AI-driven self-service warehouse distribution dynamic management method and platform

Through the AI-driven self-service warehouse distribution dynamic management method, the problem of data isolation between edge warehouses is solved, real-time data integration and optimization is realized, the efficiency and flexibility of warehouse management are improved, and the optimal warehouse distribution management solution is generated. .

CN118966970AInactive Publication Date: 2024-11-15SHANGHAI FEITUO NETWORK TECHNOLOGY CO LTD

Patent Information

Application Number
CN202411108685.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the prior art, each edge warehouse is isolated from each other, resulting in the inability to effectively integrate the storage data, resulting in information asymmetry and low warehousing management efficiency.

Method used

Using AI-driven self-service warehouse distribution dynamic management method, we collect inventory and order data at the edge warehouse, analyze based on local analysis model, and conduct unified analysis and optimization through cloud modules to generate the optimal multi-end warehouse distribution management solution.

Benefits of technology

Real-time update and analysis of data, reduce data island problems, improve the accuracy and efficiency of warehousing management, ensure the consistency and comparability of results, can adapt to market changes in real time, improve the flexibility of warehousing and distribution management and Practical application effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118966970A_ABST
    Figure CN118966970A_ABST
Patent Text Reader

Abstract

The invention provides an AI-driven self-service warehouse distribution dynamic management method and platform, and relates to the technical field of data processing, and the method comprises the steps: carrying out the collection of a plurality of inventory data sequences and a plurality of order data sequences at a plurality of edge warehouse ends to be subjected to warehouse distribution management according to a preset time period; performing storage analysis based on the plurality of edge storage analysis models to obtain a plurality of edge storage analysis results; according to a preset edge storage standard, judging whether the edge storage is qualified or not and the qualification / disqualification degree, and obtaining a plurality of edge storage judgment results; uploading to a cloud, carrying out warehouse distribution management analysis, and obtaining a multi-terminal warehouse distribution management scheme set; performing optimization to obtain an optimal multi-terminal warehouse distribution management scheme; and transmitting the information to a plurality of edge warehouse ends for warehouse allocation dynamic management. The technical problems that in the prior art, all edge warehouses are isolated from one another, storage data cannot be effectively integrated, information is asymmetric, and warehouse distribution management efficiency is low are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an AI-driven self-service warehousing and distribution dynamic management method and platform. Background Art

[0002] In modern supply chain management, warehouses in the warehousing and distribution links are usually distributed in different geographical locations, each handling a large amount of inventory and orders. In the existing technology, each edge warehouse usually operates independently, resulting in many problems. On the one hand, the data of each edge warehouse is often stored in different systems, lacking an effective data integration mechanism. This information island phenomenon makes it impossible to share inventory and order information between different warehouses in a timely manner, resulting in information asymmetry, thus affecting the efficiency of the overall supply chain; on the other hand, in the case of information isolation and difficulty in data integration, traditional methods are difficult to achieve comprehensive optimization of warehouse and distribution management solutions. Even if an edge warehouse is optimized internally, it is difficult to take into account the optimization needs of the entire supply chain. The lack of optimization capabilities from a global perspective limits the effectiveness of warehousing and distribution solutions. Summary of the invention

[0003] This application provides an AI-driven self-service warehouse and distribution dynamic management method and platform, aiming to solve the technical problems in the existing technology that various edge warehouses are isolated from each other, storage data is often not effectively integrated, resulting in information asymmetry, difficulty in comprehensive analysis, and low efficiency of warehouse and distribution management.

[0004] In a first aspect disclosed in the present application, an AI-driven self-service warehouse dynamic management method is provided, the method comprising: collecting and obtaining multiple inventory data sequences and multiple order data sequences at multiple edge warehouse terminals to be managed for warehouse according to a preset time period; performing warehouse analysis on the multiple inventory data sequences and multiple order data sequences based on multiple edge warehouse analysis models local to the multiple edge warehouse terminals, and obtaining multiple edge warehouse analysis results; judging whether the multiple edge warehouse analysis results are qualified and the degree of qualification / failure according to preset edge warehouse standards, and obtaining multiple edge warehouse judgment results; uploading the multiple edge warehouse judgment results to the cloud through the multiple edge warehouse terminals, and performing warehouse management analysis on the multiple edge warehouse judgment results based on the preset edge warehouse standards through the multi-terminal warehouse management module on the cloud, and obtaining a multi-terminal warehouse management solution set; performing optimization within the multi-terminal warehouse management solution set to obtain an optimal multi-terminal warehouse management solution; transmitting the optimal multi-terminal warehouse management solution to the multiple edge warehouse terminals, and performing warehouse dynamic management of the multiple edge warehouse terminals based on the optimal multi-terminal warehouse management solution.

