Warehouse management model determination method and system

Abstract

The present invention relates to the technical field of warehouse management models, and in particular to a method and system for determining a warehouse management model. The method comprises the following steps: obtaining inventory data; performing warehouse data preprocessing on inventory data to generate warehouse management data; performing statistical analysis on warehouse management data to generate warehouse statistical data; performing real-time warehouse data collection on warehouse statistical data to obtain real-time warehouse management data; obtaining sales cost data and historical sales data; performing numerical feature extraction on sales cost data to generate warehouse numerical feature data; performing product inventory turnover rate calculation on warehouse numerical feature data to generate overall turnover rate data; the timely feedback mechanism of the present invention helps to improve the flexibility and response speed of warehouse operations, enabling managers to respond to market changes and customer needs more quickly.

Description

Technical Field

[0001] The present invention relates to the technical field of warehouse management models, and in particular to a warehouse management model determination method and system. Background Art

[0002] In the field of modern warehouse management, the rapid development and widespread application of data-driven technology demonstrate its growing trend and great potential. With the continuous advancement of information technology and the improvement of data collection capabilities, warehouse management systems are able to collect, organize and analyze a large amount of real-time and historical data, not only limited to sales, inventory and cost data, but also multi-source complex data such as supply chain information. These data are used to establish accurate prediction models and optimization strategies, effectively improve warehouse efficiency, reduce costs, optimize inventory management, and significantly enhance the responsiveness of the supply chain. Machine learning, data mining and statistical analysis tools play a key role in modern warehouse management. By applying various algorithms such as regression analysis, cluster analysis and time series analysis, these technologies can mine hidden patterns and trends from massive data. For example, regression analysis can be used to predict product sales trends and demand, cluster analysis can identify product market segments to optimize inventory strategies, and time series analysis can help predict seasonal demand changes and help companies adjust warehouse layout and replenishment plans. These precise prediction and optimization strategies provide a scientific basis and strategic guidance for warehouse decisions. However, existing technologies still have some shortcomings in dealing with certain aspects. Data quality and consistency will affect the accuracy and stability of the model, especially the errors in the data cleaning and preprocessing stages will have a significant impact on the final analysis results. Existing models lack sufficient flexibility and adaptability in the face of complex supply chain changes and emergencies, resulting in limited prediction and decision support capabilities of the model. In addition, the subjectivity and limitations in the algorithm selection and parameter adjustment process will also affect the optimization effect and generalization ability of the model, making the model's effect in practical applications unsatisfactory. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a warehouse management model determination method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a method for determining a warehouse management model, comprising the following steps:

[0005] Step S1: Acquire inventory data; perform warehouse data preprocessing on the inventory data to generate warehouse management data; perform statistical analysis on the warehouse management data to generate warehouse statistical data; perform warehouse data real-time collection on the warehouse statistical data to obtain warehouse management real-time data;

[0006] Step S2: Acquire sales cost data and historical sales data; perform numerical feature extraction on sales cost data to generate storage numerical feature data; perform product inventory turnover rate calculation on storage numerical feature data to generate overall turnover rate data; perform classification feature extraction on historical sales data to generate storage classification feature data; perform ABC analysis on storage classification feature data to generate storage area division data; perform comprehensive optimization analysis based on overall turnover rate data and storage area division data to generate storage area allocation data;

[0007] Step S3: construct a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; perform conventional simulation tests and pressure simulation tests on the initial warehouse management model to obtain conventional warehouse simulation test data and warehouse pressure simulation test data; use the conventional warehouse simulation test data to optimize the performance of the initial warehouse management model to generate initial model performance optimization result data; use the warehouse pressure simulation test data to optimize key parameters of the initial model performance optimization result data to generate a warehouse management model;

[0008] Step S4: Upload the warehouse management model to the cloud platform to perform warehouse management operations.

[0009] The present invention firstly obtains and processes inventory data, sales cost data and historical sales data through data acquisition, preprocessing and feature extraction processes. The integration and analysis of these data provide a solid data foundation for warehouse management and can support subsequent decision-making and optimization steps. By statistically analyzing warehouse management data and real-time data acquisition system, real-time monitoring and data update of warehouse operation status can be achieved. This real-time management method helps to quickly discover problems, adjust and optimize warehouse management strategies in a timely manner, and improve operational efficiency and customer service levels. Through feature extraction, ABC analysis and comprehensive optimization analysis, the understanding of product sales characteristics and storage area operation conditions can be further deepened. Through these analyses, managers can identify high-turnover products and optimize storage layout, thereby optimizing inventory management strategies, reducing inventory costs and improving capital turnover efficiency. Model construction, simulation testing and performance optimization are carried out to introduce scientific and systematic management methods for warehouse management. Through simulation testing, the system can simulate different operating scenarios and pressure conditions in a virtual environment and evaluate the performance of the management model under different conditions. The performance optimization process further improves the model so that it can better adapt to the complex and dynamic market environment and improve the accuracy and efficiency of management decisions. The optimized warehouse management model is uploaded to the cloud platform, realizing intelligent and remote warehouse management. The application of the cloud platform allows managers to access and monitor warehouse operations anytime and anywhere, implement remote guidance and adjustments, and improve management response speed and decision-making efficiency. The timely feedback mechanism of the present invention helps to improve the flexibility and response speed of warehouse operations, enabling managers to respond to market changes and customer needs more quickly.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: collecting inventory data; performing warehouse data cleaning on the inventory data to generate warehouse cleaning data; performing data standardization processing on the warehouse cleaning data to generate warehouse standardized data; performing warehouse management data aggregation on the warehouse standardized data to generate warehouse management data;

[0012] Step S12: Performing warehouse data statistical analysis on the warehouse management data to generate warehouse statistical data;

[0013] Step S13: Collect warehouse statistical data in real time to obtain real-time warehouse management data.

[0014] The present invention ensures the quality and consistency of data through inventory data collection, cleaning and standardization. The data cleaning process removes errors, duplications or incomplete parts in the data, and the standardization process enables data of different formats and sources to be uniformly processed and compared, thereby establishing reliable warehouse cleaning data and standardized data. This processing method effectively reduces the error rate of data processing and improves the accuracy of subsequent data analysis and decision-making. The data statistical analysis stage provides in-depth data understanding and insight for warehouse management. By statistically analyzing warehouse management data, managers can identify the changing trends and laws of key indicators such as inventory volume, turnover rate, and inventory cost. This analysis helps to discover potential problems and optimization opportunities, and provides data support for managers to formulate reasonable inventory management strategies. Finally, real-time data collection ensures the acquisition and update of real-time data for warehouse management. The collection of real-time data can timely reflect changes in warehouse operation status, such as inventory changes and order processing speed, thereby helping managers to monitor and adjust operation strategies in real time. The timely feedback mechanism of the present invention helps to improve the flexibility and response speed of warehouse operations, so that managers can respond to market changes and customer needs more quickly.

[0015] Preferably, step S2 comprises the following steps:

[0016] Step S21: Acquire sales cost data and historical sales data; extract specific numerical features from the sales cost data to generate storage numerical feature data; perform storage volatility smoothing processing based on the storage numerical feature data to generate storage numerical smoothing data; calculate product inventory turnover rate based on the storage numerical feature data to generate overall turnover rate data;

[0017] Step S22: extract relevant classification features from historical sales data to generate storage classification feature data; perform binary feature conversion on the storage classification feature data to generate storage classification binary data; perform ABC analysis on the storage classification binary data to generate storage area division data;

[0018] Step S23: Perform comprehensive optimization allocation of storage areas based on the overall turnover rate data and storage area division data to generate storage area allocation data.

