A shipping method and system for intelligent multi-warehouse e-commerce business

Abstract

The present application provides an intelligent multi-warehouse shipping method and system for e-commerce business, wherein the present application generates a global inventory view based on user order information across e-commerce platforms and the corresponding logistics fulfillment requirements, and then dynamically adjusts it through preset priority rules, and generates a sorting priority sequence through the dynamically adjusted inventory allocation strategy, and then performs spectral feature matching on the mixed goods in each warehouse according to the sorting priority sequence and SKU identification information, completes the goods sorting and generates goods identification information in combination with the target warehouse lighting environment parameters corresponding to the geographical distribution of the inventory, and then uses a distributed inventory sharing model to generate cross-warehouse allocation decisions, and finally generates multi-warehouse collaborative shipping instructions based on the cross-warehouse allocation decisions and the goods identification information; the technical solution provided by the present application realizes intelligent management of the entire process from order processing to final delivery of goods, improves the efficiency of inventory management and logistics scheduling, and ensures the accuracy and timeliness of order processing.

Description

Technical Field

[0001] The present application relates to the technical field of e-commerce order management, and in particular to a method and system for intelligent multi-warehouse delivery of e-commerce business. Background Art

[0002] With the rapid development of e-commerce, the number of user orders across e-commerce platforms has exploded, placing higher demands on logistics fulfillment services.

[0003] Existing multi-warehouse shipping solutions for e-commerce businesses on the market primarily rely on traditional inventory management systems, which typically allocate inventory and schedule shipping sequences based on static rules. For example, some systems may prioritize shipments to the warehouse with the closest inventory, or perform simple inventory matching based on the product's SKU identifier. In addition, some advanced systems also introduce certain automated equipment to accelerate the sorting process. Although these methods have improved shipping efficiency to a certain extent, most lack flexibility and intelligence, and are unable to adapt to market changes and customer needs in real time, especially when faced with large amounts of heterogeneous data and complex logistics networks. Summary of the Invention

[0004] The present application provides an intelligent multi-warehouse shipping method and system for e-commerce business, which is used to solve the problems in the existing technology of incomplete inventory view, lack of flexibility and intelligence in inventory allocation strategy, and the impact of environmental factors on cargo sorting accuracy.

[0005] In a first aspect, the present application provides an e-commerce business intelligent multi-warehouse delivery method, comprising:

[0006] Generate a global inventory view based on user order information across e-commerce platforms and corresponding logistics fulfillment requirements;

[0007] Dynamically adjusting the inventory allocation strategy in the global inventory view using preset priority rules, and generating a sorting priority sequence using the dynamically adjusted inventory allocation strategy, wherein the preset priority rules include allocation logic that associates user level identifiers with inventory geographic distribution;

[0008] According to the sorting priority sequence and the SKU identification information in the global inventory view, spectral characteristics of the mixed goods in each warehouse are matched, and the goods are sorted in combination with the lighting environment parameters of the target warehouse corresponding to the inventory geographical distribution, and the goods identification information is generated;

[0009] Use pre-trained distributed inventory sharing models to coordinate inventory data exchange among warehouse nodes and generate cross-warehouse transfer decisions;

[0010] Based on the cross-warehouse transfer decision and the cargo identification information, a multi-warehouse coordinated shipment instruction is generated, wherein the instruction includes target warehouse selection logic and a sorting result verification mechanism.

[0011] Optionally, dynamically adjusting the inventory allocation strategy in the global inventory view using a preset priority rule, and generating a sorting priority sequence using the dynamically adjusted inventory allocation strategy, including:

[0012] Calculating the priority coefficient of each user's order information according to the service level agreement corresponding to the user level identifier;

[0013] Extracting the real-time distance data between the target warehouse and the user's delivery address in the inventory geographical distribution, and combining it with a preset logistics timeliness threshold to generate a timeliness coefficient corresponding to the inventory geographical distribution;

[0014] Calculating an inventory dynamic balance factor based on the matching degree between the inventory balance of each warehouse and the ordered goods in the global inventory view;

[0015] A weighted scoring model is constructed based on the priority coefficient, timeliness coefficient, and inventory dynamic balance factor, and the weighted scoring model is used to dynamically adjust the weight distribution ratio of the inventory allocation strategy to generate a dynamic inventory allocation matrix;

[0016] A sorting priority sequence is generated according to the score ranking results of each warehouse in the dynamic inventory allocation matrix, and the sorting priority sequence includes a list of goods to be sorted arranged in descending order of score.

[0017] Optionally, a sorting priority sequence is generated based on the score ranking results of each warehouse in the dynamic inventory allocation matrix. The sorting priority sequence includes a list of goods to be sorted arranged in descending order of score, including:

[0018] Sorting the scores of each warehouse in the dynamic inventory allocation matrix in descending order to generate an initial sorting order list;

[0019] Performing weighted interpolation processing on the high-level user orders in the initial sorting order list according to the real-time order waiting time corresponding to the user level identifier to generate a revised sorting order;

[0020] Based on the current sorting load rate of the target warehouse in the inventory geographical distribution, dynamically and evenly distribute the modified sorting sequence, and generate a dynamic and evenly distributed result associated with the sorting load rate;

[0021] In combination with the inventory balance of each warehouse, the dynamic balanced allocation result is incrementally adjusted, and a sorting priority sequence is generated according to the adjusted dynamic balanced allocation result.

[0022] Optionally, spectral feature matching is performed on the mixed goods in each warehouse based on the sorting priority sequence and the SKU identification information in the global inventory view, and goods sorting is completed in combination with the target warehouse lighting environment parameters corresponding to the inventory geographical distribution, and goods identification information is generated, including:

[0023] Extracting spectral feature data of mixed goods in the corresponding warehouse according to the user level weight and the sorting batch identifier in the sorting priority sequence, wherein the spectral feature data includes a standard spectral band range bound to the SKU identifier information;

[0024] Dynamically adjusting the matching threshold of the standard spectral band range based on the real-time light intensity and color temperature data in the target warehouse lighting environment parameters, and generating a dynamic matching rule adapted to the predetermined lighting conditions according to the adjusted matching threshold;

[0025] Performing multispectral scanning on the mixed goods according to the dynamic matching rules, calculating the spectral matching score between each of the goods and the SKU identification information based on the scanning results, and screening the goods with the spectral matching score higher than a preset threshold to generate a candidate sorting set;

[0026] Based on the inventory dynamic balance factor in the sorting priority sequence, the goods in the candidate sorting set are optimally allocated by sorting batches, and a sorting instruction set including a sorting path plan and a target warehouse identifier is generated according to the allocation result;

[0027] The sorting device is controlled to perform the cargo classification operation according to the sorting instruction set, and after the sorting is completed, cargo identification information bound to the SKU identification information, the sorting batch identification and the target warehouse location is generated.

[0028] Optionally, performing multispectral scanning on the mixed goods according to the dynamic matching rule, calculating the spectral matching score between each of the goods and the SKU identification information according to the scanning results, and screening the goods with the spectral matching score higher than a preset threshold to generate a candidate sorting set, including:

[0029] When performing multispectral scanning on mixed goods, the real-time color temperature offset of the target warehouse lighting environment parameters is simultaneously obtained;

[0030] Performing segmented compensation calibration on the scanning result based on the matching threshold in the dynamic matching rule to generate a compensated scanning result corresponding to the SKU identification information;

[0031] Calculate the spectral matching score of each area on the surface of the goods frame by frame based on the standard spectral band range bound to the compensated scan result and the SKU identification information;

[0032] Performing weighted correction on the spectral matching score based on the user level weight in the sorting priority sequence to generate a corrected spectral matching score;

[0033] Dynamically compensating the corrected spectrum matching score using the real-time color temperature offset, and generating an environmental adaptation score based on the compensation result;

[0034] According to the current warehouse sorting capacity limit corresponding to the inventory dynamic balance factor, the preset threshold is dynamically adjusted, and the goods with the environmental adaptation score higher than the adjusted preset threshold are screened to generate a candidate sorting set, and the confidence parameters corresponding to the candidate goods are recorded;

[0035] Based on the correlation between the confidence parameters of the goods in the candidate sorting set and the sorting batch identifiers, the candidate goods are pre-arranged in sorting order, and a candidate sorting list including sorting priority weights and confidence verification identifiers is generated;

[0036] The candidate sorting list and the sorting path plan are spatially matched and verified, and the verification result is corrected for environmental compatibility in combination with the real-time color temperature offset. Candidate goods that conflict with the target warehouse identifier in the candidate sorting set are eliminated to generate a verified candidate sorting set.

[0037] Optionally, a pre-trained distributed inventory sharing model is used to coordinate inventory data exchange among warehouse nodes and generate cross-warehouse transfer decisions, including:

[0038] Based on the sorting batch identifiers and inventory geographical distribution in the sorting priority sequence, a transfer task initialization request is initiated to each warehouse node to obtain the real-time inventory balance and sorting efficiency indicators of each warehouse node;

[0039] Through the privacy-preserving gradient aggregation mechanism in the distributed inventory sharing model, local gradient encryption calculation is performed on the inventory balance data of each warehouse node to generate a gradient aggregation result associated with the user level identifier;

[0040] Generate a cross-warehouse allocation strategy candidate set based on the gradient aggregation result and the SKU demand distribution data in the global inventory view, wherein the cross-warehouse allocation strategy candidate set includes allocation path planning, inventory balance deviation parameters, and sorting efficiency weight coefficients;

[0041] Based on the error rate statistics of the historical transfer data in the global inventory view, the transfer path plans in the warehouse transfer strategy candidate set are dynamically scored and ranked, and an optimized transfer strategy set including the transfer urgency coefficient and the cost constraint is generated according to the ranking results;

[0042] Based on the allocation urgency coefficient in the allocation strategy optimization set and the target warehouse location in the goods identification information, a multi-objective conflict resolution calculation is performed to generate a cross-warehouse allocation decision, which includes an allocation batch identifier, a priority weighting parameter, and an inventory dynamic compensation factor.

[0043] Optionally, generating a multi-warehouse coordinated shipment instruction according to the cross-warehouse transfer decision and the goods identification information includes:

[0044] Based on the sorting batch identifier in the cargo identification information, extracting the transfer urgency coefficient bound to the sorting batch from the cross-warehouse transfer decision;

[0045] extracting, from the inter-warehouse transfer decision, an inventory dynamic compensation factor associated with the warehouse's geographic location based on the target warehouse location in the cargo identification information;

[0046] Combining the allocation urgency coefficient and the inventory dynamic compensation factor, a multi-warehouse collaborative delivery constraint set including timeliness constraints, inventory balance constraints, and geographical feasibility constraints is generated;

[0047] Generate a dynamic priority-scheduled delivery task queue based on the timeliness constraints of the multi-warehouse collaborative delivery constraint set and the user level weights in the sorting priority sequence, wherein the delivery task queue includes a sorting result verification code bound to the target warehouse;

[0048] Performing consistency verification on the cargo identification information based on the sorting result verification code, and generating a shipping route planning instruction adapted to the geographical feasibility constraint according to the verification result;

[0049] Perform multi-objective optimization on the shipping route planning instructions based on the real-time load status of the target warehouse in the inventory geographical distribution and the inventory dynamic compensation factor, and generate a collaborative shipping instruction set including transportation tool switching rules;

[0050] The multi-warehouse collaborative shipping instructions are generated by the real-time path correction logic in the collaborative shipping instruction set and the logistics time commitment data in the global inventory view.

