Inventory predictive scheduling method and device based on machine learning, equipment and medium

Abstract

The invention discloses an inventory predictive scheduling method and device based on machine learning, equipment and a medium, and the method comprises the steps: carrying out the preprocessing of input historical sales data, obtaining basic data, extracting the basic data to obtain corresponding feature information, inputting the feature information into a scheduling prediction model, and training the scheduling prediction model through machine learning, and predicting the feature information based on the trained scheduling prediction model to obtain a scheduling prediction type, obtaining a target inventory strategy from an inventory strategy configuration library according to the scheduling prediction type, and configuring inventory parameters of the commodity inventory based on the target inventory strategy. According to the scheduling method, the scheduling prediction model can be trained through machine learning, and the scheduling prediction type is obtained through prediction based on the trained scheduling prediction model, so that the matched target inventory strategy is obtained to configure the inventory parameters of the commodity inventory, and accurate scheduling of the commodity inventory can be realized based on the scheduling early warning threshold.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and medium for inventory predictive scheduling based on machine learning. Background Art

[0002] Modern manufacturing and processing enterprises usually need to carry out supply chain management, and inventory management is a key link in supply chain management. Traditional inventory scheduling methods often rely on historical data and experience judgment, which is difficult to cope with the rapid changes in market demand. Due to inaccurate inventory reserve forecasts, enterprises often face problems of excess or shortage of inventory, resulting in waste of resources and reduced customer satisfaction. Therefore, the inventory scheduling methods in the prior art methods have the problem of insufficient scheduling accuracy. Summary of the invention

[0003] The embodiments of the present invention provide a method, device, equipment and medium for predictive inventory scheduling based on machine learning, aiming to solve the problem of insufficient scheduling accuracy existing in the inventory scheduling methods in the prior art methods.

[0004] In a first aspect, an embodiment of the present invention provides an inventory predictive scheduling method based on machine learning, wherein the method comprises:

[0005] If the input historical sales data is received, the historical sales data is preprocessed according to the preset preprocessing rules to obtain corresponding basic data;

[0006] Extracting corresponding feature information from the basic data according to a preset feature extraction strategy;

[0007] Inputting the characteristic information into a preset scheduling prediction model to train the scheduling prediction model through machine learning;

[0008] Predicting the characteristic information through the trained scheduling prediction model to obtain a corresponding scheduling prediction type;

[0009] Acquire a matching target inventory strategy from an inventory strategy configuration library according to the scheduling forecast type;

[0010] The inventory parameters of the commodity inventory are configured according to the target inventory strategy.

[0011] In a second aspect, an embodiment of the present invention further provides an inventory predictive scheduling device based on machine learning, wherein the device is used to execute the inventory predictive scheduling method based on machine learning as described in the first aspect above, and the device includes:

[0012] A data preprocessing unit, configured to preprocess the historical sales data according to preset preprocessing rules upon receiving the input historical sales data, so as to obtain corresponding basic data;

[0013] A feature information extraction unit, used to extract corresponding feature information from the basic data according to a preset feature extraction strategy;

[0014] A model training unit, used for inputting the feature information into a preset scheduling prediction model to train the scheduling prediction model through machine learning;

[0015] A prediction unit, used to predict the characteristic information through the trained scheduling prediction model to obtain a corresponding scheduling prediction type;

[0016] A target inventory strategy acquisition unit, configured to acquire a matching target inventory strategy from an inventory strategy configuration library according to the scheduling forecast type;

[0017] The inventory parameter configuration unit is used to configure the inventory parameters of the commodity inventory according to the target inventory strategy.

[0018] In a third aspect, an embodiment of the present invention further provides a computer device, wherein the device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0019] Memory, used to store computer programs;

[0020] The processor is used to implement the steps of the inventory predictive scheduling method based on machine learning described in the first aspect when executing the program stored in the memory.

[0021] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the inventory predictive scheduling method based on machine learning as described in the first aspect above are implemented.