[0019] The present invention ensures the integrity and diversity of data sources by acquiring sales cost data and historical sales data. By extracting specific numerical features from sales cost data, storage numerical feature data is generated, which lays a solid foundation for subsequent analysis. Next, storage numerical smoothing data is generated through storage volatility smoothing. This process helps to eliminate abnormal fluctuations in the data, making the data smoother and more reliable, and further improving the accuracy of data analysis. Subsequently, the product inventory turnover rate is calculated for the storage numerical feature data to generate overall turnover rate data, which helps managers understand inventory turnover and optimize inventory management strategies. By extracting relevant classification features from historical sales data, storage classification feature data is generated. This process makes the data structure clearer and facilitates subsequent processing and analysis. Subsequently, through binary feature conversion, the storage classification feature data is converted into storage classification binary data. This step simplifies the data structure and improves the efficiency of data processing. Next, by performing ABC analysis on the storage classification binary data, storage area division data is generated. This process helps to identify the importance of different categories of goods, thereby achieving more accurate inventory management and resource allocation. Finally, based on the overall turnover rate data and storage area division data, the storage area is comprehensively optimized and allocated to generate storage area allocation data. This process takes into account the inventory turnover rate and regional division, achieves the optimal allocation of storage resources, and improves the overall efficiency of storage management. This systematic data processing and optimization method not only improves the accuracy and efficiency of data analysis, but also enhances the flexibility and responsiveness of storage management, significantly improving the overall benefits and competitiveness of storage management. Through this process, enterprises can better manage inventory, reduce operating costs, and improve customer satisfaction, thereby maintaining their advantages in market competition.

[0020] Preferably, step S21 includes the following steps:

[0021] Step S211: Obtain sales cost data and historical sales data;

[0022] Step S212: Calculate the numerical features of the sales cost data to generate a warehouse numerical feature data set, wherein the numerical feature calculation includes average sales cost calculation, cost volatility analysis of sales cost, and change rate calculation of sales cost;

[0023] Step S213: Based on the preset storage liquidity threshold, the storage numerical feature data set is divided into high liquidity data and low liquidity data; when it is determined to be classified as high liquidity data, the physical stability of the goods is monitored by a vibration sensor to ensure that the high liquidity goods remain stable during transportation and storage; when it is determined to be classified as low liquidity data, the tiered management of the goods is planned by combining the physical characteristics of weight and volume to improve the utilization efficiency of storage space;

[0024] Step S214: Calculate the product inventory turnover rate of the warehouse numerical feature data set to generate overall turnover rate data; Calculate the product inventory turnover rate of the warehouse numerical feature data to generate overall turnover rate data.

[0025] The present invention can accurately evaluate the average value, volatility and change rate of storage costs by acquiring sales cost data and historical sales data and performing numerical feature calculation, thereby providing a scientific basis for subsequent storage management decisions. Through the preset storage liquidity threshold, the storage numerical feature data set is divided into high liquidity data and low liquidity data, further improving the accuracy of storage management. For high liquidity data, the physical stability of the goods is monitored by a vibration sensor, which can ensure that high liquidity goods remain stable during transportation and storage, reduce the risk of damage caused by vibration, and improve the safety and reliability of storage operations. At the same time, for low liquidity data, by combining the weight and volume physical properties of the goods, the hierarchical management of the goods is planned, which can maximize the use of storage space and improve storage efficiency. This data-driven method based on physical properties helps to optimize storage resource allocation, reduce invalid space waste, and improve the overall efficiency of storage operations. By calculating the product inventory turnover rate of the storage numerical feature data set and generating overall turnover rate data, the efficiency of storage operations can be fully reflected, and efficient and inefficient inventory management areas can be identified, providing key data support for optimizing inventory strategies. This method not only relies on historical data and numerical features, but also combines physical sensing technology and physical property analysis to form a comprehensive warehouse management system. This multi-dimensional and multi-level data feature extraction and analysis method not only improves the accuracy and reliability of data processing, but also provides a theoretical basis and practical guidance for the intelligent and automated warehouse management in the future. Overall, these measures significantly improve the overall efficiency and benefits of warehouse management through the combination of scientific data analysis and physical properties, and have high academic value and practical application prospects.

[0026] Preferably, step S22 includes the following steps:

[0027] Step S221: Formulate product category labels for historical sales data to generate warehouse product category data; calculate market share for historical sales data to generate warehouse market share data;

[0028] Step S222: Analyze the cargo storage location from the storage market share data; divide the storage space into high-flow calibration areas and low-flow calibration areas according to the exit distance based on the storage market share data, and optimize the physical storage layout of the high-flow calibration areas and the low-flow calibration areas according to the cargo storage location;

[0029] Step S223: Aggregate the warehouse product category data and the warehouse market share data into classification feature data to generate warehouse classification feature data;

[0030] Step S224: perform binary feature conversion on the warehouse classification feature data to generate warehouse classification binary data; perform ABC analysis on the warehouse classification binary data to generate warehouse area division data.

[0031] The present invention generates storage product category data by formulating product category labels for historical sales data, and generates storage market share data by calculating market share for historical sales data. This process lays a solid foundation for subsequent data processing and analysis, and ensures the accuracy and comprehensiveness of the data. Next, by parsing the storage location of the goods from the storage market share data, and dividing the storage space into high-flow calibration areas and low-flow calibration areas according to the export distance based on the market share data, the physical storage layout of the storage is further optimized. This not only takes into account the dynamic changes in market demand, but also combines the liquidity characteristics of the goods, making the division of high-flow calibration areas and low-flow calibration areas more scientific and reasonable. The storage product category data and the storage market share data are aggregated by classification feature data to generate storage classification feature data. This process integrates the multi-dimensional information of product categories and market shares through data aggregation technology, and provides more detailed data support for subsequent feature analysis. Finally, the storage classification feature data is converted into binary features to generate storage classification binary data, and ABC analysis is performed on it to generate storage area division data. This step uses advanced data conversion and analysis technology to convert complex classification feature data into binary data that is easy to understand and operate, and through ABC analysis, it further optimizes the division of warehouse areas, making warehouse management more efficient and refined.

[0032] Preferably, step S3 comprises the following steps:

[0033] Step S31: construct a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; use a cross-validation method to verify the initial warehouse management model to generate model performance evaluation data;

[0034] Step S32: performing a conventional simulation test and a pressure simulation test on the warehouse management initial model, thereby obtaining warehouse conventional simulation test data and warehouse pressure simulation test data;

[0035] Step S33: Use the conventional warehouse simulation test data to optimize the performance of the initial warehouse management model and generate initial model performance optimization result data; use the warehouse pressure simulation test data and model performance evaluation data to optimize the key parameters of the initial model performance optimization result data and generate a warehouse management model.

[0036] The present invention constructs a model through warehouse management real-time data and warehouse area allocation data to generate an initial warehouse management model. This step establishes an initial model that can reflect the actual warehouse management situation by integrating real-time data and area allocation data, laying a solid foundation for subsequent optimization and testing. The initial warehouse management model is verified by a cross-validation method to generate model performance evaluation data. The generalization ability of the model can be effectively evaluated through cross-validation to ensure the stability and accuracy of the model on different data sets. The initial warehouse management model is subjected to conventional simulation tests and pressure simulation tests to obtain conventional warehouse simulation test data and warehouse pressure simulation test data. Conventional simulation tests evaluate the performance of the model in daily operations by simulating warehouse operations under normal working conditions. The pressure simulation test evaluates the performance and stability of the model in dealing with emergencies and high-pressure environments by simulating operations under extreme and abnormal conditions. These two steps ensure that the model can maintain good performance in various situations through comprehensive simulation tests, thereby enhancing the reliability and adaptability of the model. The initial warehouse management model is optimized for performance using conventional warehouse simulation test data to generate initial model performance optimization result data. By analyzing conventional simulation test data, we can identify the deficiencies of the model under normal conditions, conduct targeted performance optimization, and improve the overall efficiency and accuracy of the model. Subsequently, we use the warehouse pressure simulation test data and model performance evaluation data to optimize the key parameters of the initial model performance optimization result data and generate a warehouse management model. This step combines the pressure test data and performance evaluation data to further optimize and adjust the model to ensure that the model can perform at its best in various environments.