[0005] The second aspect disclosed in the present application provides an AI-driven self-service warehouse distribution dynamic management platform, which is used for the above-mentioned AI-driven self-service warehouse distribution dynamic management method, and the platform includes: a data collection unit, which is used to collect and obtain multiple inventory data sequences and multiple order data sequences at multiple edge warehouse ends to be managed for warehouse distribution according to a preset time period; a warehouse analysis unit, which is used to perform warehouse analysis on the multiple inventory data sequences and multiple order data sequences based on multiple edge warehouse analysis models local to the multiple edge warehouse ends, and obtain multiple edge warehouse analysis results; a warehouse judgment unit, which is used to judge whether the multiple edge warehouse analysis results are qualified and qualified / unqualified according to preset edge warehouse standards. The judgment of the degree of quality is performed to obtain multiple edge warehousing judgment results; a warehousing and distribution management analysis unit, the warehousing and distribution management analysis unit is used to upload the multiple edge warehousing judgment results to the cloud through the multiple edge warehouse terminals, and through the multi-terminal warehousing and distribution management module on the cloud, based on the preset edge warehousing standards, the multiple edge warehousing judgment results are analyzed for warehousing and distribution management to obtain a set of multi-terminal warehousing and distribution management solutions; a management solution optimization unit, the management solution optimization unit is used to perform optimization within the multi-terminal warehousing and distribution management solution set to obtain an optimal multi-terminal warehousing and distribution management solution; a warehousing and distribution dynamic management unit, the warehousing and distribution dynamic management unit is used to transmit the optimal multi-terminal warehousing and distribution management solution to the multiple edge warehouse terminals, and perform warehousing and distribution dynamic management of the multiple edge warehouse terminals based on the optimal multi-terminal warehousing and distribution management solution.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: Through AI-driven self-service methods, edge warehouses can automatically collect and analyze inventory and order data, reduce manual intervention, and improve data processing efficiency and real-time response capabilities. By collecting inventory data and order data at multiple edge warehouses and analyzing them based on local edge warehouse analysis models, real-time data updates and analysis can be achieved, reducing data silos and improving the accuracy and efficiency of warehouse management. Analysis and judgment are performed based on unified preset edge warehouse standards to ensure the consistency and comparability of edge warehouse results. This standardized processing method improves data integration and reliability of analysis. Comprehensive analysis of multiple edge warehouse analysis results is performed in the cloud to generate a set of multi-end warehouse management solutions, which are optimized to adapt to market changes and warehouse status in real time. This dynamic optimization capability improves the flexibility of warehouse management. Comprehensive evaluation of warehouse costs, transportation costs, and transportation risks, establishment of fitness evaluation functions, and solution optimization can comprehensively consider the impact of various factors, obtain the best multi-end warehouse management solution, and improve the actual application effect of the overall warehouse management solution.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A flowchart of an AI-driven self-service warehouse and distribution dynamic management method provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of the AI-driven self-service warehouse and distribution dynamic management platform provided in an embodiment of the present application.

[0009] Explanation of the accompanying drawings: data collection unit 10, warehouse analysis unit 20, warehouse judgment unit 30, warehouse and distribution management analysis unit 40, management solution optimization unit 50, warehouse and distribution dynamic management unit 60. DETAILED DESCRIPTION

[0010] The embodiments of the present application provide an AI-driven self-service warehouse and distribution dynamic management method and platform to solve the technical problems in the prior art that the edge warehouses are isolated from each other, the storage data is often not effectively integrated, resulting in information asymmetry, difficulty in comprehensive analysis, and low efficiency of warehouse and distribution management.

[0011] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0012] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides an AI-driven self-service warehouse distribution dynamic management method, the method comprising: According to a preset time period, multiple inventory data sequences and multiple order data sequences are respectively collected and obtained at multiple edge warehouses to be managed for warehousing and distribution.

[0013] The preset time period refers to the data collection time interval set according to business needs. It can be daily, weekly, etc., depending on the dynamic changes in inventory and order data. For example, for fast-moving consumer goods, a shorter time period is required; for goods with stable inventory, the time period can be longer.

[0014] The edge warehouse refers to a warehouse facility close to customers or markets, which is responsible for receiving, storing and sending goods. Data collection equipment, such as IoT sensors and barcode scanners, is deployed at each edge warehouse to monitor inventory and order information in real time. Within a preset time period, multiple inventory data sequences and multiple order data sequences are collected from multiple edge warehouses. The inventory data sequence represents a data set of all inventory status and changes recorded within a preset time period, and the order data sequence represents a data set of all order information recorded within a preset time period.

[0015] Based on the multiple edge warehouse analysis models local to the multiple edge warehouses, warehouse analysis is performed on the multiple inventory data sequences and the multiple order data sequences respectively to obtain multiple edge warehouse analysis results.

[0016] The edge warehouse analysis model is an analysis model deployed locally at each edge warehouse end, which is used to process and analyze inventory and order data. These models can be trained based on historical data and have real-time analysis capabilities. The collected inventory data sequence and order data sequence are divided into multiple data blocks in chronological order and input into the corresponding edge warehouse analysis model. The model matches and analyzes the inventory data and order data and outputs multiple edge warehouse analysis results.

[0017] Furthermore, the method of performing storage analysis on the multiple inventory data sequences and the multiple order data sequences based on the multiple edge warehouse local edge storage analysis models to obtain multiple edge storage analysis results includes: Locally at the multiple edge warehouses, multiple historical inventory data sets and multiple historical order data sets are collected and obtained; multiple edge warehousing analysis models local to the multiple edge warehouses are constructed using the multiple historical inventory data sets and the multiple historical order data sets; the multiple inventory data sequences and the multiple order data sequences are divided and input into corresponding edge warehousing analysis models to obtain the multiple edge warehousing analysis results.

[0018] At multiple edge warehouses, through the warehouse management system, multiple historical inventory data sets are collected and obtained, including inventory quantity, inbound and outbound records, inventory turnover rate, inventory loss and other information. These data can be recorded by time, product category, warehouse location and other dimensions. Multiple historical order data sets are collected and obtained, including order creation time, order completion time, order quantity, customer information, order status, etc. The data are organized into a unified format to ensure data consistency and availability.