[0051] In a second aspect, this application provides an e-commerce business intelligent multi-warehouse delivery system, including:

[0052] The first generation module is used to generate a global inventory view based on user order information across e-commerce platforms and corresponding logistics fulfillment requirements;

[0053] an adjustment module, configured to dynamically adjust the inventory allocation strategy in the global inventory view using a preset priority rule, and generate a sorting priority sequence based on the dynamically adjusted inventory allocation strategy, wherein the preset priority rule includes an allocation logic that associates user level identifiers with inventory geographic distribution;

[0054] a matching module configured to perform spectral feature matching on the mixed goods in each warehouse based on the sorting priority sequence and the SKU identification information in the global inventory view, complete the goods sorting in combination with the lighting environment parameters of the target warehouse corresponding to the geographical distribution of the inventory, and generate goods identification information;

[0055] The exchange module is used to coordinate inventory data exchange among warehouse nodes using a pre-trained distributed inventory sharing model and generate cross-warehouse transfer decisions;

[0056] The second generation module is used to generate a multi-warehouse collaborative delivery instruction based on the cross-warehouse transfer decision and the cargo identification information, wherein the instruction includes a target warehouse selection logic and a sorting result verification mechanism.

[0057] In a third aspect, an embodiment of the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an intelligent multi-warehouse shipping method for e-commerce business as described in the first aspect above.

[0058] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements an e-commerce business intelligent multi-warehouse shipping method as described in the first aspect.

[0059] The embodiment of the present application generates a global inventory view by integrating user order information and logistics fulfillment requirements across e-commerce platforms, which not only improves the transparency and accuracy of inventory management, but also provides data support for subsequent dynamic adjustments. It uses preset priority rules to dynamically adjust the inventory allocation strategy, and generates a sorting priority sequence based on the adjusted strategy. It can respond to market changes and customer needs more flexibly and efficiently, completes cargo sorting based on SKU identification information and spectral feature matching technology, and ensures the accuracy of cargo sorting in combination with lighting environment parameters. In addition, the use of a pre-trained distributed inventory sharing model to coordinate data exchange among warehouse nodes can effectively improve the circulation of inventory data and the scientific nature of decision-making, and ultimately achieve collaborative delivery of multiple warehouses, thereby improving delivery efficiency and service quality.

[0060] Furthermore, by calculating the priority coefficient obtained from the service level agreement corresponding to the user level identifier, the efficiency coefficient generated by combining the real-time distance data between the target warehouse and the user's delivery address and the preset logistics efficiency threshold, and the inventory dynamic balance factor obtained by considering the matching degree between the inventory balance of each warehouse and the ordered goods, a weighted scoring model is constructed to dynamically adjust the weight distribution ratio of the inventory allocation strategy. This approach not only fully considers the influence of multiple factors such as user service level, logistics efficiency and inventory balance, but also achieves the optimal allocation of resources by generating a dynamic inventory allocation matrix. The final generated sorting priority sequence can accurately reflect the optimal sorting order under different user levels, geographical distribution and inventory balance, significantly improving the speed and accuracy of cargo processing, reducing the risk of service delays caused by unreasonable resource allocation, and greatly enhancing user experience and satisfaction.

[0061] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0063] Figure 1 A flowchart of an intelligent multi-warehouse shipping method for e-commerce business provided by this application is shown;

[0064] Figure 2 The following is a schematic diagram showing the structure of an intelligent multi-warehouse delivery system for e-commerce business provided by this application;

[0065] Figure 3 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0066] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0067] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0068] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0069] Figure 1 The present application provides a flowchart of an e-commerce business intelligent multi-warehouse delivery method, such as Figure 1 As shown, the method includes:

[0070] Step 101: Generate a global inventory view based on user order information across e-commerce platforms and corresponding logistics fulfillment requirements;

[0071] In this step, user order information includes data such as the user's purchase details, payment status, and delivery address, which is used to record and track each customer's shopping behavior; logistics fulfillment requirements cover the conditions required for the entire process from warehouse shipment to final delivery to the customer, such as estimated delivery time, transportation method selection, etc., to ensure that the goods can be delivered to the customer on time and accurately; the global inventory view is a data set that integrates the current inventory status of all warehouses, including the SKU identification information, inventory balance and geographical distribution of each warehouse, providing comprehensive data support for intelligent allocation and scheduling.

[0072] In this embodiment, the latest user order information and logistics fulfillment requirements are first obtained from cross-e-commerce platforms through API interfaces or data synchronization mechanisms and stored in a central database. Next, data analysis technology is used to parse this order information and extract key attributes such as SKU numbers and quantity requirements. At the same time, the system generates a preliminary inventory matching table based on factors such as the estimated delivery time and transportation preferences in the logistics fulfillment requirements, combined with the real-time inventory status of each warehouse. Subsequently, an optimization algorithm (such as linear programming or heuristic algorithm) is used to adjust the matching table to ensure that all orders are met while maximizing the utilization of inventory resources. Finally, the above results are integrated into a global inventory view.

[0073] For example, in a specific operation, suppose an e-commerce platform receives a batch of new orders. The system first automatically collects relevant information of these orders and identifies that 5 of them require a specific model of product X. Then, the system queries the inventory view of all warehouses and finds that warehouse A has 3 items in stock and warehouse B has 2 items. Based on the logistics requirements of the order (such as the shortest delivery time), the system decides to prioritize shipping from the warehouse that is closest and has sufficient inventory. Through this series of sub-steps, the system not only efficiently completes the matching of orders and inventory, but also ensures the best delivery route and service quality.

[0074] Step 102: dynamically adjusting the inventory allocation strategy in the global inventory view using a preset priority rule, and generating a sorting priority sequence based on the dynamically adjusted inventory allocation strategy, wherein the preset priority rule includes an allocation logic that associates user level identifiers with inventory geographic distribution.

[0075] In this step, priority rules are a set of strategies based on factors such as user level identification and inventory geographical distribution, which are used to determine the order of order processing and inventory allocation; inventory allocation strategy refers to the method of optimizing the allocation of resources in the global inventory view according to priority rules to ensure that each order can be met in the most efficient way; sorting priority sequence is a list of pending orders arranged in a certain priority order to help warehouses sort goods in order; user level identification is the different levels of different customers based on their consumption behavior and service agreements, which is used for differentiated services; inventory geographical distribution refers to the geographical location of each warehouse and the inventory status it has, which is used to calculate the optimal delivery route; allocation logic is an algorithm or rule set formed after comprehensively considering the above factors, which is used to dynamically adjust the inventory allocation strategy.

[0076] In this embodiment, the priority coefficient of each order is first calculated based on the user level identifier and service level agreement in the user order information. Then, the real-time distance data between the target warehouse and the user's delivery address in the inventory geographic distribution is extracted, and the time efficiency coefficient is generated in combination with the preset logistics time efficiency threshold. Next, the inventory balance of each warehouse in the global inventory view is analyzed with the matching degree of the ordered goods to obtain the inventory dynamic balance factor. Based on these coefficients (priority coefficient, time efficiency coefficient, and inventory dynamic balance factor), a weighted scoring model is constructed to dynamically adjust the weight distribution ratio of the inventory allocation strategy to generate a dynamic inventory allocation matrix. Finally, based on the ranking results of the scores of each warehouse in the dynamic inventory allocation matrix, a sorting priority sequence is formed, which contains a list of goods to be sorted arranged in descending order by score.

[0077] For example, in the previous step, the system has obtained user order information from cross-e-commerce platforms and generated a global inventory view; now, the system begins to dynamically adjust the inventory allocation strategy in the global inventory view using preset priority rules; for an urgent order submitted by a VIP customer, the system first identifies the high priority coefficient corresponding to its user level identifier, and finds that although the nearest warehouse A has some of the required goods, it is not enough to cover all needs, while the farther warehouse B has sufficient inventory; by calculating the actual distance from the two warehouses to the customer and the logistics time threshold, the system sets a higher time coefficient for warehouse B; after comprehensively considering the inventory dynamic balance factor, the system uses a weighted scoring model to readjust the inventory allocation strategy to ensure that VIP customers' orders can be met first; based on this, the system generates a new sorting priority sequence to guide warehouse staff to process VIP customers' orders first, thereby improving the customer service experience.

[0078] Step 103: Spectral feature matching is performed on the mixed goods in each warehouse based on the sorting priority sequence and the SKU identification information in the global inventory view. The goods are sorted in combination with the lighting environment parameters of the target warehouse corresponding to the inventory geographical distribution, and the goods identification information is generated.

[0079] In this step, SKU identification information refers to the unique code of each stock keeping unit (SKU), which contains detailed attributes such as product specifications and models, and is used to distinguish different types of goods. Mixed goods refer to a collection of various SKU products that are not classified or stored together in the warehouse. Spectral feature matching is a technology that uses spectral analysis technology to identify products, determining the specific type and characteristics of the product by comparing the target object with data in a known standard spectral database. The target warehouse lighting environment parameters refer to conditions such as light intensity and color temperature inside the warehouse, which can affect the accuracy of spectral feature matching. Goods identification information refers to the label or record generated for each item after sorting, containing detailed information such as its SKU number, batch, and location, to facilitate subsequent operations such as shipping and tracking.

[0080] In this embodiment, the orders that need to be processed are first determined according to the sorting priority sequence, and the SKU identification information corresponding to these orders is extracted from the global inventory view; then, the warehouse management system calls the spectral scanning device to perform a preliminary scan of the mixed goods in the target warehouse to obtain the basic spectral feature data of each product; based on the actual lighting environment parameters of the target warehouse (such as light intensity, color temperature, etc.), the system automatically adjusts the standard spectral band range to ensure matching accuracy; then, using the optimized dynamic matching rules, the spectral feature data obtained by the scan is compared with the standard spectral band range in the SKU identification information to calculate the spectral matching score of each product; for those products with scores higher than the preset threshold, the system will The system then marks the goods as part of the candidate sorting set. Next, the system further screens and sorts the goods in the candidate sorting set based on the inventory dynamic balancing factor in the sorting priority sequence to optimize the sorting batches. This step takes into account the warehouse's current workload and inventory levels to ensure that order demand is met without over-consuming the resources of a specific warehouse. Finally, the system automatically generates a detailed sorting instruction set, including specific sorting route planning, target warehouse location, and the final placement of the goods. Once all selected goods have been sorted, the system generates unique goods identification information for each item, including key information such as SKU number, batch number, sorting timestamp, and target warehouse identification, to facilitate subsequent logistics tracking and management.

[0081] For example, in the previous steps, priority setting for urgent orders from VIP customers has been completed, and the optimal inventory allocation plan has been determined through a weighted scoring model, generating a sorting priority sequence; now, the system begins to match the spectral features of the mixed goods in Warehouse A; for this batch of VIP customer orders, the system first obtains the SKU identification information of the required goods and adjusts the matching threshold according to the actual lighting environment parameters of Warehouse A; through a comprehensive scan of the mixed goods using multi-spectral scanning equipment, the system accurately identifies all eligible goods and adds them to the candidate sorting set; to ensure the optimal utilization of warehouse resources, the system also considers the inventory dynamic balance factor and optimizes the sorting batch allocation of goods in the candidate set; finally, the system generates detailed cargo identification information for each piece of goods, which not only improves sorting efficiency but also ensures delivery accuracy; this enables VIP customers' orders to be processed quickly and accurately, greatly improving customer satisfaction; at the same time, this process also provides solid data support for subsequent logistics distribution, ensuring the smooth operation of the entire supply chain management.

