Warehouse order quantity prediction method, device, electronic device and readable storage medium

Abstract

The present invention provides a warehouse order quantity prediction method, device, electronic device and readable storage medium, which obtains the actual order quantity of multiple historical periods from the warehouse historical data; obtains the predicted fitting value of the next period according to the actual order quantity of the multiple historical periods; obtains the order quantity prediction value of the next period according to the predicted fitting value and the preset predicted strength information corresponding to the next period. On the basis of predicting the warehouse order quantity with the actual order quantity of multiple historical periods, the predicted strength information corresponding to the next period is also introduced, the influence of human control factors on the warehouse order quantity is taken into account, and the accuracy and reliability of the warehouse order quantity prediction are improved.

Description

Technical Field

[0001] The present invention relates to the field of warehousing and logistics, and in particular to a warehouse order quantity prediction method, device, electronic equipment and readable storage medium. Background Art

[0002] With the rapid development of Internet technology, more and more Internet-based services are available, such as takeout, shopping and other e-commerce services. Based on these services, users can get the goods they need without leaving home. In order to ensure the reliability of services, suppliers need to make reasonable arrangements for warehouse inventory, personnel and other resources. For example, in special circumstances such as bad weather, special holidays, and merchants launching marketing activities, the order volume of goods usually increases sharply and the order backlog is too large. Suppliers need to pre-assign more manpower and pre-store more spare goods and other resources to cope with the possible warehouse order pressure. Therefore, it is very important to predict the warehouse order volume.

[0003] The main solution at present is usually to manually observe the order supply and demand status of each warehouse, and manually estimate the warehouse order volume based on the historical order volume to arrange warehouse personnel and inventory goods. For example, if the historical order volume during the New Year of the previous year was 4 times the average of the weekday order volume, then the order volume during the New Year of this year is estimated to be 5 times the average of the weekday order volume, and the corresponding backup resources are arranged.

[0004] However, the current warehouse order quantity forecasting method relies too much on experience, and there are still problems such as a surge in warehouse orders causing orders to be too late to ship or a decrease in orders causing warehouse staff redundancy. The existing warehouse order quantity forecasting method is not very reliable. Summary of the invention

[0005] Embodiments of the present invention provide a warehouse order quantity prediction method, device, electronic device and readable storage medium to solve the problem that the order quantity cannot be accurately predicted and the reliability of the order quantity prediction is low.

[0006] According to a first aspect of the present invention, a warehouse order quantity forecasting method is provided, comprising:

[0007] Obtain the actual order quantity of multiple historical periods from the warehouse historical data;

[0008] Obtaining a predicted fitting value for the next period based on the actual order quantities of the multiple historical periods;

[0009] According to the predicted fitting value and the preset estimated strength information corresponding to the next period, the single quantity prediction value of the next period is obtained.

[0010] Optionally, in a possible implementation manner of the first aspect, obtaining a predicted fitting value for the next period according to the actual order quantities of the multiple historical periods includes:

[0011] Determining optimized fitting values ​​of the multiple historical periods according to the actual order quantities of the multiple historical periods;

[0012] The optimized fitting values ​​of the multiple historical periods are subjected to fitting processing for the next period using a preset time series prediction algorithm to obtain a predicted fitting value.

[0013] Optionally, in another possible implementation manner of the first aspect, determining the optimized fitting values ​​of the multiple historical periods according to the actual order quantities of the multiple historical periods includes:

[0014] Using a preset time series fitting algorithm, the real order quantities of the multiple historical periods are fitted for each historical period to obtain initial fitting values ​​corresponding to the multiple historical periods;

[0015] According to the initial fitting values ​​corresponding to the multiple historical periods, the optimized fitting values ​​of the multiple historical periods are determined.

[0016] Optionally, in another possible implementation manner of the first aspect, determining the optimized fitting values ​​of the multiple historical periods according to the initial fitting values ​​corresponding to the multiple historical periods includes:

[0017] Obtaining trend fitting factors corresponding to each of the historical periods;

[0018] The optimized fitting value corresponding to each of the historical periods is determined according to the initial fitting value, the trend fitting factor and the actual order quantity corresponding to each of the historical periods.