[0019] Select key features from historical inventory data and order data, such as inventory level, order volume, sales frequency, etc., for modeling, divide the data into training sets and test sets, ensure that the training and validation of the model are representative, and build machine learning models, such as decision trees and neural networks, to process complex data patterns and predictions. Use the training set to train the selected model, adjust the model parameters to improve the prediction accuracy, use the test set to validate the model, evaluate its performance and accuracy on new data, adjust the model parameters based on the evaluation results to improve the prediction accuracy and applicability, and iterate the model adjustment process until an edge warehousing analysis model that meets the accuracy requirements is obtained, enabling it to process and analyze new inventory and order data in real time.

[0020] Divide the data series according to the time window, such as by day or week. You can choose a fixed time period or a dynamically adjusted time period. Divide the data series by quantity, such as every 100 records as a data block, or divide it according to the data change point for more detailed analysis. Merge or pair the divided inventory data series and order data series and input them into the edge warehousing analysis model. The model analyzes the input data and obtains the matching results of inventory and orders as the edge warehousing analysis results.

[0021] Furthermore, the method of using the multiple historical inventory data sets and the multiple historical order data sets to construct multiple local edge warehouse analysis models of the multiple edge warehouses includes: Based on the multiple historical inventory data sets and the multiple historical order data sets, randomly extract the first historical inventory data set and the first historical order data set of the first edge warehouse end; perform inventory and order matching analysis based on the one-to-one mapping data in the first historical inventory data set and the first historical order data set to obtain a first edge warehousing analysis result set; perform supervised data labeling on the first historical inventory data set, the first historical order data set, and the first edge warehousing analysis result set to obtain a first constructed data set; perform supervised learning based on the first constructed data set to obtain a first edge warehousing analysis model; and continue to construct multiple edge warehousing analysis models based on the data of other multiple edge warehouse ends.

[0022] Samples may be randomly selected from multiple historical inventory data sets and multiple historical order data sets. A random number generator or a sampling algorithm may be used to obtain a first historical inventory data set and a first historical order data set for a first edge warehouse end, where the first edge warehouse end is any one of the multiple edge warehouse ends.

[0023] Pair inventory data and order data according to timestamps or other relevant attributes. For example, match the inventory volume in a certain time period with the order volume in the same time period, compare the consistency of inventory and order data in terms of time, product category, warehouse location, etc., summarize the results of the matching analysis, including the matching degree of inventory and orders, the accuracy of inventory meeting orders, etc., and obtain the first edge warehousing analysis result set.

[0024] Supervised data annotation is to assign labels to data in order to provide clear learning objectives for machine learning models. Each inventory record is annotated, for example, the inventory status is classified as sufficient, insufficient or excessive. Each order record is annotated, for example, the order status is marked as satisfied, unsatisfied or excess order. The results obtained from the edge warehouse analysis model are annotated, for example, the analysis results are marked as meeting the standards or needing adjustment. Data annotation tools are used to annotate data according to the annotation standards to ensure the consistency and accuracy of the annotation. The annotated first historical inventory data set, the first historical order data set and the first edge warehouse analysis result set are integrated together to form a comprehensive first constructed data set.

[0025] Use the first constructed dataset as the training dataset. The training data contains features and annotated labels. Select a suitable supervised learning model according to the task, such as a decision tree or neural network, for the classification task. Input the annotated data into the selected model for training. The model minimizes prediction errors by adjusting internal parameters to learn patterns and rules in the data. Adjust the model's hyperparameters to optimize the model's performance, for example, select an appropriate learning rate, regularization parameters, etc. Deploy the trained edge warehousing analysis model to the actual warehousing management system for real-time analysis of various edge warehousing situations.

[0026] Based on the above model building process, we continue to analyze the data of multiple other edge warehouses and build multiple edge warehouse analysis models.

[0027] According to the preset edge warehousing standards, the plurality of edge warehousing analysis results are judged whether they are qualified and the degree of qualification / failure, and a plurality of edge warehousing judgment results are obtained.

[0028] The preset edge warehousing standards are used to evaluate whether the edge warehousing analysis results are qualified, including indicators such as inventory adequacy rate, order processing time, and inventory turnover rate. They can be formulated based on industry best practices and the company's own operational goals. A preliminary judgment is made on each edge warehousing analysis result to check whether it meets the preset edge warehousing standards. If the standards are met, it is marked as qualified; if not, it is marked as unqualified.

[0029] For edge warehousing analysis results marked as unqualified, further calculate the degree of unqualified, and calculate the deviation between the unqualified results and the preset edge warehousing standards. The deviation can be quantified into specific values, such as the percentage of insufficient inventory, the excess of order processing time, etc. According to the size of the deviation, mark the severity of the unqualified, for example, the degree of unqualified can be divided into minor, medium and severe. Combine the preliminary judgment results with the degree of unqualified to form a complete edge warehousing judgment result.

[0030] Furthermore, the method of determining whether the plurality of edge storage analysis results are qualified and the degree of qualification / failure according to the preset edge storage standard to obtain a plurality of edge storage determination results includes: Based on the warehousing management requirements, a preset edge warehousing standard is obtained; whether the multiple edge warehousing analysis results meet the preset edge warehousing standard is judged, and multiple initial edge warehousing judgment results are obtained, wherein the multiple initial edge warehousing judgment results include several unqualified judgment results; the deviations of the several unqualified judgment results and the preset edge warehousing standard are calculated respectively, and several unqualified degrees are obtained; the several unqualified degrees are added to the several unqualified judgment results, and the multiple edge warehousing judgment results are obtained.

