Inventory management method and system for intelligent warehousing

Abstract

The invention relates to the technical field of intelligent warehousing, and provides an inventory management method and system for intelligent warehousing in order to solve the technical problem of how to reduce the risk of inventory article overstock and stockout, and the method comprises the following steps: obtaining the historical data of the shipment amount of a warehouse; according to the shipment amount historical data, shipment change historical data of the warehouse is obtained, and the shipment change historical data comprises the shipment increment of the warehouse in each day in the historical time; a shipment quantity prediction model is trained according to the shipment quantity historical data and the shipment change historical data, and a loss function of the shipment quantity prediction model is defined according to the prediction difference between the shipment quantity historical data and the shipment change historical data; predicting the shipment prediction quantity of the warehouse in the next shipment period through the trained shipment quantity prediction model; obtaining the current stock amount and the safety stock proportion of the warehouse; and carrying out inventory management early warning on the warehouse according to the predicted shipment amount, the current inventory amount and the safe inventory proportion of the warehouse in the next shipment period.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent warehousing, and in particular to an inventory management method and system for intelligent warehousing. Background Art

[0002] In production and business scenarios, warehouses can be used to store production materials or business goods. Usually, the amount of production materials or sales of business goods is not fixed. If the next material or commodity demand cannot be accurately predicted, and the warehouse inventory is replenished at an unreasonable time and quantity, and production or operation is carried out at a fixed rhythm, there will often be a backlog or shortage of goods in the warehouse, reducing production or business efficiency.

[0003] Related technologies have proposed intelligent warehousing systems, which can store various data such as the entry, exit, and storage of items, and can process these data intelligently. How to use intelligent technology to effectively manage warehouse inventory and reduce the risk of overstocking and out-of-stock is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an inventory management method and system for intelligent warehousing, which can accurately predict the warehouse's shipment volume, thereby providing warehouse inventory management early warning based on the prediction results, effectively reducing the risk of backlog and out-of-stock items in the warehouse.

[0005] The technical solution adopted by the present invention is as follows: A method for inventory management of intelligent warehousing comprises the following steps: obtaining the warehouse's historical data on shipment volume, wherein the historical data on shipment volume includes the warehouse's shipment volume of each day in the historical time; obtaining the warehouse's shipment change historical data based on the historical data on shipment volume, wherein the historical data on shipment change includes the warehouse's shipment increment of each day in the historical time, and the shipment increment of each day is the difference between the shipment volume of that day and the shipment volume of the previous day; training a shipment volume prediction model based on the historical data on shipment volume and the historical data on shipment change, wherein the loss function of the shipment volume prediction model is defined based on the predicted difference between the historical data on shipment volume and the historical data on shipment change; predicting the predicted shipment volume of the warehouse in the next shipment cycle through the trained shipment volume prediction model; obtaining the warehouse's current inventory volume and safety inventory ratio; and performing inventory management early warning for the warehouse based on the predicted shipment volume of the warehouse in the next shipment cycle, the current inventory volume and the safety inventory ratio.

[0006] When training the shipment prediction model, the samples in the training sample set are constructed according to the following format: Format 1: the shipments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as the corresponding target data; Format 2: the shipment increments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as the corresponding target data, wherein i is a positive integer, i is 1, 2, 3, ..., S in sequence, S=KMN, K is the total number of days in the historical time; M is a positive integer, which is a preset fixed value; N is a positive integer greater than 1, which is a preset fixed value, equal to the number of days in a shipment cycle to be predicted.

[0007] Before defining the loss function of the shipment forecasting model, the target data in the sample and the forecast data obtained by the shipment forecasting model according to the input data are vectorized respectively, wherein, when the shipments or shipment increments from the i-th to the i+M-th day in the historical time are used as input data, the forecast data obtained by the shipment forecasting model is the forecast shipments from the i+M+1-th to the i+M+N-th day, and the loss function of the shipment forecasting model is: ; in, T 1. T 2. P 1. P 2 are respectively: when i takes the same number, the vector of target data of samples in format one, the vector of target data of samples in format two, the vector of predicted data obtained according to the input data of samples in format one, and the vector of predicted data obtained according to the input data of samples in format two; F loss is the loss value; represents the magnitude of a vector; MSE Represents the mean square error between two vectors.