[0019] Optionally, in another possible implementation manner of the first aspect, determining the optimized fitting value corresponding to each of the historical periods according to the initial fitting value corresponding to each of the historical periods, the trend fitting factor, and the actual order quantity includes:

[0020] The product of the initial fitting value corresponding to each of the historical periods and the trend fitting factor is used as the adjusted fitting value;

[0021] Obtaining the relative error between the adjusted fitting value and the actual single quantity corresponding to each of the historical periods;

[0022] If the preset stop condition is not met, the initial fitting value and the trend fitting factor are updated, and the process of multiplying the product of the initial fitting value and the trend fitting factor corresponding to each historical period is returned to be used as the adjusted fitting value;

[0023] If the preset stop condition is met, the current initial fitting value is used as the optimized fitting value.

[0024] Optionally, in another possible implementation manner of the first aspect, the preset stop condition includes at least one of the following:

[0025] When the trend fitting factor is equal to 1, the absolute value of the relative error is less than 0.15;

[0026] When the trend fitting factor is not equal to 1, the absolute value of the relative error is less than 0.1;

[0027] The update execution times reaches a preset upper limit.

[0028] Optionally, in another possible implementation manner of the first aspect, updating the initial fitting value and the trend fitting factor includes:

[0029] If the relative error is greater than or equal to 0, the trend fitting factor is reduced by a preset step size to obtain a current trend fitting factor;

[0030] If the relative error is less than 0, the trend fitting factor is increased by a preset step size to obtain a current trend fitting factor;

[0031] For each of the historical periods, the quotient of the actual order volume and the current trend fitting factor is used as the current initial fitting value.

[0032] Optionally, in another possible implementation manner of the first aspect, the acquiring the trend fitting factor corresponding to each of the historical periods includes:

[0033] Obtaining pre-stored historical order quantity forecast values ​​and historical estimated strength information for each of the historical periods;

[0034] According to the historical single quantity forecast value and the historical predicted strength information of each of the historical periods, the historical forecast fitting value of each of the historical periods is obtained;

[0035] According to the actual order quantity of each historical period and the historical forecast fitting value, the trend fitting factor corresponding to each historical period is obtained.

[0036] Optionally, in another possible implementation of the first aspect, the preset timing prediction algorithm and the preset timing prediction algorithm both include an ETS algorithm.

[0037] Optionally, in another possible implementation manner of the first aspect, before obtaining the single quantity prediction value of the next period according to the predicted fitting value and the preset predicted strength information corresponding to the next period, the method further includes:

[0038] Acquire service attribute information corresponding to the next period, where the service attribute information corresponds to preset estimated strength information;

[0039] The estimated strength information corresponding to the service attribute information is used as the estimated strength information corresponding to the next period.

[0040] Optionally, in another possible implementation manner of the first aspect, obtaining the predicted value of the number of orders for the next period according to the predicted fitting value and preset predicted strength information corresponding to the next period includes:

[0041] The product of the predicted fitting value and the expected intensity information corresponding to the next period is used as the single quantity prediction value of the next period.

[0042] According to a second aspect of the present invention, there is provided a warehouse order quantity prediction device, comprising:

[0043] The acquisition module is used to obtain the actual order quantity of multiple historical periods from the warehouse historical data;

[0044] A processing module, used for obtaining a predicted fitting value of the next period according to the actual order quantities of the multiple historical periods;

[0045] The output module is used to obtain the single quantity prediction value of the next period according to the predicted fitting value and the preset expected strength information corresponding to the next period.

[0046] Optionally, in a possible implementation of the second aspect,

[0047] The processing module is used to determine the optimized fitting values ​​of the multiple historical periods according to the actual order quantities of the multiple historical periods; and to perform fitting processing for the next period on the optimized fitting values ​​of the multiple historical periods using a preset time series prediction algorithm to obtain predicted fitting values.

[0048] Optionally, in another possible implementation of the second aspect,

[0049] The processing module is used to perform fitting processing on the actual order quantities of the multiple historical periods in each historical period using a preset time series fitting algorithm to obtain initial fitting values ​​corresponding to the multiple historical periods; and determine the optimized fitting values ​​of the multiple historical periods based on the initial fitting values ​​corresponding to the multiple historical periods.

[0050] Optionally, in yet another possible implementation of the second aspect,

[0051] The processing module is used to obtain the trend fitting factor corresponding to each of the historical periods; and determine the optimized fitting value corresponding to each of the historical periods based on the initial fitting value corresponding to each of the historical periods, the trend fitting factor and the actual order quantity.

[0052] Optionally, in another possible implementation of the second aspect,

[0053] The processing module is used to use the product of the initial fitting value corresponding to each of the historical periods and the trend fitting factor as the adjusted fitting value; obtain the relative error between the adjusted fitting value corresponding to each of the historical periods and the actual single quantity; if the preset stop condition is not met, update the initial fitting value and the trend fitting factor, and return to execute the product of the initial fitting value corresponding to each of the historical periods and the trend fitting factor as the adjusted fitting value; if the preset stop condition is met, use the current initial fitting value as the optimized fitting value.