[0022] The embodiment of the present invention provides a method, device, equipment and medium for predictive inventory scheduling based on machine learning, the method comprising: preprocessing the input historical sales data to obtain basic data and extracting corresponding feature information therefrom, inputting the feature information into a scheduling prediction model to train the scheduling prediction model through machine learning, predicting the feature information based on the trained scheduling prediction model to obtain a scheduling prediction type, obtaining a target inventory strategy from an inventory strategy configuration library according to the scheduling prediction type, and configuring the inventory parameters of the commodity inventory based on the target inventory strategy. The above-mentioned scheduling method can train the scheduling prediction model through machine learning, and obtain the scheduling prediction type based on the prediction of the trained scheduling prediction model, thereby obtaining a matching target inventory strategy to configure the inventory parameters of the commodity inventory, and can realize accurate scheduling of the inventory based on the configured scheduling warning threshold, thereby improving the reliability of inventory scheduling management. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0024] Figure 1 A method flow chart of a method for predictive inventory scheduling based on machine learning provided by an embodiment of the present invention;

[0025] Figure 2 A schematic block diagram of an inventory predictive scheduling device based on machine learning provided by an embodiment of the present invention;

[0026] Figure 3 It is a schematic block diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0029] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0030] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0031] See also Figure 1 , as shown in the figure, the embodiment of the present invention application provides an inventory predictive scheduling method based on machine learning. The method is applied to a user terminal or a management server. The method is executed by an application installed in the user terminal or the management server. The user terminal can be a laptop, desktop computer, tablet computer or mobile phone and other devices; the management server is also a server side configured within an enterprise or organization for commodity inventory scheduling management. The management server can exchange information with the client. When the method is applied to the user terminal, the user can directly input data information in the terminal, and configure the scheduling warning threshold locally through the user terminal to achieve predictive scheduling of commodity inventory; when the method is applied to the management server, the management server can provide services to the client, and the client sends data information to the management server. The management server executes the above method and configures the scheduling warning threshold on the management server, and feeds back the configuration results to the client, thereby achieving predictive scheduling of commodity inventory. As shown Figure 1 As shown, the method includes steps S110 to S160.

[0032] S110: If the input historical sales data is received, the historical sales data is preprocessed according to preset preprocessing rules to obtain corresponding basic data.

[0033] If the input historical sales data is received, the historical sales data is preprocessed according to the preset preprocessing rules to obtain the corresponding basic data. The user can input the historical sales data, and the historical sales data can be preprocessed according to the preprocessing rules after receiving the historical sales data. The preprocessing rules are specific rules for cleaning and de-noising the data information, so as to filter out the data that meets the corresponding requirements. After preprocessing, the corresponding basic data can be obtained from the historical sales data. Among them, the historical sales data contains data information of one or more commodities sold every day. The following embodiments of this application only describe the historical sales data containing data information of one commodity, and the case where the historical sales data contains data information of multiple commodities can be deduced by analogy; one commodity corresponds to at least one model, and the historical sales data can record the data of each model of the commodity sold every day.

[0034] In a specific embodiment, step S110 includes sub-steps: filtering out initial data that conforms to the data format from the historical sales data according to the data format in the preprocessing rule; and numerically filtering the initial data according to the data filtering interval in the preprocessing rule to obtain data that conforms to the data filtering interval as the corresponding basic data.

[0035] First, according to the data format configured in the preprocessing rules, initial data that conforms to the data format can be filtered out from the historical sales data. The historical sales data also includes information such as the color, style, price, etc. of each model of the product. Since it is necessary to focus on predicting the sales volume of the product in the embodiment of the present application, the data format can be set to a numerical format and a date format. Then, according to the data format, data that matches the numerical format and the date format can be filtered out from the historical sales data as the initial data.

[0036] Furthermore, the values ​​in the initial data are filtered according to the data filtering interval, so as to obtain data that conforms to the data filtering interval as the basic data. One piece of data in the initial data includes the sales value and date of the model in the product. For example, the daily sales volume of each model in the product can be set to a data filtering interval of [0, 999]. Then, it can be determined whether each value in the initial data is within the above data filtering interval. If it is, the corresponding sales value and the corresponding date are retained. If it is not within the data filtering interval, the corresponding sales value and the corresponding date are deleted. Finally, the sales value, date and combination obtained are retained as the basic data.

