Cloud storage intelligent inventory prediction and dynamic allocation system based on AI

Abstract

The invention relates to the technical field of cloud storage intelligent management, in particular to an AI-based cloud storage intelligent inventory prediction and dynamic allocation system, which can realize second-level synchronization of all storage ends corresponding to cloud storage through global coverage of an Internet of Things sensor, deduces inventory demands of the storage ends according to a full-link synchronization model, and improves the storage efficiency. The method comprises the following steps of: forming two-dimensional demand verification with an inventory demand prediction model constructed by an LSTM neural network, accurately capturing an inventory gap and capturing an inventory demand, converting a plurality of sets of artificially defined dynamic allocation rules into executable instructions by a system through a rule engine when an allocation demand condition exists, performing multi-target optimization verification in a full-link simulation environment, and performing multi-target optimization verification in a full-link simulation environment. The meteorological limitation of the allocation path and the storage end is deeply analyzed, the allocation strategy with the optimal comprehensive effect is screened out by integrating the virtual path and actual transportation, the experience decision bottleneck of traditional cloud storage is broken through, and the agility, the anti-risk capability and the sustainable development level of cloud storage management are remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the field of cloud warehousing intelligent management technology, and more specifically, to an AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system. Background Art

[0002] As a core node in the modern supply chain, cloud warehousing's inventory management and allocation efficiency directly impact enterprise costs, customer experience, and sustainable development capabilities. Traditional cloud warehousing management models, which rely primarily on manual experience and lag-free system records, struggle to cope with high-frequency fluctuations in market demand and complex and volatile supply chain environments.

[0003] With the development of the Internet of Things, AI, and digital twin technologies, some companies are attempting to collect warehouse data through sensors. However, these approaches only connect physical devices and fail to establish a closed loop of data-driven intelligent decision-making. Existing solutions generally suffer from two major technical bottlenecks: First, the fusion of multi-source data (inventory, weather, and traffic) lacks precision, making it impossible to accurately map warehouse status; second, allocation route optimization fails to deeply incorporate environmental factors, resulting in poor strategy robustness.

[0004] Therefore, how to build an intelligent system covering the entire process of "data collection-intelligent prediction-dynamic allocation-simulation verification" to achieve accurate capture of inventory demand, environmental adaptation of allocation strategies and balanced optimization of cost-timeliness-carbon emissions across the entire chain has become the core technical challenge for the intelligent upgrade of cloud warehousing. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide an AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system.

[0006] To achieve the above object, the present invention provides the following technical solutions: AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system, including a full-link fusion system construction unit, used to build a cloud warehousing full-link fusion system; The warehouse inventory forecasting unit regularly determines the actual inventory demand of each warehouse end corresponding to the cloud warehouse, and further determines the inventory demand forecast of each warehouse end; The inventory transfer initiation unit controls the cloud warehousing full-link fusion system to conduct a simulation for H duration. After the simulation is completed, the actual inventory demand deduced from each warehouse in the cloud warehousing full-link fusion system is obtained, and the dynamic allocation demand value of the cloud warehousing is determined. Based on the comparison result of the dynamic allocation demand value and the demand threshold, it is determined whether to initiate the inventory transfer step. After determining to initiate the inventory transfer step, the inventory transfer execution unit manually formulates multiple dynamic transfer rules, converts each dynamic transfer rule into an executable command, and imports it into the cloud warehouse full-link integration system. It further determines the transfer feedback value of each dynamic transfer rule, selects the transfer execution rule from all dynamic transfer rules, and performs inventory transfers based on the transfer execution rule to the cloud warehouse.

[0007] Furthermore, the method for determining the inventory demand forecast quantity at the warehouse end is as follows: a warehouse end is identified, the inventory time series set of the warehouse end is obtained, the inventory time series set is imported into the inventory demand forecast model corresponding to the warehouse end, and the inventory demand forecast model derives the inventory demand forecast quantity of the warehouse end; To obtain the inventory time series set of the warehouse end, the steps are as follows: obtain the actual inventory demand quantity of a warehouse end in the previous t time length, and integrate the actual inventory demand quantity in the t time length into the inventory time series set in the form of a time series set.