[0054] Optionally, in another possible implementation manner of the second aspect, the preset stop condition includes at least one of the following:

[0055] When the trend fitting factor is equal to 1, the absolute value of the relative error is less than 0.15;

[0056] When the trend fitting factor is not equal to 1, the absolute value of the relative error is less than 0.1;

[0057] The update execution times reaches a preset upper limit.

[0058] Optionally, in another possible implementation of the second aspect,

[0059] The processing module is used to reduce the trend fitting factor by a preset step size to obtain a current trend fitting factor if the relative error is greater than or equal to 0; if the relative error is less than 0, increase the trend fitting factor by a preset step size to obtain the current trend fitting factor; for each historical period, take the quotient of the actual order volume and the current trend fitting factor as the current initial fitting value.

[0060] Optionally, in another possible implementation of the second aspect,

[0061] The processing module is used to obtain the pre-stored historical order quantity forecast values ​​and historical expected strength information of each historical period; obtain the historical forecast fitting value of each historical period based on the historical order quantity forecast value and the historical expected strength information of each historical period; and obtain the trend fitting factor corresponding to each historical period based on the actual order quantity of each historical period and the historical forecast fitting value.

[0062] Optionally, in another possible implementation of the second aspect,

[0063] The output module is also used to obtain the business attribute information corresponding to the next period before obtaining the single quantity prediction value of the next period based on the predicted fitting value and the preset expected strength information corresponding to the next period, and the business attribute information corresponds to the preset expected strength information; and use the expected strength information corresponding to the business attribute information as the expected strength information corresponding to the next period.

[0064] Optionally, in another possible implementation of the second aspect,

[0065] The output module is used to take the product of the predicted fitting value and the expected intensity information corresponding to the next period as the single quantity prediction value of the next period.

[0066] According to a third aspect of the present invention, there is provided an electronic device comprising: a memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the warehouse order quantity forecasting method of the first aspect of the present invention and various possible designs of the first aspect.

[0067] According to a fourth aspect of the present invention, there is provided a readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it is used to implement the warehouse order quantity forecasting method of the first aspect of the present invention and various possible designs of the first aspect.

[0068] The present invention provides a warehouse order quantity prediction method, device, electronic device and readable storage medium, which obtains the actual order quantity of multiple historical periods from the warehouse historical data; obtains the predicted fitting value of the next period according to the actual order quantity of the multiple historical periods; obtains the order quantity prediction value of the next period according to the predicted fitting value and the preset predicted strength information corresponding to the next period. On the basis of predicting the warehouse order quantity with the actual order quantity of multiple historical periods, the predicted strength information corresponding to the next period is also introduced, the influence of human control factors on the warehouse order quantity is taken into account, and the accuracy and reliability of the warehouse order quantity prediction are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of the present invention;

[0070] Figure 2 It is a flow chart of a warehouse order quantity prediction method provided by an embodiment of the present invention;

[0071] Figure 3 is a schematic diagram of a process for determining an optimized fitting value provided by an embodiment of the present invention;

[0072] Figure 4It is a structural schematic diagram of a warehouse order quantity prediction device provided by an embodiment of the present invention;

[0073] Figure 5 It is a schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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.

[0075] The size of the serial numbers of the processes in the specification and claims of the embodiments of the present invention and the above-mentioned drawings does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0076] It should be understood that in the embodiments of the present invention, "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or apparatus.

[0077] It should be understood that in the embodiments of the present invention, "plurality" refers to two or more than two.

[0078] It should be understood that in the embodiments of the present invention, "B corresponding to A", "B corresponding to A", "A corresponds to B", or "B corresponds to A" means that B is associated with A, and B can be determined according to A. Determining B according to A does not mean that B is determined only according to A, and B can also be determined according to A and / or other information.

[0079] Depending on the context, "if" as used herein may be interpreted as "when" or "when" or "in response to determining" or "in response to detecting."

[0080] It should be understood that in the embodiment of the present invention, the historical period is a calculation period with actual order quantity, which can be understood as a calculation period for making warehouse order quantity forecasts before, such as one day, one week or one month.