[0037] In a specific embodiment, after selecting the initial data conforming to the data format from the historical sales data according to the data format in the preprocessing rule, it also includes: determining whether the initial data contains missing data; if the initial data contains missing data, calculating the interpolation corresponding to the missing data according to the interpolation calculation formula in the preprocessing rule and filling the missing data.

[0038] After obtaining the initial data, it can be determined whether the initial data contains missing data. If so, the interpolation corresponding to the missing data is calculated according to the interpolation calculation formula. The interpolation calculation formula can be a mean calculation formula, which calculates the average value between the adjacent values ​​before and after the missing data as the interpolation value (sales value) corresponding to the missing data for numerical filling. The values ​​adjacent to the missing data are the sales values ​​with dates one date before and one date after the missing data. In addition, the interpolation calculation formula can also be a linear interpolation basis function constructed based on Lagrange's mean value theorem. The adjacent values ​​before and after the missing data can be calculated based on the interpolation calculation formula, and the obtained values ​​are used as the interpolation value (sales value) corresponding to the missing data for numerical filling.

[0039] S120: Extract corresponding feature information from the basic data according to a preset feature extraction strategy.

[0040] The corresponding feature information is extracted from the basic data according to the preset feature extraction strategy. Furthermore, the basic data can be subjected to feature extraction according to the feature extraction strategy, thereby obtaining the corresponding feature information.

[0041] In a specific embodiment, step S120 includes sub-steps: performing statistics on the values ​​in the basic data according to the statistical items in the feature extraction strategy to obtain statistical values ​​corresponding to each statistical item; calculating the feature conversion value corresponding to the statistical value according to the conversion calculation formula in the feature extraction strategy to obtain corresponding feature information.

[0042] First, the values ​​in the basic data can be counted according to the statistical items set in the feature extraction rules. The statistical items can include the monthly sales of each model of the product, the month-on-month sales growth rate of each model of the product, the mean square error between the monthly sales of each model of the product, etc. The values ​​in the basic data can be counted by the set statistical items, so as to obtain the statistical value corresponding to each statistical item. The month-on-month sales growth rate of a certain model of the product can be calculated based on the monthly sales value of the model, such as calculating the growth rate of the sales value of the same model in two consecutive months as the month-on-month sales growth rate; the monthly sales mean square error between each model can be calculated based on the monthly sales of each model.

[0043] The statistical values ​​of each statistical item can be converted according to the conversion calculation formula in the feature extraction strategy, and each statistical value can be converted into a corresponding feature conversion value. The conversion calculation formula can be: r = ln (e + (xt) / t); wherein t is the parameter value pre-configured by the conversion calculation formula, x is the statistical value, and r is the calculated feature conversion value corresponding to the statistical value. By performing the above feature extraction steps, the corresponding feature information can be extracted from the basic data, and the feature information includes feature conversion values ​​corresponding to multiple months.

[0044] S130: Input the feature information into a preset scheduling prediction model to train the scheduling prediction model through machine learning.

[0045] The feature information is input into a preset scheduling prediction model to train the scheduling prediction model through machine learning. The scheduling prediction model can be trained through the feature information. Specifically, part of the feature conversion values ​​in the feature information can be used as input information, and the feature conversion values ​​(or statistical values) of subsequent months can be used as verification information of the training results to train the scheduling prediction model through machine learning.

[0046] For example, S y-6 To S y-1 , six months of feature conversion values ​​as input information, where S y-6 Represents the feature conversion value corresponding to the month y-6, and y represents the month number. The scheduling prediction model is a neural network model built based on artificial intelligence. The feature values ​​of multiple months are input into the scheduling prediction model, so that the node values ​​corresponding to each output node in the scheduling prediction model can be obtained. Each output node corresponds to a prediction type, and the node value of each output node is the matching degree between the input information and the prediction type of the output node. The prediction type of the output node with the highest matching degree is obtained, and the deviation between the basic parameter value in the prediction type and the feature conversion value corresponding to the month y is calculated. Machine learning is used to optimize and adjust the parameters in the scheduling prediction model based on the deviation, such as using the gradient descent algorithm to optimize and adjust the parameters in the scheduling prediction model, so as to realize the training of the scheduling prediction model. Furthermore, S can be obtained. y-5 To S y-5 The feature conversion value of month y+1 is used as input information, and the feature conversion value of month y+1 is used as verification information of the training result, and the scheduling prediction model is trained again.