[0031] Determine the various indicators and standards that need to be met in the warehouse management process, including inventory levels, order processing time, inventory turnover rate, etc. These requirements can be set based on the company's business needs, industry standards, and business goals, including inventory standards, such as minimum inventory, inventory safety level, inventory turnover rate, etc., and order standards, such as order completion time, order processing accuracy, order delays, etc.

[0032] The analysis results of multiple edge warehouses are compared with the preset edge warehouse standards, including comparing whether the actual inventory is higher or lower than the preset minimum inventory standard, and comparing whether the order processing time exceeds the preset time range. Qualified and unqualified judgment results are generated based on the comparison results, and the unqualified judgment results include the specific circumstances of each unqualified result, such as the quantity of insufficient inventory, the time of order delay, etc.

[0033] The deviations between several unqualified judgment results and the preset marginal warehousing standards are calculated respectively, that is, the gap between the actual results and the preset standards is calculated to quantify the severity of the problem. For example, the specific deviation value is calculated for each insufficient inventory situation. The deviation is the actual inventory quantity minus the preset minimum inventory quantity. The obtained deviation value is used to describe the severity of the insufficient inventory. The deviations such as inventory backlog and order delay are calculated in the same way to obtain several degrees of unqualifiedness.

[0034] Each unqualified judgment result and its corresponding unqualified degree are summarized to form the final edge storage judgment result, providing a basis for subsequent warehouse management improvements and decision-making.

[0035] The multiple edge warehouse judgment results are uploaded to the cloud through the multiple edge warehouse terminals, and through the multi-terminal warehouse distribution management module on the cloud, based on the preset edge warehouse standards, the multiple edge warehouse judgment results are analyzed for warehouse distribution management to obtain a set of multi-terminal warehouse distribution management solutions.

[0036] At each edge warehouse, the respective warehousing judgment results are organized into a standardized data format, and the data is uploaded to the cloud through the preset API interface. The multi-terminal warehousing and distribution management module on the cloud processes the judgment results uploaded from each edge warehouse. This module provides integrated analysis, decision support and solution optimization functions, and conducts integrated analysis on all edge warehousing judgment results to identify the overall warehousing management situation, including inventory distribution, order processing efficiency, etc. It analyzes data trends and identifies potential global problems, such as the frequency of insufficient inventory in certain areas, the pattern of order processing delays, etc. Based on the analysis results, multiple warehousing and distribution management solutions are generated so that multiple edge warehouses meet the preset edge warehousing standards.

[0037] An optimization is performed within the set of multi-terminal warehouse distribution management solutions to obtain the optimal multi-terminal warehouse distribution management solution.

[0038] Clarify the optimization goals, such as minimizing total warehousing costs, minimizing transportation costs, minimizing transportation losses, etc. Collect the warehousing management cost data related to each plan, including warehousing costs, transportation costs, labor costs, etc. Collect the benefit data related to each plan, including inventory utilization, order processing efficiency, etc. According to the optimization goals, evaluate and compare each plan in various aspects. Based on the comparison results, obtain the optimal multi-terminal warehousing and distribution management plan. The optimal plan is the best choice that combines factors such as cost, benefit, and risk.

[0039] The optimal multi-end warehouse distribution management plan is transmitted to the multiple edge warehouse ends, and the warehouse distribution dynamic management of the multiple edge warehouse ends is performed based on the optimal multi-end warehouse distribution management plan.

[0040] The optimal multi-end warehouse management plan is organized into a standardized data format and transmitted to each edge warehouse end. Each edge warehouse end receives the optimal management plan, formulates an implementation plan based on the optimal plan, and clarifies specific management measures and time nodes, such as inventory adjustment, order processing optimization, resource reallocation, etc., to ensure that various adjustment measures are implemented as planned. Based on real-time data and optimization plans, warehousing and distribution are dynamically adjusted and managed to cope with changing needs and conditions.

[0041] Furthermore, the method of searching for the best solution in the set of multi-terminal warehouse distribution management solutions to obtain the best multi-terminal warehouse distribution management solution includes: Collect multiple warehousing management cost information of the multiple edge warehouse ends to obtain multiple warehousing cost coefficients; collect the scheduling distance information between the multiple edge warehouse ends, perform transportation cost analysis based on the scheduling distance information, and obtain multiple transportation cost coefficients; retrieve the historical transportation data of the target goods, perform transportation risk analysis of the target goods, and obtain transportation risk coefficients; and search for the best solution in the multi-end warehousing and distribution management solution set based on the multiple warehousing cost coefficients, the multiple transportation cost coefficients, and the transportation risk coefficient.

[0042] Warehouse management cost information refers to all cost data related to edge warehouse management, including but not limited to warehouse rental, labor costs, equipment maintenance costs, inventory management costs, etc. The warehouse cost coefficient represents the ratio of unit inventory management cost and is used to measure the cost-effectiveness of different warehouses. For example, the total warehouse management cost is divided by the total inventory volume to calculate the warehouse cost coefficient.

[0043] The dispatch distance information refers to the distance data between different edge warehouses, which is used to calculate the transportation cost. The dispatch distance includes the actual distance of the transportation path, including the distance adjusted according to factors such as traffic conditions and road conditions. The transportation cost coefficient represents the transportation cost per unit distance, which is used to estimate the transportation cost of different paths. For example, the transportation cost coefficient is obtained by dividing the total transportation cost by the dispatch distance.

