Intelligent inventory early warning and dynamic replenishment strategy integration method and system

Abstract

The application relates to an integrated method of intelligent inventory early warning and dynamic replenishment strategy, relates to the field of data processing specially used for administrative, commercial, financial, management, supervision or prediction purposes, and the method comprises the following steps: adopting an intelligent prediction model to intelligently predict the total sales of various retail commodities corresponding to each set of sales in a current time segment of a set distribution site according to various basic data selected in a targeted manner; and formulating an inventory early warning mechanism and a dynamic replenishment strategy for the set distribution site in the current time segment based on the intelligent prediction result. The application also relates to an integrated system of intelligent inventory early warning and dynamic replenishment strategy. Through the application, an artificial intelligence model with a customized structure can be constructed to intelligently predict the sales data of various retail commodities of a distribution site in a future time segment of each block, and then a corresponding inventory early warning mechanism and a dynamic replenishment strategy can be formulated, so that the commodity sales efficiency and the platform operation cost are taken into account.

Description

Technical Field

[0001] This invention relates to the field of data processing, which is specifically applicable to administrative, commercial, financial, management, supervisory, or forecasting purposes, and in particular to an integrated method and system for intelligent inventory early warning and dynamic replenishment strategies. Background Technology

[0002] With the rapid development of electronic finance and e-commerce, online sales of various retail goods have become a growing trend in retail. More and more physical retail stores are being replaced by online sales platforms, such as online sales apps. These online sales platforms typically set up a single distribution point in a fixed area of ​​a city, such as multiple city blocks, to supply online orders for retail goods to residents in those blocks. This reduces the operating costs of physical stores while balancing the economic interests of both the buyers (residents) and the online sales platform operators.

[0003] For example, Chinese invention patent publication CN111125140A proposes an automatic inventory replenishment system based on daily sales coefficients. The system includes an inventory dynamic adjustment module, whose signal output is electrically connected to a computer. The computer's signal output is electrically connected to a dynamic replenishment module, a replenishment switch module, an inventory quantity calculation module, an inventory depth calculation module, a safety stock calculation module, and a safety stock vs. inventory comparison module. The inventory dynamic adjustment module can more accurately calculate the inventory quantity of products on the product page, thus accurately displaying the dynamic inventory parameters on the product page. This improves the accuracy of product inventory values, providing staff with more accurate replenishment information, enabling them to replenish products in advance and avoiding losses to the store due to untimely replenishment.

[0004] For example, Chinese invention patent publication CN118171991A discloses a method, apparatus, electronic device, and storage medium for inventory control of goods. The method includes: determining the total fulfillment rate of goods across all categories; determining at least one category of goods to be replenished based on the total fulfillment rate and / or the replenishment point quantity corresponding to each category; determining the replenishment quantity of the current category of goods to be replenished based on historical unsold inventory data; and determining the target replenishment quantity for each category of goods to be replenished based on the target inventory quantity, the replenishment quantity, and the existing inventory quantity. This technical solution solves the problem of untimely replenishment of goods categories, which leads to the inability to provide corresponding services to users. It achieves timely determination of the replenishment quantity of corresponding goods categories and, based on the replenishment quantity and the corresponding target inventory quantity, determines the target replenishment quantity, thus achieving the technical effect of dynamically replenishing goods based on shipment information.

[0005] Therefore, the inventory status analysis and replenishment strategy formulation mentioned in the above-mentioned existing technologies, when mentioned in terms of timeliness, only refer to the timely calculation and analysis of the current inventory and sales volume of various products. Obviously, the results of such calculation and analysis are all lagging data, which cannot accurately obtain the sales data corresponding to each product in future time segments. Naturally, it is impossible to determine whether the inventory data corresponding to each product on the online sales platform can keep up with the sales data based on the sales data corresponding to each product in future time segments. As a result, the inventory early warning mechanism and dynamic replenishment strategy of the online sales platform cannot be effectively and reliably implemented due to the lack of reliable basic data, which can easily lead to the sales dilemma of insufficient or excessive supply of a certain type of product. Summary of the Invention

[0006] To address the technical deficiencies in existing technologies, this invention provides an integrated method and system for intelligent inventory early warning and dynamic replenishment strategies. It can construct a customized artificial intelligence model for the same distribution station serving multiple districts simultaneously. Based on comprehensive and sufficient basic data selected in a targeted manner, it intelligently predicts the sales volume of various retail goods at the distribution station within the current time segment (i.e., a future time segment). This allows for the formulation of corresponding inventory early warning mechanisms and dynamic replenishment strategies, achieving an organic integration of an inventory early warning mechanism based on intelligent prediction results and a dynamic replenishment strategy. This avoids the predicament of insufficient or excessive supply of a certain type of product, improves the operational efficiency of online sales platforms, and reduces their operating costs.

[0007] According to a first aspect of the present invention, an integrated method for intelligent inventory early warning and dynamic replenishment strategy is provided, the method comprising:

[0008] Get the number of retail product types sold at the set distribution site, the maximum storage volume, the number of blocks in each managed block, the cumulative value of each resident population corresponding to each managed block, and the geographical area occupied by each managed block, and output them as multiple configuration information of the set distribution site.

[0009] Get the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and get the sales data of a single retail product corresponding to the past simultaneous time segment of the set distribution station that was in the same time segment position as the current time segment on the previous day.

[0010] Perform multiple training operations on the feedforward neural network to obtain a feedforward neural network after multiple training operations, and output the feedforward neural network after multiple training operations as an intelligent prediction model.

[0011] The intelligent prediction model uses the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each past time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the past same segment to intelligently predict the total sales volume of each retail product of the distribution station in the current time segment.

[0012] When there is a type of retail product whose total sales volume is less than the total inventory volume in the various sales volumes of the intelligently predicted retail products, the type of retail product will be used as an inventory warning type.

[0013] When there is a category of retail goods for sale whose total sales volume is less than the total inventory volume in the various sales volumes of the intelligently predicted retail goods, the category of retail goods for sale is designated as a dynamic replenishment category, and replenishment is performed on the dynamic replenishment category based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment category.

[0014] According to a second aspect of the present invention, an integrated system for intelligent inventory early warning and dynamic replenishment strategy is provided, the system comprising:

[0015] The first capturing mechanism is used to obtain the number of retail product types sold at the set distribution station, the maximum storage volume, the number of blocks in each managed block, the cumulative value of each resident population corresponding to each managed block, and the geographical area occupied by each managed block, and output them as multiple configuration information of the set distribution station.

[0016] The second capturing mechanism is used to acquire the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and to acquire the sales data of a single retail product corresponding to the past simultaneous segment of the set distribution station that was in the same time segment position as the current time segment on the previous day.

[0017] The object assembly mechanism is used to perform multiple training operations on the feedforward neural network to obtain the feedforward neural network after multiple training operations, and output the feedforward neural network after multiple training operations as the intelligent prediction model.