[0047] S140. Predict the characteristic information using the trained scheduling prediction model to obtain a corresponding scheduling prediction type.

[0048] The characteristic information is predicted by the trained scheduling prediction model to obtain a corresponding scheduling prediction type. The trained scheduling prediction model can be used to predict the characteristic information, thereby outputting a scheduling prediction type corresponding to the characteristic information.

[0049] In a specific embodiment, step S140 includes sub-steps: intercepting partial feature data of a preset duration from the feature information; inputting the intercepted partial feature data into the scheduling prediction model for prediction to obtain a corresponding scheduling prediction type.

[0050] Specifically, partial feature data of a preset duration can be extracted from the feature information according to the current month. For example, if the current month is y+5 and the preset duration is 6 months, partial feature data of the six months y-1 to y+4 can be correspondingly extracted from the feature information.

[0051] The intercepted part of feature data is input into the scheduling prediction model for prediction, so as to obtain the prediction type of an output node with the highest matching degree as the scheduling prediction type.

[0052] The forecast types include flat sales, slow sales growth, rapid sales growth, slow sales decline, and rapid sales decline.

[0053] S150. Acquire a matching target inventory strategy from an inventory strategy configuration library according to the scheduling forecast type.

[0054] According to the scheduling forecast type, a matching target inventory strategy is obtained from the inventory strategy configuration library. An inventory strategy configuration library is pre-configured, and the inventory strategy configuration library includes multiple inventory strategies. According to the scheduling forecast type, a matching target inventory strategy can be obtained from the inventory strategy configuration library.

[0055] In a specific embodiment, step S150 includes sub-steps: obtaining a matching alternative inventory strategy from an inventory strategy configuration library according to the commodity category of the historical sales data; and obtaining an alternative inventory strategy that matches the scheduling forecast type from the alternative inventory strategy as a target inventory strategy.

[0056] Each commodity in the historical sales data corresponds to a commodity category, such as shoes and hats, clothing, daily necessities, stationery, etc. The commodity category can be matched with the inventory strategy in the inventory strategy configuration library to obtain the inventory strategy that matches the commodity category of the historical sales data as the alternative inventory strategy.

[0057] Furthermore, the obtained alternative inventory strategy is matched with the scheduling forecast type, so as to obtain an alternative inventory strategy that matches the scheduling forecast as the target inventory strategy. The obtained target inventory strategy includes storage time and warning storage location. The storage time can be long-term inventory, medium- and long-term inventory, medium-term inventory, medium- and short-term inventory or short-term inventory. The deposit time corresponding to long-term inventory is 12 months, medium- and long-term inventory corresponds to 9 months, medium-term inventory corresponds to 6 months, medium- and short-term inventory corresponds to 3 months, and short-term inventory corresponds to 1 month. The warning storage location can be high warning, medium warning or low warning.

[0058] S160: configuring inventory parameters of commodity inventory according to the target inventory strategy.

[0059] The inventory parameters of the commodity inventory are configured according to the target inventory strategy. Each model of a commodity needs to be configured with corresponding inventory parameters, and the inventory parameters of each model include the maximum inventory and scheduling warning threshold. The inventory parameters of the commodity inventory can be configured according to the target inventory strategy, thereby realizing the intelligent configuration of the inventory scheduling plan for each model in the commodity.

[0060] In a specific embodiment, step S160 includes sub-steps: calculating the maximum inventory of each model according to the storage time in the target inventory strategy and the average sales volume of each model of the goods in the basic data; accumulating the maximum inventory of each model to obtain the corresponding accumulated total inventory; calculating the corresponding warning ratio according to the warning storage location in the target inventory strategy and the accumulated total inventory; calculating the parameter value of each model according to the warning ratio and the maximum inventory and configuring the scheduling warning threshold corresponding to each model in the commodity inventory.