[0008] Furthermore, the dynamic allocation demand value of cloud warehousing is determined by: obtaining the predicted inventory deduced satisfaction quantity of each warehouse end, summing up the predicted inventory deduced satisfaction quantities of each warehouse end and taking the average value to calculate the average predicted inventory deduced satisfaction quantity, comparing each warehouse end pairwise, calculating the absolute difference between the predicted inventory deduced satisfaction quantities of the two compared warehouse ends, calculating the deduced satisfaction gap quantity, summing up all the deduced satisfaction gap quantities and taking the average value to calculate the average deduced satisfaction gap quantity, calculating the ratio of the average deduced satisfaction gap quantity to the average predicted inventory deduced satisfaction quantity, and calculating the dynamic allocation demand value of cloud warehousing.

[0009] Furthermore, the method for obtaining the predicted inventory deduction satisfaction quantity at the warehouse end is: select a warehouse end, calculate the difference between the cloud warehouse inventory demand forecast quantity and the actual inventory demand deduction quantity of the warehouse end, and calculate the predicted inventory deduction satisfaction quantity of the warehouse end.

[0010] Furthermore, the allocation feedback value of the dynamic allocation rule is determined as follows: After a dynamic allocation rule is converted into an executable command and imported into the cloud warehousing full-link fusion system, the cloud warehousing full-link fusion system performs a simulation for H duration. During the simulation, for each allocation path generated, the path meteorological impact value corresponding to the allocation path is determined, and the starting warehouse end and terminal warehouse end of each allocation path are simultaneously determined. When all allocation paths are generated, all allocation paths are sorted in the order of generation to generate a path sequence. The allocation restriction value of each warehouse end is further determined, and the allocation restriction value of each warehouse end is summed and averaged to calculate the average allocation restriction value. After the simulation is completed, an allocation basic data set is generated and imported into the allocation basic quantitative model. The allocation basic quantitative model derives the allocation basic quantitative value. The dynamic allocation demand value of the cloud warehouse after the simulation is simultaneously obtained. The allocation feedback value of the dynamic allocation rule is calculated based on the average allocation restriction value, the allocation basic quantitative value, and the dynamic allocation demand value.

[0011] Furthermore, the allocation restriction value of the warehouse end is determined by selecting a warehouse end, marking all allocation paths existing in the path sequence of the warehouse end as associated paths, sorting all associated paths in the order of generation, matching every two adjacent associated paths after sorting into a warehouse associated path group, obtaining the associated restriction value of each warehouse associated path group, summing up the associated restriction values of all warehouse associated path groups and taking the average value to calculate the allocation restriction value of the warehouse end.

[0012] Furthermore, the method for obtaining the associated restricted value of the warehouse associated path group is: calculating the sum of the path meteorological impact values of the two associated paths in the warehouse associated path group to calculate the meteorological impact sum value, calculating the absolute difference between the path meteorological impact values of the two associated paths to calculate the associated restricted gap value, calculating the ratio of the meteorological impact sum value to the associated restricted gap value to calculate the associated restricted value of the warehouse associated path group.

[0013] Furthermore, the path meteorological impact value of the allocation path is determined by obtaining predicted meteorological data on the allocation path, importing the predicted meteorological data into a meteorological evaluation model, and using the meteorological evaluation model to derive the path meteorological impact value of the allocation path.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The system of the present invention, through the full coverage of IoT sensors, can achieve second-level synchronization of cloud warehousing with all warehousing ends, and deduce the inventory demand of the warehousing end based on the full-link synchronization model, forming a two-dimensional demand verification with the inventory demand prediction model constructed by the LSTM neural network, accurately capturing inventory gaps and capturing inventory demand. When there is an allocation demand, the system converts multiple sets of manually defined dynamic allocation rules into executable instructions through the rule engine, and performs multi-objective optimization verification (balancing the three dimensions of cost, timeliness, and carbon emissions) in a full-link simulation environment, deeply analyzes the allocation path and the meteorological limitations of the warehousing end, and comprehensively selects the allocation strategy with the best overall effect by combining virtual paths and actual transportation, breaking through the experience-based decision-making bottleneck of traditional cloud warehousing, and significantly enhancing the agility, risk resistance and sustainable development level of cloud warehousing management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is the principle block diagram of the AI-based cloud warehouse intelligent inventory forecasting and dynamic allocation system; Figure 2 This is the operation flow chart of the AI-based cloud warehouse intelligent inventory forecasting and dynamic allocation system; Figure 3 This is a flowchart for building a full-link integrated cloud warehousing system. DETAILED DESCRIPTION