[0081] It should be understood that in the embodiment of the present invention, the time series prediction algorithm is to predict and analyze the future change trend based on the time series of historical statistical data. Generally speaking, the time series is composed of four change components, such as long-term trend changes, seasonal changes, cyclical changes and random fluctuations. Simple prediction algorithms can be used to predict the first three trend changes, such as exponential smoothing algorithms, moving average algorithms, etc.

[0082] It should be understood that in the embodiment of the present invention, the ETS algorithm is a time series prediction algorithm implemented in R language. The author Rob Hyndman named the algorithm with the three letters E, T, and S, which can be understood as Error, Trend, and Seasonality, and can also be interpreted as Exponential Smoothing. The former reveals the three components of the ETS algorithm, and the latter describes the working principle of the ETS algorithm. To be precise, the ETS algorithm is actually a whole series of algorithms that can be combined arbitrarily based on these three components.

[0083] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0084] See also Figure 1 , is a schematic diagram of an application scenario provided by an embodiment of the present invention. Figure 1 In the application scenario shown, the user initiates an online order through the client 1. After receiving the orders from multiple clients 1, the server 2 sends a prompt to the warehouse staff to process the orders, and stores the received orders in the database 3 for future order quantity prediction. For example, if one day is regarded as a cycle, the order quantity received every day is regarded as the real order quantity of a historical cycle. The server 2 also presets an estimated strength information for each cycle in the database 3 to reflect the promotion strength or the purchase demand strength. The server 2 predicts the warehouse order quantity for a period of time in the future based on the real order quantity of the past few days and the future estimated strength information. Then the predicted order quantity can be provided to the warehouse management personnel to adjust the commodity replenishment quantity, personnel arrangement adjustment, system network adjustment and other processing. For example, before the arrival of the shopping festival, the warehouse order quantity of this year's shopping festival is accurately and reliably predicted by this embodiment, and then the corresponding preparations for the shopping festival are carried out to reduce the possibility of insufficient manpower or insufficient inventory during the shopping festival. The warehouse in this embodiment can be understood as a warehouse for physical services, or as a backup resource management system in a virtual service. In this embodiment, the server 2 executes the warehouse order quantity prediction method in the following various embodiments, thereby improving the accuracy and reliability of the warehouse order quantity prediction.

[0085] See also Figure 2 , is a flow chart of a warehouse order quantity prediction method provided by an embodiment of the present invention, Figure 2 The execution subject of the method shown may be software and / or hardware devices, and the following is an example using a server as the execution subject. Figure 2 The method shown mainly includes steps S101 to S103, which are specifically as follows:

[0086] S101, obtaining the actual order quantity of multiple historical periods from the warehouse historical data.

[0087] It can be understood that when receiving warehouse orders from customers, order information is stored. It can be based on multiple historical cycles, and the pre-stored real order quantity is obtained after each cycle to predict the order quantity of the next cycle in the future. The real order quantity refers to the real number of orders. The historical cycle can be based on one day, one week, one month or other preset time periods. For example, taking one day as a cycle, before preparing to start the shopping festival promotion, the real order quantity of the last 8 weeks, that is, 56 days, is obtained as the real order quantity of multiple historical cycles for order quantity prediction.

[0088] S102, obtaining a predicted fitting value for the next period according to the actual order quantities of the multiple historical periods.

[0089] It can be understood as calculating the trend of the actual order quantity over time based on the actual order quantity of multiple historical periods to obtain a fitting curve. Each data point in the fitting curve corresponds to a time point and an order quantity fitting value. The last time point is extended one time point later according to the trend of the fitting curve to obtain the predicted fitting value of the next period.

[0090] In one implementation, the server may determine the optimized fitting values ​​of the multiple historical periods based on the actual order quantities of the multiple historical periods. For example, the actual order quantities of the multiple historical periods are first fitted for each historical period using a preset time series fitting algorithm to obtain initial fitting values ​​corresponding to the multiple historical periods. Then, based on the initial fitting values ​​corresponding to the multiple historical periods, the optimized fitting values ​​of the multiple historical periods are determined. After obtaining the optimized fitting values, the optimized fitting values ​​of the multiple historical periods are fitted for the next period using a preset time series prediction algorithm to obtain predicted fitting values. Wherein, the time series prediction algorithm and the time series prediction algorithm can both be ETS algorithms.