[0018] The prediction execution mechanism is connected to the first capture mechanism, the second capture mechanism, and the object assembly mechanism, respectively. It is used to use an intelligent prediction model to intelligently predict the total sales volume of each type of retail product of the set distribution station in the current time segment based on the duration of each time segment, multiple configuration information of the set distribution station, the sales data of each retail product corresponding to each past time segment before the current time segment, and the sales data of a single retail product corresponding to the set distribution station in the past same segment.

[0019] An inventory warning mechanism, connected to the forecast execution mechanism, is used to identify a product category whose total sales volume is less than the total inventory volume in each of the various retail products for sale in the intelligent forecast.

[0020] A dynamic replenishment mechanism, connected to the forecast execution mechanism, is used to identify, when there is a type of retail product whose total sales volume is less than the total inventory volume in each of the various sales volumes corresponding to the intelligent forecasted retail products, as a dynamic replenishment type, and replenish the dynamic replenishment type based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type.

[0021] Therefore, it can be seen that the present invention has at least the following key inventive points:

[0022] The first step: For designated distribution stations that are responsible for supplying retail goods to various blocks, a customized intelligent prediction model is constructed for the intelligent prediction of the sales quantity of each type of retail goods sold at the designated distribution station in the current time segment (i.e., a future time segment). The intelligent prediction model is a feedforward neural network that has undergone multiple training operations, and the number of training operations is monotonically positively correlated with the types and quantities of retail goods sold at the designated distribution station. Thus, different intelligent prediction models with different structures are designed for different distribution stations, ensuring the effectiveness and stability of the intelligent prediction results.

[0023] The second aspect involves targeted filtering of various basic data for intelligent prediction of the sales volume of each type of retail product sold at a distribution station in the current time segment. These basic data include the number of retail product types sold at the distribution station, the maximum storage capacity, the number of blocks under its management, the cumulative number of permanent residents corresponding to each block under its management, the total geographical area occupied by each block under its management, the sales data of each retail product corresponding to each previous time segment before the current time segment, and the sales data of a single retail product corresponding to the same time segment as the current time segment on the previous day. The thorough and comprehensive filtering of the above basic data further ensures the effectiveness and stability of the intelligent prediction results.

[0024] Thirdly: In each training operation performed on the feedforward neural network, the total sales volume of each type of retail product sold at the designated distribution station within a certain past time segment is used as the output of the feedforward neural network. The duration of each time segment, multiple configuration information of the designated distribution station, the sales data of each retail product corresponding to each past time segment before the designated distribution station, and the sales data of a single retail product corresponding to the designated distribution station in the past simultaneous segments of the designated distribution station are used as the input of the feedforward neural network. This ensures the training effect of each training operation of the feedforward neural network.

[0025] Fourthly: When there is a type of retail product whose total sales volume is less than the total inventory in each segment of the current time period corresponding to various retail products sold at the intelligently predicted distribution station, this type of product is designated as an inventory warning type and an inventory warning is issued. Similarly, when there is a type of retail product whose total sales volume is less than the total inventory in each segment of the current time period corresponding to various retail products sold at the intelligently predicted distribution station, this type of product is designated as a dynamic replenishment type. Replenishment is then carried out on this dynamic replenishment type based on the difference between its total sales volume and total inventory. The quantity of products replenished for a dynamic replenishment type is equal to the difference between its total sales volume and total inventory. This process, based on intelligent prediction results, completes inventory warnings and dynamic replenishment for various retail products sold, achieving an organic integration of an inventory warning mechanism and a dynamic replenishment strategy based on intelligent prediction results, thus improving the intelligence and automation levels of the distribution station management. Attached Figure Description

[0026] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein:

[0027] Figure 1The present invention provides a technical flowchart of the integrated method and system for intelligent inventory early warning and dynamic replenishment strategies.

[0028] Figure 2 The following is a flowchart illustrating the steps of an integrated method for intelligent inventory early warning and dynamic replenishment strategy according to Embodiment 1 of the present invention.

[0029] Figure 3 The following is a flowchart illustrating the steps of an integrated method for intelligent inventory early warning and dynamic replenishment strategy according to Embodiment 2 of the present invention.

[0030] Figure 4 The following is a flowchart illustrating the steps of an integrated method for intelligent inventory early warning and dynamic replenishment strategy according to Embodiment 3 of the present invention.

[0031] Figure 5 This is an internal structure diagram of an integrated system for intelligent inventory early warning and dynamic replenishment strategies according to Embodiment 4 of the present invention.

[0032] Figure 6 This is an internal structure diagram of an integrated system for intelligent inventory early warning and dynamic replenishment strategies according to Embodiment 5 of the present invention.

[0033] Figure 7 This is an internal structure diagram of an integrated system for intelligent inventory early warning and dynamic replenishment strategies according to Embodiment 6 of the present invention. Detailed Implementation

[0034] like Figure 1 As shown, a technical flowchart of the integrated method and system for intelligent inventory early warning and dynamic replenishment strategy according to the present invention is presented.

[0035] like Figure 1 As shown, the specific technical process of the present invention is as follows:

[0036] The first technical process: For the designated distribution stations that are responsible for supplying retail goods to various blocks, a customized intelligent prediction model is built for the intelligent prediction of the sales quantity of each batch of various retail goods for sale in the current time segment (i.e., a future time segment) of the designated distribution station.

[0037] exist Figure 1 In the process, each block feeds back sales order information to the designated distribution station, which in turn feeds back to the intelligent prediction model, so as to develop an inventory early warning mechanism and dynamic replenishment strategy for future time segments for the designated distribution station;

[0038] And in Figure 1In this process, the first, second, and third technical processes are carried out at the intelligent prediction model, while the fourth technical process is carried out on the online sales platform above the intelligent prediction model, such as an online sales APP.

[0039] Specifically, the structural customization of the intelligent prediction model is mainly reflected in the following aspects:

[0040] First: The intelligent prediction model is a feedforward neural network that has undergone multiple training operations;

[0041] Secondly, the number of training operations of the feedforward neural network is monotonically positively correlated with the number of retail product types sold at the designated distribution stations, thereby designing intelligent prediction models with different structures for different distribution stations and ensuring the effectiveness and stability of the intelligent prediction results.

[0042] Finally: In each training operation performed on the feedforward neural network, the total sales volume of each type of retail product sold at the designated distribution station within a certain past time segment is used as the output of the feedforward neural network. The duration of each time segment, multiple configuration information of the designated distribution station, the sales data of each retail product corresponding to each past time segment before the designated distribution station, and the sales data of a single retail product corresponding to the designated distribution station in the past simultaneous segments of the designated distribution station are used as the input of the feedforward neural network. This ensures the training effect of each training operation of the feedforward neural network.

[0043] The second technical process involves intelligently filtering various basic data to predict the sales volume of each type of retail product for the current time segment at the distribution station.

[0044] Specifically, the basic data includes the number of retail product types sold at the designated distribution station, the maximum storage capacity, the number of blocks in each managed district, the cumulative number of permanent residents corresponding to each managed district, and the total geographical area occupied by each managed district.

