0156] The strategy acquisition unit is used to adjust the storage location of the goods to be shipped out through a pre-built warehouse digital twin model according to the intermediate order information to obtain the shipping strategy information.

[0157] In an exemplary embodiment, the sorting unit includes:

[0158] A model input subunit, used for inputting the product order information into a sorting model, and outputting a plurality of sales order data arranged in sequence through the sorting model, wherein the plurality of sales order data arranged in sequence are sorted according to priority;

[0159] The training process of the sorting model is as follows:

[0160] Obtain a first training set, the first training set comprising a plurality of training data, each of the training data comprising a plurality of historical sales order data of a sample and labeled data of sequence information corresponding to the plurality of historical sales order data; each of the historical sales order data comprising an order number, a product ID, a product quantity, and an expected delivery time;

[0161] The preset deep learning model is trained using the first training set to obtain the sorting model.

[0162] In an exemplary embodiment, the method for obtaining the preset duration includes:

[0163] Obtain a duration matching relationship between duration and time period; using the duration matching relationship, calculate the duration corresponding to the time period for periodically obtaining product order information from the e-commerce platform and use it as the preset duration; the duration matching relationship is used to indicate the expected order processing capacity in different time periods.

[0164] In an exemplary embodiment, the warehouse digital twin model is constructed in the following manner:

[0165] Obtaining the structural information of the warehouse and the information of the goods in the warehouse; the goods information includes the inventory quantity, inventory location and goods size of each product; the structural information is used to indicate the physical layout of the warehouse, including shelf data and channel data;

[0166] According to the structural information, a three-dimensional model of the warehouse is obtained using three-dimensional modeling software, the product information is associated with the three-dimensional model, and a digital twin model of the warehouse is constructed.

[0167] In an exemplary embodiment, the actual execution information includes the actual execution time information and actual operation information of each link, the scheduled execution information includes the scheduled execution time information and scheduled operation information of each link, and the early warning judgment module includes:

[0168] A score acquisition unit, used to acquire a completion score according to actual execution time information and scheduled execution time information corresponding to two adjacent links;

[0169] A score comparison unit, used for comparing the scores of the next two adjacent links when the completion score meets the preset first completion requirement;

[0170] A first information production unit is used to generate a first warning message and stop the delivery of the current sales order when the completion score does not meet the first completion requirement; and send the first warning message to the user device;

[0171] A link confirmation unit, configured to use the corresponding link as a target link when the completion score meets a first completion requirement but does not meet a second completion requirement; the second completion requirement is higher than the first completion requirement;

[0172] An actual operation information acquisition unit, configured to acquire actual operation information of a target link of the sales order when all links have completed position adjustment and completion scores have met the first completion requirement;

[0173] The second information production unit is used to generate a second warning information if it is judged that the target link has been subjected to repeated operations or the operation has been omitted based on the actual operation information and the predetermined operation data in the predetermined operation information corresponding to the target link; the second warning information includes the task information of repeated execution or omitted execution; and the second warning information is sent to the user device.

[0174] In an exemplary embodiment, the steps in the storage location adjustment process include: picking and checking step, packaging step, loading step and outbound inspection and registration step.

[0175] In an exemplary embodiment, the score acquisition unit includes:

[0176] The scoring acquisition subunit is used to input the actual execution time information and the scheduled execution time information corresponding to two adjacent links into a similarity model, obtain the similarity between the actual execution time information and the scheduled execution time information corresponding to the two links and use it as a completion score, wherein the similarity model is obtained by training using a preset deep learning neural network.

[0177] System Embodiments.

[0178] This embodiment provides an intelligent warehouse management system suitable for e-commerce, the system includes the intelligent warehouse management device described in the device embodiment; the system also includes a user device, the user device is used to receive warning information and perform warning operations. Its specific implementation method is consistent with the implementation method and technical effects recorded in the embodiment of the above method, and some contents are not repeated here.

[0179] The terms "first", "second", "third", etc. (if any) in the specification and claims of the present application and the drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "corresponding to" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0180] This application is explained from the perspectives of purpose of use, effectiveness, progress and novelty, and has met the functional enhancement and usage requirements emphasized by the Patent Law. The above description and drawings of this application are only the preferred embodiments of this application, and are not intended to limit this application. Therefore, all structures, devices, features, etc. that are similar or identical to this application, that is, all equivalent replacements or modifications made in accordance with the scope of the patent application of this application, should fall within the scope of protection of the patent application of this application.

Claims

1. An intelligent warehouse management method suitable for e-commerce, characterized in that: The method comprises: Step S100, periodically obtaining product order information from an e-commerce platform, wherein the product order information includes sales order data of multiple sales orders obtained by the e-commerce platform within the same time period; the sales order data includes order number, product ID, product quantity, customer information, and expected delivery time; Step S200, adjusting the storage location of the goods to be shipped out through a pre-built warehouse digital twin model according to the goods order information, and obtaining shipping strategy information; Step S300, for each sales order corresponding to the product order information, controlling the shifting device to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information, and obtaining the actual execution information corresponding to each link of the product shifting process of the sales order; Step S400, obtaining the scheduled execution information corresponding to each link of the sales order in the commodity position adjustment process according to the outbound strategy information; Step S500, judging whether the completion degree of the sales order meets the early warning requirement according to the actual execution information and the scheduled execution information; when the early warning requirement is met, stopping the delivery of the sales order, and generating early warning information to be sent to the user device.