[0082] Step 104: Use the pre-trained distributed inventory sharing model to coordinate inventory data exchange among warehouse nodes and generate cross-warehouse transfer decisions;

[0083] In this step, the pre-trained distributed inventory sharing model refers to a model trained using a machine learning algorithm that can effectively coordinate the exchange of inventory data between different warehouses to achieve optimal resource allocation. Warehouse nodes refer to geographically distributed storage facilities that participate in inventory management and scheduling as independent data processing units. Inventory data exchange refers to the process of sharing information such as inventory status and demand forecasts in real time between warehouse nodes to facilitate more accurate allocation decisions. Cross-warehouse allocation decisions are made based on the results of inventory data exchange and are strategies or plans aimed at balancing inventory levels between warehouses to ensure that all orders are met in a timely manner.

[0084] In this embodiment, the system first initiates a transfer task initialization request to each warehouse node based on the sorting batch identification and inventory geographical distribution in the sorting priority sequence, and obtains the real-time inventory balance and sorting efficiency indicators of each warehouse node; then, through the privacy-preserving gradient aggregation mechanism in the pre-trained distributed inventory sharing model, the inventory balance data of each warehouse node is locally encrypted by gradient calculation to generate a gradient aggregation result associated with the user level identification; next, the system generates a candidate set of cross-warehouse transfer strategies based on the gradient aggregation results and the SKU demand distribution data in the global inventory view; in order to further optimize these strategies, the system dynamically scores and ranks the transfer path planning in the candidate set based on the error rate statistics of historical transfer data, and finally generates an optimized set of transfer strategies that includes the transfer urgency coefficient and cost constraints. Finally, the system performs multi-objective conflict resolution calculations based on the target warehouse location in the cargo identification information to determine the optimal cross-warehouse transfer decision, ensuring logistics timeliness and service quality. Specifically, when the system detects that Warehouse A is short of a certain commodity while Warehouse B has sufficient inventory, it will activate a distributed inventory sharing model to automatically analyze factors such as the current workload, inventory levels, and logistics costs of the two warehouses. If the system finds that transferring part of the inventory from Warehouse B to Warehouse A can not only relieve the pressure on Warehouse A but also reduce overall transportation costs, the system will generate a corresponding transfer instruction. During this process, the system will also consider external factors such as real-time traffic conditions and weather forecasts to ensure the feasibility and cost-effectiveness of the transfer decision. Finally, the system will convey the detailed transfer plan to the relevant warehouses and guide the relevant personnel to execute the transfer task according to the established plan.

[0085] For example, in the previous steps, the spectral feature matching and cargo sorting work for the urgent orders of VIP customers have been completed, and now it is necessary to ensure that there is enough inventory to support the subsequent delivery needs; suppose the system finds that the inventory of a certain key SKU in warehouse A is about to run out, while warehouse B has sufficient inventory reserves; at this time, the system starts the pre-trained distributed inventory sharing model, sends transfer task initialization requests to warehouses A and B respectively, and collects the real-time inventory balance and sorting efficiency indicators of both; after the encrypted calculation of the privacy-preserving gradient aggregation mechanism, the system generates a gradient aggregation result, which shows that part of the inventory has been transferred from warehouse B. Transferring the inventory to Warehouse A is the optimal option. The system then scores and ranks possible transfer paths based on error rate statistics from historical transfer data, and selects the most appropriate route based on the day's traffic conditions and weather forecast. Finally, the system generates a detailed cross-warehouse transfer decision, including information such as transfer batch identifiers, priority weighting parameters, and inventory dynamic compensation factors. This transfer decision not only resolves the inventory shortage issue at Warehouse A but also ensures the efficient operation of the entire supply chain. For example, the system recommends transferring during off-peak hours at night to avoid traffic congestion during peak hours, saving time and costs.

[0086] Step 105: Generate a multi-warehouse coordinated shipment instruction based on the cross-warehouse transfer decision and the cargo identification information. The instruction includes target warehouse selection logic and a sorting result verification mechanism.

[0087] In this step, the multi-warehouse collaborative shipping instructions are detailed operational guidelines generated based on cross-warehouse transfer decisions and cargo identification information, guiding warehouses on how to collaborate efficiently to complete shipping tasks. These instructions not only include specific shipping plans but also cover key aspects such as target warehouse selection logic and sorting result verification mechanisms, ensuring a smooth and error-free shipping process. The sorting result verification mechanism is a set of rules or procedures used to confirm the accuracy of the sorting process, ensuring the correctness of the goods by comparing the actual sorting results with the system's preset targets. The target warehouse selection logic is a set of rules based on factors such as inventory distribution, geographical distance, and logistics costs to determine which warehouse is most suitable for handling the shipping task of a specific order.

[0088] In this embodiment, the system first extracts the transfer urgency coefficient associated with the sorting batch from the cross-warehouse transfer decision based on the sorting batch identifier in the cargo identification information. Next, the system extracts the inventory dynamic compensation factor associated with the warehouse's geographic location based on the target warehouse location in the cargo identification information. Combining the transfer urgency coefficient and the inventory dynamic compensation factor, the system generates a multi-warehouse collaborative shipment constraint set that includes timeliness constraints, inventory balance constraints, and geographic feasibility constraints. Next, based on the timeliness constraints in the multi-warehouse collaborative shipment constraint set and the user level weights in the sorting priority sequence, the system generates a dynamically prioritized shipment task queue and assigns a corresponding sorting result verification code to each task. To further ensure accuracy, the system also verifies the consistency of the cargo identification information based on the sorting result verification code and generates shipment routing instructions that are compatible with the geographic feasibility constraints. Finally, the system performs multi-objective optimization on the shipment routing instructions based on the real-time load status of the target warehouse in the inventory geographic distribution and the inventory dynamic compensation factor, ultimately generating a collaborative shipment instruction set that includes transportation tool switching rules, ensuring that all warehouses can work seamlessly together and successfully complete the shipment task.

[0089] For example, in the previous steps, the cross-warehouse transfer decision was completed, successfully resolving the issue of insufficient inventory of a key SKU at Warehouse A. Now, the system needs to generate multi-warehouse coordinated shipping instructions to ensure that VIP customers' orders can be shipped on time. The system first identifies the high-priority orders that need to be processed first based on the sorting batch identifier in the goods identification information and extracts the transfer urgency coefficient from it. Considering that some goods transferred from Warehouse B to Warehouse A are about to arrive, the system also analyzes Warehouse A's current inventory dynamic compensation factor to ensure sufficient inventory to meet these orders. Next, the system comprehensively considers timeliness constraints, inventory balance constraints, and geographical feasibility constraints to develop a detailed set of multi-warehouse coordinated shipping constraints. For example, the system discovered that although Warehouse A had partially replenished its inventory, there was still a small gap, so it decided to allocate the remaining required goods from Warehouse C; the system then generated a delivery task queue with dynamic priority scheduling and assigned a sorting result verification code to each task; to ensure the accuracy of the sorting results, the system also performed multiple verifications, comparing the actual sorted goods with the data recorded by the system to ensure that there were no errors; finally, the system generated a collaborative delivery instruction set that included rules for switching between means of transport, clarifying the specific operating steps and schedule for each warehouse; for example, the system recommended using fast transportation to transport goods from Warehouse C to Warehouse A, and then Warehouse A would uniformly package and ship them to VIP customers, thus ensuring that the entire delivery process was both efficient and accurate.

[0090] In existing multi-warehouse delivery systems, inventory allocation strategies are typically adjusted based on static rules, making it difficult to respond to user needs and market changes in real time. For example, orders from certain high-priority customers may be delayed due to long warehouse distances or insufficient inventory, resulting in decreased customer satisfaction. Furthermore, uneven resource utilization between warehouses in different geographic locations can easily lead to inventory backlogs in some warehouses and shortages in others. Based on this, in some embodiments, according to step 102, the inventory allocation strategy in the global inventory view is dynamically adjusted using preset priority rules, and a sorting priority sequence is generated using the dynamically adjusted inventory allocation strategy, including:

[0091] Step 201, calculating the priority coefficient of each user's order information according to the service level agreement corresponding to the user level identifier;

[0092] In this step, the Service Level Agreement (SLA) is a service quality agreement between the e-commerce platform and the customer, which specifies specific terms such as delivery speed and customer service response time. The priority coefficient is a numerical value used to quantify the importance and urgency of each order, ensuring that high-priority orders are processed first.

[0093] In this embodiment, the system first extracts the user's level identification and service level agreement information from the user database; then, through a predefined algorithm or rule set, this information is converted into a specific priority coefficient; for example, orders from VIP customers may be assigned a higher priority coefficient, while orders from ordinary customers have a lower priority coefficient; in specific implementation, the system will query the corresponding SLA terms based on the user level identification of each order and calculate the corresponding priority coefficient; this coefficient will be combined with other factors in subsequent steps to jointly determine the inventory allocation strategy and sorting order.

[0094] Step 202: extract the real-time distance data between the target warehouse and the user's delivery address in the inventory geographic distribution, and generate a time efficiency coefficient corresponding to the inventory geographic distribution in combination with a preset logistics time efficiency threshold;

[0095] In this step, the logistics timeliness threshold is a pre-set minimum delivery time requirement, used to measure whether the order can be delivered on time. The timeliness coefficient corresponding to the inventory geographic distribution is a value that reflects the actual transportation distance and logistics timeliness from the target warehouse to the user's delivery address, and is used to assess the urgency and feasibility of the order.

[0096] In this embodiment, the system first calls the map service API to obtain real-time distance data between the target warehouse and the user's delivery address; then, the system calculates the time efficiency coefficient for each order based on the preset logistics time efficiency threshold; specifically, if the distance from a warehouse to the user's delivery address is far and the logistics time efficiency threshold is short, the system will generate a lower time efficiency coefficient for the order; conversely, if the distance is short and the time efficiency requirements are loose, a higher time efficiency coefficient will be generated.

[0097] Step 203: Calculate the inventory dynamic balance factor based on the matching degree between the inventory balance of each warehouse and the ordered goods in the global inventory view;

[0098] In this step, the inventory dynamic balance factor is a value that reflects the degree of matching between each warehouse's inventory balance and ordered items. It is used to assess the current inventory health of the warehouse and help the system optimize inventory allocation strategies to avoid situations where some warehouses have overstocked inventory while others are out of stock.

[0099] In this embodiment, the system first obtains the inventory balance data of each warehouse from the global inventory view and matches it with the product requirements of the current order; specifically, the system analyzes the inventory status of the products required for each order in each warehouse, and calculates the matching degree between the inventory balance of each warehouse and the order products; then, based on these matching degree data, the system generates an inventory dynamic balance factor; for example, if the inventory balance of a warehouse is sufficient and highly matched with the order requirements, the system will generate a higher inventory dynamic balance factor for the warehouse; conversely, if the inventory is tight or mismatched, a lower factor will be generated.

[0100] Step 204: construct a weighted scoring model based on the priority coefficient, timeliness coefficient, and inventory dynamic balance factor, and dynamically adjust the weight distribution ratio of the inventory allocation strategy using the weighted scoring model to generate a dynamic inventory allocation matrix;

[0101] In this step, the weighted scoring model is a mathematical model based on multiple factors (such as priority coefficient, timeliness coefficient, and inventory dynamic balance factor). It is used to comprehensively evaluate the suitability of each warehouse and generate a dynamic inventory allocation matrix. The weight allocation ratio refers to the proportion of different factors in the scoring model and is used to adjust the impact of each factor on the final score. The dynamic inventory allocation matrix is ​​a data structure that contains the scores of each warehouse and their ranking results, which is used to guide inventory allocation and sorting order.