[0016] Reference Figures 1 to 3 , an AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system, including a full-link fusion system construction unit, a warehousing inventory forecasting unit, an inventory allocation startup unit, and an inventory allocation execution unit.

[0017] The full-link fusion system construction unit builds a full-link fusion system for cloud warehousing. First, physical equipment networking, business system docking and external data integration are carried out. Physical equipment networking includes sensor deployment at each storage end and transportation end (sensors must be deployed at each storage end corresponding to cloud warehousing), and Node-RED is used to build an edge data gateway, which supports access to protocols such as MQTT and OPCUA. Sensors required to be deployed at the storage end include RFID readers (shelf entrances and exits, collecting inventory in and out data), smart meters (deployed in different areas, collecting lighting and equipment energy consumption, with an accuracy of ±1%), temperature and humidity sensors (such as fresh food warehouses, one is deployed every 50 square meters, with an accuracy of ±0.5°C), etc.; sensors required to be deployed at the transportation end include GPS + Beidou dual-mode positioning (real-time vehicle location, error ≤ 5 meters), OBD interface (collecting vehicle fuel consumption, mileage, and carbon emission conversion), etc.; the business systems connected include ERP / WMS (used to obtain inventory ledgers, order waves, and supplier information), TMS (used to obtain transportation orders, logistics provider quotations, and route planning data), carbon footprint platform (used to connect to third-party data services (such as CarbonTrust) and obtain transportation mode carbon factor libraries (such as 0.015kgCO2 / km・ton for railway)). The integrated external data includes weather API (used to obtain regional precipitation and wind speed data (affecting road transportation timeliness, delay coefficient ±10-30%)) and policy data (used to access carbon emission trading market prices (such as 1kgCO2=50 yuan, used for cost-carbon emission conversion)). Build 3D physical modeling of each warehouse end, which includes geometric modeling (building an accurate digital mirror of the warehouse space at the warehouse end, such as using lidar scanning to obtain warehouse point cloud data, using Blender / B IM (Revit) builds warehouse structures (shelves, aisles, loading and unloading ports) and equipment models (AGVs, stackers). Unity's ProBuilder plug-in is used to simplify the model face count (≤ 100,000 triangles), ensure a browser rendering frame rate of ≥ 30 fps, add attribute tags to each model (such as shelf ID, AGV number, and carbon emission factor), and dynamically bind attributes (associating the 3D model with the real-time status, business data, and calculated attributes of the physical entity. Real-time status data items include inventory quantity, equipment operating status (operating / faulty), location coordinates, energy consumption values, and temperature and humidity. Business data items include shelf SKU information, order priority, equipment model, maintenance cycle, and operator ID. Calculated attribute data items include inventory turnover rate, carbon emissions, energy consumption, equipment utilization rate, and order response time). Ultimately, a cloud warehousing full-link fusion system is built (the cloud warehousing full-link fusion system is updated in real time to ensure that its functions, performance, security, and business adaptability are always consistent with the actual cloud warehousing situation).

[0018] The warehouse inventory forecasting unit regularly determines the actual inventory demand of each warehouse end corresponding to the cloud warehouse (the length of the regular time interval is determined according to the inventory management requirements of the cloud warehouse, and the actual inventory demand is the actual inventory demand of the warehouse end based on the ongoing business activities), and further determines the inventory demand forecast of each warehouse end (the inventory demand forecast is the actual inventory demand of the warehouse end after H time (which can be 1 hour or 2 hours)).