[0045] Specifically, the basic data also includes the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and the sales data of a single retail product corresponding to the past simultaneous time segment of the set distribution station that was in the same time segment position as the current time segment on the previous day.

[0046] In this way, the thorough and comprehensive screening of the above-mentioned basic data further ensures the effectiveness and stability of the intelligent prediction results;

[0047] The third technical process: The intelligent prediction model with the customized structure of the first technical process intelligently predicts and sets the sales volume of each type of retail product at the current time segment of the distribution station based on the basic data fully and comprehensively screened by the second technical process.

[0048] Specifically, since the distribution stations are set up to serve each neighborhood, the intelligent prediction of the total sales volume of each type of retail product at the current time segment of the distribution station reflects the demand for each type of retail product in the orders placed by residents of each neighborhood as sales users on the same online sales platform in the current time segment.

[0049] Here, since the current time segment starts from the current moment, the current time segment is actually a type of future time segment;

[0050] The fourth technical process: Based on the intelligent forecasting results of the third technical process, the inventory early warning mechanism and dynamic replenishment strategy for the current time segment are customized for the distribution station, thereby completing the organic integration of the inventory early warning mechanism and dynamic replenishment strategy based on the intelligent forecasting results.

[0051] Specifically, when there is a type of retail product whose total sales volume is less than the total inventory volume in each segment of the current time period of the intelligent prediction distribution station, the type of retail product will be used as an inventory warning type for warning purposes.

[0052] Specifically, when there is a type of retail product whose total sales volume is less than the total inventory volume in each segment of the current time period of the intelligent prediction distribution station, the type of retail product is designated as a dynamic replenishment type. Replenishment is carried out on the dynamic replenishment type based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type. The number of products replenished for the dynamic replenishment type is equal to the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type.

[0053] In this way, based on the intelligent prediction results, inventory warnings and dynamic replenishment are completed for various retail goods on sale, thereby improving the level of intelligence and automation of the management of distribution sites.

[0054] The key points of this invention are: an intelligent prediction model with a customized structure for intelligent prediction of the sales quantity of various retail goods in future time segments corresponding to the designated distribution stations responsible for supplying retail goods to various blocks simultaneously; multiple basic data for targeted selection for intelligent prediction; targeted design of each training mechanism of the feedforward neural network; and the organic integration of inventory early warning mechanism and dynamic replenishment strategy for retail goods in sale based on intelligent prediction results.

[0055] The integration method and system of intelligent inventory early warning and dynamic replenishment strategy of the present invention will be specifically described below by way of embodiments.

[0056] Example 1

[0057] Figure 2 The following is a flowchart illustrating the steps of an integrated method for intelligent inventory early warning and dynamic replenishment strategy according to Embodiment 1 of the present invention.

[0058] like Figure 2 As shown, the integration method of intelligent inventory early warning and dynamic replenishment strategy includes the following steps:

[0059] Step S201: Obtain the number of retail product types sold at the set distribution station, the maximum storage volume, the number of blocks in each managed block, the cumulative value of each permanent resident population corresponding to each managed block, and the geographical area occupied by each managed block, and output them as multiple configuration information of the set distribution station.

[0060] Specifically, the system obtains the number of retail product types sold at the designated distribution station, the maximum storage capacity, the number of blocks in each managed district, the cumulative value of each resident population corresponding to each managed district, and the geographical area occupied by each managed district. This information is then output as multiple configuration information for the designated distribution station, including: using multiple information parsing components to obtain the number of retail product types sold at the designated distribution station, the maximum storage capacity, the number of blocks in each managed district, the cumulative value of each resident population corresponding to each managed district, and the geographical area occupied by each managed district.

[0061] Step S202: Obtain the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and obtain the sales data of a single retail product corresponding to the past simultaneous time segment of the set distribution station that was in the same time segment position as the current time segment on the previous day.

[0062] For example, obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and obtaining the sales data of a single retail product corresponding to the past simultaneous segment of the set distribution station that is in the same time segment position as the current time segment on the previous day, includes: if the current time segment is 11:00 AM to 11:30 AM, then the past time segments before the current time segment are 10:30 AM to 11:00 AM, 10:00 AM to 10:30 AM, 9:30 AM to 10:00 AM, 9:00 AM to 9:30 AM, 8:30 AM to 9:00 AM, 8:00 AM to 8:30 AM, 7:30 AM to 8:00 AM, and 7:00 AM to 7:30 AM, a total of 8 past time segments, and the past simultaneous segment is 11:00 AM to 11:30 AM on the previous day;

[0063] Step S203: Perform multiple training operations on the feedforward neural network to obtain the feedforward neural network after multiple training operations, and output the feedforward neural network after multiple training operations as the intelligent prediction model.

[0064] For example, performing multiple training operations on a feedforward neural network to obtain a feedforward neural network after multiple training operations, and using the feedforward neural network after multiple training operations as the output of an intelligent prediction model, includes: testing and simulating the process of performing multiple training operations on a feedforward neural network to obtain a feedforward neural network after multiple training operations, and using the feedforward neural network after multiple training operations as the output of an intelligent prediction model using a numerical simulation mode.

[0065] Step S204: Using an intelligent prediction model, based on the duration of each time segment, multiple configuration information of the distribution station, the sales data of each retail product corresponding to each past time segment before the current time segment, and the sales data of a single retail product corresponding to the distribution station in the past same segment, the model intelligently predicts the total sales volume of each retail product of the distribution station in the current time segment.

[0066] Step S205: When there is a type of retail product whose total sales volume is less than the total inventory volume in each of the various sales volumes corresponding to the intelligent prediction of retail products for sale, the type of retail product for sale shall be designated as an inventory warning type.

[0067] Step S206: When there is a type of retail product whose total sales volume is less than the total inventory volume in each of the various sales volumes corresponding to the intelligently predicted retail products, the type of retail product is designated as a dynamic replenishment type, and replenishment is performed on the dynamic replenishment type based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type.

[0068] Among them, performing multiple training operations on the feedforward neural network to obtain the feedforward neural network after completing multiple training operations, and outputting the feedforward neural network after completing multiple training operations as the intelligent prediction model includes: the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of retail product types sold at the set distribution station.

[0069] For example, the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of retail product types sold at the set distribution station, including: when the number of retail product types sold at the set distribution station is 200, the number of training operations completed by the feedforward neural network is 100; when the number of retail product types sold at the set distribution station is 300, the number of training operations completed by the feedforward neural network is 150; when the number of retail product types sold at the set distribution station is 400, the number of training operations completed by the feedforward neural network is 200, and so on.