[0091] Specifically, the ETS algorithm can be understood as an algorithm executed by the preset ETS model in the server. The server may use the preset ETS model to perform data fitting on the real single quantities of multiple historical periods to smooth the data, thereby obtaining the initial fitting values ​​corresponding to each historical period. The server then adjusts the initial fitting value sequence and the fitting factors of each initial fitting value to obtain the adjusted optimized fitting value sequence. Finally, the obtained optimized fitting value sequence is fitted for the next period using the ETS model described above to obtain the predicted fitting value. It can be understood as adding a time point to the existing optimized fitting value sequence for predictive calculation of the fitting value. Fitting and predicting with the same ETS algorithm ensures that trend fitting and prediction are performed with the same algorithm, thereby improving the consistency and accuracy of the predicted fitting value.

[0092] S103, obtaining a predicted value of the order quantity of the next period according to the predicted fitting value and the preset estimated strength information corresponding to the next period.

[0093] It can be understood that before obtaining the predicted value of the number of orders for the next period, the estimated strength information corresponding to the next period can also be determined by the following steps, specifically:

[0094] When the server needs to predict the order volume of the next cycle, it first obtains the business attribute information corresponding to the next cycle, and the business attribute information indicates the promotion information of the next cycle. Then the expected strength information corresponding to the business attribute information is used as the expected strength information corresponding to the next cycle. For example, the next cycle includes a shopping festival, and the business attribute information includes a large-scale promotion plan. Each promotion level corresponds to an expected strength information. If the pre-planned promotion level is too high, the expected strength information is determined to be 3. For another example, because the weather forecast for the next cycle is a typhoon weather period, the business attribute information includes a plan to close the market, and the expected strength information is set to 0.1. Or, the next cycle is a normal working day, and the business attribute information has no special information, then the expected strength information is set to 1. It can also be understood that different types of business attribute information can have a one-to-one mapping correspondence with the expected strength information, such as a preset correspondence list, and then the corresponding expected strength information is queried according to the business attribute information.

[0095] After the estimated strength information is determined, a specific implementation of the above step S103 may be that the server multiplies the predicted fitting value by the estimated strength information corresponding to the next period as the order quantity forecast value for the next period. Thus, in the order quantity forecast value for the next period, the estimated strength information reflects the impact of human control factors on future order quantities.

[0096] The warehouse order quantity prediction method provided in the present embodiment obtains the actual order quantity of multiple historical periods from the warehouse historical data; obtains the predicted fitting value of the next period based on the actual order quantity of the multiple historical periods; obtains the order quantity prediction value of the next period based on the predicted fitting value and the preset predicted strength information corresponding to the next period. On the basis of predicting the warehouse order quantity with the actual order quantity of multiple historical periods, the predicted strength information corresponding to the next period is also introduced, the influence of human control factors on the warehouse order quantity is taken into account, and the accuracy and reliability of the warehouse order quantity prediction are improved.

[0097] exist Figure 2 In the illustrated embodiment, before step S102 obtains the predicted fitting value of the next period, there are multiple ways to determine the optimized fitting values ​​of the multiple historical periods based on the initial fitting values ​​corresponding to the multiple historical periods. In an optional implementation, the server may first obtain the trend fitting factor corresponding to each of the historical periods. Then, based on the initial fitting value, the trend fitting factor, and the actual order quantity corresponding to each of the historical periods, the optimized fitting value corresponding to each of the historical periods is determined. Specifically, see Figure 3 , is a schematic diagram of a flow chart for determining an optimized fitting value provided by an embodiment of the present invention, Figure 3 The method shown can achieve the determination of the optimized fitting value through the following steps S201 to S207.

[0098] S201, obtaining the pre-stored historical order quantity forecast value and historical estimated strength information of each of the historical periods.

[0099] The historical order quantity forecast value can be understood as the order quantity forecast value of the historical period. Each time the order quantity forecast value is obtained, it is stored for subsequent forecasting. The historical forecast strength information can be determined based on the business attribute information of each historical period, or it can be directly read from the historical stored data.

[0100] S202: Obtain the historical forecast fitting value of each of the historical periods according to the historical single quantity forecast value and the historical expected strength information of each of the historical periods.

[0101] It can be understood that the historical forecast fitting value can be the quotient of the historical single volume forecast value and the historical expected strength information. The obtained historical forecast fitting value is a sequence corresponding to each historical period.

[0102] S203, obtaining a trend fitting factor corresponding to each of the historical periods according to the actual order quantity of each of the historical periods and the historical forecast fitting value.

[0103] It can be understood that the trend fitting factor can be the quotient of the actual single quantity of the historical period and the historical forecast fitting value. Each historical period corresponds to a trend fitting factor, thereby obtaining a sequence of trend fitting factors corresponding to each of the historical periods.