[0019] Method for determining the inventory demand forecast at the warehouse end: identify a warehouse end, obtain the inventory time series set of the warehouse end, import the inventory time series set into the inventory demand forecast model corresponding to the warehouse end, and use the inventory demand forecast model to derive the inventory demand forecast for the warehouse end.

[0020] To obtain the inventory time series set of the warehouse end, the steps are as follows: obtain the actual inventory demand quantity of a warehouse end in the previous t time length, and integrate the actual inventory demand quantity in the t time length into the inventory time series set in the form of a time series set.

[0021] Each warehouse end corresponds to an independent inventory demand forecasting model. All inventory demand forecasting models are built based on the LSTM model. This embodiment takes warehouse end A as an example to disclose the construction process of the inventory demand forecasting model: collect multiple inventory time series sets of warehouse end A, build an LSTM model, use the inventory time series set as the basic data, and train the built LSTM model. In this process, an inventory demand forecast is assigned to each inventory time series set. The inventory demand forecast is the actual inventory demand of warehouse end A after H time. Then, the multiple inventory time series sets are divided into training set, validation set and test set according to a specific ratio. The specific division ratio is determined to be 70%:15%:15%. The training set accounts for 70% to ensure that the model has sufficient data learning mode, and the validation set and test set account for 70%. The sets each account for 15%, which balances parameter tuning and generalization ability evaluation. For time series data, it needs to be divided in chronological order (such as January-July 2023 as the training set, August-September as the validation set, and October-December as the test set) to avoid future information leakage. The role of the validation set: use the validation set to evaluate model performance (such as MSE, MAE) after each round of training to avoid overfitting; hyperparameter tuning: adjust parameters such as the number of LSTM layers, the number of neurons, the learning rate, and the sequence length to find the optimal combination; overfitting prevention: add a Dropout layer (such as adding 20% dropout after the LSTM layer) and an early stopping mechanism (EarlyStopping); use the test set data that did not participate in the training to calculate the generalization performance indicators of the model. Finally, the inventory demand forecasting model of warehouse end A is completed.

[0022] The inventory transfer initiation unit controls the cloud warehousing full-link fusion system to perform a simulation for H duration. After the simulation, it obtains the actual inventory demand deduction of each warehouse end in the cloud warehousing full-link fusion system (that is, the cloud warehousing full-link fusion system performs time-dimensional deduction according to conventional allocation, replenishment, distribution and other strategies, and analyzes the actual inventory demand of each warehouse end in the cloud warehousing full-link fusion system after H duration under conventional strategies). Then, the dynamic allocation demand value of the cloud warehousing is determined. When the dynamic allocation demand value of the cloud warehousing is higher than the demand threshold (the demand threshold is the key trigger condition for initiating inventory transfer), the inventory transfer step is determined to be initiated.

[0023] The dynamic allocation demand value of cloud warehousing is determined by obtaining the predicted inventory fulfillment of each warehouse end, summing the predicted inventory fulfillment of each warehouse end and taking the average value to calculate the average predicted inventory fulfillment (when the average predicted inventory fulfillment is calculated to be 0, adjust the value of the average predicted inventory fulfillment to 0.05). Each warehouse end is compared pairwise, and the predicted inventory fulfillment of the two compared warehouse ends is calculated by taking the absolute difference to calculate the deduced fulfillment gap. All the deduced fulfillment gaps are summed and taken the average value to calculate the average deduced fulfillment gap. The dynamic allocation demand value of cloud warehousing is calculated by calculating the ratio of the average deduced fulfillment gap to the average predicted inventory fulfillment.

[0024] The method for obtaining the predicted inventory satisfaction quantity at the warehouse end is as follows: select a warehouse end, calculate the difference between the cloud warehouse inventory demand forecast quantity and the actual inventory demand deduction quantity of the warehouse end, and calculate the predicted inventory satisfaction quantity of the warehouse end (the larger the value of the predicted inventory satisfaction quantity, the stronger the actual inventory demand coverage ability of the predicted inventory demand under the deduction of the current rules).