[0102] In this embodiment, the system first constructs a weighted scoring model based on the priority coefficient, timeliness coefficient, and inventory dynamic balance factor; in specific implementation, the system determines the weight distribution ratio of each factor based on business needs and historical data analysis; for example, for high-priority orders, the system may assign a higher weight to the priority coefficient; and for orders with strict timeliness requirements, the timeliness coefficient is assigned a higher weight; then, the system uses the weighted scoring model to score each warehouse and generate a dynamic inventory allocation matrix; this process not only takes into account the current inventory status, but also combines future forecasted demand to ensure the optimal allocation of resources.

[0103] Step 205: Generate a sorting priority sequence based on the ranking results of the scores of each warehouse in the dynamic inventory allocation matrix, wherein the sorting priority sequence includes a list of goods to be sorted in descending order of scores;

[0104] In this step, the score sorting result refers to the result of sorting each warehouse according to its score in the dynamic inventory allocation matrix; the descending score sorting method is to sort each warehouse in descending order of score; the goods to be sorted list is a list of warehouses and their corresponding goods sorted by score, which is used to guide the warehouse in sorting operations;

[0105] In this embodiment, the system first sorts the warehouses according to their scores in the dynamic inventory allocation matrix to generate an initial sorting order list. Next, the system selects the warehouse with the highest score as the target warehouse for processing the current order based on the principle of descending score. In specific implementation, the system will further optimize the sorting order by combining factors such as user level weight and inventory dynamic balance factor. Finally, the system generates a list of goods to be sorted, listing in detail each target warehouse and its corresponding goods information, to ensure that the warehouse can complete the sorting task efficiently and accurately.

[0106] In existing inventory allocation and sorting sequence generation mechanisms, while orders can be prioritized based on certain rules, they often overlook the real-time order waiting time corresponding to user levels and the warehouse's current sorting load rate. This results in actual operations, whereby high-priority customer orders may be delayed due to inability to process them promptly. At the same time, some warehouses may be inefficient due to excessive concentration of order processing, leading to delays or errors. Based on this, as another embodiment, according to step 205, a sorting priority sequence is generated based on the ranking results of the scores of each warehouse in the dynamic inventory allocation matrix. The sorting priority sequence includes a list of goods to be sorted, sorted in descending order of score, including:

[0107] Step 301: sort the scores of each warehouse in the dynamic inventory allocation matrix in descending order to generate an initial sorting order list;

[0108] In this step, the dynamic inventory allocation matrix is ​​a data structure generated by a scoring model based on the priority coefficient, timeliness coefficient, and inventory dynamic balance factor. It is used to evaluate the suitability of each warehouse. The initial sorting order list is a sequence list generated by sorting the warehouses in the dynamic inventory allocation matrix from high to low scores. It is used to guide warehouses to perform sorting operations in order.

[0109] In this embodiment, the system first extracts the scores of each warehouse from the dynamic inventory allocation matrix and sorts the warehouses in descending order of scores to generate an initial sorting order list. In specific implementation, the system assigns a target warehouse to each order based on the scoring results and sorts them in descending order of scores. For example, if Warehouse A has the highest score, the system selects Warehouse A as the target warehouse for processing the current order and places it at the top of the initial sorting order list.

[0110] Step 302: Perform weighted interpolation processing on the high-level user orders in the initial sorting order list according to the waiting time of the real-time orders corresponding to the user level identifiers to generate a revised sorting order;

[0111] In this step, order waiting time refers to the time interval between a user placing an order and the start of order processing; high-level user orders refer to orders from high-level users, which typically have higher service level agreements (SLAs); weighted interpolation is a mathematical method that adjusts data by assigning different weights to different factors; and the revised sorting order refers to the optimized sorting order list after weighted interpolation, ensuring that orders from high-level users are processed first.

[0112] In this embodiment, the system first queries the real-time order waiting time corresponding to each order based on the user level identifier; then, the system performs weighted interpolation processing on the orders of high-level users in the initial sorting order list to ensure that these orders can be processed in the shortest time possible; in specific implementation, the system calculates a weight coefficient based on the user level and the order waiting time, and applies it to the corresponding orders in the initial sorting order list; for example, the system may assign a higher weight coefficient to the orders of VIP customers, so that they are ranked higher in the revised sorting order; in this way, the system can more flexibly adjust the sorting order to ensure that orders from high-level users are processed first.

[0113] Step 303: Based on the current sorting load rate of the target warehouse in the inventory geographical distribution, dynamically balance the modified sorting sequence and generate a dynamic balance distribution result associated with the sorting load rate.

[0114] In this step, the current sorting load rate refers to the ratio of the target warehouse's current sorting tasks to its maximum processing capacity. Dynamic balancing is an algorithm or mechanism used to balance the workload of each warehouse, preventing inefficiency caused by excessive concentration of orders in some warehouses. The dynamic balancing result associated with the sorting load rate refers to the optimized sorting order list after dynamic balancing, ensuring a more balanced workload across warehouses.

[0115] In this embodiment, the system first obtains the current sorting load rate of each target warehouse and dynamically balances the distribution based on the modified sorting order. In specific implementation, the system adjusts the modified sorting order according to the current load of each warehouse to ensure a relatively balanced workload. For example, if the sorting load rate of a warehouse is close to saturation, the system will reduce the amount of tasks assigned to the warehouse and reallocate some orders to other warehouses with lower loads. In this way, the system can effectively avoid a decrease in overall efficiency due to overloading of a warehouse.

[0116] Step 304: incrementally adjust the dynamic balanced allocation result based on the inventory balance of each warehouse, and generate a sorting priority sequence based on the adjusted dynamic balanced allocation result;

[0117] In this step, the inventory balance of each warehouse refers to the number of goods currently held by each warehouse; incremental adjustment is a gradual adjustment method that optimizes the final result by gradually increasing or decreasing resource allocation; the sorting priority sequence refers to the final optimized sorting order list after incremental adjustment, ensuring that the inventory balance of each warehouse is reasonably utilized;

[0118] The system first obtains the inventory balance of each warehouse and makes incremental adjustments based on the dynamic balanced allocation results. In specific implementation, the system will gradually optimize the sorting order according to the inventory balance of each warehouse to ensure that warehouses with sufficient inventory balances are given priority in processing orders. For example, if a warehouse has sufficient inventory balance and it is highly matched with the needs of multiple orders, the system will give priority to assigning this warehouse to process related orders. On the contrary, if the inventory is tight or mismatched, the amount of tasks assigned to this warehouse will be reduced. In this way, the system can more reasonably utilize the inventory resources of each warehouse to ensure that all orders can be met in a timely manner.

[0119] In the existing goods sorting process, manual recognition or simple barcode scanning technology is usually relied on to match SKU identification information. This is not only inefficient but also prone to errors, especially when handling large quantities of mixed goods. In addition, environmental factors such as changes in lighting conditions can also affect sorting accuracy. Based on this, in some embodiments, according to step 103, spectral characteristics of mixed goods in each warehouse are matched according to the sorting priority sequence and the SKU identification information in the global inventory view. The goods are sorted in combination with the lighting environment parameters of the target warehouse corresponding to the inventory geographical distribution, and the goods identification information is generated, including:

[0120] Step 401: extracting spectral feature data of mixed goods in the corresponding warehouse according to the user level weight and the sorting batch identifier in the sorting priority sequence, wherein the spectral feature data includes a standard spectral band range bound to the SKU identifier information;

[0121] In this step, user level weight refers to the priority value assigned to different users based on their service level agreement (SLA); sorting batch identification is a set of unique identifiers used to distinguish different sorting tasks, helping the system manage and track each sorting task; spectral feature data is optical property information obtained from the surface of the goods, including the standard spectral band range bound to the SKU identification information, which is used to identify and classify goods;

[0122] In this embodiment, the system first extracts the user level weight and sorting batch identifier from the sorting priority sequence to determine the order currently to be processed and its priority; then, the system extracts the spectral feature data of the mixed goods from the target warehouse based on this information; in specific implementation, the system will call a multi-spectral scanning device to perform a preliminary scan of the mixed goods and compare the scan results with the standard spectral band range to ensure that the SKU identification information of each item can be accurately identified; for example, for a high-priority order from a VIP customer, the system will prioritize extracting the spectral feature data of the goods in the corresponding sorting batch to ensure that the sorting task is completed quickly and accurately.

[0123] Step 402: dynamically adjusting a matching threshold of the standard spectral band range based on real-time light intensity and color temperature data in the target warehouse lighting environment parameters, and generating a dynamic matching rule adapted to the predetermined lighting conditions according to the adjusted matching threshold;

[0124] In this step, the target warehouse's lighting environment parameters, including real-time light intensity and color temperature data, are used to describe the lighting conditions inside the warehouse. The dynamic matching rule is a matching threshold for a standard spectral band range adjusted according to real-time lighting conditions, ensuring high-precision cargo identification under different lighting conditions.

[0125] In this embodiment, the system first obtains the real-time light intensity and color temperature data of the target warehouse, and dynamically adjusts the matching threshold of the standard spectral band range based on these parameters. In specific implementation, the system will update the matching threshold in real time according to changes in lighting conditions and generate dynamic matching rules that adapt to the current lighting conditions. For example, if the light intensity in the warehouse is low, the system will appropriately relax the matching threshold to improve recognition accuracy. Conversely, if the lighting conditions are good, a stricter matching threshold can be used. In this way, the system can maintain high-precision cargo recognition capabilities under various lighting conditions.

[0126] Step 403: Perform multispectral scanning on the mixed goods according to the dynamic matching rule, calculate the spectral matching score between each goods and the SKU identification information based on the scanning results, and select goods with spectral matching scores higher than a preset threshold to generate a candidate sorting set;

[0127] In this step, the dynamic matching rule is the matching threshold of the standard spectral band range adjusted according to real-time lighting conditions; the candidate sorting set is a set of goods selected by the system with spectral matching scores higher than the preset threshold, which is used to further optimize the sorting batches;

[0128] In this embodiment, the system performs multispectral scanning on mixed goods according to dynamic matching rules and calculates the spectral matching score between each item and the SKU identification information. In specific implementation, the system analyzes the scanning results frame by frame, calculates the spectral matching score for each item, and filters out items with scores above a preset threshold to generate a candidate sorting set. For example, the system may set a minimum threshold for the spectral matching score, and only items that meet or exceed this threshold will be included in the candidate sorting set. In this way, the system can efficiently and accurately identify goods that meet the requirements.

[0129] Step 404: Based on the inventory dynamic balance factor in the sorting priority sequence, the goods in the candidate sorting set are optimally allocated by sorting batches, and a sorting instruction set including a sorting path plan and a target warehouse identifier is generated according to the allocation result.

[0130] In this step, the inventory dynamic balance factor is a value that reflects the degree of match between each warehouse's inventory balance and ordered goods, and is used to assess the warehouse's current inventory health. The sorting instruction set is a set of detailed sorting operation guidelines, including sorting path planning and target warehouse identification, which is used to guide the warehouse to perform sorting operations in sequence. The candidate sorting set is a set of goods selected by the system with a spectral matching score higher than a preset threshold.

[0131] In this embodiment, the system first optimizes the sorting batch allocation of the goods in the candidate sorting set based on the inventory dynamic balance factor in the sorting priority sequence; in specific implementation, the system will optimize the sorting order based on the inventory balance and order demand of each warehouse to ensure maximum resource utilization; for example, if the inventory balance of a warehouse is sufficient and highly matched with multiple order demands, the system will give priority to assigning this warehouse to process the relevant orders; conversely, if the inventory is tight or mismatched, the amount of tasks assigned to this warehouse will be reduced; finally, the system generates a sorting instruction set including sorting path planning and target warehouse identification, and lists each target warehouse and its corresponding goods information in detail.