[0070] Among them, obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the designated distribution station, and obtaining the sales data of a single retail product corresponding to the past simultaneous segment of the designated distribution station that was in the same time segment position as the current time segment on the previous day, includes: the number of time segments of each past time segment before the current time segment is proportional to the number of blocks of each street managed by the designated distribution station;

[0071] For example, the number of time segments in each previous time segment before the current time segment is proportional to the number of blocks in each district managed by the distribution station, including: the number of blocks in each district managed by the distribution station is 3, the number of time segments in each previous time segment before the current time segment is 6, the number of blocks in each district managed by the distribution station is 4, the number of time segments in each previous time segment before the current time segment is 8, the number of blocks in each district managed by the distribution station is 5, the number of time segments in each previous time segment before the current time segment is 10, and so on;

[0072] The process of obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the designated distribution station, and obtaining the sales data of a single retail product corresponding to the past simultaneous segment of the designated distribution station that was in the same time segment position as the current time segment on the previous day, also includes: the position of the past simultaneous segment on the time axis on the previous day is the same as the position of the current time segment on the time axis on the current day, and the current time segment is a time segment starting from the current moment;

[0073] The process of obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the designated distribution station, and obtaining the sales data of a single retail product corresponding to the same time segment of the designated distribution station on the previous day and the current time segment, further includes: the duration of each time segment is equal, and the sales data of a single retail product corresponding to each time segment of the designated distribution station is the total sales volume of each type of retail product sold by the designated distribution station within the time segment.

[0074] The process of performing multiple training operations on the feedforward neural network to obtain a feedforward neural network after multiple training operations, and using the feedforward neural network after multiple training operations as the output of the intelligent prediction model, further includes: in each training operation performed on the feedforward neural network, using the known total sales volume of each type of retail product sold at a designated distribution station within a certain past time segment as the output content of the feedforward neural network, and using the duration of each time segment, multiple configuration information of the designated distribution station, the sales data of each retail product corresponding to each past time segment before the designated distribution station, and the sales data of a single retail product corresponding to the designated distribution station in the past simultaneous segments of the designated past time segment as the input content of the feedforward neural network, and performing this training operation.

[0075] Example 2

[0076] Figure 3 The following is a flowchart illustrating the steps of an integrated method for intelligent inventory early warning and dynamic replenishment strategy according to Embodiment 2 of the present invention.

[0077] like Figure 3 As shown, when there are product categories whose total sales volume is less than the total inventory volume among the various retail products for which intelligent prediction is used, after designating these product categories as inventory warning categories, i.e., after step S205, the integration method of intelligent inventory warning and dynamic replenishment strategy further includes:

[0078] Step S301: Receive each type of inventory warning and use a giant screen display mechanism to complete the on-site display of each type of inventory warning;

[0079] Specifically, optical alarm mechanisms or acoustic alarm mechanisms can be selected to perform on-site early warning operations for each type of inventory warning.

[0080] Example 3

[0081] Figure 4 The following is a flowchart illustrating the steps of an integrated method for intelligent inventory early warning and dynamic replenishment strategy according to Embodiment 3 of the present invention.

[0082] like Figure 4 As shown, after performing multiple training operations on the feedforward neural network to obtain a feedforward neural network after multiple training operations, and outputting the feedforward neural network after multiple training operations as the intelligent prediction model, that is, after step S203, the integration method of intelligent inventory early warning and dynamic replenishment strategy further includes:

[0083] Step S401: Receive various model parameters of the intelligent prediction model using a data storage chip, and complete the model storage of the intelligent prediction model by storing the various model parameters of the intelligent prediction model;

[0084] For example, the data storage chip is used to receive various model parameters of the intelligent prediction model, and the model storage of the intelligent prediction model is completed by storing various model parameters of the intelligent prediction model. The data storage chip can be selected as a CF memory chip, a TF memory chip, or an MMC memory chip.

[0085] In any of the above embodiments 1-3, optionally, in the integrated method of intelligent inventory early warning and dynamic replenishment strategy:

[0086] When there is a type of retail product whose total sales volume is less than the total inventory volume in each of the various sales volumes corresponding to the intelligent prediction of retail products for sale, the retail product for sale shall be regarded as a dynamic replenishment type, and replenishment shall be carried out on the dynamic replenishment type based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type. This includes: the quantity of products replenished for the dynamic replenishment type is equal to the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type.

[0087] For example, the quantity of goods replenished for a dynamic replenishment category is equal to the difference between the total sales and total inventory corresponding to the dynamic replenishment category. This includes cases where, when garbage bags are determined to be a dynamic replenishment category for the current time segment, the predicted total sales of garbage bags in the current time segment is 50 units, and the current inventory of garbage bags is 35 units. In this case, the quantity of goods replenished for garbage bags is equal to the difference between the total sales and total inventory corresponding to garbage bags, i.e., the quantity of goods replenished for garbage bags for the current time segment is 15 units.

[0088] In each training operation performed on the feedforward neural network, the total sales volume of each type of retail product sold at the designated distribution station within a certain past time segment is used as the output of the feedforward neural network. The duration of each time segment, multiple configuration information of the designated distribution station, sales data of each retail product corresponding to each past time segment before the designated distribution station, and sales data of a single retail product corresponding to the designated distribution station in the past simultaneous segment of the designated past time segment are used as the input of the feedforward neural network. The training operation includes ensuring that the position of the past simultaneous segment of the designated past time segment on the time axis of the day before the past day of the designated past time segment is the same as the position of the designated past time segment on the time axis of the past day of the designated past time segment.

[0089] The intelligent prediction model uses the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each previous time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the previous simultaneous time segment to intelligently predict the total sales volume of each retail product of the distribution station in the current time segment. This includes: performing numerical normalization processing on the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each previous time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the previous simultaneous time segment before synchronously inputting them into the intelligent prediction model.

[0090] The intelligent prediction model uses the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each past time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the past same segment to intelligently predict the total sales volume of each retail product of the distribution station in the current time segment. It also includes: the total sales volume of each retail product of the distribution station in the current time segment obtained by intelligent prediction is a normalized representation.

[0091] The monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the designated distribution station includes: using a numerical mapping function to represent the numerical mapping relationship between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the designated distribution station.

[0092] The numerical mapping function used to represent the monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the set distribution station includes: in the numerical mapping function, the set retail product types sold at the distribution station are the input content of the numerical mapping function.

[0093] Specifically, the numerical mapping function used to represent the monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the set distribution station also includes: the option to use the MATLAB toolbox to test and simulate the running process of the numerical mapping function.

[0094] Furthermore, the numerical mapping function used to represent the monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the set distribution station also includes: in the numerical mapping function, the number of training operations completed by the feedforward neural network corresponding to the retail product types sold at the set distribution station is the output of the numerical mapping function.

[0095] Example 4

[0096] Figure 5 This is an internal structure diagram of an integrated system for intelligent inventory early warning and dynamic replenishment strategies according to Embodiment 4 of the present invention.

[0097] like Figure 5 As shown, the integrated system for intelligent inventory early warning and dynamic replenishment strategies includes the following components:

[0098] The first capturing mechanism is used to obtain the number of retail product types sold at the set distribution station, the maximum storage volume, the number of blocks in each managed block, the cumulative value of each resident population corresponding to each managed block, and the geographical area occupied by each managed block, and output them as multiple configuration information of the set distribution station.