[0044] Obtain historical transportation data from the logistics management system and transportation management system. The historical transportation data is all previous transportation records of the target goods, including transportation time, route, weather conditions, loss conditions, etc., which are used to analyze the risks that may be encountered during transportation. The transportation risk coefficient is used to quantify the risk level during transportation, including statistical analysis of historical transportation losses, weather conditions, route problems, etc., quantifying risk factors, and predicting the risk level of future transportation.

[0045] Determine the optimization factors. Specifically, calculate the storage cost of each warehouse, including rent, equipment maintenance, labor costs, etc., calculate the costs of different transportation routes, including fuel costs, labor costs, vehicle depreciation, etc., calculate the risks in the transportation process, including the costs of potential losses, delays and other factors, define a fitness evaluation function to measure the overall performance of each plan, comprehensively consider the storage cost, transportation cost and risk factor, use the fitness evaluation function to evaluate the storage cost coefficient, transportation cost coefficient and transportation risk coefficient of all plans, calculate the fitness of each plan, and find the optimal solution. Compare and obtain the plan with the largest value of the fitness evaluation function, which is the optimal multi-terminal warehouse distribution management plan.

[0046] Furthermore, the method of retrieving the historical transportation data of the target goods, performing transportation risk analysis on the target goods, and obtaining the transportation risk coefficient includes: Acquire historical transportation data, wherein the historical transportation data includes a historical meteorological data set and a historical transportation loss data set; construct a transportation loss analyzer using the historical meteorological data set and the historical transportation loss data set; acquire a transportation time interval for the transportation of goods to be carried out, and acquire predicted meteorological data for the transportation time interval; analyze the predicted meteorological data based on the transportation loss analyzer to obtain predicted transportation loss data; and calculate the transportation risk coefficient based on the predicted transportation loss data.

[0047] The historical meteorological data set contains meteorological information over a period of time in the past, such as temperature, humidity, wind speed, precipitation, etc. These data are obtained from meteorological departments or meteorological service providers. The historical transportation loss data set contains losses recorded during past transportation, such as cargo damage, delays, etc., and is provided by logistics companies and transportation carriers.

[0048] Extract relevant meteorological features, such as temperature and humidity, from historical meteorological data as input features, extract relevant features, such as loss amount and loss type, from historical transportation loss data, select the model type. For predicting the loss amount, you can use linear regression, decision tree regression, etc. For predicting the probability or type of loss, you can use logistic regression, support vector machine, random forest, etc. Use historical meteorological data and transportation loss data to train the model until the model converges to obtain a transportation loss analyzer.

[0049] Determine the time range for the transportation of goods, which can be one day or one week in the future, and use meteorology to obtain the forecasted meteorological data for the transportation time interval, including temperature, humidity, wind speed, precipitation, etc. Extract features from the forecasted meteorological data, which may affect the transportation loss, and input the extracted meteorological features into the transportation loss analyzer. The forecasted data of the transportation loss is calculated based on the transportation loss analyzer, including the expected loss amount or the probability of loss, to obtain the forecasted transportation loss data.

[0050] The transportation risk coefficient is a comprehensive indicator used to quantify the risks in the transportation process. For example, the transportation risk coefficient is calculated by multiplying the expected loss amount by the loss probability. A high risk coefficient indicates that the loss that may be encountered during transportation is large, and a low risk coefficient indicates that the risk is relatively low. The transportation risk coefficient can help optimize transportation plans, such as choosing a safer transportation route.

[0051] Furthermore, the method of searching for the best solution in the set of multi-terminal warehouse distribution management solutions to obtain the best multi-terminal warehouse distribution management solution includes: Construct a fitness evaluation function, the fitness evaluation function is as follows: ; ; ; ; Among them, C is the total fitness, Csc is the storage cost fitness, ω sc Storage cost weight, Ctc is the transportation cost fitness, ω tc is the transportation cost weight, Ctr is the transportation risk adaptability, ω tr is the transportation risk weight, SC i is the storage cost coefficient of the i-th edge warehouse in the multi-end warehouse distribution management solution, w i is the weight of the ith edge warehouse, TC i is the transportation cost coefficient of the jth transportation route in the multi-end warehouse distribution management solution, w j is the weight of the jth transport route, TR i is the transportation risk coefficient of the target goods, and α is the adjustment coefficient; according to the fitness evaluation function, the scheme fitness evaluation is performed within the multi-terminal warehouse distribution management scheme set, and the scheme with the largest total fitness is obtained as the optimal multi-terminal warehouse distribution management scheme.

[0052] Specifically, a fitness evaluation function is constructed, and the fitness evaluation function is as follows: ; ; ; ; Among them, the total fitness is a comprehensive indicator for evaluating the warehouse management plan, which indicates the overall pros and cons of the plan. The calculation formula integrates the storage cost, transportation cost and transportation risk. The storage cost fitness is used to measure the pros and cons of the storage cost in the plan. The higher the storage cost fitness, the lower the storage cost and the better the plan; the transportation cost fitness is used to measure the pros and cons of the transportation cost in the plan. The higher the transportation cost fitness, the lower the transportation cost and the better the plan; the transportation risk fitness is used to measure the pros and cons of the transportation risk in the plan. The higher the transportation risk fitness, the lower the transportation risk and the better the plan. The total fitness is calculated by weighted summation based on the storage cost fitness, transportation cost fitness and transportation risk fitness, which is used to evaluate and optimize the multi-terminal warehouse management plan.