[0132] Step 405: Control the sorting device to perform the cargo sorting operation according to the sorting instruction set, and generate cargo identification information bound to the SKU identification information, the sorting batch identification, and the target warehouse location after the sorting is completed;

[0133] In this step, the sorting instruction set is a set of detailed sorting operation guidelines, including sorting route planning and target warehouse identification, which is used to guide the warehouse to carry out sorting operations in sequence. The cargo identification information is a unique identifier for each cargo, including key data such as SKU identification information, sorting batch identification, and target warehouse location, which facilitates subsequent delivery and tracking.

[0134] In this embodiment, the system controls the sorting equipment to perform cargo classification operations according to the sorting instruction set and generates cargo identification information after the sorting is completed. In specific implementation, the system will guide the sorting equipment to process the cargo in sequence according to the path planning and target warehouse identification in the sorting instruction set. After the sorting is completed, each piece of cargo will be affixed with a label or record containing SKU identification information, sorting batch identification and target warehouse location to ensure the smooth progress of subsequent delivery and tracking work. For example, the system will automatically generate a cargo identification information file containing all necessary information for use by the logistics department.

[0135] When using spectral feature matching technology to sort goods, environmental factors such as changes in lighting conditions (especially color temperature shift) can significantly affect the accuracy of scanning results. In addition, existing systems typically use fixed thresholds to screen goods with high matching degrees, failing to fully consider the warehouse's current workload and inventory dynamic balance factors. This can lead to sorting errors or inefficiencies under high load conditions. Based on this, as another embodiment, according to step 403, multispectral scanning is performed on mixed goods according to the dynamic matching rules. Based on the scanning results, a spectral matching score between each item and the SKU identification information is calculated, and goods with spectral matching scores above a preset threshold are screened to generate a candidate sorting set, including:

[0136] Step 501: When performing multispectral scanning on mixed goods, synchronously obtain the real-time color temperature offset of the target warehouse lighting environment parameters;

[0137] In this step, the real-time color temperature offset refers to the deviation of the color temperature in the target warehouse's lighting environment parameters from the standard color temperature. This parameter is used to describe the current lighting conditions in the warehouse and helps the system adjust the accuracy of spectral feature matching.

[0138] In this embodiment, when the system performs multispectral scanning on mixed goods, it simultaneously obtains the real-time color temperature offset of the target warehouse. In specific implementation, the system uses environmental sensors to monitor the lighting conditions in the warehouse in real time, compares this data with the standard color temperature, and calculates the real-time color temperature offset. For example, if the color temperature in the warehouse is warmer, the system will record this offset and apply it to the subsequent spectral feature matching process to ensure high-precision recognition even under different lighting conditions.

[0139] Step 502: performing segmented compensation calibration on the scan result based on the matching threshold in the dynamic matching rule to generate a compensated scan result corresponding to the SKU identification information;

[0140] In this step, segmented compensation calibration is a technology that adjusts the scan results based on real-time color temperature offset, aiming to improve the accuracy of spectral feature matching. The compensated scan result corresponding to the SKU identification information is the scan result after segmented compensation calibration, ensuring that the SKU information of each item can be accurately identified.

[0141] In this embodiment, the system performs segmented compensation calibration on the scanning results based on the matching threshold in the dynamic matching rules to generate compensated scanning results corresponding to the SKU identification information; in specific implementation, the system adjusts the matching threshold of each band according to the real-time color temperature offset, and optimizes the scanning results segment by segment; for example, if the color temperature offset of a certain band is large, the system will appropriately relax the matching threshold of the band to ensure the accuracy of the scanning results; in this way, the system can maintain consistent recognition accuracy under different lighting conditions.

[0142] Step 503 , calculating the spectral matching score of each area on the surface of the goods frame by frame based on the compensated scanning result and the standard spectral band range bound to the SKU identification information;

[0143] In this step, the standard spectral band range associated with the SKU identification information refers to the specific spectral band range reflected or emitted by each SKU product on its surface. The spectral matching score is a value calculated frame by frame based on the compensated scan results and the standard spectral band range, and is used to evaluate the degree of match between the product and the SKU identification information.

[0144] In this embodiment, the system calculates the spectral matching score of each area on the surface of the goods frame by frame based on the compensated scanning results and the standard spectral band range bound to the SKU identification information. In specific implementation, the system analyzes each frame of the image, calculates its similarity with the standard spectral band range, and generates a spectral matching score. For example, the system compares each frame of the image with the standard spectral band in the database and calculates the matching score for each frame of the image. In this way, the system can efficiently and accurately identify the specific SKU information of each item of goods.

[0145] Step 504 , performing weighted correction on the spectrum matching score based on the user level weight in the sorting priority sequence to generate a corrected spectrum matching score;

[0146] In this step, the user level weight refers to the priority value assigned to different users based on their service level agreement (SLA); the modified spectrum matching score is the weighted and modified value of the original spectrum matching score, ensuring that orders from high-priority users are processed first;

[0147] In this embodiment, the system performs a weighted correction on the spectral matching score based on the user level weight in the sorting priority sequence to generate a corrected spectral matching score. In specific implementation, the system adjusts the spectral matching score based on the user level weight to ensure that orders from high-priority users receive higher scores. For example, the system may assign a higher weight coefficient to orders from VIP customers, resulting in a higher spectral matching score after correction. In this way, the system can more flexibly adjust the sorting order to ensure that orders from high-level users are processed first.

[0148] Step 505: dynamically compensate the corrected spectrum matching score using the real-time color temperature offset, and generate an environment adaptation score based on the compensation result;

[0149] In this step, the environmental adaptation score is the value obtained by dynamically compensating the corrected spectral matching score, ensuring high-precision recognition under different lighting conditions. The real-time color temperature offset is used to describe the current lighting conditions in the warehouse and helps the system adjust the accuracy of spectral feature matching.

[0150] In this embodiment, the system uses the real-time color temperature offset to dynamically compensate the corrected spectral matching score to generate an environmental adaptation score. In specific implementation, the system adjusts the corrected spectral matching score based on the real-time color temperature offset to ensure consistent recognition accuracy under different lighting conditions. For example, if the color temperature offset in the warehouse is large, the system will appropriately adjust the score to ensure that it adapts to the current lighting conditions. In this way, the system can maintain high-precision recognition capabilities in various lighting environments.

[0151] Step 506: Dynamically adjust the preset threshold according to the current warehouse sorting capacity limit corresponding to the inventory dynamic balance factor, select goods with environmental adaptation scores higher than the adjusted preset threshold to generate a candidate sorting set, and record the confidence parameters corresponding to the candidate goods;

[0152] In this step, the current warehouse sorting capacity limit refers to the current maximum processing capacity of each warehouse; the confidence parameter is a value generated by the system for each candidate item, which is used to measure the accuracy and reliability of its recognition; the candidate sorting set is a set of items selected by the system with an environmental adaptation score higher than the preset threshold, which is used to further optimize the sorting batches;

[0153] In this embodiment, the system dynamically adjusts the preset threshold according to the current warehouse sorting capacity limit corresponding to the inventory dynamic balance factor, and screens goods with an environmental adaptation score higher than the adjusted preset threshold to generate a candidate sorting set; in specific implementation, the system will optimize the screening conditions based on the current workload and inventory status of each warehouse; for example, if the workload of a warehouse is close to saturation, the system will appropriately increase the preset threshold and reduce the amount of tasks assigned to the warehouse; conversely, if the warehouse load is low, a lower threshold can be used to increase the amount of tasks; the system will also generate a confidence parameter for each candidate product to record its recognition accuracy and reliability.

[0154] Step 507: Based on the association between the confidence parameters of the goods in the candidate sorting set and the sorting batch identifiers, pre-arrange the sorting order of the candidate goods and generate a candidate sorting list including sorting priority weights and confidence verification identifiers;

[0155] In this step, the sorting priority weight is a numerical value assigned to each item based on factors such as user level and order urgency, which is used to determine its priority in the sorting process. The confidence verification mark is a mark used to measure the accuracy and reliability of the identification of each candidate item, ensuring that only high-confidence items enter the final sorting process. The candidate sorting list is a pre-organized and detailed list of all candidate items and their related information (such as sorting priority weight and confidence verification mark), which guides the warehouse to carry out sorting operations in sequence.

[0156] In this embodiment, the system first pre-arranges the sorting order of candidate goods based on the association between the confidence parameters of the goods in the candidate sorting set and the sorting batch identifiers; in specific implementation, the system will combine the confidence parameters and sorting batch identifiers of each piece of goods to calculate a comprehensive sorting priority weight; for example, for goods with higher confidence, the system will assign a higher sorting priority weight; while for goods with lower confidence, their priority may need further verification or adjustment; then, the system integrates this information together to generate a candidate sorting list containing sorting priority weights and confidence verification identifiers; this list not only contains the specific information of each piece of goods, but also clarifies their priority in the sorting process, ensuring that the warehouse can complete the sorting task efficiently and orderly.

[0157] Step 508: Perform spatial matching verification on the candidate sorting list and the sorting path plan, perform environmental compatibility correction on the verification result based on the real-time color temperature offset, eliminate candidate goods that conflict with the target warehouse identifier in the candidate sorting set, and generate a verified candidate sorting set;

[0158] In this step, spatial matching verification compares the cargo information in the candidate sorting list with the pre-planned sorting routes to ensure that each piece of cargo is accurately sorted and transported along the optimal route. Environmental compatibility correction adjusts the verification results based on environmental parameters such as real-time color temperature offset to ensure that all operations are feasible and efficient in the current environment. Through these two processes, the system can eliminate cargo that cannot be smoothly processed due to environmental or route conflicts and generate an optimized, verified candidate sorting set.

[0159] In this embodiment, the system first performs a spatial matching check between the candidate sorting list and the sorting path planning, that is, checks whether each item can be efficiently processed on the predetermined sorting path; in specific implementation, the system uses an algorithm to simulate the path of each item from the current location in the warehouse to the target location, and evaluates whether there is a path conflict or inefficiency; for example, if two items have the same target warehouse identifier but overlapping sorting paths, it may lead to a decrease in sorting efficiency. In this case, the system needs to replan the path or adjust the sorting order; then, the system combines the real-time color temperature offset to make an environmental compatibility correction for the verification result; this means that the system needs to consider whether the current lighting conditions in the warehouse are suitable for executing the predetermined sorting task; for example, if the color temperature offset in a certain area is large, it may cause the item to be sorted in a different way. It will affect the accuracy of the multi-spectral scanning equipment, and the system will adjust the tasks in these areas, such as postponing the processing time or replacing the processing equipment; in this way, the system can ensure that all sorting tasks are performed under optimal conditions; finally, the system eliminates the candidate goods that conflict with the target warehouse identification in the candidate sorting set, and generates a verified candidate sorting set; specifically, the system will check whether the target warehouse identification of each goods conflicts with other goods, such as multiple goods are assigned to the same warehouse that is already running at full capacity; for these conflicts, the system will reallocate tasks to ensure that the workload of each warehouse is balanced and reasonable; for example, if it is found that Warehouse A has reached its maximum processing capacity, the system will reallocate some tasks to Warehouse B or other warehouses with spare processing capacity.