[0061] Specifically, the corresponding maximum inventory is calculated based on the storage time in the target inventory strategy and the average sales volume of each model of the product in the basic data. The basic data contains the average monthly sales volume of each model of the product. The average monthly sales volume of a single model of the product is multiplied by the length of the month corresponding to the storage time to obtain the maximum inventory of the model. For example, if the storage time is mid-term inventory, the corresponding length of the month is 6 months, and the average monthly sales volume of the model is 80, then the corresponding maximum inventory of the model is 480.

[0062] The maximum inventory of each model is accumulated to obtain the accumulated total inventory. The corresponding warning ratio is further calculated based on the warning warehouse location and the accumulated total inventory in the target inventory strategy, where the warning ratio can be calculated using the formula p = b / ln(q / m), where p is the calculated warning ratio, b is the basic ratio set in the warning warehouse location, q is the accumulated total inventory, and m is the parameter value pre-set in the formula. For example, if b is 0.4, q is 2000, and m is 500, then p can be calculated to be 0.289.

[0063] The parameter value of each model is further calculated based on the warning ratio and the maximum inventory of each model. The warning ratio is multiplied by the maximum inventory of each model, and the product result is rounded down to the parameter value. For example, the warning ratio is 0.289, the maximum inventory of a certain model in the product is 480, and the calculated parameter value is 138.

[0064] The dispatch warning value of each model in the commodity inventory is configured according to the parameter value of each model, and the dispatch warning value of this model can be configured to 138. If the available quantity of this model is less than 138, the inventory dispatch warning is triggered, and the user can dispatch the inventory according to the inventory dispatch warning to meet the sales demand of each model of the commodity.

[0065] In the inventory predictive scheduling method based on machine learning disclosed in the above embodiment, the method includes: preprocessing the input historical sales data to obtain basic data and extracting corresponding feature information therefrom, inputting the feature information into the scheduling prediction model to train the scheduling prediction model through machine learning, predicting the feature information based on the trained scheduling prediction model to obtain the scheduling prediction type, obtaining the target inventory strategy from the inventory strategy configuration library according to the scheduling prediction type, and configuring the inventory parameters of the commodity inventory based on the target inventory strategy. The above scheduling method can train the scheduling prediction model through machine learning, and obtain the scheduling prediction type based on the trained scheduling prediction model prediction, so as to obtain the matching target inventory strategy to configure the inventory parameters of the commodity inventory, and the scheduling warning threshold based on the configuration can realize the accurate scheduling of the inventory, thereby improving the reliability of inventory scheduling management.

[0066] The embodiment of the present invention also provides a machine learning-based inventory predictive scheduling device, which can be configured in a user terminal or a management server, and is used to execute any embodiment of the aforementioned machine learning-based inventory predictive scheduling method. Figure 2 , Figure 2 A schematic block diagram of an inventory predictive scheduling device based on machine learning provided in an embodiment of the present invention.

[0067] like Figure 2 As shown, the inventory predictive scheduling device 100 based on machine learning includes a data preprocessing unit 110, a feature information extraction unit 120, a model training unit 130, a prediction unit 140, a target inventory strategy acquisition unit 150 and an inventory parameter configuration unit 160.

[0068] The data preprocessing unit 110 is used to preprocess the historical sales data according to preset preprocessing rules to obtain corresponding basic data when the historical sales data is received.

[0069] The feature information extraction unit 120 is used to extract corresponding feature information from the basic data according to a preset feature extraction strategy.

[0070] The model training unit 130 is used to input the feature information into a preset scheduling prediction model to train the scheduling prediction model through machine learning.

[0071] The prediction unit 140 is used to predict the characteristic information through the trained scheduling prediction model to obtain a corresponding scheduling prediction type.

[0072] The target inventory policy acquisition unit 150 is used to acquire a matching target inventory policy from an inventory policy configuration library according to the scheduling forecast type.