[0037] Preferably, step S31 includes the following steps:

[0038] Step S311: perform continuous variable prediction on the warehouse management real-time data to generate warehouse continuous variable prediction data; perform regression analysis based on the warehouse continuous variable prediction data to generate a warehouse model regression module;

[0039] Step S312: Perform storage cluster analysis on the storage area allocation data to generate storage area clustering results; perform storage classification processing based on the storage area clustering results to generate a storage model classification module;

[0040] Step S313: construct a warehouse model according to the warehouse model regression module and the warehouse model classification module to generate an initial warehouse management model.

[0041] The present invention generates warehouse continuous variable prediction data by performing continuous variable prediction on warehouse management real-time data. The beneficial effect of this step is that it can predict the changing trends and key variables that will appear in the warehouse management process in advance, so that warehouse management can be more forward-looking and predictable. Through continuous variable prediction, the warehouse management system can foresee the inventory fluctuations and demand change factors that will appear in the future, so as to prepare countermeasures in advance, reduce the risk of excessive or insufficient inventory, and improve the efficiency and accuracy of inventory management. Regression analysis is performed based on the warehouse continuous variable prediction data to generate a warehouse model regression module. Regression analysis is a commonly used statistical method for determining the relationship between variables. Through regression analysis, the key factors affecting warehouse management can be accurately identified and the corresponding regression model can be established. The beneficial effect of this step is that it can provide quantitative analysis results of warehouse management, making management decisions more scientific and data-driven. The introduction of the regression analysis module enables the warehouse management model to more accurately reflect the actual situation and improve the prediction ability and applicability of the model.

[0042] Perform storage cluster analysis on the storage area allocation data to generate storage area clustering results. Cluster analysis is a technique that divides a data set into multiple similar subsets. Through this technique, different regional characteristics and patterns in storage management can be identified. The beneficial effect of this step is that it can effectively divide and manage storage areas according to actual conditions and improve the utilization efficiency of storage resources. Through cluster analysis, the specific needs and characteristics of different regions can be discovered, so as to carry out targeted management and optimization. Based on the storage area clustering results, storage classification processing is carried out to generate a storage model classification module. Classification processing can further refine and optimize the management of storage areas, making storage management more refined and personalized. The beneficial effect of this step is that it can formulate corresponding management strategies and measures according to the characteristics of different regions, and improve the flexibility and response speed of storage management. The introduction of the classification module enables the warehouse management system to better adapt to the needs of different regions and provide more accurate and efficient services. Finally, the warehouse model is constructed according to the warehouse model regression module and the warehouse model classification module to generate the initial warehouse management model. The beneficial effect of this comprehensive step is that, by combining the results of regression analysis and classification processing, a comprehensive and scientific initial model of warehouse management is established. This model can not only reflect the overall situation of warehouse management, but also conduct specific analysis and optimization for different areas and variables, thus improving the overall efficiency and accuracy of warehouse management.

[0043] Preferably, step S32 includes the following steps:

[0044] Step S321: Create a regular simulation environment condition for the warehouse management initial model to generate a normal virtual warehouse simulation environment, wherein the normal virtual warehouse simulation environment includes a normal order processing flow virtual warehouse simulation environment, an inventory replenishment flow virtual warehouse simulation environment, and a return processing flow virtual warehouse simulation environment;

[0045] Step S322: applying the warehouse management initial model to a normal virtual warehouse simulation environment to perform normal environment simulation and generate a warehouse conventional simulation test data set;

[0046] Step S323: creating stress simulation environment conditions for the initial warehouse management model to generate a stress virtual warehouse simulation environment, wherein the stress virtual warehouse simulation environment includes a stress order processing process virtual warehouse simulation environment, a logistics delay process virtual warehouse simulation environment, and a supply chain interruption virtual warehouse simulation environment;

[0047] Step S324: Apply the initial warehouse management model to the pressure virtual warehouse simulation environment to generate a warehouse pressure simulation test data set.

[0048] The present invention generates a normal virtual warehouse simulation environment by creating a conventional simulation environment condition for the warehouse management initial model, including a virtual simulation environment of a normal order processing flow, an inventory replenishment flow, and a return processing flow. The beneficial effect of this step is that by establishing a simulation environment of a conventional operation flow in the real world, the performance of the warehouse management system in daily operation can be simulated. This helps to identify and solve problems that may occur under normal operating conditions, and ensure that the warehouse management system can operate stably and efficiently under normal conditions. The warehouse management initial model is applied to the normal virtual warehouse simulation environment for normal environment simulation, and a warehouse conventional simulation test data set is generated. The beneficial effect of this step is that by running the initial model in a normal virtual environment, detailed simulation test data can be collected to evaluate the performance of the model in different conventional operation scenarios. These data provide a basis for subsequent model optimization, ensuring the reliability and efficiency of the model under normal operating conditions. The warehouse management initial model is subjected to a stress simulation environment condition to generate a stress virtual warehouse simulation environment, including a virtual simulation environment of a stress order processing flow, a logistics delay flow, and a supply chain interruption. The beneficial effect of this step is that by creating a simulation environment under various extreme and stress conditions, the performance of the warehouse management system in abnormal and emergency situations can be simulated. This helps to identify potential risks and weaknesses of the system under high pressure and extreme conditions, so as to carry out targeted optimization and improvement. Finally, the initial warehouse management model is applied to the pressure virtual warehouse simulation environment to generate a warehouse pressure simulation test data set. The beneficial effect of this step is that by running the initial model in the pressure simulation environment, detailed data of the system under high load and extreme conditions can be collected. These data help to comprehensively evaluate the performance and stability of the system in various extreme situations, ensuring that the warehouse management system can maintain stable and efficient operation in the face of emergencies.

[0049] Preferably, step S33 includes the following steps:

[0050] Step S331: constructing a warehouse data optimization strategy for warehouse conventional simulation test data to generate a warehouse model optimization strategy;

[0051] Step S332: Optimizing the performance of the initial warehouse management model based on the warehouse model optimization strategy, and generating initial model performance optimization result data;

[0052] Step S333: Utilize the warehouse pressure simulation test data to optimize key parameters of the initial model performance optimization result data to generate a warehouse management model.

[0053] The present invention generates a warehouse model optimization strategy by constructing a warehouse data optimization strategy for conventional warehouse simulation test data. The beneficial effect of this step is that, through in-depth analysis of conventional simulation test data and formulation of optimization strategies, potential improvement points and optimization directions of the warehouse management system under conventional operation can be identified. The warehouse data optimization strategy not only provides scientific guidance and basis for subsequent performance optimization, but also ensures the systematicness and effectiveness of the optimization process, thereby improving the efficiency and reliability of the warehouse management system in daily operation. Based on the warehouse model optimization strategy, the initial model of warehouse management is optimized to generate initial model performance optimization result data. The beneficial effect of this step is that, through a systematic performance optimization process, the overall performance of the initial model of warehouse management can be significantly improved. The optimization process makes full use of the warehouse model optimization strategy constructed in the early stage to ensure the pertinence and effectiveness of the optimization. The performance optimization result data not only reflects the optimization effect of the model under conventional operation, but also provides a basis and reference for subsequent key parameter optimization. Finally, the warehouse pressure simulation test data is used to optimize the key parameters of the initial model performance optimization result data to generate a warehouse management model. The beneficial effect of this step is that through further analysis of the initial model performance optimization result data and adjustment of key parameters, the adaptability and stability of the warehouse management model under extreme and stress conditions can be improved. The pressure simulation test data provides a wealth of performance data under extreme conditions. Through in-depth analysis of these data and optimization and adjustment of key parameters, potential problems of the system under high pressure and extreme conditions can be identified and solved, ensuring the reliability and stability of the warehouse management model under various operating conditions.