[0160] In existing multi-warehouse inventory management systems, inventory data exchange between warehouse nodes typically relies on manual operations or simple automated scripts, lacking an effective coordination mechanism to dynamically adjust inventory allocation and transfer decisions. This can lead to situations where some warehouses are overstocked while others are out of stock when facing order peaks or inventory imbalances, thereby affecting overall logistics efficiency and service quality. Furthermore, due to the lack of privacy protection measures, direct sharing of inventory data may lead to data leakage risks. Therefore, in some embodiments, according to step 104, a pre-trained distributed inventory sharing model is used to coordinate inventory data exchange between warehouse nodes and generate cross-warehouse transfer decisions, including:

[0161] Step 601: Based on the sorting batch identifiers and inventory geographical distribution in the sorting priority sequence, a transfer task initialization request is sent to each warehouse node to obtain the real-time inventory balance and sorting efficiency index of each warehouse node;

[0162] In this step, the transfer task initialization request refers to the notification sent by the system to each warehouse node to start the transfer task. It contains information such as the sorting batch identifier and inventory geographical distribution, and is used to obtain real-time data from each warehouse node. The sorting efficiency index is a numerical value that measures the ability of each warehouse to complete the sorting task within a unit of time. It usually includes key performance indicators such as the number of orders processed per hour and the average sorting time.

[0163] In this embodiment, the system first initiates a transfer task initialization request to each warehouse node based on the sorting batch identifier and inventory geographical distribution in the sorting priority sequence. In specific implementation, the system determines the tasks that need to be transferred based on the current order demand and priority, and packages this information into a transfer task initialization request and sends it to the relevant warehouse node. For example, if a VIP customer's urgent order requires goods to be transferred from multiple warehouses, the system will generate a transfer task initialization request containing the order's sorting batch identifier and the warehouse node information involved. Then, the system waits for each warehouse node to respond to the transfer task initialization request. And return its real-time inventory balance and sorting efficiency indicators; to ensure the accuracy and timeliness of the data, the system may set a timeout mechanism. If a warehouse node fails to respond within the specified time, the system will automatically mark the node as unavailable and consider other alternatives; for example, suppose the system initiates a transfer task initialization request to three warehouse nodes A, B, and C. Warehouse A responds quickly and provides the latest inventory balance and sorting efficiency indicators (such as 200 orders can be processed per hour), while Warehouse B responds late due to network problems. The system will mark it as unavailable after the timeout and increase its dependence on Warehouse C.

[0164] Step 602: Perform local gradient encryption calculation on the inventory balance data of each warehouse node through the privacy-preserving gradient aggregation mechanism in the distributed inventory sharing model to generate a gradient aggregation result associated with the user level identifier;

[0165] In this step, the privacy-preserving gradient aggregation mechanism is a distributed computing method designed to protect the data privacy of each warehouse node while allowing the system to aggregate and analyze inventory balance data. Through this method, each warehouse node only uploads its local gradient encryption calculation results, rather than the original data, ensuring data privacy and security. Local gradient encryption calculation refers to the encryption operation performed locally on each warehouse node. By encrypting the inventory balance data, it generates an encrypted value that cannot directly identify the specific inventory information. The gradient aggregation result associated with the user level identifier is a comprehensive data obtained after processing by the privacy-preserving gradient aggregation mechanism. It not only contains the inventory information of each warehouse node, but also combines the user level identifier to facilitate the subsequent adjustment of allocation strategies based on different user priorities.

[0166] In this embodiment, the system first activates the privacy-preserving gradient aggregation mechanism in the distributed inventory sharing model and sends instructions to each warehouse node, requiring them to perform local gradient encryption calculations on their inventory balance data. In specific implementation, each warehouse node will use a predefined encryption algorithm to encrypt its own inventory balance data to generate a local gradient encrypted value. For example, assuming that Warehouse A currently has 1,000 items in stock, Warehouse B has 800 items, and Warehouse C has 1,200 items, these three warehouses will encrypt these numbers and generate their own local gradient encrypted values. Next, each warehouse node will upload the generated local gradient encrypted value to a central server or an aggregation node in the distributed network. In this process, due to the use of encrypted transmission and storage, even if the data is intercepted, the specific inventory information cannot be directly deciphered, thereby effectively protecting the data privacy of each warehouse node. After the central server or aggregation node receives the local gradient encrypted values ​​uploaded by all warehouse nodes, it uses a preset aggregation algorithm to summarize and calculate these encrypted values ​​to generate a gradient aggregation result associated with the user level identifier.

[0167] Step 603: Generate a cross-warehouse allocation strategy candidate set based on the gradient aggregation result and the SKU demand distribution data in the global inventory view, wherein the cross-warehouse allocation strategy candidate set includes an allocation path plan, an inventory balance deviation parameter, and a sorting efficiency weight coefficient;

[0168] In this step, SKU demand distribution data refers to the demand and distribution of different SKUs (stock keeping units) throughout the supply chain network. This data helps the system understand the demand status of each SKU in different warehouses, so as to optimize inventory allocation and allocation strategies. The cross-warehouse allocation strategy candidate set is a set of possible allocation plans, including allocation path planning, inventory balance deviation parameters, and sorting efficiency weight coefficients, which are used to guide subsequent allocation decisions. Allocation path planning refers to designing the optimal transportation path for each allocation task based on the current inventory distribution and logistics network structure. The inventory balance deviation parameter is a numerical value that measures the difference between the actual inventory and the ideal inventory of each warehouse and is used to evaluate the necessity and urgency of allocation. The sorting efficiency weight coefficient is a value set according to the sorting capacity and efficiency of each warehouse and is used to optimize the allocation strategy to improve overall sorting efficiency.

[0169] In this embodiment, the system first obtains the gradient aggregation results and SKU demand distribution data from the global inventory view. During implementation, the system combines these two data sources to analyze the demand status of each SKU in each warehouse and identify which warehouses have insufficient or excessive inventory. For example, if the system finds that a specific SKU has high demand but insufficient inventory in Warehouse A, while it has sufficient inventory but lower demand in Warehouse B, the system will consider allocating this SKU from Warehouse B to Warehouse A. Next, the system generates a candidate set of cross-warehouse allocation strategies based on the above analysis. Specifically, the system designs a detailed allocation path plan for each potential allocation task to ensure that the goods can be delivered to the target warehouse in the shortest time and at the lowest cost. At the same time, the system also calculates the inventory balance deviation parameter for each warehouse to assess the necessity and urgency of the allocation. For example, if the inventory balance deviation of a certain SKU in Warehouse A is large and the demand is urgent, the system will give priority to including it in the allocation plan. In addition, the system will set a sorting efficiency weight coefficient based on the sorting efficiency of each warehouse to ensure that the allocation task is executed first in the warehouse with higher sorting efficiency.

[0170] Step 604: Based on the error rate statistics of the historical transfer data in the global inventory view, the transfer path plans in the warehouse transfer strategy candidate set are dynamically scored and ranked, and an optimized transfer strategy set including the transfer urgency coefficient and the cost constraint is generated according to the ranking results.

[0171] In this step, historical transfer data refers to the records of all warehouse transfer operations over a period of time, including detailed information such as transfer time, route, and quantity. Error rate statistics are the frequency and type of errors, such as delays, losses, or misdeliveries, derived from analyzing this historical transfer data. The transfer urgency coefficient is a numerical value used to measure the urgency of each transfer task, set based on factors such as order priority and customer demand. Cost constraints are budget limits set for each transfer task to ensure the economic feasibility of the transfer plan. The optimized transfer strategy set is a set of scored, ranked, and optimized transfer plans, containing key information such as transfer path planning, transfer urgency coefficient, and cost constraints, to guide the final transfer decision.

[0172] In this embodiment, the system first extracts historical transfer data from the global inventory view and performs error rate statistical analysis on it. In specific implementation, the system uses data analysis tools to conduct a comprehensive review of historical transfer data to identify common error types and their frequency of occurrence. For example, the system finds that transfer tasks on certain specific paths are prone to delays, while other paths often experience cargo loss. Based on these error rate statistics, the system assesses the risk of each transfer path and generates a corresponding score. Next, the system dynamically scores and ranks the transfer path plans in the warehouse transfer strategy candidate set based on the above scores. Specifically, the system assigns a score to each path based on its historical performance (such as error rate, delay, etc.). Paths with higher scores will be considered more reliable options. For example, assuming that the error rate of path A was 2% in the past month, while the error rate of path B was 5%, the system will give path A a higher score and prioritize it in allocation path planning; on this basis, the system further combines the allocation urgency coefficient and cost constraints to generate an optimized set of allocation strategies; in specific implementation, the system will set an allocation urgency coefficient for each allocation task to reflect its urgency and priority; for example, for urgent orders from VIP customers, the system will assign a higher allocation urgency coefficient to ensure that it can be processed quickly; at the same time, the system will also consider cost constraints to ensure that each allocation plan is within the budget; for example, if the cost of a certain allocation plan exceeds the budget, the system may adjust the path selection or reallocate resources to find a more cost-effective alternative.

[0173] Step 605: Perform multi-objective conflict resolution based on the allocation urgency coefficient in the allocation strategy optimization set and the target warehouse location in the goods identification information to generate a cross-warehouse allocation decision. The allocation decision includes an allocation batch identifier, a priority weighting parameter, and an inventory dynamic compensation factor.

[0174] In this step, multi-objective conflict resolution is a mathematical or algorithmic method used to resolve multiple conflicting objectives (such as transfer urgency, cost, and time). By weighing the importance of different objectives, an optimal or near-optimal solution is found to ensure that all key factors are properly considered. The cross-warehouse transfer decision is the final determination of the specific transfer plan based on various parameters in the transfer strategy optimization set (such as transfer urgency coefficient and target warehouse location). This decision includes not only transfer batch identification and priority weighting parameters, but also inventory dynamic compensation factors to adjust and optimize resource allocation during the transfer process.

[0175] In this embodiment, the system first performs multi-objective conflict resolution calculations based on the allocation urgency coefficient and the target warehouse location in the goods identification information in the allocation strategy optimization set. In specific implementation, the system will use one or more types of optimization algorithms (such as linear programming, genetic algorithm, etc.) to comprehensively evaluate each allocation task and find the best allocation path and sequence. For example, assuming that the system has several urgent orders that need to allocate goods from multiple warehouses to different target warehouses, the system will calculate a comprehensive score based on the allocation urgency coefficient of each order (reflecting its urgency), the location of the target warehouse, and the current inventory status of each warehouse. For those tasks with high allocation urgency and close distance to the target warehouse, the system may assign higher priority; for those tasks that are urgent but For tasks that are farther away, it may be necessary to re-plan the route or adjust resource allocation; then, the system generates a cross-warehouse transfer decision; specifically, based on multi-objective conflict resolution calculations, the system assigns a unique transfer batch ID to each transfer task and sets corresponding priority weighting parameters; for example, for urgent orders from VIP customers, the system may assign them a high-priority transfer batch ID and assign them a higher priority weighting parameter to ensure that they are handled first; in addition, the system will also combine the inventory dynamic balance factor to further optimize the transfer decision; for example, if the inventory balance of a warehouse is low and the demand is high in the short term, the system may increase the inventory replenishment of the warehouse while reducing the transfer quantity of other warehouses to maintain the balance of the overall inventory.

[0176] In the multi-warehouse coordinated delivery process, traditional delivery instruction generation methods often lack comprehensive consideration of timeliness, inventory balance, and geographical feasibility. This results in delayed delivery of certain high-priority orders due to remote warehouse locations or insufficient inventory, affecting customer satisfaction. Furthermore, uneven resource utilization among different warehouses can also cause inventory backlogs in some warehouses while others are out of stock. Furthermore, existing systems typically fail to optimize transportation tool selection and route planning in real time, making them unable to respond to emergencies or changes in traffic conditions. Therefore, in some embodiments, according to step 105, generating a multi-warehouse coordinated delivery instruction based on the cross-warehouse transfer decision and the cargo identification information includes:

[0177] Step 701: extracting the transfer urgency coefficient bound to the sorting batch from the cross-warehouse transfer decision based on the sorting batch identifier in the cargo identification information;

[0178] In this embodiment, the system first extracts the allocation urgency coefficient bound to the sorting batch from the cross-warehouse allocation decision based on the sorting batch identifier in the goods identification information; in specific implementation, the system queries the corresponding allocation urgency coefficient based on each sorting batch identifier and records it; for example, if the goods identification information of a batch shows that it is an urgent order for a VIP customer, the system will extract the corresponding high allocation urgency coefficient from the cross-warehouse allocation decision and associate it with the batch; this ensures that these urgent orders can be given priority in the subsequent processing process, thereby improving response speed and service quality.