[0053] For each warehousing and distribution management plan in the set, the constructed fitness evaluation function is used to evaluate the plan. This function comprehensively evaluates the fitness of warehousing cost, transportation cost and transportation risk, and applies the corresponding weights to calculate the total fitness of each plan. Then, the total fitness of all plans is compared, and the plan with the largest fitness is selected as the optimal plan. This plan has the best comprehensive performance in warehousing cost, transportation cost and transportation risk. The plan with the largest total fitness is used as the final recommended multi-terminal warehousing and distribution management plan.

[0054] In summary, the AI-driven self-service warehouse distribution dynamic management method provided in the embodiments of the present application has the following technical effects: Through AI-driven self-service methods, edge warehouses can automatically collect and analyze inventory and order data, reduce manual intervention, and improve data processing efficiency and real-time response capabilities. By collecting inventory data and order data at multiple edge warehouses and analyzing them based on local edge warehouse analysis models, real-time data updates and analysis can be achieved, reducing data silos and improving the accuracy and efficiency of warehouse management. Analysis and judgment are performed based on unified preset edge warehouse standards to ensure the consistency and comparability of edge warehouse results. This standardized processing method improves data integration and reliability of analysis. Comprehensive analysis of multiple edge warehouse analysis results is performed in the cloud to generate a set of multi-end warehouse management solutions, which are optimized to adapt to market changes and warehouse status in real time. This dynamic optimization capability improves the flexibility of warehouse management. Comprehensive evaluation of warehouse costs, transportation costs, and transportation risks, establishment of fitness evaluation functions, and solution optimization can comprehensively consider the impact of various factors, obtain the best multi-end warehouse management solution, and improve the actual application effect of the overall warehouse management solution.

[0055] Embodiment 2 is based on the same inventive concept as the AI-driven self-service warehouse distribution dynamic management method in the above embodiment. Figure 2 As shown, the embodiment of the present application provides an AI-driven self-service warehouse distribution dynamic management platform, the platform comprising: A data collection unit 10, the data collection unit 10 is used to collect and obtain multiple inventory data sequences and multiple order data sequences at multiple edge warehouse ends to be managed for warehousing and distribution according to a preset time period; a warehouse analysis unit 20, the warehouse analysis unit 20 is used to perform warehouse analysis on the multiple inventory data sequences and multiple order data sequences based on multiple edge warehouse analysis models local to the multiple edge warehouse ends, and obtain multiple edge warehouse analysis results; a warehouse judgment unit 30, the warehouse judgment unit 30 is used to judge whether the multiple edge warehouse analysis results are qualified and the degree of qualification / failure according to preset edge warehouse standards, and obtain multiple edge warehouse judgment results; a warehouse distribution management analysis unit 40, the The warehouse and distribution management analysis unit 40 is used to upload the multiple edge warehousing judgment results to the cloud through the multiple edge warehouse terminals, and through the multi-terminal warehouse and distribution management module on the cloud, based on the preset edge warehousing standards, the multiple edge warehousing judgment results are subjected to warehouse and distribution management analysis to obtain a set of multi-terminal warehouse and distribution management solutions; the management solution optimization unit 50, the management solution optimization unit 50 is used to perform optimization within the set of multi-terminal warehouse and distribution management solutions to obtain the optimal multi-terminal warehouse and distribution management solution; the warehouse and distribution dynamic management unit 60, the warehouse and distribution dynamic management unit 60 is used to transmit the optimal multi-terminal warehouse and distribution management solution to the multiple edge warehouse terminals, and perform warehouse and distribution dynamic management of the multiple edge warehouse terminals based on the optimal multi-terminal warehouse and distribution management solution.

[0056] Furthermore, the platform also includes an edge storage analysis result acquisition module to perform the following operation steps: Locally at the multiple edge warehouses, multiple historical inventory data sets and multiple historical order data sets are collected and obtained; multiple edge warehousing analysis models local to the multiple edge warehouses are constructed using the multiple historical inventory data sets and the multiple historical order data sets; the multiple inventory data sequences and the multiple order data sequences are divided and input into corresponding edge warehousing analysis models to obtain the multiple edge warehousing analysis results.

[0057] Furthermore, the platform also includes an edge storage analysis model building module to perform the following operation steps: Based on the multiple historical inventory data sets and the multiple historical order data sets, randomly extract the first historical inventory data set and the first historical order data set of the first edge warehouse end; perform inventory and order matching analysis based on the one-to-one mapping data in the first historical inventory data set and the first historical order data set to obtain a first edge warehousing analysis result set; perform supervised data labeling on the first historical inventory data set, the first historical order data set, and the first edge warehousing analysis result set to obtain a first constructed data set; perform supervised learning based on the first constructed data set to obtain a first edge warehousing analysis model; and continue to construct multiple edge warehousing analysis models based on the data of other multiple edge warehouse ends.

[0058] Furthermore, the platform also includes an edge storage judgment result acquisition module to perform the following operation steps: Based on the warehousing management requirements, a preset edge warehousing standard is obtained; whether the multiple edge warehousing analysis results meet the preset edge warehousing standard is judged, and multiple initial edge warehousing judgment results are obtained, wherein the multiple initial edge warehousing judgment results include several unqualified judgment results; the deviations of the several unqualified judgment results and the preset edge warehousing standard are calculated respectively, and several unqualified degrees are obtained; the several unqualified degrees are added to the several unqualified judgment results, and the multiple edge warehousing judgment results are obtained.