[0099] Specifically, the system obtains the number of retail product types sold at the designated distribution station, the maximum storage capacity, the number of blocks in each managed district, the cumulative value of each resident population corresponding to each managed district, and the geographical area occupied by each managed district. This information is then output as multiple configuration information for the designated distribution station, including: using multiple information parsing components to obtain the number of retail product types sold at the designated distribution station, the maximum storage capacity, the number of blocks in each managed district, the cumulative value of each resident population corresponding to each managed district, and the geographical area occupied by each managed district.

[0100] The second capturing mechanism is used to acquire the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and to acquire the sales data of a single retail product corresponding to the past simultaneous segment of the set distribution station that was in the same time segment position as the current time segment on the previous day.

[0101] For example, obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the set distribution station, and obtaining the sales data of a single retail product corresponding to the past simultaneous segment of the set distribution station that is in the same time segment position as the current time segment on the previous day, includes: if the current time segment is 11:00 AM to 11:30 AM, then the past time segments before the current time segment are 10:30 AM to 11:00 AM, 10:00 AM to 10:30 AM, 9:30 AM to 10:00 AM, 9:00 AM to 9:30 AM, 8:30 AM to 9:00 AM, 8:00 AM to 8:30 AM, 7:30 AM to 8:00 AM, and 7:00 AM to 7:30 AM, a total of 8 past time segments, and the past simultaneous segment is 11:00 AM to 11:30 AM on the previous day;

[0102] The object assembly mechanism is used to perform multiple training operations on the feedforward neural network to obtain the feedforward neural network after multiple training operations, and output the feedforward neural network after multiple training operations as the intelligent prediction model.

[0103] For example, performing multiple training operations on a feedforward neural network to obtain a feedforward neural network after multiple training operations, and using the feedforward neural network after multiple training operations as the output of an intelligent prediction model, includes: testing and simulating the process of performing multiple training operations on a feedforward neural network to obtain a feedforward neural network after multiple training operations, and using the feedforward neural network after multiple training operations as the output of an intelligent prediction model using a numerical simulation mode.

[0104] The prediction execution mechanism is connected to the first capture mechanism, the second capture mechanism, and the object assembly mechanism, respectively. It is used to use an intelligent prediction model to intelligently predict the total sales volume of each type of retail product of the set distribution station in the current time segment based on the duration of each time segment, multiple configuration information of the set distribution station, the sales data of each retail product corresponding to each past time segment before the current time segment, and the sales data of a single retail product corresponding to the set distribution station in the past same segment.

[0105] An inventory warning mechanism, connected to the forecast execution mechanism, is used to identify a product category whose total sales volume is less than the total inventory volume in each of the various retail products for sale in the intelligent forecast.

[0106] A dynamic replenishment mechanism, connected to the forecast execution mechanism, is used to identify, when there is a type of retail product whose total sales volume is less than the total inventory volume in each of the various sales volumes corresponding to the intelligent forecast retail products, as a dynamic replenishment type, and to replenish the dynamic replenishment type based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type.

[0107] Among them, performing multiple training operations on the feedforward neural network to obtain the feedforward neural network after completing multiple training operations, and outputting the feedforward neural network after completing multiple training operations as the intelligent prediction model includes: the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of retail product types sold at the set distribution station.

[0108] For example, the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of retail product types sold at the set distribution station, including: when the number of retail product types sold at the set distribution station is 200, the number of training operations completed by the feedforward neural network is 100; when the number of retail product types sold at the set distribution station is 300, the number of training operations completed by the feedforward neural network is 150; when the number of retail product types sold at the set distribution station is 400, the number of training operations completed by the feedforward neural network is 200, and so on.

[0109] Among them, obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the designated distribution station, and obtaining the sales data of a single retail product corresponding to the past simultaneous segment of the designated distribution station that was in the same time segment position as the current time segment on the previous day, includes: the number of time segments of each past time segment before the current time segment is proportional to the number of blocks of each street managed by the designated distribution station;

[0110] For example, the number of time segments in each previous time segment before the current time segment is proportional to the number of blocks in each district managed by the distribution station, including: the number of blocks in each district managed by the distribution station is 3, the number of time segments in each previous time segment before the current time segment is 6, the number of blocks in each district managed by the distribution station is 4, the number of time segments in each previous time segment before the current time segment is 8, the number of blocks in each district managed by the distribution station is 5, the number of time segments in each previous time segment before the current time segment is 10, and so on;

[0111] The process of obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the designated distribution station, and obtaining the sales data of a single retail product corresponding to the past simultaneous segment of the designated distribution station that was in the same time segment position as the current time segment on the previous day, also includes: the position of the past simultaneous segment on the time axis on the previous day is the same as the position of the current time segment on the time axis on the current day, and the current time segment is a time segment starting from the current moment;

[0112] The process of obtaining the sales data of each retail product corresponding to each past time segment before the current time segment of the designated distribution station, and obtaining the sales data of a single retail product corresponding to the same time segment of the designated distribution station on the previous day and the current time segment, further includes: the duration of each time segment is equal, and the sales data of a single retail product corresponding to each time segment of the designated distribution station is the total sales volume of each type of retail product sold by the designated distribution station within the time segment.

[0113] The process of performing multiple training operations on the feedforward neural network to obtain a feedforward neural network after multiple training operations, and using the feedforward neural network after multiple training operations as the output of the intelligent prediction model, further includes: in each training operation performed on the feedforward neural network, using the known total sales volume of each type of retail product sold at a designated distribution station within a certain past time segment as the output content of the feedforward neural network, and using the duration of each time segment, multiple configuration information of the designated distribution station, the sales data of each retail product corresponding to each past time segment before the designated distribution station, and the sales data of a single retail product corresponding to the designated distribution station in the past simultaneous segments of the designated past time segment as the input content of the feedforward neural network, and performing this training operation.

[0114] Example 5

[0115] Figure 6 This is an internal structure diagram of an integrated system for intelligent inventory early warning and dynamic replenishment strategies according to Embodiment 5 of the present invention.

[0116] like Figure 6 As shown, the integrated system of intelligent inventory early warning and dynamic replenishment strategy also includes:

[0117] A giant screen display mechanism is connected to the inventory warning mechanism to receive each type of inventory warning and to display each type of inventory warning on-site.

[0118] Specifically, optical alarm mechanisms or acoustic alarm mechanisms can be selected to perform on-site early warning operations for each type of inventory warning.

[0119] Example 6

[0120] Figure 7 This is an internal structure diagram of an integrated system for intelligent inventory early warning and dynamic replenishment strategies according to Embodiment 6 of the present invention.

[0121] like Figure 7 As shown, the integrated system of intelligent inventory early warning and dynamic replenishment strategy also includes:

[0122] A model storage mechanism, connected to the object assembly mechanism, is used for various model parameters of the intelligent prediction model, and completes the model storage of the intelligent prediction model by storing various model parameters of the intelligent prediction model.

[0123] For example, a model storage mechanism, connected to the object assembly mechanism, is used for various model parameters of the intelligent prediction model, and the model storage of the intelligent prediction model is completed by storing various model parameters of the intelligent prediction model. The model storage mechanism can be selected as a CF memory chip, a TF memory chip, or an MMC memory chip.