[0179] Step 702: extracting an inventory dynamic compensation factor associated with the warehouse's geographic location from the inter-warehouse transfer decision based on the target warehouse location in the cargo identification information;

[0180] In this embodiment, the system extracts the inventory dynamic compensation factor associated with the warehouse's geographical location from the cross-warehouse transfer decision based on the target warehouse location in the goods identification information. In specific implementation, the system will query its inventory dynamic compensation factor based on the location information of the target warehouse and record it. For example, assuming that a batch of goods needs to be sent to Warehouse A, and Warehouse A currently has a tight inventory, the system will extract the corresponding low inventory dynamic compensation factor from the cross-warehouse transfer decision and associate it with the batch. This helps the system to reasonably allocate resources in the subsequent scheduling process, ensuring that warehouses with insufficient inventory can be replenished in a timely manner to avoid affecting order processing efficiency.

[0181] Step 703: Combine the transfer urgency coefficient and the inventory dynamic compensation factor to generate a multi-warehouse coordinated delivery constraint set including timeliness constraints, inventory balance constraints, and geographical feasibility constraints.

[0182] In this embodiment, the system combines the allocation urgency coefficient with the inventory dynamic compensation factor to generate a set of multi-warehouse collaborative shipment constraint conditions that include timeliness constraints, inventory balance constraints, and geographical feasibility constraints. During implementation, the system will comprehensively consider the urgency of each allocation task (allocation urgency coefficient) and the inventory status of the target warehouse (inventory dynamic compensation factor) to formulate detailed shipment constraints. For example, for a task with high allocation urgency and tight inventory at the target warehouse, the system will set strict timeliness constraints and high inventory balance constraints to ensure that the task can be processed first and inventory can be replenished in time. At the same time, the system will also consider geographical feasibility constraints, such as transportation distance and traffic conditions, to determine the optimal shipment route and schedule. In this way, the system can ensure that all shipment tasks are executed smoothly while ensuring high efficiency.

[0183] Step 704: Generate a dynamic priority-scheduled delivery task queue based on the timeliness constraints of the multi-warehouse collaborative delivery constraint set and the user level weights in the sorting priority sequence. The delivery task queue includes a sorting result verification code bound to the target warehouse.

[0184] In this step, the dynamic priority scheduling shipping task queue is a priority-ordered task list that includes a sorting result verification code bound to the target warehouse. The sorting result verification code is a unique identifier generated for each item after sorting, used to verify the correctness and integrity of the item.

[0185] In this embodiment, the system generates a shipping task queue with dynamic priority scheduling based on the timeliness constraints of the multi-warehouse collaborative shipping constraint set and the user level weights in the sorting priority sequence; in specific implementation, the system will combine the timeliness requirements and user level weights of each shipping task to calculate a comprehensive priority score, and generate a shipping task queue accordingly; for example, assuming that a batch of goods comes from VIP customers and has a high allocation urgency coefficient, the system will assign it a higher priority score and put it at the forefront of the shipping task queue; in addition, the system will also generate a sorting result verification code bound to the target warehouse for each task to ensure that the integrity and correctness of the goods can be accurately verified before shipment; for example, in actual operation, when a batch of goods is ready for shipment, the system will generate a unique sorting result verification code and bind it to the shipping task of the batch; this will not only improve shipping efficiency, but also effectively reduce the situation of erroneous shipments, thereby improving customer satisfaction and service quality.

[0186] Step 705: Verify the consistency of the cargo identification information based on the sorting result verification code, and generate a shipping route planning instruction adapted to the geographical feasibility constraint according to the verification result;

[0187] In this step, consistency verification involves checking the cargo identification information using the sorting result verification code to ensure the correctness and completeness of the cargo. Specifically, it involves comparing the actual sorted cargo information with the information recorded in the system to confirm whether the two are consistent. The shipping route planning instructions adapted to geographical feasibility constraints are specific operational guidelines generated based on factors such as warehouse location, transportation capacity, and traffic conditions, ensuring that the cargo can be delivered from the current warehouse to the target warehouse or customer along the optimal route.

[0188] In this embodiment, the system first performs consistency verification on the goods identification information based on the sorting result verification code. Specifically, when implemented, the system reads the sorting result verification code of each piece of goods and compares it with the goods identification information (such as SKU number, batch identification, etc.) stored in the system. For example, assume that a batch of goods contains 100 items. The system scans and verifies the sorting result verification code of each item to ensure that the information such as the SKU number and batch identification of these goods is consistent with the system records. If any inconsistency is found (such as quantity mismatch, SKU error, etc.), the system immediately issues an alarm and suspends the relevant task for manual verification and correction. Then, based on the results of the consistency verification, the system generates a shipping route planning instruction adapted to the geographical feasibility constraints. Specifically, the system formulates a detailed route plan for each shipping task by considering factors such as the geographical location of the warehouse, the real-time traffic conditions, and the transportation capacity. For example, assume that the system has verified that the information of a batch of goods is completely correct and is ready for shipment. The system selects an optimal transportation route according to the location of the target warehouse and the traffic conditions around it. If the target warehouse is relatively close and the traffic conditions are good, the system may choose the shortest route. If the target warehouse is far away or the traffic is busy, the system may choose a route that bypasses less congested areas. In addition, the system also considers the selection of transportation tools (such as trucks, drones, etc.) to ensure that the shipping task can be completed within the specified time.

[0189] Step 706: According to the real-time load status of the target warehouse and the inventory dynamic compensation factor in the inventory geographical distribution, perform multi-objective optimization on the shipping route planning instruction, and generate a collaborative shipping instruction set including transportation tool switching rules.

[0190] In this step, the real-time load status refers to the current workload situation of the target warehouse, including information such as the quantity of tasks being processed, equipment utilization rate, and personnel busyness. Multi-objective optimization is a mathematical or algorithmic method aimed at simultaneously optimizing multiple conflicting objectives (such as cost, time, resource utilization efficiency, etc.) to find an optimal or near-optimal solution. The transportation tool switching rules are a set of guidelines for when and how to change transportation tools, ensuring that goods can be delivered to the destination efficiently and economically. The collaborative shipping instruction set is a set of detailed shipping operation guides, including transportation routes, transportation tool selection, and switching rules, etc., used to guide the logistics team to execute the shipping task.

[0191] In this embodiment, the system first performs multi-objective optimization on the delivery route planning instructions based on the real-time load status of the target warehouse in the inventory geographical distribution and the inventory dynamic compensation factor; in specific implementation, the system will collect the real-time load data of the target warehouse, including information such as the number of tasks currently in progress, equipment utilization rate, and personnel busyness; for example, assuming that the target warehouse A is currently in a high-load operation state, the system will recognize this situation and adjust the delivery route planning instructions accordingly; in addition, the system will also combine the inventory dynamic compensation factor to evaluate the inventory balance and demand matching of each warehouse to further optimize the delivery strategy; for example, if a warehouse has tight inventory and high demand in the short term, the system may give priority to the task of replenishing inventory and adjust the relevant delivery routes; then, the system performs multi-objective optimization on the delivery route planning instructions based on the above analysis results; specifically, the system will use optimization algorithms (such as linear programming, genetic algorithms, etc.) to comprehensively consider multiple objectives, such as minimum The system can minimize transportation time and cost, maximize the utilization rate of transportation tools, etc., and generate the optimal delivery route and scheduling plan; for example, in actual operation, the system finds that the direct transportation route from warehouse B to the target warehouse A is the shortest but has serious traffic congestion, while transshipment through the intermediate warehouse C can avoid peak traffic sections and reduce transportation time; in this case, the system will select the latter as the optimized delivery route and formulate a corresponding scheduling plan; then, the system generates a collaborative delivery instruction set containing transportation tool switching rules; in specific implementation, the system will set detailed transportation tool switching rules according to different transportation stages and conditions; for example, for long-distance transportation, the system may first choose trucks for large-scale transportation, and then switch to electric delivery vehicles in urban areas close to the target warehouse to improve flexibility and environmental protection; or in an emergency, the system will choose drones for fast delivery; these rules will be integrated into the collaborative delivery instruction set to form a complete set of operating guidelines.

[0192] Step 707: Generate a multi-warehouse coordinated shipping instruction by combining the real-time path correction logic in the coordinated shipping instruction set with the logistics time commitment data in the global inventory view.

[0193] In this step, real-time route correction logic is a set of rules and algorithms used to adjust shipping routes based on real-time traffic conditions, warehouse load status, and other dynamic factors to ensure that goods can be delivered to the destination in the most efficient manner. Logistics time commitment data refers to the specific commitments made by the company to customers regarding order processing and delivery time, including the time requirements from order placement to shipment and final delivery.

[0194] In this embodiment, the system first generates multi-warehouse collaborative shipping instructions through the real-time path correction logic in the collaborative shipping instruction set, combined with the logistics time commitment data in the global inventory view. In specific implementation, the system will continuously monitor the real-time traffic conditions on each transportation route and dynamically adjust the shipping route based on this information. For example, if a route is severely congested due to a traffic accident, the system will immediately identify this situation and select an alternative route through the real-time path correction logic to avoid delays. At the same time, the system will also refer to the logistics time commitment data in the global inventory view to ensure that all shipment tasks are completed on time. Specifically, the system will check the logistics time commitment corresponding to each shipment task and evaluate whether the current route can complete the delivery within the specified time. If it is found that the current route cannot meet the time commitment, the system will further optimize the route or adjust the transportation method. For example, for an urgent order that needs to be delivered within 24 hours, the system will give priority to the fastest transportation method (such as air or high-speed rail) and avoid any routes that may cause delays.

[0195] Figure 2 The present application provides a structural diagram of an e-commerce business intelligent multi-warehouse delivery system, such as Figure 2 As shown, the system includes:

[0196] A first generation module 21 is used to generate a global inventory view based on user order information across e-commerce platforms and corresponding logistics fulfillment requirements;

[0197] an adjustment module 22 for dynamically adjusting the inventory allocation strategy in the global inventory view using a preset priority rule, and generating a sorting priority sequence based on the dynamically adjusted inventory allocation strategy, wherein the preset priority rule includes an allocation logic that associates user level identifiers with inventory geographic distribution;

[0198] a matching module 23 configured to perform spectral feature matching on the mixed goods in each warehouse based on the sorting priority sequence and the SKU identification information in the global inventory view, complete the goods sorting based on the target warehouse lighting environment parameters corresponding to the inventory geographical distribution, and generate goods identification information;

[0199] The exchange module 24 is used to coordinate the inventory data exchange among the warehouse nodes using the pre-trained distributed inventory sharing model and generate cross-warehouse transfer decisions;

[0200] The second generating module 25 is used to generate a multi-warehouse coordinated delivery instruction according to the cross-warehouse transfer decision and the cargo identification information, wherein the instruction includes a target warehouse selection logic and a sorting result verification mechanism.