[0008] The shipment volume prediction model is an LSTM (Long Short-Term Memory) model.

[0009] According to the forecast shipment quantity of the warehouse in the next shipment cycle, the current inventory quantity and the safety stock ratio, the inventory management warning of the warehouse is carried out, specifically including: if the difference between the current inventory quantity of the warehouse and the forecast shipment quantity of the next shipment cycle is less than X% of the total inventory of the warehouse, a replenishment warning is issued; if the difference between the current inventory quantity of the warehouse and the forecast shipment quantity of the next shipment cycle is greater than Y% of the current inventory, a clearance warning is issued, wherein X% and Y% are the lower limit and upper limit of the safety stock ratio respectively.

[0010] An inventory management system for intelligent warehousing includes: a first acquisition module, the first acquisition module is used to acquire the warehouse's shipment history data, wherein the shipment history data includes the warehouse's shipments of each day in the historical time; a second acquisition module, the second acquisition module is used to acquire the warehouse's shipment change history data based on the shipment history data, wherein the shipment change history data includes the warehouse's shipment increments of each day in the historical time, and each day's shipment increment is the difference between the shipment of the day and the shipment of the previous day; a training module, the training module is used to train the training module based on the shipment history data and the shipment history data. A shipment volume prediction model is trained based on the shipment volume change historical data, wherein the loss function of the shipment volume prediction model is defined according to the prediction difference between the shipment volume historical data and the shipment change historical data; a prediction module is used to predict the shipment forecast of the warehouse in the next shipment cycle through the trained shipment volume prediction model; a third acquisition module is used to obtain the current inventory volume and the safety stock ratio of the warehouse; an early warning module is used to perform inventory management early warning of the warehouse according to the shipment forecast volume of the warehouse in the next shipment cycle, the current inventory volume and the safety stock ratio.

[0011] When the training module trains the shipment prediction model, the samples in the training sample set used are constructed according to the following format: Format 1: the shipments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as corresponding target data; Format 2: the shipment increments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as corresponding target data, wherein i is a positive integer, i is 1, 2, 3, ..., S in sequence, S=KMN, K is the total number of days in the historical time; M is a positive integer, which is a preset fixed value; N is a positive integer greater than 1, which is a preset fixed value, equal to the number of days in a shipment cycle to be predicted.

[0012] Before defining the loss function of the shipment forecasting model, the target data in the sample and the forecast data obtained by the shipment forecasting model according to the input data are vectorized respectively, wherein, when the shipments or shipment increments from the i-th to the i+M-th day in the historical time are used as input data, the forecast data obtained by the shipment forecasting model is the forecast shipments from the i+M+1-th to the i+M+N-th day, and the loss function of the shipment forecasting model is: ; in, T 1. T 2. P 1. P2 are respectively: when i takes the same number, the vector of target data of samples in format one, the vector of target data of samples in format two, the vector of predicted data obtained according to the input data of samples in format one, and the vector of predicted data obtained according to the input data of samples in format two; F loss is the loss value; represents the magnitude of a vector; MSE Represents the mean square error between two vectors.

[0013] The shipment volume prediction model is an LSTM model.

[0014] The early warning module is specifically used to: issue a replenishment warning when the difference between the current inventory of the warehouse and the forecast shipment amount of the next shipment cycle is less than X% of the current inventory; issue a clearance warning when the difference between the current inventory of the warehouse and the forecast shipment amount of the next shipment cycle is greater than Y% of the current inventory, wherein X% and Y% are the lower limit and upper limit of the safety stock ratio, respectively.

[0015] Beneficial effects of the present invention: The present invention derives the historical shipment increment through the historical shipment quantity, then trains the shipment prediction model based on these two data, and defines the loss function based on the predicted difference between the two data, and then predicts the shipment volume of the warehouse in the next shipment cycle according to the trained shipment prediction model, and finally performs inventory management early warning of the warehouse based on the shipment prediction volume, inventory volume and safety inventory ratio. In this way, the shipment volume of the warehouse can be accurately predicted, and the inventory management early warning of the warehouse can be performed according to the prediction results, which effectively reduces the risk of backlog and out-of-stock of items in the warehouse. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the inventory management method of intelligent warehousing according to an embodiment of the present invention; Figure 2 A block diagram of an inventory management system for intelligent warehousing according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] 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 only 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.