[0124] And, optionally, in any of the embodiments 4-6 above, in the integrated system of intelligent inventory early warning and dynamic replenishment strategy:

[0125] When there is a type of retail product whose total sales volume is less than the total inventory volume in each of the various sales volumes corresponding to the intelligent prediction of retail products for sale, the retail product for sale shall be regarded as a dynamic replenishment type, and replenishment shall be carried out on the dynamic replenishment type based on the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type. This includes: the quantity of products replenished for the dynamic replenishment type is equal to the difference between the total sales volume and the total inventory volume corresponding to the dynamic replenishment type.

[0126] For example, the quantity of goods replenished for a dynamic replenishment category is equal to the difference between the total sales and total inventory corresponding to the dynamic replenishment category. This includes cases where, when garbage bags are determined to be a dynamic replenishment category for the current time segment, the predicted total sales of garbage bags in the current time segment is 50 units, and the current inventory of garbage bags is 35 units. In this case, the quantity of goods replenished for garbage bags is equal to the difference between the total sales and total inventory corresponding to garbage bags, i.e., the quantity of goods replenished for garbage bags for the current time segment is 15 units.

[0127] In each training operation performed on the feedforward neural network, the total sales volume of each type of retail product sold at the designated distribution station within a certain past time segment is used as the output of the feedforward neural network. The duration of each time segment, multiple configuration information of the designated distribution station, sales data of each retail product corresponding to each past time segment before the designated distribution station, and sales data of a single retail product corresponding to the designated distribution station in the past simultaneous segment of the designated past time segment are used as the input of the feedforward neural network. The training operation includes ensuring that the position of the past simultaneous segment of the designated past time segment on the time axis of the day before the past day of the designated past time segment is the same as the position of the designated past time segment on the time axis of the past day of the designated past time segment.

[0128] The intelligent prediction model uses the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each previous time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the previous simultaneous time segment to intelligently predict the total sales volume of each retail product of the distribution station in the current time segment. This includes: performing numerical normalization processing on the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each previous time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the previous simultaneous time segment before synchronously inputting them into the intelligent prediction model.

[0129] The intelligent prediction model uses the duration of each time segment, multiple configuration information of the distribution station, sales data of each retail product corresponding to each past time segment before the current time segment, and sales data of a single retail product corresponding to the distribution station in the past same segment to intelligently predict the total sales volume of each retail product of the distribution station in the current time segment. It also includes: the total sales volume of each retail product of the distribution station in the current time segment obtained by intelligent prediction is a normalized representation.

[0130] The monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the designated distribution station includes: using a numerical mapping function to represent the numerical mapping relationship between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the designated distribution station.

[0131] The numerical mapping function used to represent the monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the set distribution station includes: in the numerical mapping function, the set retail product types sold at the distribution station are the input content of the numerical mapping function.

[0132] Specifically, the numerical mapping function used to represent the monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the set distribution station also includes: the option to use the MATLAB toolbox to test and simulate the running process of the numerical mapping function.

[0133] Furthermore, the numerical mapping function used to represent the monotonically positive correlation between the number of training operations completed by the feedforward neural network and the number of retail product types sold at the set distribution station also includes: in the numerical mapping function, the number of training operations completed by the feedforward neural network corresponding to the retail product types sold at the set distribution station is the output of the numerical mapping function.

[0134] Furthermore, in the integrated method and system of intelligent inventory early warning and dynamic replenishment strategy according to the present invention:

[0135] The process of performing numerical normalization on the duration of each time segment, the configuration information of the distribution station, the sales data of each retail product corresponding to each previous time segment before the current time segment, and the sales data of a single retail product corresponding to the distribution station in the previous simultaneous time segment before synchronously inputting them into the intelligent prediction model includes: using a numerical processing component to perform numerical normalization on the duration of each time segment, the configuration information of the distribution station, the sales data of each retail product corresponding to each previous time segment before the current time segment, and the sales data of a single retail product corresponding to the distribution station in the previous simultaneous time segment;

[0136] For example, the numerical processing component is used to perform numerical normalization processing on the duration of each time segment, setting multiple configuration information of the distribution station, setting the sales data of each retail product corresponding to each past time segment before the current time segment, and obtaining the sales data of a single retail product corresponding to the distribution station in the past same segment. The numerical processing component is a CPLD device designed using VHDL language.

[0137] The process of performing numerical normalization on the duration of each time segment, the configuration information of the distribution station, the sales data of each retail product corresponding to each past time segment before the current time segment, and the sales data of a single retail product corresponding to the distribution station in the past simultaneous segments before synchronously inputting them into the intelligent prediction model further includes: using a synchronization control component connected to the numerical processing component to complete the synchronous input of the duration of each time segment after numerical normalization, the configuration information of the distribution station, the sales data of each retail product corresponding to each past time segment before the current time segment, and the sales data of a single retail product corresponding to the distribution station in the past simultaneous segments into the intelligent prediction model;

[0138] For example, a synchronization control component connected to the numerical processing component is used to complete the synchronization input of the intelligent prediction model to the following: the duration of each time segment after numerical normalization processing is performed, multiple configuration information of the distribution station is set, the sales data of each retail product corresponding to each past time segment before the current time segment is set, and the sales data of the single retail product corresponding to the distribution station in the past simultaneous segments is obtained. The numerical normalization processing is a hexadecimal numerical conversion process.

[0139] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of embodiments of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0140] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0141] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An integrated method of intelligent inventory alert and dynamic replenishment strategy, characterized in that, The method comprises: obtaining the number of retail commodity categories sold by the set distribution site, the maximum storage volume, the number of blocks managed by the set distribution site, the cumulative number of each block corresponding to the population of the set distribution site, and the geographical area occupied by each block managed by the set distribution site, and outputting them as multiple configuration information of the set distribution site; obtaining retail commodity sales data corresponding to each past time segment before the current time segment of the set distribution site, and obtaining single retail commodity sales data corresponding to the past time segment of the set distribution site at the same time segment position as the current time segment of the previous day; performing multiple training operations on the feedforward neural network to obtain a feedforward neural network after completing the multiple training operations, and outputting the feedforward neural network after completing the multiple training operations as an intelligent prediction model; using the intelligent prediction model to intelligently predict the total sales of each retail commodity corresponding to each retail commodity sold by the set distribution site in the current time segment based on the duration of each time segment, the multiple configuration information of the set distribution site, the retail commodity sales data corresponding to each past time segment before the current time segment of the set distribution site, and the single retail commodity sales data corresponding to the past time segment of the set distribution site; when there is a retail commodity sold with a total sales less than the total inventory among the total sales of each retail commodity corresponding to each retail commodity sold intelligently predicted, the retail commodity sold is taken as an inventory warning category; when there is a retail commodity sold with a total sales less than the total inventory among the total sales of each retail commodity corresponding to each retail commodity sold intelligently predicted, the retail commodity sold is taken as a dynamic replenishment category, and the dynamic replenishment category is replenished based on the difference between the total sales and the total inventory corresponding to the dynamic replenishment category; wherein the numerical processing component is used to perform numerical normalization processing on the duration of each time segment, the multiple configuration information of the set distribution site, the retail commodity sales data corresponding to each past time segment before the current time segment of the set distribution site, and the single retail commodity sales data corresponding to the past time segment of the set distribution site, respectively, the numerical processing component is a CPLD device designed using VHDL language, and the numerical normalization processing is hexadecimal numerical conversion processing; wherein the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of retail commodities sold by the set distribution site. In each training operation performed on the feedforward neural network, the total sales of each type of retail product sold in a certain past time segment of the known set distribution site is taken as the output of the feedforward neural network, and the duration of each time segment, the configuration information of the set distribution site, the retail product sales data of each past time segment before the certain past time segment of the set distribution site, and the single retail product sales data of the past time segment corresponding to the certain past time segment of the set distribution site are taken as the input of the feedforward neural network, and the training operation is performed. The number of time segments before the current time segment is proportional to the number of blocks managed by the set distribution site.