[0201] Figure 2 The intelligent multi-warehouse delivery system for e-commerce business can execute Figure 1 The implementation principles and technical effects of the intelligent multi-warehouse shipping method for e-commerce business described in the illustrated embodiment will not be repeated here. The specific manner in which each module and unit performs operations in the intelligent multi-warehouse shipping system for e-commerce business in the above embodiment has been described in detail in the embodiment of the method and will not be elaborated on here.

[0202] In one possible design, Figure 2 The embodiment of the present invention is an e-commerce business intelligent multi-warehouse delivery system that can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0203] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0204] The processing component 32 is used for the above Figure 1 The embodiment provides an intelligent multi-warehouse shipping method for e-commerce business.

[0205] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0206] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0207] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0208] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0209] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0210] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0211] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The embodiment shown is an intelligent multi-warehouse shipping method for e-commerce business.

[0212] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0213] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0214] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0215] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent multi-warehouse delivery method for e-commerce business, characterized by: include: Generate a global inventory view based on user order information across e-commerce platforms and corresponding logistics fulfillment requirements; Dynamically adjusting the inventory allocation strategy in the global inventory view using preset priority rules, and generating a sorting priority sequence using the dynamically adjusted inventory allocation strategy, wherein the preset priority rules include allocation logic that associates user level identifiers with inventory geographic distribution; According to the sorting priority sequence and the SKU identification information in the global inventory view, spectral characteristics of the mixed goods in each warehouse are matched, and the goods are sorted in combination with the lighting environment parameters of the target warehouse corresponding to the inventory geographical distribution, and the goods identification information is generated; Based on the sorting batch identification and inventory geographical distribution in the sorting priority sequence, a transfer task initialization request is initiated to each warehouse node to obtain the real-time inventory balance and sorting efficiency index of each warehouse node; through the privacy protection gradient aggregation mechanism in the pre-trained distributed inventory sharing model, the inventory balance data of each warehouse node is locally encrypted and calculated to generate a gradient aggregation result associated with the user level identification; based on the gradient aggregation result and the SKU demand distribution data in the global inventory view, a cross-warehouse transfer strategy candidate set is generated, wherein the cross-warehouse transfer strategy candidate set includes transfer Path planning, inventory balance deviation parameters, and sorting efficiency weight coefficients; based on the error rate statistics of historical allocation data in the global inventory view, dynamically score and sort the allocation path plans in the warehouse allocation strategy candidate set, and generate an allocation strategy optimization set containing allocation urgency coefficients and cost constraints based on the sorting results; perform multi-objective conflict resolution calculations based on the allocation urgency coefficients in the allocation strategy optimization set and the target warehouse location in the goods identification information to generate a cross-warehouse allocation decision, which includes an allocation batch identifier, priority weighting parameters, and an inventory dynamic compensation factor; Based on the cross-warehouse transfer decision and the cargo identification information, a multi-warehouse coordinated shipment instruction is generated, wherein the instruction includes target warehouse selection logic and a sorting result verification mechanism.

2. The method according to claim 1, characterized in that Dynamically adjusting the inventory allocation strategy in the global inventory view using a preset priority rule, and generating a sorting priority sequence using the dynamically adjusted inventory allocation strategy, including: Calculating the priority coefficient of each user's order information according to the service level agreement corresponding to the user level identifier; Extracting the real-time distance data between the target warehouse and the user's delivery address in the inventory geographical distribution, and combining it with a preset logistics timeliness threshold to generate a timeliness coefficient corresponding to the inventory geographical distribution; Calculating an inventory dynamic balance factor based on the matching degree between the inventory balance of each warehouse and the ordered goods in the global inventory view; A weighted scoring model is constructed based on the priority coefficient, timeliness coefficient, and inventory dynamic balance factor, and the weighted scoring model is used to dynamically adjust the weight distribution ratio of the inventory allocation strategy to generate a dynamic inventory allocation matrix; A sorting priority sequence is generated according to the score ranking results of each warehouse in the dynamic inventory allocation matrix, and the sorting priority sequence includes a list of goods to be sorted arranged in descending order of score.

3. The method according to claim 2, characterized in that According to the ranking results of the scores of each warehouse in the dynamic inventory allocation matrix, a sorting priority sequence is generated. The sorting priority sequence includes a list of goods to be sorted in descending order of scores, including: Sorting the scores of each warehouse in the dynamic inventory allocation matrix in descending order to generate an initial sorting order list; Performing weighted interpolation processing on the high-level user orders in the initial sorting order list according to the real-time order waiting time corresponding to the user level identifier to generate a revised sorting order; Based on the current sorting load rate of the target warehouse in the inventory geographical distribution, dynamically and evenly distribute the modified sorting sequence, and generate a dynamic and evenly distributed result associated with the sorting load rate; In combination with the inventory balance of each warehouse, the dynamic balanced allocation result is incrementally adjusted, and a sorting priority sequence is generated according to the adjusted dynamic balanced allocation result.

4. The method according to claim 1, wherein According to the sorting priority sequence and the SKU identification information in the global inventory view, spectral characteristics of the mixed goods in each warehouse are matched, and the goods are sorted in combination with the target warehouse lighting environment parameters corresponding to the inventory geographical distribution, and the goods identification information is generated, including: Extracting spectral feature data of mixed goods in the corresponding warehouse according to the user level weight and the sorting batch identifier in the sorting priority sequence, wherein the spectral feature data includes a standard spectral band range bound to the SKU identifier information; Dynamically adjusting the matching threshold of the standard spectral band range based on the real-time light intensity and color temperature data in the target warehouse lighting environment parameters, and generating a dynamic matching rule adapted to the predetermined lighting conditions according to the adjusted matching threshold; Performing multispectral scanning on the mixed goods according to the dynamic matching rules, calculating the spectral matching score between each of the goods and the SKU identification information based on the scanning results, and screening the goods with the spectral matching score higher than a preset threshold to generate a candidate sorting set; Based on the inventory dynamic balance factor in the sorting priority sequence, the goods in the candidate sorting set are optimally allocated by sorting batches, and a sorting instruction set including a sorting path plan and a target warehouse identifier is generated according to the allocation result; The sorting device is controlled to perform the cargo classification operation according to the sorting instruction set, and after the sorting is completed, cargo identification information bound to the SKU identification information, the sorting batch identification and the target warehouse location is generated.

5. The method according to claim 4, characterized in that Performing multispectral scanning on the mixed goods according to the dynamic matching rule, calculating the spectral matching score between each of the goods and the SKU identification information according to the scanning results, and screening the goods with the spectral matching score higher than a preset threshold to generate a candidate sorting set, including: When performing multispectral scanning on mixed goods, the real-time color temperature offset of the target warehouse lighting environment parameters is simultaneously obtained; Performing segmented compensation calibration on the scanning result based on the matching threshold in the dynamic matching rule to generate a compensated scanning result corresponding to the SKU identification information; Calculate the spectral matching score of each area on the surface of the goods frame by frame based on the standard spectral band range bound to the compensated scan result and the SKU identification information; Performing weighted correction on the spectral matching score based on the user level weight in the sorting priority sequence to generate a corrected spectral matching score; Dynamically compensating the corrected spectrum matching score using the real-time color temperature offset, and generating an environmental adaptation score based on the compensation result; According to the current warehouse sorting capacity limit corresponding to the inventory dynamic balance factor, the preset threshold is dynamically adjusted, and the goods with the environmental adaptation score higher than the adjusted preset threshold are screened to generate a candidate sorting set, and the confidence parameters corresponding to the candidate goods are recorded; Based on the correlation between the confidence parameters of the goods in the candidate sorting set and the sorting batch identifiers, the candidate goods are pre-arranged in sorting order, and a candidate sorting list including sorting priority weights and confidence verification identifiers is generated; The candidate sorting list and the sorting path plan are spatially matched and verified, and the verification result is corrected for environmental compatibility in combination with the real-time color temperature offset. Candidate goods that conflict with the target warehouse identifier in the candidate sorting set are eliminated to generate a verified candidate sorting set.

6. The method according to claim 1, characterized in that Generate a multi-warehouse coordinated shipment instruction based on the cross-warehouse transfer decision and the cargo identification information, including: Based on the sorting batch identifier in the cargo identification information, extracting the transfer urgency coefficient bound to the sorting batch from the cross-warehouse transfer decision; extracting, from the inter-warehouse transfer decision, an inventory dynamic compensation factor associated with the warehouse's geographic location based on the target warehouse location in the cargo identification information; Combining the allocation urgency coefficient and the inventory dynamic compensation factor, a multi-warehouse collaborative delivery constraint set including timeliness constraints, inventory balance constraints, and geographical feasibility constraints is generated; Generate a dynamic priority-scheduled delivery task queue based on the timeliness constraints of the multi-warehouse collaborative delivery constraint set and the user level weights in the sorting priority sequence, wherein the delivery task queue includes a sorting result verification code bound to the target warehouse; Performing consistency verification on the cargo identification information based on the sorting result verification code, and generating a shipping route planning instruction adapted to the geographical feasibility constraint according to the verification result; Perform multi-objective optimization on the shipping route planning instructions based on the real-time load status of the target warehouse in the inventory geographical distribution and the inventory dynamic compensation factor, and generate a collaborative shipping instruction set including transportation tool switching rules; The multi-warehouse collaborative shipping instructions are generated by the real-time path correction logic in the collaborative shipping instruction set and the logistics time commitment data in the global inventory view.

7. An intelligent multi-warehouse delivery system for e-commerce business, characterized by: include: The first generation module is used to generate a global inventory view based on user order information across e-commerce platforms and corresponding logistics fulfillment requirements; an adjustment module, configured to dynamically adjust the inventory allocation strategy in the global inventory view using a preset priority rule, and generate a sorting priority sequence based on the dynamically adjusted inventory allocation strategy, wherein the preset priority rule includes an allocation logic that associates user level identifiers with inventory geographic distribution; a matching module configured to perform spectral feature matching on the mixed goods in each warehouse based on the sorting priority sequence and the SKU identification information in the global inventory view, complete the goods sorting in combination with the lighting environment parameters of the target warehouse corresponding to the geographical distribution of the inventory, and generate goods identification information; The exchange module is used to initiate a transfer task initialization request to each warehouse node based on the sorting batch identification and inventory geographical distribution in the sorting priority sequence, and obtain the real-time inventory balance and sorting efficiency index of each warehouse node; through the privacy-preserving gradient aggregation mechanism in the pre-trained distributed inventory sharing model, perform local gradient encryption calculation on the inventory balance data of each warehouse node, and generate a gradient aggregation result associated with the user level identification; based on the gradient aggregation result and the SKU demand distribution data in the global inventory view, generate a cross-warehouse transfer strategy candidate set, wherein the cross-warehouse transfer strategy candidate set The system includes allocation path planning, inventory balance deviation parameters, and sorting efficiency weight coefficients. Based on the error rate statistics of historical allocation data in the global inventory view, the allocation path planning in the warehouse allocation strategy candidate set is dynamically scored and ranked, and an allocation strategy optimization set including allocation urgency coefficients and cost constraints is generated based on the ranking results. Based on the allocation urgency coefficients in the allocation strategy optimization set and the target warehouse location in the cargo identification information, a multi-objective conflict resolution calculation is performed to generate a cross-warehouse allocation decision, which includes an allocation batch identification, priority weighting parameters, and an inventory dynamic compensation factor. The second generation module is used to generate a multi-warehouse collaborative delivery instruction based on the cross-warehouse transfer decision and the cargo identification information, wherein the instruction includes a target warehouse selection logic and a sorting result verification mechanism.

8. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an e-commerce business intelligent multi-warehouse shipping method as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, an e-commerce business intelligent multi-warehouse delivery method as described in any one of claims 1 to 6 is implemented.

Images