[0018] like Figure 1 As shown, the inventory management method of intelligent warehousing in the embodiment of the present invention includes the following steps: S1, obtaining the historical data of the shipment volume of the warehouse, wherein the historical data of the shipment volume is time series data, including the shipment volume of the warehouse on each day in the historical time.

[0019] In an embodiment of the present invention, the historical data of the warehouse's shipment volume can be directly retrieved from the intelligent warehousing system. The warehouse's daily shipment volume is the number of items shipped out of the warehouse every day due to production or sales.

[0020] S2, obtaining the historical data of shipment changes of the warehouse based on the historical data of shipment volume, wherein the historical data of shipment changes is time series data, including the shipment increment of the warehouse each day in the historical time, and the shipment increment of each day is the difference between the shipment volume of that day and the shipment volume of the previous day.

[0021] Since there is no previous day for the first day, its shipment increment can be directly set to 0.

[0022] S3, training a shipment prediction model according to the shipment history data and the shipment change history data, wherein a loss function of the shipment prediction model is defined according to the prediction difference between the shipment history data and the shipment change history data.

[0023] In one embodiment of the present invention, the shipment volume prediction model may be an LSTM model.

[0024] When training the shipment prediction model, the samples in the training sample set are constructed according to the following format: Format 1: The shipments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as the corresponding target data; Format 2: The shipment increments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as the corresponding target data. Among them, i is a positive integer, i is 1, 2, 3, ..., S in sequence, S=KMN, K is the total number of days in the historical time; M is a positive integer, which is a preset fixed value; N is a positive integer greater than 1, which is a preset fixed value, equal to the number of days in a shipment cycle to be predicted.

[0025] For example, the historical shipment data of a warehouse is {46, 55, 34, 37, 62, 53, 71, 46, 45, 36, 51, 59, 63, 47, 34, 49, 55, ...}, and the historical shipment change data of the warehouse is {0, 9, -21, 3, 25, -9, 18, -25, -1, -9, 15, 8, 4, -16, -13, 15, 6, ...}. Assuming that M is 4 and N is 3, the samples in the training sample set can be divided into two types as shown in Table 1.

[0026] Table 1

[0027] It should be noted that the above-mentioned assumed values ​​of M and N are assumptions given for the convenience of understanding the solutions of the embodiments of the present invention. In some specific embodiments, M is a relatively large value, and M is a relatively small value (the larger M is and the smaller N is, the more accurate the prediction result of the shipment volume prediction model is). Usually, M is greater than 30 and N is less than 10.

[0028] During the training process of the shipment prediction model, the basic process is as follows: after the sample is input into the shipment prediction model, it first undergoes forward propagation to obtain the predicted data based on the input data; the loss function is used to calculate the difference between the predicted data and the target data, that is, the loss value, and then back propagation is performed to optimize and update the parameters in the model according to the loss value, in order to reduce future loss values.

[0029] In the embodiment of the present invention, when training the shipment prediction model, not only the data in two formats is input, but also the loss function is defined in combination with the prediction difference of the two data. In one embodiment of the present invention, before defining the loss function of the shipment prediction model, the target data in the sample and the prediction data obtained by the shipment prediction model according to the input data can also be vectorized respectively, wherein, when the shipment or shipment increment from the i-th to the i+M-th day in the historical time is used as the input data, the prediction data obtained by the shipment prediction model is the shipment prediction quantity from the i+M+1th to the i+M+Nth day. Since both the target data and the prediction data are time series data, there are many inconveniences in calculation. Therefore, the embodiment of the present invention converts the time series data into a vector that is easy to calculate by vectorization, and the components of the vector correspond to the elements of the time series data in terms of value and position. For example, the time series data {53,71,46} can be converted into a three-dimensional vector (53,71,46), that is, the vector of the target data of sample 1-1 in the above Table 1 is (53,71,46).

[0030] In one embodiment of the present invention, the loss function of the shipment volume prediction model is: ; in, T 1. T 2. P 1. P 2 are respectively: when i takes the same number, the vector of target data of samples in format one, the vector of target data of samples in format two, the vector of predicted data obtained according to the input data of samples in format one, and the vector of predicted data obtained according to the input data of samples in format two; F loss is the loss value; represents the magnitude of a vector; MSE Represents the mean square error between two vectors.