[0025] The inventory transfer execution unit, after determining to start the inventory transfer step, manually formulates multiple dynamic transfer rules, converts each dynamic transfer rule into an executable command and imports it into the cloud warehousing full-link integration system (the way to convert dynamic transfer rules into executable commands includes decomposing rule elements, digitally expressing rules (using structured language (JSON / XML) or decision table (Decision The system then performs the following steps: defining rules using a table (e.g., defining rules using a rule engine), parsing the rule engine (using the rule engine to convert natural language rules into system-executable logic), and standardizing the instruction format (generating a system-recognizable instruction set, typically including: transfer orders (including outbound and inbound warehouses, SKUs, quantities, and transportation methods); path planning instructions (transferring routes and capacity allocation to the TMS); and equipment scheduling instructions (notifying the WMS to dispatch equipment such as AGVs and sorting robots). Rule elements are broken down into conditions, actions, parameters, and constraints. For example, conditions include: inventory at the warehouse end ≤ safety line and inventory at the adjacent warehouse end ≥ transfer quantity; action: transfer a specified SKU from the adjacent warehouse end; parameter: transfer quantity = safety line - current inventory; constraint: single transfer cost ≤ 500 yuan, timeliness ≤ 12 hours). The system then determines the transfer feedback value for each dynamic transfer rule, marks the dynamic transfer rule with the largest transfer feedback value as the transfer execution rule, and performs inventory transfers for the cloud warehouse based on the transfer execution rule.

[0026] The method for determining the allocation feedback value of the dynamic allocation rule is as follows: After a dynamic allocation rule is converted into an executable command and imported into the cloud warehousing full-link fusion system, the cloud warehousing full-link fusion system performs a simulation for H time. During the simulation, each allocation path is generated (allocation from warehouse end A to warehouse end D, that is, an allocation path is generated), the path weather impact value corresponding to the allocation path is determined, and the starting warehouse end and the terminal warehouse end of each allocation path are simultaneously determined. When all allocation paths are generated, all allocation paths are sorted in the order of generation, and a path sequence is generated (in the sorting order). Determine the allocation restriction value of each warehouse end, sum the allocation restriction values of each warehouse end and take the average value to calculate the average allocation restriction value and mark it as EnC. After the simulation is completed, generate the allocation basic data set (the allocation basic data set includes the transportation cost, storage cost, and total carbon emissions of the cloud warehousing full-link integration system during the entire simulation process). Import the allocation basic data set into the allocation basic quantitative model. The allocation basic quantitative model derives the allocation basic quantitative value and marks it as Dsp. Simultaneously obtain the dynamic allocation demand value of the cloud warehouse after the simulation is completed and mark it as Wzr. Use the formula Calculate the allocation feedback value of the dynamic allocation rule; ag1 represents the first correction coefficient (which can be 1.15), and ag2 represents the second correction coefficient (which can be 1.08).

[0027] The method for determining the transfer restriction value at the warehouse end is as follows: select a warehouse end, mark all transfer paths that exist in the path sequence to this warehouse end as associated paths (regardless of whether this warehouse end is the starting warehouse end or the terminal warehouse end in the transfer path, it is marked as an associated path), sort all associated paths in the order of generation, match every two adjacent associated paths after sorting into a warehouse associated path group, obtain the associated restriction value of each warehouse associated path group, sum and average the associated restriction values of all warehouse associated path groups, and calculate the transfer restriction value of the warehouse end.

[0028] The method for obtaining the association restricted value of the warehouse association path group is as follows: the path meteorological impact values of the two association paths in the warehouse association path group are summed to calculate the meteorological impact sum value, the path meteorological impact values of the two association paths are calculated to perform absolute difference calculations to calculate the association restricted gap value (when the calculated association restricted gap value is 0, the association restricted gap value is adjusted to 0.1), and the ratio of the meteorological impact sum value to the association restricted gap value is calculated to calculate the association restricted value of the warehouse association path group.