[0054] In this specification, a warehouse management model determination system is provided, which is used to execute the above-mentioned warehouse management model determination method. The warehouse management model determination system includes:

[0055] Data preparation and preprocessing module: used to obtain inventory data; perform warehouse data preprocessing on inventory data to generate warehouse management data; perform statistical analysis on warehouse management data to generate warehouse statistical data; perform real-time warehouse data collection on warehouse statistical data to obtain real-time warehouse management data;

[0056] Optimization and analysis module: used to obtain sales cost data and historical sales data; extract numerical features from sales cost data to generate warehouse numerical feature data; calculate product inventory turnover rate from warehouse numerical feature data to generate overall turnover rate data; extract classification features from historical sales data to generate warehouse classification feature data; perform ABC analysis on warehouse classification feature data to generate warehouse area division data; perform comprehensive optimization analysis based on overall turnover rate data and warehouse area division data to generate warehouse area allocation data.

[0057] Model building and optimization module: used to build models based on real-time warehouse management data and warehouse area allocation data to generate an initial warehouse management model; perform conventional simulation tests and pressure simulation tests on the initial warehouse management model to obtain conventional warehouse simulation test data and warehouse pressure simulation test data; use conventional warehouse simulation test data to optimize the performance of the initial warehouse management model to generate initial model performance optimization result data; use warehouse pressure simulation test data to optimize key parameters of the initial model performance optimization result data to generate a warehouse management model;

[0058] Cloud platform integration and deployment module: used to upload the warehouse management model to the cloud platform to perform warehouse management operations.

[0059] The beneficial effect of the present invention is that the accuracy and integrity of the data are ensured by obtaining inventory data, performing data preprocessing and statistical analysis. This module can not only generate warehouse management data and warehouse statistical data, but also provide real-time dynamic information of warehouse management through real-time data collection. This lays a solid data foundation for subsequent optimization analysis and model construction, and helps to improve the scientificity and effectiveness of warehouse management decisions. By obtaining sales cost data and historical sales data, numerical feature extraction and classification feature extraction are performed to generate warehouse numerical feature data and warehouse classification feature data. The processing and analysis of these data can not only calculate the overall turnover rate data, but also generate warehouse area division data through ABC analysis. Based on these data, comprehensive optimization analysis is performed, and the generated warehouse area allocation data can effectively optimize the utilization rate of warehouse space and improve the efficiency and accuracy of warehouse management. By constructing a model based on warehouse management real-time data and warehouse area allocation data, an initial warehouse management model is generated. Through conventional simulation tests and pressure simulation tests, the obtained test data can be optimized for performance and generate initial model performance optimization result data. By using these data to optimize key parameters, the generated warehouse management model can improve the adaptability and stability of the system under various operating conditions. This module ensures the efficient operation and continuous optimization of the model, and improves the overall performance of the warehouse management system. Finally, by uploading the optimized warehouse management model to the cloud platform to perform warehouse management operations, the high availability and scalability of the system are ensured. The integration and deployment of the cloud platform can not only provide efficient computing resources, but also realize remote management and real-time monitoring of the warehouse management system, improving the flexibility and response speed of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of the steps of a method for determining a warehouse management model according to the present invention;

[0061] Figure 2 for Figure 1 Detailed step flow diagram of step S2;

[0062] Figure 3 for Figure 2 Detailed step flow diagram of step S21;

[0063] Figure 4 for Figure 2 Detailed step flow diagram of step S22;

[0064] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0065] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0066] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0067] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0068] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for determining a warehouse management model, comprising the following steps:

[0069] Step S1: Acquire inventory data; perform warehouse data preprocessing on the inventory data to generate warehouse management data; perform statistical analysis on the warehouse management data to generate warehouse statistical data; perform warehouse data real-time collection on the warehouse statistical data to obtain warehouse management real-time data;

[0070] Step S2: Acquire sales cost data and historical sales data; perform numerical feature extraction on sales cost data to generate storage numerical feature data; perform product inventory turnover rate calculation on storage numerical feature data to generate overall turnover rate data; perform classification feature extraction on historical sales data to generate storage classification feature data; perform ABC analysis on storage classification feature data to generate storage area division data; perform comprehensive optimization analysis based on overall turnover rate data and storage area division data to generate storage area allocation data.

[0071] Step S3: construct a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; perform conventional simulation tests and pressure simulation tests on the initial warehouse management model to obtain conventional warehouse simulation test data and warehouse pressure simulation test data; use the conventional warehouse simulation test data to optimize the performance of the initial warehouse management model to generate initial model performance optimization result data; use the warehouse pressure simulation test data to optimize key parameters of the initial model performance optimization result data to generate a warehouse management model;

[0072] Step S4: Upload the warehouse management model to the cloud platform to perform warehouse management operations.

[0073] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a warehouse management model determination method of the present invention. In this example, the warehouse management model determination method includes the following steps:

[0074] Step S1: Acquire inventory data; perform warehouse data preprocessing on the inventory data to generate warehouse management data; perform statistical analysis on the warehouse management data to generate warehouse statistical data; perform warehouse data real-time collection on the warehouse statistical data to obtain warehouse management real-time data;

[0075] In an embodiment of the present invention, inventory data is obtained through an automated system or manual input, and these data include basic information such as product name, quantity, and location. Subsequently, these inventory data are preprocessed for warehouse data, including data cleaning, deduplication, normalization, and structuring, to ensure the accuracy and consistency of the data, and to generate warehouse management data. Statistical analysis methods are used to conduct in-depth analysis of warehouse management data, and warehouse statistical data are generated through descriptive statistics, trend analysis, and anomaly detection methods. These statistical data can reveal key indicators of inventory status, turnover rate, and demand forecasting. In order to achieve real-time and high efficiency of warehouse management, warehouse statistical data is further collected in real time. This process is usually implemented by deploying RFID Internet of Things devices and using the stream processing framework Apache Kafka big data technology to obtain real-time data for warehouse management. The collection and processing of real-time data can not only provide the real-time status of the current inventory, but also support dynamic decision-making and timely response to inventory changes.

[0076] Step S2: Acquire sales cost data and historical sales data; perform numerical feature extraction on sales cost data to generate storage numerical feature data; perform product inventory turnover rate calculation on storage numerical feature data to generate overall turnover rate data; perform classification feature extraction on historical sales data to generate storage classification feature data; perform ABC analysis on storage classification feature data to generate storage area division data; perform comprehensive optimization analysis based on overall turnover rate data and storage area division data to generate storage area allocation data.

[0077] In an embodiment of the present invention, sales cost data and historical sales data are obtained through an enterprise resource planning (ERP) system. For sales cost data, numerical feature extraction technology is applied, and key numerical features such as average cost and cost volatility are extracted from the data using statistical methods or machine learning algorithms to generate storage numerical feature data. Next, the product inventory turnover rate is calculated for the storage numerical feature data. The inventory turnover rate is calculated by the formula: inventory turnover rate = sales cost / average inventory value to generate overall turnover rate data, which reflects the efficiency of inventory use and capital occupation. Classification feature extraction is performed on historical sales data, and K-means cluster analysis is used to classify sales data according to different dimensions such as product category, sales area, and customer type to generate storage classification feature data. Then, ABC analysis is performed based on the storage classification feature data, and the Pareto principle (80 / 20 rule) is adopted to divide products into three categories: Class A (high value, high frequency), Class B (medium value and frequency) and Class C (low value, low frequency), and generate storage area division data. Finally, comprehensive optimization analysis is performed based on the overall turnover rate data and storage area division data. In this process, linear programming, integer programming or heuristic algorithm optimization techniques are used to comprehensively consider inventory turnover rate and product category characteristics, and to rationally allocate storage space to improve the efficiency and response speed of warehouse management, and ultimately generate storage area allocation data.