[0031] The above-mentioned “when i takes the same number” can also be understood as: the process of training the shipment prediction model using the shipment volume and shipment increment from the i-th to the i+M-th day.

[0032] It should be noted that when the training process involves multiple iterations, the total loss is calculated over multiple iterations. F loss sum.

[0033] The embodiment of the present invention derives the historical shipment increment from the historical shipment quantity, and then trains the shipment prediction model based on the two data, and defines the loss function based on the prediction difference between the two data. Therefore, the shipment prediction model can also optimize the model parameters and improve the prediction accuracy by narrowing the prediction difference between the two data. Finally, the difference between the results predicted by the trained model based on the two data is small, and the prediction results obtained based on the two data are more accurate.

[0034] S4, predicting the warehouse’s shipment forecast volume in the next shipment cycle through the trained shipment forecast model.

[0035] In one embodiment of the present invention, the prediction behavior can be performed after the shipment stops on the day before the next shipment cycle and before the shipment starts on the first day of the next shipment cycle. For example, if the shipment starts at 8:00 a.m. every day and stops after 10:00 p.m., and the first day of the next shipment cycle is tomorrow, the shipment volume of the next shipment cycle can be predicted between 10:00 p.m. today and 8:00 a.m. tomorrow. Any data of the shipment volume and the shipment increment M+1 days before the next shipment cycle can be input into the trained shipment volume prediction model to obtain the shipment prediction volume for the next shipment cycle for a total of N days.

[0036] In one embodiment of the present invention, the shipment volume M+1 days before the next shipment cycle can also be input into the trained shipment volume forecasting model to obtain a first forecast result of the shipment forecast volume of the next shipment cycle, and the shipment increment M+1 days before the next shipment cycle can be input into the trained shipment volume forecasting model to obtain a second forecast result of the shipment forecast volume of the next shipment cycle. Then, the first forecast result and the second forecast result are averaged to obtain the final forecast result. This joint forecasting strategy can further improve the forecast accuracy.

[0037] S5, obtain the current inventory and safety stock ratio of the warehouse.

[0038] In one embodiment of the present invention, the current inventory level of the warehouse may be the inventory level after shipments are stopped one day before the next shipment cycle and before shipments start on the first day of the next shipment cycle.

[0039] In one embodiment of the present invention, the safety stock ratio may be a preset interval [X%, Y%], where X% is the lower limit of the safety stock ratio. If the inventory is lower than X% of the total inventory of the warehouse, it is considered that there is a risk of out-of-stock; Y% is the upper limit of the safety stock ratio. If the inventory is higher than Y% of the total inventory of the warehouse, it is considered that there is a risk of overstocking. The total inventory here may be the maximum inventory that the warehouse can accommodate.

[0040] S6, based on the warehouse's forecasted shipment volume in the next shipment cycle, current inventory volume and safety stock ratio, provides an inventory management warning for the warehouse.

[0041] Specifically, if the difference between the current inventory of the warehouse and the forecast shipment amount of the next shipment cycle is less than X% of the total inventory of the warehouse, a replenishment warning is issued; if the difference between the current inventory of the warehouse and the forecast shipment amount of the next shipment cycle is greater than Y% of the current inventory, a clearance warning is issued.

[0042] For production scenarios, when the warehouse stores materials for production, it can replenish stocks in time or reduce the production rhythm in the next shipment cycle according to the replenishment warning; it can suspend replenishment or speed up the production rhythm in the next shipment cycle according to the clearance warning. For business scenarios, when the warehouse stores goods for sale, it can replenish stocks in time or tighten operations in the next shipment cycle according to the replenishment warning, such as reducing advertising investment or raising prices for goods; it can suspend replenishment or expand operations in the next shipment cycle according to the clearance warning, such as increasing advertising investment or reducing prices for goods.

[0043] If the difference between the current inventory level in the warehouse and the forecasted shipment level for the next shipment cycle is between X% and Y%, no warning will be issued and normal production or operation can be continued.