2. The intelligent warehouse management method applicable to e-commerce according to claim 1 is characterized in that: The step S200 includes: Step S210, putting the deferred sales order data into the product order information, and obtaining the priority sorting results of the corresponding multiple sales orders according to the product order information; Step S220, obtaining the total delivery forecast duration of the sales orders corresponding to the product order information; when the total delivery forecast duration is greater than the preset duration, according to the priority sorting result, the sales order data corresponding to the lowest N sales orders are used as new deferred sales order data and removed from the product order information, so that the total delivery forecast duration is less than the preset duration; Step S230, taking the multiple sales order data corresponding to the sales orders remaining after the removal and the priority sorting results as the intermediate order information; Step S240: According to the intermediate order information, the storage location of the goods to be shipped out is adjusted through the pre-built warehouse digital twin model to obtain the shipping strategy information.

3. The intelligent warehouse management method applicable to e-commerce according to claim 2 is characterized in that: In step S210, obtaining the priority ranking results of the corresponding multiple sales orders according to the product order information includes: Inputting the product order information into a sorting model, and outputting a plurality of sales order data arranged in sequence through the sorting model, wherein the plurality of sales order data arranged in sequence are sorted according to priority; The training process of the sorting model is as follows: Obtain a first training set, the first training set comprising a plurality of training data, each of the training data comprising a plurality of historical sales order data of a sample and labeled data of sequence information corresponding to the plurality of historical sales order data; each of the historical sales order data comprising an order number, a product ID, a product quantity, and an expected delivery time; The preset deep learning model is trained using the first training set to obtain the sorting model.

4. The intelligent warehouse management method applicable to e-commerce according to claim 3 is characterized in that: The method for obtaining the preset duration in step S220 includes: Obtain a duration matching relationship between duration and time period; using the duration matching relationship, calculate the duration corresponding to the time period for periodically obtaining product order information from the e-commerce platform and use it as the preset duration; the duration matching relationship is used to indicate the expected order processing capacity in different time periods.

5. The intelligent warehouse management method applicable to e-commerce according to claim 2 is characterized in that: The construction method of the warehouse digital twin model includes: Obtaining warehouse structure information and product information in the warehouse; the product information includes the inventory quantity, inventory location and product size of each product; the structure information is used to indicate the physical layout of the warehouse, including shelf data and channel data; According to the structural information, a three-dimensional model of the warehouse is obtained using three-dimensional modeling software, the product information is associated with the three-dimensional model, and a digital twin model of the warehouse is constructed.

6. The intelligent warehouse management method applicable to e-commerce according to claim 1, characterized in that: The actual execution information includes the actual execution time information and the actual operation information of each link, and the scheduled execution information includes the scheduled execution time information and the scheduled operation information of each link. The step S500 includes: Step S510, obtaining a completion score according to the actual execution time information and the scheduled execution time information corresponding to two adjacent links; Step S520, when the completion score meets the preset first completion requirement, the completion scores of the next two adjacent links are compared; Step S530: when the completion score does not meet the first completion requirement, generate a first warning message and stop the delivery of the current sales order; send the first warning message to the user device; Step S540: when the completion score meets the first completion requirement but does not meet the second completion requirement, the corresponding link is used as the target link; the second completion requirement is higher than the first completion requirement; Step S550, when all links have completed position adjustment and the completion scores all meet the first completion requirement, obtaining actual operation information of the target link of the sales order; Step S560: If it is determined that the target link is subjected to repeated operations or the operations are omitted based on the actual operation information and the predetermined operation data in the predetermined operation information corresponding to the target link, a second warning information is generated; the second warning information includes the task information of repeated execution or omitted execution; and the second warning information is sent to the user device.

7. The intelligent warehouse management method applicable to e-commerce according to claim 6 is characterized in that: The intelligent warehouse management method is run on the edge computing device of the warehouse; The links in the storage location adjustment process include: picking and checking, packaging, loading and outbound inspection and registration.

8. The intelligent warehouse management method applicable to e-commerce according to claim 7 is characterized in that: The step S510 includes: The actual execution time information and the scheduled execution time information corresponding to two adjacent links are input into a similarity model to obtain the similarity between the actual execution time information and the scheduled execution time information corresponding to the two links and use it as a completion score, wherein the similarity model is obtained by training using a preset deep learning neural network.

9. An intelligent warehouse management device suitable for e-commerce, characterized in that: The device comprises: An order acquisition module is used to periodically acquire product order information from an e-commerce platform, wherein the product order information includes sales order data of multiple sales orders acquired by the e-commerce platform within the same time period; the sales order data includes order number, product ID, product quantity, customer information, and expected delivery time; A strategy acquisition module is used to adjust the storage location of the goods to be shipped out through a pre-built warehouse digital twin model according to the goods order information to obtain the shipping strategy information; A position adjustment module, for controlling the shifting device to adjust the position of the goods in the intelligent warehouse according to the outbound strategy information for each sales order corresponding to the goods order information, and obtaining the actual execution information corresponding to each link of the goods shifting process of the sales order; A reservation execution information acquisition module, used to acquire the reservation execution information corresponding to each link of the sales order in the commodity position adjustment process according to the outbound strategy information; The early warning judgment module is used to judge whether the completion degree of the sales order meets the early warning requirements based on the actual execution information and the scheduled execution information; when the early warning requirements are met, the delivery of the sales order is stopped, and an early warning message is generated and sent to the user device.

10. An intelligent warehouse management system suitable for e-commerce, characterized in that: The system includes the intelligent warehouse management device according to claim 9; the system also includes a user device, and the user device is used to receive warning information and perform warning operations.

Images