[0078] Step S3: construct a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; perform conventional simulation tests and pressure simulation tests on the initial warehouse management model to obtain conventional warehouse simulation test data and warehouse pressure simulation test data; use the conventional warehouse simulation test data to optimize the performance of the initial warehouse management model to generate initial model performance optimization result data; use the warehouse pressure simulation test data to optimize key parameters of the initial model performance optimization result data to generate a warehouse management model;

[0079] In the embodiment of the present invention, the model is constructed based on the real-time data of warehouse management and the warehouse area allocation data, and the initial model of warehouse management is constructed by using machine learning and data mining technology. The model reflects the basic operation status and operation rules of the warehouse system by analyzing and modeling the real-time data of warehouse and the area allocation data. The initial model of warehouse management is subjected to conventional simulation test and stress simulation test. The conventional simulation test simulates the operation of the warehouse system under normal working conditions, and uses AnyLogic simulation software to generate conventional warehouse simulation test data to evaluate the accuracy and stability of the model. The stress simulation test simulates the operation of the warehouse system under high load and extreme conditions, evaluates the performance and coping ability of the system under stress conditions, and generates warehouse stress simulation test data. These two types of simulation test data provide important reference for subsequent model optimization. The initial model of warehouse management is optimized by using conventional warehouse simulation test data. By analyzing the conventional simulation test data, the bottlenecks and deficiencies in the model are identified, and the genetic algorithm optimization algorithm is used to adjust and improve it, and the initial model performance optimization result data is generated. The data reflects the improvement effect and performance improvement of the optimized model under normal conditions. The key parameters of the initial model performance optimization result data are optimized using warehouse pressure simulation test data. In this process, the key parameters of the model are further optimized through sensitivity analysis and parameter adjustment, so that it can also show high performance and stability under pressure conditions, and finally generate a warehouse management model.

[0080] Step S4: Upload the warehouse management model to the cloud platform to perform warehouse management operations.

[0081] In the embodiment of the present invention, the optimized warehouse management model is prepared to ensure that its format and structure meet the requirements of the cloud platform. Then, the warehouse management model is uploaded to the cloud platform through the secure HTTPS data transmission protocol. This step ensures that the model is not interfered with by the outside world during the transmission process, and ensures the integrity and security of the data. After the model is uploaded to the cloud platform, it is deployed and executed using cloud computing technology. The cloud platform uses Docker virtualization technology and containerization technology to achieve efficient management and allocation of resources, thereby supporting the deployment and operation of the warehouse management model. On the cloud platform, the warehouse management model can use the computing resources provided by the cloud, such as processors, memory, and storage, to perform real-time calculations and analysis to achieve automation and intelligence of warehouse management operations. The execution of the warehouse management model on the cloud platform includes multiple levels. The model will receive and process various data from the warehouse site in real time, such as inventory data, order data, and logistics data, and realize rapid data processing and response through ApacheKafka data stream processing technology. Using the computing power and storage capacity of the cloud platform, a large amount of data is analyzed and mined in real time to generate optimization decisions and operation instructions for warehouse operations. These decisions and instructions are fed back to the warehouse site through the cloud platform to achieve dynamic scheduling and management of warehouse operations. The cloud platform also provides high availability and fault tolerance to ensure that the warehouse management model can respond to various emergencies during operation and maintain the stability and continuity of the system. The elastic expansion capability of the cloud platform enables the warehouse management model to dynamically adjust computing resources according to actual needs, achieving efficient resource utilization and cost control.

[0082] Preferably, step S1 comprises the following steps:

[0083] Step S11: collecting inventory data; performing warehouse data cleaning on the inventory data to generate warehouse cleaning data; performing data standardization processing on the warehouse cleaning data to generate warehouse standardized data; performing warehouse management data aggregation on the warehouse standardized data to generate warehouse management data;

[0084] Step S12: Performing warehouse data statistical analysis on the warehouse management data to generate warehouse statistical data;

[0085] Step S13: Collect warehouse statistical data in real time to obtain real-time warehouse management data.

[0086] In an embodiment of the present invention, inventory data is collected by an automated system or manual input. These raw data usually contain basic information such as product name, quantity, and location. Next, the inventory data is cleaned, and data cleaning techniques such as missing value filling, outlier detection and processing, and duplicate data deletion are used to remove or correct errors and inconsistencies in the data, thereby generating higher quality warehouse cleaned data. After the warehouse cleaned data is generated, data standardization is performed. Standardization processing technology converts the scale of the data to make it conform to a unified standard and format, such as normalization and Z-score standardization methods, to ensure that the data is consistent and comparable in different systems and algorithms, thereby generating warehouse standardized data. Subsequently, warehouse management data aggregation is performed on the warehouse standardized data. Data aggregation techniques, such as group aggregation, rolling calculation, and window function, summarize and calculate the standardized data according to specific dimensions, such as time and product category, to generate more meaningful warehouse management data. Statistical analysis methods are used to conduct in-depth analysis of warehouse management data. Descriptive statistical analysis, such as mean, variance, frequency distribution, and advanced analytical techniques, trend analysis, and correlation analysis, are applied to warehouse management data to generate warehouse statistics. These statistics can reveal key indicators of inventory status, inventory turnover, and demand forecasting, providing data support for decision-making. Finally, warehouse statistics are collected in real time. By deploying RFID IoT devices and using Apache Kafka big data technology, warehouse statistics are monitored and collected in real time to obtain real-time warehouse management data. These real-time data not only provide the real-time status of the current inventory, but also support dynamic decision-making and timely response to inventory changes to ensure the efficient operation of the warehouse system.

[0087] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0088] Step S21: Acquire sales cost data and historical sales data; extract specific numerical features from the sales cost data to generate storage numerical feature data; perform storage volatility smoothing processing based on the storage numerical feature data to generate storage numerical smoothing data; calculate product inventory turnover rate based on the storage numerical feature data to generate overall turnover rate data;

[0089] Step S22: extract relevant classification features from historical sales data to generate storage classification feature data; perform binary feature conversion on the storage classification feature data to generate storage classification binary data; perform ABC analysis on the storage classification binary data to generate storage area division data;

[0090] Step S23: Perform comprehensive optimization allocation of storage areas based on the overall turnover rate data and storage area division data to generate storage area allocation data.

[0091] In an embodiment of the present invention, sales cost data and historical sales data are obtained through an enterprise resource planning (ERP) system. For sales cost data, a specific numerical feature extraction technology is applied, and key numerical features such as average cost and cost volatility are extracted from the data using statistical methods or machine learning algorithms to generate warehouse numerical feature data. Warehouse volatility smoothing is performed based on these numerical feature data, and smoothing techniques such as exponential smoothing and moving average are used to reduce data fluctuations and generate warehouse numerical smoothing data. This step helps to eliminate noise in the data and improve the stability of the data. Product inventory turnover rate is calculated for warehouse numerical feature data. The inventory turnover rate is calculated by the formula: inventory turnover rate = sales cost / average inventory value to generate overall turnover rate data, which reflects the efficiency of inventory use and capital occupation. Relevant classification features are extracted for historical sales data. K-means cluster analysis is used to classify sales data according to different dimensions, such as product category, sales area, and customer type, to generate warehouse classification feature data. Then, these classification feature data are converted into binary features, and a One-Hot encoding method is used to convert the classification features into binary feature data to generate warehouse classification binary data. This step converts discrete classification features into binary data that can be directly used for calculation and analysis. Perform ABC analysis based on storage classification binary data. Using the Pareto principle (80 / 20 rule), products are divided into three categories: Class A (high value, high frequency), Class B (medium value and frequency) and Class C (low value, low frequency), and generate storage area division data. This analysis helps to optimize inventory management, focusing on high-value and high-frequency products. Comprehensive optimization allocation of storage areas is performed based on overall turnover rate data and storage area division data. Using linear programming, integer programming or heuristic algorithm optimization techniques, comprehensively consider inventory turnover rate and product category characteristics, and reasonably allocate storage space to improve the efficiency and response speed of warehouse management, and generate storage area allocation data.