[0044] To summarize, according to the inventory management method of intelligent warehousing in an embodiment of the present invention, the historical shipment increment is derived from the historical shipment quantity, and then the shipment prediction model is trained based on these two data, and the loss function is defined based on the prediction difference between the two data, and then the shipment volume of the warehouse in the next shipment cycle is predicted according to the trained shipment prediction model. Finally, the warehouse inventory management warning is performed based on the shipment prediction quantity, inventory quantity and safety stock ratio. In this way, the warehouse shipment volume can be accurately predicted, and the warehouse inventory management warning can be performed according to the prediction results, which effectively reduces the risk of backlog and out-of-stock items in the warehouse.

[0045] Corresponding to the inventory management method for intelligent warehousing of the above-mentioned embodiment, the present invention also proposes an inventory management system for intelligent warehousing.

[0046] like Figure 2As shown, the inventory management system for intelligent warehousing according to an embodiment of the present invention includes: a first acquisition module 10, a second acquisition module 20, a training module 30, a prediction module 40, a third acquisition module 50 and an early warning module 60. The first acquisition module 10 is used to acquire the warehouse's historical shipment data, wherein the shipment history data includes the warehouse's shipments of each day in the historical time; the second acquisition module 20 is used to acquire the warehouse's shipment change history data based on the shipment history data, wherein the shipment change history data includes the warehouse's shipment increments of each day in the historical time, and each day's shipment increment is the difference between the day's shipment and the previous day's shipment; the training module 30 is used to train a shipment prediction model based on the shipment history data and the shipment change history data, wherein the loss function of the shipment prediction model is defined based on the predicted difference between the shipment history data and the shipment change history data; the prediction module 40 is used to predict the warehouse's shipment forecast for the next shipment cycle using the trained shipment prediction model; the third acquisition module 50 is used to acquire the warehouse's current inventory and safety inventory ratio; the early warning module 60 is used to perform inventory management early warning for the warehouse based on the warehouse's shipment forecast, current inventory and safety inventory ratio in the next shipment cycle.

[0047] In an embodiment of the present invention, the historical data of the warehouse's shipment volume can be directly retrieved from the intelligent warehousing system. The warehouse's daily shipment volume is the number of items shipped out of the warehouse every day due to production or sales.

[0048] Since there is no previous day for the first day, its shipment increment can be directly set to 0.

[0049] In one embodiment of the present invention, the shipment volume prediction model may be an LSTM model.

[0050] When training the shipment prediction model in the training module 30, the samples in the training sample set used are constructed according to the following format: Format 1: the shipments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as the corresponding target data; Format 2: the shipment increments from the i-th to the i+M-th day in the historical time are used as input data, and the shipments from the i+M+1-th to the i+M+N-th day are used as the corresponding target data. Wherein, i is a positive integer, i is 1, 2, 3, ..., S in sequence, S=KMN, K is the total number of days in the historical time; M is a positive integer, which is a preset fixed value; N is a positive integer greater than 1, which is a preset fixed value, equal to the number of days in a shipment cycle to be predicted.

[0051] For example, the historical shipment data of a warehouse is {46, 55, 34, 37, 62, 53, 71, 46, 45, 36, 51, 59, 63, 47, 34, 49, 55, ...}, and the historical shipment change data of the warehouse is {0, 9, -21, 3, 25, -9, 18, -25, -1, -9, 15, 8, 4, -16, -13, 15, 6, ...}. Assuming that M is 4 and N is 3, the samples in the training sample set can be divided into two types as shown in Table 1.

[0052] It should be noted that the above-mentioned assumed values ​​of M and N are assumptions given for the convenience of understanding the solutions of the embodiments of the present invention. In some specific embodiments, M is a relatively large value, and M is a relatively small value (the larger M is and the smaller N is, the more accurate the prediction result of the shipment volume prediction model is). Usually, M is greater than 30 and N is less than 10.

[0053] During the training process of the shipment prediction model, the basic process is as follows: after the sample is input into the shipment prediction model, it first undergoes forward propagation to obtain the predicted data based on the input data; the loss function is used to calculate the difference between the predicted data and the target data, that is, the loss value, and then back propagation is performed to optimize and update the parameters in the model according to the loss value, in order to reduce future loss values.