[0059] Furthermore, the platform also includes an optimal multi-terminal warehouse distribution management solution acquisition module to perform the following operation steps: Collect multiple warehousing management cost information of the multiple edge warehouse ends to obtain multiple warehousing cost coefficients; collect the scheduling distance information between the multiple edge warehouse ends, perform transportation cost analysis based on the scheduling distance information, and obtain multiple transportation cost coefficients; retrieve the historical transportation data of the target goods, perform transportation risk analysis of the target goods, and obtain transportation risk coefficients; and search for the best solution in the multi-end warehousing and distribution management solution set based on the multiple warehousing cost coefficients, the multiple transportation cost coefficients, and the transportation risk coefficient.

[0060] Furthermore, the platform also includes a transportation risk coefficient acquisition module to perform the following operation steps: Acquire historical transportation data, wherein the historical transportation data includes a historical meteorological data set and a historical transportation loss data set; construct a transportation loss analyzer using the historical meteorological data set and the historical transportation loss data set; acquire a transportation time interval for the transportation of goods to be carried out, and acquire predicted meteorological data for the transportation time interval; analyze the predicted meteorological data based on the transportation loss analyzer to obtain predicted transportation loss data; and calculate the transportation risk coefficient based on the predicted transportation loss data.

[0061] Furthermore, the platform also includes an optimal solution acquisition module to perform the following operation steps: Construct a fitness evaluation function, the fitness evaluation function is as follows: ; ; ; ; Among them, C is the total fitness, Csc is the storage cost fitness, ω sc Storage cost weight, Ctc is the transportation cost fitness, ω tc is the transportation cost weight, Ctr is the transportation risk adaptability, ω tr is the transportation risk weight, SC i is the storage cost coefficient of the i-th edge warehouse in the multi-end warehouse distribution management solution, w i is the weight of the ith edge warehouse, TC i is the transportation cost coefficient of the jth transportation route in the multi-end warehouse distribution management solution, w j is the weight of the jth transport route, TR i is the transportation risk coefficient of the target goods, and α is the adjustment coefficient; according to the fitness evaluation function, the scheme fitness evaluation is performed within the multi-terminal warehouse distribution management scheme set, and the scheme with the largest total fitness is obtained as the optimal multi-terminal warehouse distribution management scheme.

[0062] Through the above-mentioned detailed description of the AI-driven self-service warehouse and distribution dynamic management method in this specification, technical personnel in this field can clearly understand the AI-driven self-service warehouse and distribution dynamic management platform in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0063] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The AI-driven self-service warehouse distribution dynamic management method is characterized by: The method comprises: According to a preset time period, multiple inventory data sequences and multiple order data sequences are respectively collected and obtained at multiple edge warehouses to be managed for warehouse distribution; Based on the multiple edge warehouse analysis models local to the multiple edge warehouses, respectively perform warehouse analysis on the multiple inventory data sequences and the multiple order data sequences to obtain multiple edge warehouse analysis results; According to the preset edge storage standard, the plurality of edge storage analysis results are judged whether they are qualified and the degree of qualification / failure, and a plurality of edge storage judgment results are obtained; The plurality of edge warehouse judgment results are uploaded to the cloud through the plurality of edge warehouse terminals, and a multi-terminal warehouse distribution management module on the cloud performs warehouse distribution management analysis on the plurality of edge warehouse judgment results based on the preset edge warehouse standards to obtain a multi-terminal warehouse distribution management solution set; Searching for the best solution in the multi-terminal warehouse distribution management solution set to obtain the best multi-terminal warehouse distribution management solution; The optimal multi-end warehouse distribution management plan is transmitted to the multiple edge warehouse ends, and the warehouse distribution dynamic management of the multiple edge warehouse ends is performed based on the optimal multi-end warehouse distribution management plan.

2. The AI-driven self-service warehouse distribution dynamic management method according to claim 1, characterized in that: The method of performing storage analysis on the multiple inventory data sequences and the multiple order data sequences based on the multiple local edge warehouse end edge storage analysis models to obtain multiple edge storage analysis results includes: Locally at the multiple edge warehouses, multiple historical inventory data sets and multiple historical order data sets are collected and obtained; Using the multiple historical inventory data sets and the multiple historical order data sets, construct multiple edge warehouse analysis models local to the multiple edge warehouses; The multiple inventory data sequences and the multiple order data sequences are divided and input into corresponding edge warehouse analysis models to obtain the multiple edge warehouse analysis results.

3. The AI-driven self-service warehouse distribution dynamic management method according to claim 2, characterized in that: The method of using the multiple historical inventory data sets and the multiple historical order data sets to construct multiple local edge warehouse analysis models at the multiple edge warehouses includes: Based on the multiple historical inventory data sets and the multiple historical order data sets, randomly extracting a first historical inventory data set and a first historical order data set at a first edge warehouse end; Performing a matching analysis of inventory and orders based on the one-to-one mapping data in the first historical inventory data set and the first historical order data set to obtain a first edge warehouse analysis result set; Performing supervised data labeling on the first historical inventory data set, the first historical order data set, and the first edge warehouse analysis result set to obtain a first constructed data set; Perform supervised learning based on the first constructed data set to obtain a first edge storage analysis model; Continue to build multiple edge warehousing analysis models based on data from multiple other edge warehouses.