[0073] The inventory parameter configuration unit 160 is used to configure the inventory parameters of the commodity inventory according to the target inventory strategy.

[0074] The machine learning-based inventory predictive scheduling device provided in the embodiment of the present invention applies the above-mentioned machine learning-based inventory predictive scheduling method, pre-processes the input historical sales data to obtain basic data and extracts corresponding feature information therefrom, inputs the feature information into the scheduling prediction model to train the scheduling prediction model through machine learning, predicts the feature information based on the trained scheduling prediction model to obtain the scheduling prediction type, obtains the target inventory strategy from the inventory strategy configuration library according to the scheduling prediction type, and configures the inventory parameters of the commodity inventory based on the target inventory strategy. The above-mentioned scheduling method can train the scheduling prediction model through machine learning, and obtain the scheduling prediction type based on the prediction of the trained scheduling prediction model, thereby obtaining the matching target inventory strategy to configure the inventory parameters of the commodity inventory, and can realize accurate scheduling of the inventory based on the configured scheduling warning threshold, thereby improving the reliability of inventory scheduling management.

[0075] The above-mentioned inventory predictive scheduling device based on machine learning can be implemented in the form of a computer program, and the computer program can be used in Figure 3 Runs on the computer device shown.

[0076] See also Figure 3 , Figure 3 is a schematic block diagram of a computer device provided by an embodiment of the present invention. The computer device may be a user terminal or a management server for executing a machine learning-based inventory predictive scheduling method to implement commodity inventory predictive scheduling.

[0077] See also Figure 3 The computer device 500 includes a processor 502 , a memory and a communication interface 505 connected via a communication bus 501 , wherein the memory may include a storage medium 503 and an internal memory 504 .

[0078] The storage medium 503 may store an operating system 5031 and a computer program 5032. When the computer program 5032 is executed, the processor 502 may execute an inventory predictive scheduling method based on machine learning, wherein the storage medium 503 may be a volatile storage medium or a non-volatile storage medium.

[0079] The processor 502 is used to provide computing and control capabilities to support the operation of the entire computer device 500 .

[0080] The internal memory 504 provides an environment for the operation of the computer program 5032 in the storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute an inventory predictive scheduling method based on machine learning.

[0081] The communication interface 505 is used for network communication, such as providing data information transmission, etc. Those skilled in the art will appreciate that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device 500 to which the solution of the present invention is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0082] The processor 502 is used to run a computer program 5032 stored in the memory to implement the corresponding functions in the above-mentioned inventory predictive scheduling method based on machine learning.

[0083] Those skilled in the art will understand that Figure 3 The embodiments of the computer device shown in the figure do not constitute a limitation on the specific composition of the computer device. In other embodiments, the computer device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. For example, in some embodiments, the computer device may only include a memory and a processor. In such embodiments, the structure and function of the memory and the processor are the same as those of the embodiment of the present invention. Figure 3 The embodiments shown are consistent and will not be described again here.

[0084] It should be understood that in the embodiment of the present invention, the processor 502 may be a central processing unit (CPU), and the processor 502 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0085] In another embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium may be a volatile or non-volatile computer-readable storage medium. The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps included in the above-mentioned inventory predictive scheduling method based on machine learning are implemented.

[0086] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the above-described equipment, devices and units can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. Those of ordinary skill in the art can appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0087] In the several embodiments provided by the present invention, it should be understood that the disclosed equipment, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. Units with the same function may also be combined into one unit. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices or units, or may be an electrical, mechanical or other form of connection.

[0088] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0089] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0090] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a computer-readable storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned computer-readable storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a magnetic disk or an optical disk.

[0091] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for predictive inventory scheduling based on machine learning, characterized in that: The method comprises: If the input historical sales data is received, the historical sales data is preprocessed according to the preset preprocessing rules to obtain corresponding basic data; Extracting corresponding feature information from the basic data according to a preset feature extraction strategy; Inputting the characteristic information into a preset scheduling prediction model to train the scheduling prediction model through machine learning; Predicting the characteristic information through the trained scheduling prediction model to obtain a corresponding scheduling prediction type; Acquire a matching target inventory strategy from an inventory strategy configuration library according to the scheduling forecast type; The inventory parameters of the commodity inventory are configured according to the target inventory strategy.