[0092] As an example of the present invention, refer to Figure 3 As shown, in this example, step S21 includes:

[0093] Step S211: Obtain sales cost data and historical sales data;

[0094] Step S212: Calculate the numerical features of the sales cost data to generate a warehouse numerical feature data set, wherein the numerical feature calculation includes average sales cost calculation, cost volatility analysis of sales cost, and change rate calculation of sales cost;

[0095] Step S213: Based on the preset storage liquidity threshold, the storage numerical feature data set is divided into high liquidity data and low liquidity data; when it is determined to be classified as high liquidity data, the physical stability of the goods is monitored by a vibration sensor to ensure that the high liquidity goods remain stable during transportation and storage; when it is determined to be classified as low liquidity data, the tiered management of the goods is planned by combining the physical characteristics of weight and volume to improve the utilization efficiency of storage space;

[0096] Step S214: Calculate the product inventory turnover rate of the warehouse numerical feature data set to generate overall turnover rate data; Calculate the product inventory turnover rate of the warehouse numerical feature data to generate overall turnover rate data.

[0097] In the embodiment of the present invention, by acquiring sales cost data and historical sales data, a large amount of basic data is collected using data acquisition technology, which provides rich original materials for subsequent analysis. The numerical feature calculation of the sales cost data includes the average sales cost calculation, the volatility analysis of the sales cost, and the change rate calculation of the sales cost. This process uses technical means such as statistical analysis and time series analysis. Through deep mining of the data, a warehouse numerical feature data set is generated, thereby revealing the dynamic characteristics and change trends of the sales cost. The warehouse numerical feature data set is divided into high-liquidity data and low-liquidity data by a preset warehouse liquidity threshold. This step uses threshold division and cluster analysis technology to classify the data set according to the preset liquidity standard, thereby providing a basis for the management of different liquidity data. For high-liquidity data, the physical stability of the goods is monitored by a vibration sensor. This process combines sensor technology and Internet of Things technology. By real-time monitoring of the vibration of the goods during transportation and storage, the physical stability and safety of high-liquidity goods are ensured. For low-liquidity data, the hierarchical management of goods is planned by combining the physical characteristics of weight and volume. This process uses physical characteristic analysis and optimization algorithms. Through the analysis of weight and volume data, the storage location and method of goods are reasonably planned, and the utilization efficiency of storage space is improved. The product inventory turnover rate is calculated for the warehouse numerical feature data set to generate overall turnover rate data. This process uses inventory management and logistics analysis technology. Through the calculation of numerical feature data, the inventory turnover of products is accurately reflected, providing data support for optimizing inventory management.

[0098] As an example of the present invention, refer to Figure 4 As shown, in this example, step S22 includes:

[0099] Step S221: Formulate product category labels for historical sales data to generate warehouse product category data; calculate market share for historical sales data to generate warehouse market share data;

[0100] Step S222: Analyze the cargo storage location from the storage market share data; divide the storage space into high-flow calibration areas and low-flow calibration areas according to the exit distance based on the storage market share data, and optimize the physical storage layout of the high-flow calibration areas and the low-flow calibration areas according to the cargo storage location;

[0101] Step S223: Aggregate the warehouse product category data and the warehouse market share data into classification feature data to generate warehouse classification feature data;

[0102] Step S224: perform binary feature conversion on the warehouse classification feature data to generate warehouse classification binary data; perform ABC analysis on the warehouse classification binary data to generate warehouse area division data.

[0103] In an embodiment of the present invention, by formulating product category labels for historical sales data, storage product category data is generated. This process uses classification algorithms and label generation technology to ensure the accuracy and consistency of historical sales data in category labels. Market share calculation is performed on historical sales data to generate storage market share data. Market analysis and proportion calculation technology are used. Through quantitative analysis of sales data, the market share of each type of product is accurately reflected. The storage location of the goods is parsed from the storage market share data. Based on the storage market share data, the storage space is divided into high-flow calibration areas and low-flow calibration areas according to the export distance. This process uses spatial data analysis and optimization algorithms. By parsing the market share data, the storage location of the goods is reasonably allocated, and the physical storage layout of the high-flow calibration area and the low-flow calibration area is optimized in combination with the physical characteristics of the export distance. This layout optimization technology not only takes into account the differences in cargo mobility, but also improves the utilization efficiency of storage space through space optimization algorithms. The warehouse product category data and warehouse market share data are aggregated into classified feature data to generate warehouse classification feature data. Feature aggregation technology is used to integrate multi-dimensional data to form comprehensive warehouse classification feature data, which provides rich feature information for subsequent analysis. Binary feature conversion is performed on warehouse classification feature data to generate warehouse classification binary data. This process utilizes data binarization technology to simplify the data structure by converting classification feature data into binary form, which is convenient for subsequent analysis and processing. Then, ABC analysis is performed on warehouse classification binary data to generate warehouse area division data. The ABC classification method is used to divide the warehouse area into different importance levels through analysis of binary data, thereby realizing the refinement and scientificization of warehouse management.

[0104] Preferably, step S3 comprises the following steps:

[0105] Step S31: construct a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; use a cross-validation method to verify the initial warehouse management model to generate model performance evaluation data;

[0106] Step S32: performing a conventional simulation test and a pressure simulation test on the warehouse management initial model, thereby obtaining warehouse conventional simulation test data and warehouse pressure simulation test data;

[0107] Step S33: Use the conventional warehouse simulation test data to optimize the performance of the initial warehouse management model and generate initial model performance optimization result data; use the warehouse pressure simulation test data and model performance evaluation data to optimize the key parameters of the initial model performance optimization result data and generate a warehouse management model.

[0108] In the embodiment of the present invention, the model is constructed based on the real-time data of warehouse management and the warehouse area allocation data, and the initial warehouse management model is created using machine learning and data mining technology. These real-time data and allocation data contain key information of warehouse operations, such as inventory levels, product locations, and order requirements. The model simulates the operation and decision-making process of the warehouse system by analyzing these data. The initial warehouse management model is verified using the cross-validation method. Cross-validation is a commonly used model evaluation technique. By dividing the data set into a training set and a validation set, the model is trained and validated in a loop to generate model performance evaluation data. This process can effectively evaluate the accuracy, stability, and generalization ability of the model and ensure that the model performs consistently on different data sets. The initial warehouse management model is subjected to conventional simulation tests and stress simulation tests. Conventional simulation tests simulate the operation of the warehouse system under normal working conditions, and use AnyLogic simulation software to generate warehouse conventional simulation test data to evaluate the accuracy and stability of the model. Stress simulation tests simulate the operation of the warehouse system under high load and extreme conditions, evaluate the performance and coping ability of the system under stress conditions, and generate warehouse stress simulation test data. These two types of simulation test data provide important references for subsequent model optimization. The performance of the initial warehouse management model is optimized using conventional warehouse simulation test data. By analyzing conventional simulation test data, the bottlenecks and deficiencies in the model are identified, and the genetic algorithm optimization algorithm is used to adjust and improve it to generate the initial model performance optimization result data. This data reflects the improvement effect and performance improvement of the optimized model under conventional conditions. The key parameters of the initial model performance optimization result data are optimized using warehouse pressure simulation test data and model performance evaluation data. In this process, through sensitivity analysis and parameter adjustment, the key parameters of the model are further optimized so that it can also show high performance and stability under pressure conditions, and finally a warehouse management model is generated.

[0109] Preferably, step S31 includes the following steps:

[0110] Step S311: perform continuous variable prediction on the warehouse management real-time data to generate warehouse continuous variable prediction data; perform regression analysis based on the warehouse continuous variable prediction data to generate a warehouse model regression module;

[0111] Step S312: Perform storage cluster analysis on the storage area allocation data to generate storage area clustering results; perform storage classification processing based on the storage area clustering results to generate a storage model classification module;

[0112] Step S313: construct a warehouse model according to the warehouse model regression module and the warehouse model classification module to generate an initial warehouse management model.