4. The AI-driven self-service warehouse distribution dynamic management method according to claim 1, characterized in that: The method of determining whether the plurality of edge storage analysis results are qualified and the degree of qualification / failure according to the preset edge storage standard to obtain a plurality of edge storage judgment results includes: Based on warehouse management requirements, obtain preset edge warehouse standards; Determine whether the plurality of edge storage analysis results meet the preset edge storage standard, and obtain a plurality of initial edge storage judgment results, wherein the plurality of initial edge storage judgment results include a plurality of unqualified judgment results; Calculating the deviations of the plurality of unqualified judgment results from the preset marginal storage standard respectively to obtain a plurality of unqualified degrees; The plurality of unqualified degrees are added to the plurality of unqualified judgment results to obtain the plurality of edge storage judgment results.

5. The AI-driven self-service warehouse distribution dynamic management method according to claim 1, characterized in that: The method of searching for the best solution in the set of multi-terminal warehouse distribution management solutions to obtain the best multi-terminal warehouse distribution management solution includes: Collecting multiple storage management cost information of the multiple edge warehouse ends to obtain multiple storage cost coefficients; Collecting dispatch distance information between the plurality of edge warehouse terminals, performing transportation cost analysis based on the dispatch distance information, and obtaining a plurality of transportation cost coefficients; Retrieving historical transportation data of target goods, conducting transportation risk analysis of the target goods, and obtaining a transportation risk coefficient; According to the multiple warehousing cost coefficients, the multiple transportation cost coefficients, and the transportation risk coefficient, an optimization is performed within the set of multi-terminal warehousing and distribution management solutions to obtain an optimal multi-terminal warehousing and distribution management solution.

6. The AI-driven self-service warehouse distribution dynamic management method according to claim 5, characterized in that: The method of retrieving the historical transportation data of the target goods, performing transportation risk analysis on the target goods, and obtaining the transportation risk coefficient includes: Acquire historical transportation data, wherein the historical transportation data includes a historical meteorological data set and a historical transportation loss data set; Using the historical meteorological data set and the historical transportation loss data set to construct a transportation loss analyzer; Obtaining a transportation time interval for the cargo to be transported, and obtaining forecasted meteorological data for the transportation time interval; Based on the transportation loss analyzer, the predicted meteorological data is analyzed to obtain predicted transportation loss data; The transportation risk coefficient is calculated based on the predicted transportation loss data.

7. The AI-driven self-service warehouse distribution dynamic management method according to claim 5, characterized in that: The method of searching for the best solution in the set of multi-terminal warehouse distribution management solutions to obtain the best multi-terminal warehouse distribution management solution includes: Construct a fitness evaluation function, the fitness evaluation function is as follows: ; ; ; ; Among them, C is the total fitness, Csc is the storage cost fitness, ω sc Storage cost weight, Ctc is the transportation cost fitness, ω tc is the transportation cost weight, Ctr is the transportation risk adaptability, ω tr is the transportation risk weight, SC i is the storage cost coefficient of the i-th edge warehouse in the multi-end warehouse distribution management solution, w i is the weight of the ith edge warehouse, TC i is the transportation cost coefficient of the jth transportation route in the multi-end warehouse distribution management solution, w j is the weight of the jth transport route, TR i is the transportation risk coefficient of the target goods, and α is the adjustment coefficient; According to the fitness evaluation function, the fitness of the solutions is evaluated within the set of multi-end warehouse distribution management solutions, and the solution with the largest total fitness is obtained as the optimal multi-end warehouse distribution management solution.

8. AI-driven self-service warehouse and distribution dynamic management platform, characterized by: The platform is used to implement the AI-driven self-service warehouse distribution dynamic management method according to any one of claims 1 to 7, wherein the platform comprises: A data collection unit, wherein the data collection unit is used to collect and obtain multiple inventory data sequences and multiple order data sequences at multiple edge warehouses to be managed for warehouse distribution according to a preset time period; A warehouse analysis unit, the warehouse analysis unit is used to perform warehouse analysis on the multiple inventory data sequences and the multiple order data sequences respectively based on the multiple edge warehouse local edge warehouse analysis models to obtain multiple edge warehouse analysis results; A storage judgment unit, the storage judgment unit is used to judge whether the plurality of edge storage analysis results are qualified and the degree of qualification / failure according to a preset edge storage standard, and obtain a plurality of edge storage judgment results; A warehouse distribution management analysis unit, wherein the warehouse distribution management analysis unit is used to upload the plurality of edge warehouse judgment results to the cloud through the plurality of edge warehouse terminals, and perform warehouse distribution management analysis on the plurality of edge warehouse judgment results based on the preset edge warehouse standards through the multi-terminal warehouse distribution management module on the cloud, so as to obtain a multi-terminal warehouse distribution management solution set; A management solution optimization unit, the management solution optimization unit is used to optimize the multi-terminal warehouse distribution management solution set to obtain the optimal multi-terminal warehouse distribution management solution; The warehouse and distribution dynamic management unit is used to transmit the optimal multi-end warehouse and distribution management plan to the multiple edge warehouse ends, and perform warehouse and distribution dynamic management of the multiple edge warehouse ends based on the optimal multi-end warehouse and distribution management plan.

Citation Information

Patent Citations

  • Warehouse cargo intelligent management system and intelligent management method

    CN112651694A

  • Warehouse management method and system based on informatization means

    CN113837610A

  • A warehouse management method and system based on Internet of Things (IoT) technology

    CN114936830A

  • Logistics distribution business optimization method and system based on Internet of Things technology

    CN115062850A

  • Storage layout optimization method and system based on big data analysis

    CN116976484A

Cited By

  • Supply chain management method and system for cloud storage

    CN120198055A