2. The inventory predictive scheduling method based on machine learning according to claim 1, characterized in that: The preprocessing of the historical sales data according to the preset preprocessing rules to obtain the corresponding basic data includes: Filtering out initial data conforming to the data format from the historical sales data according to the data format in the preprocessing rule; The initial data is numerically screened according to the data screening interval in the preprocessing rule to obtain data that conforms to the data screening interval as corresponding basic data.

3. The inventory predictive scheduling method based on machine learning according to claim 2, characterized in that: After the initial data conforming to the data format is screened out from the historical sales data according to the data format in the preprocessing rule, the method further includes: Determining whether the initial data contains missing data; If the initial data includes missing data, an interpolation value corresponding to the missing data is calculated according to an interpolation calculation formula in the preprocessing rule and the missing data is filled.

4. The inventory predictive scheduling method based on machine learning according to claim 1, characterized in that: The extracting corresponding feature information from the basic data according to a preset feature extraction strategy includes: Performing statistics on the values ​​in the basic data according to the statistical items in the feature extraction strategy to obtain statistical values ​​corresponding to each of the statistical items; The feature conversion value corresponding to the statistical value is calculated according to the conversion calculation formula in the feature extraction strategy to obtain the corresponding feature information.

5. The inventory predictive scheduling method based on machine learning according to claim 1, characterized in that: The trained scheduling prediction model is used to predict the characteristic information to obtain a corresponding scheduling prediction type, including: Extracting part of the characteristic data of a preset duration from the characteristic information; The intercepted part of feature data is input into the scheduling prediction model for prediction to obtain the corresponding scheduling prediction type.

6. The inventory predictive scheduling method based on machine learning according to claim 1, characterized in that: The acquiring a matching target inventory strategy from an inventory strategy configuration library according to the scheduling forecast type includes: Obtaining a matching alternative inventory strategy from an inventory strategy configuration library according to the commodity category of the historical sales data; An alternative inventory strategy matching the scheduling forecast type is obtained from the alternative inventory strategies as a target inventory strategy.

7. The inventory predictive scheduling method based on machine learning according to claim 1, characterized in that: The configuring of inventory parameters of commodity inventory according to the target inventory strategy includes: Calculate the maximum inventory of each model according to the storage time in the target inventory strategy and the average sales volume of each model of the product in the basic data; The maximum inventory of each model is accumulated to obtain the corresponding accumulated total inventory; The corresponding warning ratio is calculated according to the warning stock location in the target inventory strategy and the accumulated total inventory; The parameter value of each model is calculated according to the warning ratio and the maximum inventory, and the scheduling warning threshold corresponding to each model in the commodity inventory is configured.

8. A predictive inventory scheduling device based on machine learning, characterized in that: The inventory predictive scheduling device based on machine learning is used to execute the inventory predictive scheduling method based on machine learning according to any one of claims 1 to 7, and the device includes: A data preprocessing unit, configured to preprocess the historical sales data according to preset preprocessing rules upon receiving the input historical sales data, so as to obtain corresponding basic data; A feature information extraction unit, used to extract corresponding feature information from the basic data according to a preset feature extraction strategy; A model training unit, used for inputting the feature information into a preset scheduling prediction model to train the scheduling prediction model through machine learning; A prediction unit, used to predict the characteristic information through the trained scheduling prediction model to obtain a corresponding scheduling prediction type; A target inventory strategy acquisition unit, configured to acquire a matching target inventory strategy from an inventory strategy configuration library according to the scheduling forecast type; The inventory parameter configuration unit is used to configure the inventory parameters of the commodity inventory according to the target inventory strategy.

9. A computer device, characterized in that: The device includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; A processor is used to implement the steps of the inventory predictive scheduling method based on machine learning described in any one of claims 1-7 when executing the program stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the inventory predictive scheduling method based on machine learning are implemented as described in any one of claims 1-7.

Images