[0054] In the embodiment of the present invention, when training the shipment prediction model, not only the data in two formats is input, but also the loss function is defined in combination with the prediction difference of the two data. In one embodiment of the present invention, before defining the loss function of the shipment prediction model, the target data in the sample and the prediction data obtained by the shipment prediction model according to the input data can also be vectorized respectively, wherein, when the shipment or shipment increment from the i-th to the i+M-th day in the historical time is used as the input data, the prediction data obtained by the shipment prediction model is the shipment prediction quantity from the i+M+1th to the i+M+Nth day. Since both the target data and the prediction data are time series data, there are many inconveniences in calculation. Therefore, the embodiment of the present invention converts the time series data into a vector that is easy to calculate by vectorization, and the components of the vector correspond to the elements of the time series data in terms of value and position. For example, the time series data {53,71,46} can be converted into a three-dimensional vector (53,71,46), that is, the vector of the target data of sample 1-1 in the above Table 1 is (53,71,46).

[0055] In one embodiment of the present invention, the loss function of the shipment volume prediction model is: ; in, T 1. T 2. P 1. P2 are respectively: when i takes the same number, the vector of target data of samples in format one, the vector of target data of samples in format two, the vector of predicted data obtained according to the input data of samples in format one, and the vector of predicted data obtained according to the input data of samples in format two; F loss is the loss value; represents the magnitude of a vector; MSE Represents the mean square error between two vectors.

[0056] The above-mentioned “when i takes the same number” can also be understood as: the process of training the shipment prediction model using the shipment volume and shipment increment from the i-th to the i+M-th day.

[0057] It should be noted that when the training process involves multiple iterations, the total loss is calculated over multiple iterations. F loss sum.

[0058] The embodiment of the present invention derives the historical shipment increment from the historical shipment quantity, and then trains the shipment prediction model based on the two data, and defines the loss function based on the prediction difference between the two data. Therefore, the shipment prediction model can also optimize the model parameters and improve the prediction accuracy by narrowing the prediction difference between the two data. Finally, the difference between the results predicted by the trained model based on the two data is small, and the prediction results obtained based on the two data are more accurate.

[0059] In one embodiment of the present invention, the prediction behavior can be performed after the shipment stops on the day before the next shipment cycle and before the shipment starts on the first day of the next shipment cycle. For example, if the shipment starts at 8:00 a.m. every day and stops after 10:00 p.m., and the first day of the next shipment cycle is tomorrow, the shipment volume of the next shipment cycle can be predicted between 10:00 p.m. today and 8:00 a.m. tomorrow. Any data of the shipment volume and the shipment increment M+1 days before the next shipment cycle can be input into the trained shipment volume prediction model to obtain the shipment prediction volume for the next shipment cycle for a total of N days.

[0060] In one embodiment of the present invention, the shipment volume M+1 days before the next shipment cycle can also be input into the trained shipment volume forecasting model to obtain a first forecast result of the shipment forecast volume of the next shipment cycle, and the shipment increment M+1 days before the next shipment cycle can be input into the trained shipment volume forecasting model to obtain a second forecast result of the shipment forecast volume of the next shipment cycle. Then, the first forecast result and the second forecast result are averaged to obtain the final forecast result. This joint forecasting strategy can further improve the forecast accuracy.

[0061] In one embodiment of the present invention, the current inventory level of the warehouse may be the inventory level after shipments are stopped one day before the next shipment cycle and before shipments start on the first day of the next shipment cycle.

[0062] In one embodiment of the present invention, the safety stock ratio may be a preset interval [X%, Y%], where X% is the lower limit of the safety stock ratio. If the inventory is lower than X% of the total inventory of the warehouse, it is considered that there is a risk of out-of-stock; Y% is the upper limit of the safety stock ratio. If the inventory is higher than Y% of the total inventory of the warehouse, it is considered that there is a risk of overstocking. The total inventory here may be the maximum inventory that the warehouse can accommodate.

[0063] The early warning module 60 can specifically issue a replenishment warning when the difference between the current inventory of the warehouse and the forecast shipment quantity of the next shipment cycle is less than X% of the total inventory of the warehouse; and issue a clearance warning when the difference between the current inventory of the warehouse and the forecast shipment quantity of the next shipment cycle is greater than Y% of the current inventory.