[0113] In the embodiment of the present invention, firstly, continuous variable prediction is performed on the real-time data of warehouse management. In this step, the continuous variables in the real-time data are predicted by the time series ARIMA analysis method to generate warehouse continuous variable prediction data. These continuous variables include key indicators such as inventory level, demand, and replenishment time. By predicting these variables, potential problems and trends can be identified in advance. Regression analysis is performed based on the warehouse continuous variable prediction data. Linear regression analysis technology is used to explore the relationship between continuous variables, identify important factors affecting warehouse management, and establish corresponding mathematical models to generate a warehouse model regression module. This module can effectively explain and predict key variables in the warehouse system and provide a basis for optimizing decisions. Warehouse cluster analysis is performed on the warehouse area allocation data. K-means and hierarchical cluster analysis techniques are used to group the warehouse area allocation data according to similarity and generate warehouse area clustering results. These clustering results can reveal the natural grouping and potential patterns of the warehouse area, providing a basis for subsequent classification processing. Warehouse classification processing is performed based on the warehouse area clustering results, and the clustering results are further refined and classified using a classification algorithm to generate a warehouse model classification module. This module improves the level of refinement of warehouse management by accurately classifying data from different warehouse areas. The warehouse model is constructed based on the warehouse model regression module and the warehouse model classification module. During the model construction process, the results of the regression module and the classification module are comprehensively utilized, and the ensemble learning method is adopted to integrate the prediction and classification functions of different modules to generate the initial warehouse management model. This initial model realizes comprehensive modeling and optimization of the warehouse system through multi-level and multi-angle data analysis and processing.

[0114] Preferably, step S32 includes the following steps:

[0115] Step S321: Create a regular simulation environment condition for the warehouse management initial model to generate a normal virtual warehouse simulation environment, wherein the normal virtual warehouse simulation environment includes a normal order processing flow virtual warehouse simulation environment, an inventory replenishment flow virtual warehouse simulation environment, and a return processing flow virtual warehouse simulation environment;

[0116] Step S322: applying the warehouse management initial model to a normal virtual warehouse simulation environment to perform normal environment simulation and generate a warehouse conventional simulation test data set;

[0117] Step S323: creating stress simulation environment conditions for the initial warehouse management model to generate a stress virtual warehouse simulation environment, wherein the stress virtual warehouse simulation environment includes a stress order processing process virtual warehouse simulation environment, a logistics delay process virtual warehouse simulation environment, and a supply chain interruption virtual warehouse simulation environment;

[0118] Step S324: Apply the initial warehouse management model to the pressure virtual warehouse simulation environment to generate a warehouse pressure simulation test data set.

[0119] In an embodiment of the present invention, a conventional simulation environment condition is created for the initial model of warehouse management. This includes creating a normal virtual warehouse simulation environment, which includes a virtual warehouse simulation environment for a normal order processing process, an inventory replenishment process, and a return processing process. The normal virtual warehouse simulation environment generates simulation environment data that conforms to the actual situation by simulating typical operating processes and activities, such as order receipt, inventory update, and return processing. These data sets reflect the operation of the warehouse system under normal operating conditions and provide basic data for subsequent model application and performance evaluation. The initial model of warehouse management is applied to the normal virtual warehouse simulation environment for normal environment simulation. By embedding the model into the normal virtual warehouse simulation environment, the performance of the model under actual operating conditions is simulated and recorded. This process generates a conventional warehouse simulation test data set for evaluating the accuracy, efficiency, and stability of the model under standard operating conditions and verifying the actual application capability of the model. The initial model of warehouse management is subjected to stress simulation environment conditions. The stress virtual warehouse simulation environment includes a virtual warehouse simulation environment for stress order processing processes, logistics delay processes, and supply chain interruption processes. These environments simulate the operation of warehouse systems under abnormal or high-load conditions, such as sudden increase in orders, logistics problems, and supply chain disruptions, and generate simulation environment data that meets actual challenges. These data sets are used to simulate and test the response capabilities and performance of warehouse management models under stress conditions. The initial warehouse management model is applied to the stress virtual warehouse simulation environment to generate a warehouse stress simulation test data set. Through the application and simulation testing of the model in a stress environment, the performance and response capabilities of the model when facing stressful situations are recorded and analyzed.

[0120] Preferably, step S33 includes the following steps:

[0121] Step S331: constructing a warehouse data optimization strategy for warehouse conventional simulation test data to generate a warehouse model optimization strategy;

[0122] Step S332: Optimizing the performance of the initial warehouse management model based on the warehouse model optimization strategy, and generating initial model performance optimization result data;

[0123] Step S333: Utilize the warehouse pressure simulation test data to optimize key parameters of the initial model performance optimization result data to generate a warehouse management model.

[0124] In the embodiment of the present invention, a warehouse data optimization strategy is constructed by using conventional warehouse simulation test data. This step involves using data analysis and optimization algorithms, such as data cleaning, feature selection, and anomaly detection technology, to deeply analyze and process the simulation test data and construct an optimization strategy suitable for the warehouse management model. The establishment of the optimization strategy aims to improve the quality and consistency of the data, reduce noise and deviation, and thus ensure the effectiveness of subsequent performance optimization and model application. The performance of the initial warehouse management model is optimized based on the warehouse model optimization strategy. In this stage, the initial warehouse management model is carefully adjusted and improved using the optimized simulation data to improve the model's prediction ability, decision support ability, and operational efficiency. The performance optimization process includes the adjustment of model parameters, the optimization of algorithms, and the improvement of model structure to ensure that the model can more accurately reflect the operating laws and changing trends of the warehouse management system in actual applications. Finally, the warehouse pressure simulation test data is used to optimize the key parameters of the initial model performance optimization result data. Pressure simulation test data usually includes simulation data in the face of abnormal situations, such as high-load order processing and supply chain interruptions, which are used to evaluate and adjust the performance of the model under pressure conditions.

[0125] In this specification, a warehouse management model determination system is provided, which is used to execute the above-mentioned warehouse management model determination method. The warehouse management model determination system includes:

[0126] Data preparation and preprocessing module: obtain inventory data; perform warehouse data preprocessing on inventory data to generate warehouse management data; perform statistical analysis on warehouse management data to generate warehouse statistical data; perform real-time warehouse data collection on warehouse statistical data to obtain real-time warehouse management data;

[0127] Optimization and analysis module: obtain sales cost data and historical sales data; extract numerical features of sales cost data to generate warehouse numerical feature data; calculate product inventory turnover rate based on warehouse numerical feature data to generate overall turnover rate data; extract classification features of historical sales data to generate warehouse classification feature data; perform ABC analysis on warehouse classification feature data to generate warehouse area division data; perform comprehensive optimization analysis based on overall turnover rate data and warehouse area division data to generate warehouse area allocation data.

[0128] Model building and optimization module: Model building is performed based on real-time warehouse management data and warehouse area allocation data to generate an initial warehouse management model; conventional simulation tests and pressure simulation tests are performed on the initial warehouse management model to obtain conventional warehouse simulation test data and warehouse pressure simulation test data; the performance of the initial warehouse management model is optimized using the conventional warehouse simulation test data to generate initial model performance optimization result data; key parameters of the initial model performance optimization result data are optimized using the warehouse pressure simulation test data to generate a warehouse management model;

[0129] Cloud platform integration and deployment module: upload the warehouse management model to the cloud platform to perform warehouse management operations.