2. The integrated method of intelligent inventory warning and dynamic replenishment strategy according to claim 1, wherein: The retail product sales data of each past time segment before the current time segment of the set distribution site and the single retail product sales data of the past time segment corresponding to the same time segment position of the previous day and the current time segment of the set distribution site further include that the position of the past time segment on the time axis of the previous day is the same as the position of the current time segment on the time axis of the current day, and the current time segment is a time segment starting from the current time. The retail product sales data of each past time segment before the current time segment of the set distribution site and the single retail product sales data of the past time segment corresponding to the same time segment position of the previous day and the current time segment of the set distribution site further include that the duration of each time segment is equal, and the single retail product sales data of each time segment of the set distribution site is the total sales of each type of retail product sold in the time segment.

3. The method of claim 2, wherein the intelligent inventory alert and dynamic replenishment strategy integration method is characterized by, When there is a type of retail product sold with a total sales less than the total inventory in the total sales of each type of retail product sold in the intelligent prediction, the method further includes: Receiving each inventory warning type and displaying each inventory warning type on a large screen display mechanism.

4. The method of claim 2, wherein the intelligent inventory alert and dynamic replenishment strategy integration method is characterized by, After performing multiple training operations on the feedforward neural network to obtain the feedforward neural network after the multiple training operations, and taking the feedforward neural network after the multiple training operations as the intelligent prediction model output, the method further includes: Receiving the model parameters of the intelligent prediction model using a data storage chip, and storing the model parameters of the intelligent prediction model to complete the model storage of the intelligent prediction model.

5. The integrated method of intelligent inventory warning and dynamic replenishment strategy according to any one of claims 2-4, wherein: In the case that the total sales of each of the various types of on-sale retail goods in each time segment is less than the total inventory of the corresponding type of on-sale retail goods, the type of on-sale retail good is taken as a dynamic replenishment type, and the dynamic replenishment type is replenished based on the difference between the total sales and the total inventory of the dynamic replenishment type, including: the number of goods replenished for the dynamic replenishment type is equal to the difference between the total sales and the total inventory of the dynamic replenishment type; In each training operation of the feedforward neural network, the known total sales of each of the various types of on-sale retail goods in a certain past time segment is taken as the output of the feedforward neural network, and the duration of each time segment, the configuration information of the set distribution site, the retail sales data of each past time segment before the certain past time segment, and the single retail sales data of the past time segment corresponding to the certain past time segment of the set distribution site are taken as the input of the feedforward neural network, and the execution of the training operation includes: the position of the past time segment in the certain past time segment on the time axis of the day before the day of the certain past time segment is the same as the position of the certain past time segment on the time axis of the day of the certain past time segment; In the case that the total sales of each of the various types of on-sale retail goods in each time segment is less than the total inventory of the corresponding type of on-sale retail goods, the type of on-sale retail good is taken as a dynamic replenishment type, and the dynamic replenishment type is replenished based on the difference between the total sales and the total inventory of the dynamic replenishment type, including: the number of goods replenished for the dynamic replenishment type is equal to the difference between the total sales and the total inventory of the dynamic replenishment type; In the case that the total sales of each of the various types of on-sale retail goods in each time segment is less than the total inventory of the corresponding type of on-sale retail goods, the type of on-sale retail good is taken as a dynamic replenishment type, and the dynamic replenishment type is replenished based on the difference between the total sales and the total inventory of the dynamic replenishment type, including: the number of goods replenished for the dynamic replenishment type is equal to the difference between the total sales and the total inventory of the dynamic replenishment type; In the case that the total sales of each of the various types of on-sale retail goods in each time segment is less than the total inventory of the corresponding type of on-sale retail goods, the type of on-sale retail good is taken as a dynamic replenishment type, and the dynamic replenishment type is replenished based on the difference between the total sales and the total inventory of the dynamic replenishment type, including: the number of goods replenished for the dynamic replenishment type is equal to the difference between the total sales and the total inventory of the dynamic replenishment type; The number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of selling retail product categories of the set distribution site, and the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of selling retail product categories of the set distribution site includes: using a numerical mapping function to represent the numerical mapping relationship between the number of training operations completed by the feedforward neural network and the number of selling retail product categories of the set distribution site. The number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of selling retail product categories of the set distribution site, and the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of selling retail product categories of the set distribution site includes: in the numerical mapping function, the number of selling retail product categories of the set distribution site is the input content of the numerical mapping function. The number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of selling retail product categories of the set distribution site, and the number of training operations completed by the feedforward neural network is monotonically positively correlated with the number of selling retail product categories of the set distribution site includes: in the numerical mapping function, the number of training operations completed by the feedforward neural network corresponding to the number of selling retail product categories of the set distribution site is the output content of the numerical mapping function.