[0029] The method for determining the path meteorological impact value of the allocation path is as follows: (Based on the meteorological system) obtain the predicted meteorological data on the allocation path (the predicted meteorological data is the meteorological data of the allocation path within a time period H predicted by the meteorological system, including precipitation, visibility, and wind speed), import the predicted meteorological data into the meteorological evaluation model, and the meteorological evaluation model derives the path meteorological impact value of the allocation path.

[0030] The method of building a meteorological evaluation model is as follows: a neural network model is built, multiple predicted meteorological data are collected, and the predicted meteorological data are used as the basic data to train the built neural network model. In this process, a path meteorological impact value is assigned to each predicted meteorological data. The value range of the path meteorological impact value is set between 1 and 20. The larger the value, the greater the impact of the predicted meteorological data on the allocation process of the allocation path. For example, the greater the precipitation, the greater the driving risk on the allocation path, and the greater the impact on the allocation process of the allocation path. Then, the multiple predicted meteorological data are divided into training set, validation set and test set according to a specific ratio. The specific division ratio is determined to be 70%:15%:15%. The meteorological evaluation model is completed through neural network training.

[0031] The method of building the allocation basic quantitative model is as follows: build a neural network model, collect multiple allocation basic data sets, use the allocation basic data sets as the basic data, and train the built neural network model. In this process, each allocation basic data set is assigned an allocation basic quantitative value. The value range of the allocation basic quantitative value is set between 1 and 100. The larger the value, the better the allocation basic effect of the allocation basic data set, such as lower transportation costs, lower storage costs, and less total carbon emissions. Then, the multiple allocation basic quantitative values are divided into training sets, validation sets, and test sets according to a specific ratio. The specific division ratio is determined to be 60%:20%:20%. The weighted mean square error (WMSE) is used to give higher weights to high-risk samples (index > 80), optimize the model structure, and make it develop in a more accurate and stable direction. Finally, the allocation basic quantitative model is completed.

[0032] Based on the above system, through the full coverage of IoT sensors, cloud warehousing can achieve second-level synchronization of all warehousing ends, and the inventory demand of the warehousing end can be deduced according to the full-link synchronization model, forming a two-dimensional demand verification with the inventory demand prediction model constructed by the LSTM neural network, accurately capturing inventory gaps and capturing inventory demand. When there is a demand for allocation, the system converts multiple sets of manually defined dynamic allocation rules into executable instructions through the rule engine, and performs multi-objective optimization verification (balance of cost, timeliness, and carbon emissions) in a full-link simulation environment. It deeply analyzes the meteorological limitations of the allocation path and the warehousing end, and comprehensively combines virtual paths and actual transportation to select the allocation strategy with the best overall effect, breaking through the experience decision-making bottleneck of traditional cloud warehousing, and significantly enhancing the agility, risk resistance and sustainable development level of cloud warehousing management.

[0033] The above formulas are all dimensionless and numerically calculated, and the preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0034] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0035] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0036] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0037] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0038] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0039] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system, characterized by: Includes a full-link fusion system construction unit for building a cloud warehousing full-link fusion system; The warehouse inventory forecasting unit regularly determines the actual inventory demand of each warehouse end corresponding to the cloud warehouse, and further determines the inventory demand forecast of each warehouse end; The inventory transfer initiation unit controls the cloud warehousing full-link fusion system to conduct a simulation for H duration. After the simulation is completed, the actual inventory demand deduced from each warehouse in the cloud warehousing full-link fusion system is obtained, and the dynamic allocation demand value of the cloud warehousing is determined. Based on the comparison result of the dynamic allocation demand value and the demand threshold, it is determined whether to initiate the inventory transfer step. After determining to initiate the inventory transfer step, the inventory transfer execution unit manually formulates multiple dynamic transfer rules, converts each dynamic transfer rule into an executable command, and imports it into the cloud warehouse full-link integration system. It further determines the transfer feedback value of each dynamic transfer rule, selects the transfer execution rule from all dynamic transfer rules, and performs inventory transfers based on the transfer execution rule to the cloud warehouse.

2. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 1 is characterized in that: Method for determining the inventory demand forecast at the warehouse end: Identify a warehouse end, obtain the inventory time series set at the warehouse end, import the inventory time series set into the inventory demand forecast model corresponding to the warehouse end, and use the inventory demand forecast model to derive the inventory demand forecast for the warehouse end; To obtain the inventory time series set of the warehouse end, the steps are as follows: obtain the actual inventory demand quantity of a warehouse end in the previous t time length, and integrate the actual inventory demand quantity in the t time length into the inventory time series set in the form of a time series set.

3. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 1 is characterized in that: The method for determining the dynamic allocation demand value of cloud warehousing is as follows: obtain the predicted inventory deduced satisfaction quantity of each warehouse end, sum and average the predicted inventory deduced satisfaction quantities of each warehouse end, calculate the average predicted inventory deduced satisfaction quantity, compare each warehouse end pairwise, calculate the absolute difference between the predicted inventory deduced satisfaction quantities of the two compared warehouse ends, calculate the deduced satisfaction gap quantity, sum and average all the deduced satisfaction gap quantities, calculate the average deduced satisfaction gap quantity, calculate the ratio of the average deduced satisfaction gap quantity to the average predicted inventory deduced satisfaction quantity, and calculate the dynamic allocation demand value of cloud warehousing.

4. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 3 is characterized in that: The method for obtaining the predicted inventory satisfaction quantity at the warehouse end is: select a warehouse end, calculate the difference between the cloud warehouse inventory demand forecast quantity and the actual inventory demand deduction quantity of the warehouse end, and calculate the predicted inventory satisfaction quantity of the warehouse end.

5. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 1 is characterized in that: The allocation feedback value for a dynamic allocation rule is determined as follows: After a dynamic allocation rule is converted into an executable command and imported into the cloud warehousing full-link fusion system, the cloud warehousing full-link fusion system performs a simulation for a duration of H. During the simulation, for each allocation path generated, the path weather impact value corresponding to the allocation path is determined. The starting and ending warehouse ends of each allocation path are also determined. Once all allocation paths have been generated, they are sorted in the order of generation to generate a path sequence. The allocation restriction value for each warehouse end is further determined. The restriction values for each warehouse end are summed and averaged to calculate the average allocation restriction value. After the simulation, an allocation basic dataset is generated and imported into the allocation basic quantitative model. The allocation basic quantitative model derives the allocation basic quantitative value. The dynamic allocation demand value of the cloud warehouse after the simulation is simultaneously obtained. The allocation feedback value for the dynamic allocation rule is calculated based on the average allocation restriction value, the allocation basic quantitative value, and the dynamic allocation demand value.

6. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 5 is characterized in that: The method for determining the transfer restriction value at the warehouse end is as follows: select a warehouse end, mark all transfer paths existing in the path sequence to this warehouse end as associated paths, sort all associated paths in the order of generation, match every two adjacent associated paths after sorting into a warehouse associated path group, obtain the associated restriction value of each warehouse associated path group, sum and average the associated restriction values of all warehouse associated path groups, and calculate the transfer restriction value of the warehouse end.

7. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 6 is characterized in that: The method for obtaining the association restricted value of the warehouse association path group is as follows: the path meteorological impact values of the two association paths in the warehouse association path group are summed to calculate the meteorological impact sum value, the path meteorological impact values of the two association paths are calculated by performing absolute difference calculation to calculate the association restricted gap value, and the ratio of the meteorological impact sum value to the association restricted gap value is calculated to calculate the association restricted value of the warehouse association path group.

8. The AI-based cloud warehousing intelligent inventory forecasting and dynamic allocation system according to claim 7 is characterized in that: The path meteorological impact value of the allocation path is determined by obtaining the forecast meteorological data on the allocation path, importing the forecast meteorological data into the meteorological evaluation model, and deriving the path meteorological impact value of the allocation path from the meteorological evaluation model.

Images