[0130] The beneficial effect of the present invention is that the accuracy and integrity of the data are ensured by obtaining inventory data, performing data preprocessing and statistical analysis. This module can not only generate warehouse management data and warehouse statistical data, but also provide real-time dynamic information of warehouse management through real-time data collection. This lays a solid data foundation for subsequent optimization analysis and model construction, and helps to improve the scientificity and effectiveness of warehouse management decisions. By obtaining sales cost data and historical sales data, numerical feature extraction and classification feature extraction are performed to generate warehouse numerical feature data and warehouse classification feature data. The processing and analysis of these data can not only calculate the overall turnover rate data, but also generate warehouse area division data through ABC analysis. Based on these data, comprehensive optimization analysis is performed, and the generated warehouse area allocation data can effectively optimize the utilization rate of warehouse space and improve the efficiency and accuracy of warehouse management. By constructing a model based on warehouse management real-time data and warehouse area allocation data, an initial warehouse management model is generated. Through conventional simulation tests and pressure simulation tests, the obtained test data can be optimized for performance and generate initial model performance optimization result data. By using these data to optimize key parameters, the generated warehouse management model can improve the adaptability and stability of the system under various operating conditions. This module ensures the efficient operation and continuous optimization of the model, and improves the overall performance of the warehouse management system. Finally, by uploading the optimized warehouse management model to the cloud platform to perform warehouse management operations, the high availability and scalability of the system are ensured. The integration and deployment of the cloud platform can not only provide efficient computing resources, but also realize remote management and real-time monitoring of the warehouse management system, improving the flexibility and response speed of the system.

[0131] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points of view, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the same elements of the application documents are included in the present invention.

[0132] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. 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 invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for determining a warehouse management model, characterized in that: The following steps are involved: Step S1: Obtain inventory data; Perform warehouse data preprocessing on inventory data to generate warehouse management data; Conduct statistical analysis on warehouse management data to generate warehouse statistical data; collect warehouse statistical data in real time to obtain real-time warehouse management data; Step S2: Obtain sales cost data and historical sales data; extract numerical features from the sales cost data to generate storage numerical feature data; perform storage volatility smoothing based on the storage numerical feature data to generate storage numerical smoothing data; calculate product inventory turnover rate based on the storage numerical feature data to generate overall turnover rate data; extract classification features from the historical sales data to generate storage classification feature data; Perform ABC analysis on the warehouse classification feature data to generate warehouse area division data; perform comprehensive optimization analysis based on the overall turnover rate data and warehouse area division data to generate warehouse area allocation data; Step S3: constructing a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; Conduct conventional simulation tests and pressure simulation tests on the initial warehouse management model to obtain conventional warehouse simulation test data and warehouse pressure simulation test data; use the conventional warehouse simulation test data to optimize the performance of the initial warehouse management model and generate initial model performance optimization result data; The warehouse pressure simulation test data is used to optimize the key parameters of the initial model performance optimization result data to generate a warehouse management model; step S3 includes the following steps: Step S31: construct a model based on the real-time warehouse management data and the warehouse area allocation data to generate an initial warehouse management model; use a cross-validation method to verify the initial warehouse management model to generate model performance evaluation data; Step S31 includes the following steps: Step S311: perform continuous variable prediction on the warehouse management real-time data to generate warehouse continuous variable prediction data; perform regression analysis based on the warehouse continuous variable prediction data to generate a warehouse model regression module; Step S312: Perform storage cluster analysis on the storage area allocation data to generate storage area clustering results; perform storage classification processing based on the storage area clustering results to generate a storage model classification module; Step S313: constructing a warehouse model according to the warehouse model regression module and the warehouse model classification module to generate an initial warehouse management model; Step S32: Performing a conventional simulation test and a pressure simulation test on the warehouse management initial model, thereby obtaining warehouse conventional simulation test data and warehouse pressure simulation test data; Step S32 includes the following steps: Step S321: Create a regular simulation environment condition for the warehouse management initial model to generate a normal virtual warehouse simulation environment, wherein the normal virtual warehouse simulation environment includes a normal order processing flow virtual warehouse simulation environment, an inventory replenishment flow virtual warehouse simulation environment, and a return processing flow virtual warehouse simulation environment; Step S322: applying the warehouse management initial model to a normal virtual warehouse simulation environment to perform normal environment simulation and generate a warehouse conventional simulation test data set; Step S323: creating stress simulation environment conditions for the initial warehouse management model to generate a stress virtual warehouse simulation environment, wherein the stress virtual warehouse simulation environment includes a stress order processing process virtual warehouse simulation environment, a logistics delay process virtual warehouse simulation environment, and a supply chain interruption virtual warehouse simulation environment; Step S324: applying the warehouse management initial model to the pressure virtual warehouse simulation environment to generate a warehouse pressure simulation test data set; Step S33: Optimize the performance of the initial warehouse management model using conventional warehouse simulation test data to generate initial model performance optimization result data; optimize key parameters of the initial model performance optimization result data using warehouse pressure simulation test data and model performance evaluation data to generate a warehouse management model; Step S33 includes the following steps: Step S331: constructing a warehouse data optimization strategy for warehouse conventional simulation test data to generate a warehouse model optimization strategy; Step S332: Optimizing the performance of the initial warehouse management model based on the warehouse model optimization strategy to generate initial model performance optimization result data; Step S333: Optimize key parameters of the initial model performance optimization result data using the warehouse pressure simulation test data to generate a warehouse management model; Step S4: Upload the warehouse management model to the cloud platform to perform warehouse management operations.

2. The method for determining a warehouse management model according to claim 1, wherein: Step S1 includes the following steps: Step S11: collecting inventory data; performing warehouse data cleaning on the inventory data to generate warehouse cleaning data; performing data standardization processing on the warehouse cleaning data to generate warehouse standardized data; performing warehouse management data aggregation on the warehouse standardized data to generate warehouse management data; Step S12: Performing warehouse data statistical analysis on the warehouse management data to generate warehouse statistical data; Step S13: Collect warehouse statistical data in real time to obtain real-time warehouse management data.

3. The method for determining a warehouse management model according to claim 1, wherein: Step S2 includes the following steps: Step S21: Acquire sales cost data and historical sales data; extract specific numerical features from the sales cost data to generate storage numerical feature data; perform storage volatility smoothing processing based on the storage numerical feature data to generate storage numerical smoothing data; calculate product inventory turnover rate based on the storage numerical feature data to generate overall turnover rate data; Step S22: extract relevant classification features from historical sales data to generate storage classification feature data; perform binary feature conversion on the storage classification feature data to generate storage classification binary data; perform ABC analysis on the storage classification binary data to generate storage area division data; Step S23: Perform comprehensive optimization allocation of storage areas based on the overall turnover rate data and storage area division data to generate storage area allocation data.

4. The method for determining a warehouse management model according to claim 3, wherein: Step S21 includes the following steps: Step S211: Obtain sales cost data and historical sales data; Step S212: Calculate the numerical features of the sales cost data to generate a warehouse numerical feature data set, wherein the numerical feature calculation includes average sales cost calculation, cost volatility analysis of sales cost, and change rate calculation of sales cost; Step S213: Based on the preset storage liquidity threshold, the storage numerical feature data set is divided into high liquidity data and low liquidity data; when it is determined to be classified as high liquidity data, the physical stability of the goods is monitored by a vibration sensor to ensure that the high liquidity goods remain stable during transportation and storage; when it is determined to be classified as low liquidity data, the tiered management of the goods is planned by combining the physical characteristics of weight and volume to improve the utilization efficiency of storage space; Step S214: Calculate the product inventory turnover rate of the warehouse numerical feature data set to generate overall turnover rate data; Calculate the product inventory turnover rate of the warehouse numerical feature data to generate overall turnover rate data.

5. The method for determining a warehouse management model according to claim 3, wherein: Step S22 includes the following steps: Step S221: Formulate product category labels for historical sales data to generate warehouse product category data; calculate market share for historical sales data to generate warehouse market share data; Step S222: Analyze the cargo storage location from the storage market share data; divide the storage space into high-flow calibration areas and low-flow calibration areas according to the exit distance based on the storage market share data, and optimize the physical storage layout of the high-flow calibration areas and the low-flow calibration areas according to the cargo storage location; Step S223: Aggregate the warehouse product category data and the warehouse market share data into classification feature data to generate warehouse classification feature data; Step S224: perform binary feature conversion on the warehouse classification feature data to generate warehouse classification binary data; perform ABC analysis on the warehouse classification binary data to generate warehouse area division data.

Images