6. An integrated system of intelligent inventory early warning and dynamic replenishment strategy, characterized in that, The system comprises: The first capturing mechanism is used to acquire the number of selling retail product categories, the maximum storage volume, the number of blocks managed by each block, the cumulative value of each resident population corresponding to each block managed by each block, and the geographical area jointly occupied by each block managed by the set distribution site, and output as multiple configuration information of the set distribution site. The second capturing mechanism is used to acquire each retail product sales data corresponding to each past time segment before the current time segment of the set distribution site and to acquire single retail product sales data corresponding to the past same time segment of the set distribution site on the previous day and the current time segment. The object building mechanism is used to perform multiple training operations on the feedforward neural network to obtain the feedforward neural network after completing the multiple training operations, and output the feedforward neural network after completing the multiple training operations as an intelligent prediction model. The prediction execution mechanism is respectively connected with the first capturing mechanism, the second capturing mechanism and the object building mechanism, and is used to intelligently predict the total sales amount of each selling retail product in the current time segment of the set distribution site according to the duration of each time segment, the multiple configuration information of the set distribution site, each retail product sales data corresponding to each past time segment before the current time segment of the set distribution site, and the single retail product sales data corresponding to the past same time segment of the set distribution site. The inventory warning mechanism is connected with the prediction execution mechanism, and is used to take the selling retail product category as an inventory warning category when there is a selling retail product category with a total sales amount less than the total inventory amount in the intelligent prediction of the total sales amount of each selling retail product. The dynamic replenishment mechanism is connected with the prediction execution mechanism, and is used for replenishing the on-sale retail commodity category as a dynamic replenishment category when the total sales amount of each on-sale retail commodity category is less than the total inventory amount in the total sales amount of each on-sale retail commodity category predicted by the intelligent prediction, and replenishing the dynamic replenishment category based on the difference between the total sales amount and the total inventory amount corresponding to the dynamic replenishment category; The numerical processing assembly is used to complete the numerical normalization processing of the duration of each time segment, the multiple configuration information of the set distribution site, the sales data of each retail commodity corresponding to each past time segment before the current time segment, and the sales data of each retail commodity corresponding to each past time segment of the set distribution site, the numerical processing assembly is a CPLD device designed by using a VHDL language, and the numerical normalization processing is a hexadecimal numerical conversion processing. The number of training operations completed by the feedforward neural network is monotonically and positively correlated with the number of on-sale retail commodity categories of the set distribution site. In each training operation of the feedforward neural network, the total sales amount of each on-sale retail commodity category of the set distribution site in a certain past time segment is used as the output content of the feedforward neural network, and the duration of each time segment, the multiple configuration information of the set distribution site, the sales data of each retail commodity corresponding to each past time segment before the certain past time segment of the set distribution site, and the sales data of each retail commodity corresponding to each past time segment of the set distribution site in the certain past time segment are used as the input content of the feedforward neural network. The number of time segments of each past time segment before the current time segment is proportional to the number of blocks managed by the set distribution site.

7. The integrated system of the intelligent inventory early warning and dynamic replenishment strategy according to claim 6, wherein: The sales data of each retail commodity corresponding to each past time segment before the current time segment of the set distribution site and the sales data of each retail commodity corresponding to each past time segment of the set distribution site in the same time segment position as the current time segment on the previous day further include that the position of the past time segment on the time axis of the previous day is the same as the position of the current time segment on the time axis of the current day, and the current time segment is a time segment with the current time as a starting point. The duration of each time segment is equal, and the sales data of each retail commodity corresponding to each time segment of the set distribution site is the total sales amount of each on-sale retail commodity category in the time segment.

8. The integrated system of intelligent inventory alert and dynamic replenishment strategy as claimed in claim 7, wherein, The system further comprises: The giant screen display mechanism is connected with the inventory early warning mechanism and is used for receiving each inventory early warning category and completing on-site display of each inventory early warning category.

9. The integrated system of intelligent inventory alert and dynamic replenishment strategy as claimed in claim 7, wherein, The system further comprises: The model storage mechanism is connected with the object group mechanism and is used for various model parameters of the intelligent prediction model and completes model storage of the intelligent prediction model by storing the various model parameters of the intelligent prediction model. 10.The integrated system of intelligent inventory early warning and dynamic replenishment strategy according to any one of claims 7-9, wherein: when there is a sold retail commodity category with a total sales amount less than a total inventory amount in each total sales amount of various sold retail commodities corresponding to the intelligent prediction, the sold retail commodity category is taken as a dynamic replenishment category, and replenishment of the dynamic replenishment category based on a difference between the total sales amount and the total inventory amount corresponding to the dynamic replenishment category includes: a quantity of goods replenished for the dynamic replenishment category is equal to the difference between the total sales amount and the total inventory amount corresponding to the dynamic replenishment category; wherein, in each training operation performed on the feedforward neural network, each total sales amount of various sold retail commodities corresponding to a known set distribution site in a certain past time segment is taken as output content of the feedforward neural network, a duration of each time segment, multiple configuration information of the set distribution site, each total retail commodity sales data of each past time segment corresponding to the set distribution site before the certain past time segment, and single retail commodity sales data corresponding to a past time segment of the set distribution site in the certain past time segment are taken as input content of the feedforward neural network, and the training operation includes: a position of the past time segment in the certain past time segment on a time axis of a day before a day of the certain past time segment is the same as a position of the certain past time segment on the time axis of the day of the certain past time segment; wherein, the intelligent prediction of each total sales amount of various sold retail commodities corresponding to the set distribution site in the current time segment by the intelligent prediction model according to the duration of each time segment, the multiple configuration information of the set distribution site, each total retail commodity sales data of each past time segment corresponding to the set distribution site before the current time segment, and the single retail commodity sales data corresponding to the past time segment of the set distribution site includes: the duration of each time segment, the multiple configuration information of the set distribution site, each total retail commodity sales data of each past time segment corresponding to the set distribution site before the current time segment, and the single retail commodity sales data corresponding to the past time segment of the set distribution site are synchronously input to the intelligent prediction model after being subjected to value normalization processing respectively; wherein, the intelligent prediction of each total sales amount of various sold retail commodities corresponding to the set distribution site in the current time segment by the intelligent prediction model according to the duration of each time segment, the multiple configuration information of the set distribution site, each total retail commodity sales data of each past time segment corresponding to the set distribution site before the current time segment, and the single retail commodity sales data corresponding to the past time segment of the set distribution site includes: the duration of each time segment, the multiple configuration information of the set distribution site, each total retail commodity sales data of each past time segment corresponding to the set distribution site before the current time segment, and the single retail commodity sales data corresponding to the past time segment of the set distribution site are synchronously input to the intelligent prediction model after being subjected to value normalization processing respectively; The intelligent prediction model is used to intelligently predict the total sales of each type of retail product in the current time segment of the set distribution site according to the duration of each time segment, the multiple configuration information of the set distribution site, the sales data of each type of retail product corresponding to each past time segment before the current time segment of the set distribution site, and the sales data of each type of retail product corresponding to the past time segment of the set distribution site. The total sales of each type of retail product in the current time segment of the set distribution site obtained by intelligent prediction is in a numerical normalized form. The number of training operations completed by the feedforward neural network is monotonically and positively associated with the number of types of retail products sold by the set distribution site, and the number of training operations completed by the feedforward neural network is monotonically and positively associated with the number of types of retail products sold by the set distribution site includes that a numerical mapping function is used to represent a numerical mapping relationship between the number of training operations completed by the feedforward neural network and the number of types of retail products sold by the set distribution site. The number of training operations completed by the feedforward neural network is monotonically and positively associated with the number of types of retail products sold by the set distribution site, and the number of training operations completed by the feedforward neural network is monotonically and positively associated with the number of types of retail products sold by the set distribution site includes that in the numerical mapping function, the number of types of retail products sold by the set distribution site is input content of the numerical mapping function. The number of training operations completed by the feedforward neural network is monotonically and positively associated with the number of types of retail products sold by the set distribution site, and the number of training operations completed by the feedforward neural network is monotonically and positively associated with the number of types of retail products sold by the set distribution site includes that in the numerical mapping function, the number of training operations completed by the feedforward neural network corresponding to the number of types of retail products sold by the set distribution site is output content of the numerical mapping function.

Images