[0064] For production scenarios, when the warehouse stores materials for production, it can replenish stocks in time or reduce the production rhythm in the next shipment cycle according to the replenishment warning; it can suspend replenishment or speed up the production rhythm in the next shipment cycle according to the clearance warning. For business scenarios, when the warehouse stores goods for sale, it can replenish stocks in time or tighten operations in the next shipment cycle according to the replenishment warning, such as reducing advertising investment or raising prices for goods; it can suspend replenishment or expand operations in the next shipment cycle according to the clearance warning, such as increasing advertising investment or reducing prices for goods.

[0065] If the difference between the current inventory of the warehouse and the forecasted shipment amount of the next shipment cycle is between X% and Y%, the warning module 60 does not issue a warning and can continue normal production or operation.

[0066] To summarize, according to the inventory management system of the intelligent warehousing according to the embodiment of the present invention, the historical shipment increment is derived from the historical shipment quantity, and then the shipment prediction model is trained based on these two data, and the loss function is defined based on the prediction difference between the two data, and then the shipment volume of the warehouse in the next shipment cycle is predicted according to the trained shipment prediction model. Finally, the warehouse inventory management warning is carried out based on the shipment prediction quantity, inventory quantity and safety stock ratio. In this way, the warehouse shipment volume can be accurately predicted, and the warehouse inventory management warning can be carried out according to the prediction results, which effectively reduces the risk of backlog and out-of-stock items in the warehouse.

[0067] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. "Multiple" means two or more, unless otherwise clearly and specifically defined.

[0068] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0069] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0070] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0071] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0072] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0073] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0074] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0075] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0076] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. An inventory management method for intelligent warehousing, characterized in that: The following steps are involved: Obtaining historical shipment data of the warehouse, wherein the historical shipment data includes the shipment volume of the warehouse on each day during the historical period; Acquire the shipment change history data of the warehouse according to the shipment volume history data, wherein the shipment change history data includes the shipment increment of the warehouse on each day in the historical time, and the shipment increment of each day is the difference between the shipment volume of the day and the shipment volume of the previous day; Training a shipment prediction model according to the shipment history data and the shipment change history data, wherein a loss function of the shipment prediction model is defined according to a prediction difference between the shipment history data and the shipment change history data; Predicting the shipment forecast volume of the warehouse in the next shipment cycle through the trained shipment forecast model; Obtain the current inventory and safety stock ratio of the warehouse; An inventory management warning for the warehouse is performed based on the forecasted shipment quantity of the warehouse in the next shipment cycle, the current inventory quantity and the safety inventory ratio.

2. The inventory management method for intelligent warehousing according to claim 1 is characterized in that: When training the shipment volume prediction model, the samples in the training sample set used are constructed according to the following format: Format 1: The shipment volume from the i-th to the i+M-th day in the historical time is used as the input data, and the shipment volume from the i+M+1-th to the i+M+N-th day is used as the corresponding target data; Format 2: The shipment increment from the i-th to the i+M-th day in the historical time is used as the input data, and the shipment volume from the i+M+1-th to the i+M+N-th day is used as the corresponding target data. Wherein, i is a positive integer, i is 1, 2, 3, ..., S in sequence, S=KMN, K is the total number of days in the historical time; M is a positive integer, which is a preset fixed value; N is a positive integer greater than 1, which is a preset fixed value, equal to the number of days of a delivery cycle to be predicted.

3. The inventory management method for intelligent warehousing according to claim 2 is characterized in that: Before defining the loss function of the shipment forecasting model, the target data in the sample and the forecast data obtained by the shipment forecasting model according to the input data are vectorized respectively, wherein when the shipments or shipment increments from the i-th to the i+M-th day in the historical time are used as input data, the forecast data obtained by the shipment forecasting model is the forecast shipments from the i+M+1-th to the i+M+N-th day, The loss function of the shipment volume prediction model is: ； in, T 1. T 2. P 1. P 2 are respectively: when i takes the same number, the vector of target data of samples in format one, the vector of target data of samples in format two, the vector of predicted data obtained according to the input data of samples in format one, and the vector of predicted data obtained according to the input data of samples in format two; F loss is the loss value; represents the magnitude of a vector; MSE Represents the mean square error between two vectors.

4. The inventory management method for intelligent warehousing according to any one of claims 1 to 3, characterized in that: The shipment volume prediction model is an LSTM model.

5. The inventory management method for intelligent warehousing according to claim 1 is characterized in that: According to the forecasted shipment amount of the warehouse in the next shipment cycle, the current inventory amount and the safety inventory ratio, an inventory management early warning of the warehouse is performed, specifically including: If the difference between the current inventory of the warehouse and the forecasted shipment amount of the next shipment cycle is less than X% of the total inventory of the warehouse, a replenishment warning is issued; If the difference between the current inventory of the warehouse and the forecasted shipment amount of the next shipment cycle is greater than Y% of the current inventory, a stock clearance warning is issued, where X% and Y% are the lower limit and upper limit of the safety stock ratio respectively.

6. An intelligent warehousing inventory management system, characterized in that: include: A first acquisition module, the first acquisition module is used to acquire the historical shipment data of the warehouse, wherein the historical shipment data includes the shipment volume of the warehouse on each day in the historical time; A second acquisition module, the second acquisition module is used to acquire the shipment change history data of the warehouse according to the shipment volume history data, wherein the shipment change history data includes the shipment increment of each day of the warehouse in the historical time, and the shipment increment of each day is the difference between the shipment volume of the day and the shipment volume of the previous day; A training module, the training module is used to train a shipment prediction model according to the shipment history data and the shipment change history data, wherein a loss function of the shipment prediction model is defined according to a prediction difference between the shipment history data and the shipment change history data; A prediction module, the prediction module is used to predict the shipment forecast volume of the warehouse in the next shipment cycle through the trained shipment forecast model; A third acquisition module, the third acquisition module is used to obtain the current inventory and safety stock ratio of the warehouse; An early warning module is used to provide an inventory management early warning for the warehouse based on the forecast shipment quantity of the warehouse in the next shipment cycle, the current inventory quantity and the safety inventory ratio.

7. The intelligent warehouse inventory management system according to claim 6, characterized in that: When the training module trains the shipment volume prediction model, the samples in the training sample set used are constructed according to the following format: Format 1: The shipment volume from the i-th to the i+M-th day in the historical time is used as the input data, and the shipment volume from the i+M+1-th to the i+M+N-th day is used as the corresponding target data; Format 2: The shipment increment from the i-th to the i+M-th day in the historical time is used as the input data, and the shipment volume from the i+M+1-th to the i+M+N-th day is used as the corresponding target data. Wherein, i is a positive integer, i is 1, 2, 3, ..., S in sequence, S=KMN, K is the total number of days in the historical time; M is a positive integer, which is a preset fixed value; N is a positive integer greater than 1, which is a preset fixed value, equal to the number of days of a delivery cycle to be predicted.

8. The intelligent warehouse inventory management system according to claim 7, characterized in that: Before defining the loss function of the shipment forecasting model, the target data in the sample and the forecast data obtained by the shipment forecasting model according to the input data are vectorized respectively, wherein when the shipments or shipment increments from the i-th to the i+M-th day in the historical time are used as input data, the forecast data obtained by the shipment forecasting model is the forecast shipments from the i+M+1-th to the i+M+N-th day, The loss function of the shipment volume prediction model is: ； in, T 1. T 2. P 1. P 2 are respectively: when i takes the same number, the vector of target data of samples in format one, the vector of target data of samples in format two, the vector of predicted data obtained according to the input data of samples in format one, and the vector of predicted data obtained according to the input data of samples in format two; F loss is the loss value; represents the magnitude of a vector; MSE Represents the mean square error between two vectors.

9. The intelligent warehouse inventory management system according to any one of claims 6 to 8, characterized in that: The shipment volume prediction model is an LSTM model.

10. The intelligent warehouse inventory management system according to claim 6, characterized in that: The early warning module is specifically used for: When the difference between the current inventory of the warehouse and the forecasted shipment amount of the next shipment cycle is less than X% of the current inventory, a replenishment warning is issued; When the difference between the current inventory of the warehouse and the forecasted shipment amount of the next shipment cycle is greater than Y% of the current inventory, a stock clearance warning is issued, where X% and Y% are the lower limit and upper limit of the safety stock ratio respectively.

Images