[0104] S204: taking the product of the initial fitting value corresponding to each of the historical periods and the trend fitting factor as the adjusted fitting value.

[0105] It can be understood that in order to improve the matching degree between the optimized fitting value and the data trend, so as to make the multiple optimized fitting values ​​used to reflect the historical period order volume smoother, this embodiment performs numerical optimization by cyclically adjusting the initial fitting value and the trend fitting factor. The product of the initial fitting value and the trend fitting factor obtains the adjusted fitting value that better matches the data change trend.

[0106] S205, obtaining the relative error between the adjusted fitting value corresponding to each of the historical periods and the actual single quantity.

[0107] Among them, the relative error corresponding to each historical period can be understood as the percentage of the ratio of the difference between the adjusted fitting value and the actual order quantity to the actual order quantity.

[0108] S206, if the preset stop condition is not met, the initial fitting value and the trend fitting factor are updated, and the process of returning to execute the product of the initial fitting value corresponding to each historical period and the trend fitting factor is used as the adjusted fitting value.

[0109] The specific method of updating the trend fitting factor may be: if the relative error is greater than or equal to 0, indicating that the current initial fitting value is equal to or greater than the true value, then the trend fitting factor is reduced by a preset step size to obtain the current trend fitting factor; if the relative error is less than 0, indicating that the current initial fitting value is less than the true value, then the trend fitting factor is increased by a preset step size to obtain the current trend fitting factor. The preset step size may be, for example, 0.01.

[0110] The specific method of updating the initial fitting value in the loop may be: for each of the historical periods, the quotient of the actual order volume and the current trend fitting factor is used as the current initial fitting value, so that the initial fitting value can be used in the next loop.

[0111] The preset stop condition may be at least one of the following:

[0112] When the trend fitting factor is equal to 1, the absolute value of the relative error is less than 0.15;

[0113] When the trend fitting factor is not equal to 1, the absolute value of the relative error is less than 0.1;

[0114] The update execution times reaches a preset upper limit.

[0115] Among them, limiting the number of executions can prevent overfitting.

[0116] S207: If the preset stop condition is met, the current initial fitting value is used as the optimized fitting value.

[0117] This embodiment achieves gradual smooth optimization of the initial fitting value by cyclically adjusting the initial fitting value and the trend fitting factor, and obtains each optimized fitting value with a changing trend close to the true value, thereby improving the accuracy of the single quantity prediction value.

[0118] Table 1 is an example of warehouse order quantity prediction process data provided by an embodiment of the present invention. In the example shown in Table 1 below, various historical data of the warehouse are pre-stored in the database, and the server obtains the pre-stored historical order quantity prediction value 600 and the historical predicted strength information 1.2 from the database.

[0119] Table 1 only shows the historical order volume forecast value and historical expected strength information for one day as an example, but in reality, the historical data of the previous 56 days should be obtained: 56 historical order volume forecast values, 56 historical expected strength information, and 56 real orders. The same calculation is performed on each historical data to obtain 56 trend fitting factors, and the 56 real orders are input into the ETS model to obtain the initial fitting value of the next cycle. If the next cycle is one day, only one initial fitting value of 570 is obtained; if the next cycle is 7 days, 7 initial fitting values ​​can be obtained, and the initial fitting value of the first day of the next cycle is 570.

[0120] Specifically, this embodiment obtains a historical forecast fitting value of 500 by taking the quotient of 600 and 1.2 for the historical order volume forecast value 600 and the historical estimated strength information 1.2 for the example date. The actual order volume 700 in the historical period of the example date is obtained, and the historical trend fitting factor 1.4 of the example date is obtained by taking the quotient of the actual order volume 700 and the historical forecast fitting value 500. The 56 actual order volumes corresponding to the previous 56 days are input into the preset ETS model for data fitting to obtain an initial fitting value or the first initial fitting value 570 of the next period. The adjusted fitting value 798 is then obtained from each initial fitting value and the corresponding trend fitting factor. The relative error of 0.14 between the adjusted fitting value 798 and the actual order volume is obtained, which does not meet the stop condition, so the initial fitting value and the trend fitting factor are adjusted in a cycle until the initial fitting value and the trend fitting factor that meet the stop condition are obtained, and then the adjustment is stopped. For example, if the initial fitting value of 570 is adjusted three times to obtain 503.6, which meets the preset stop condition, 503.6 will be used as the final optimized fitting value and input into the ETS model to perform order quantity forecasting for the next period to obtain the order quantity forecast value.

[0121] Table 1

[0122]

[0123]

[0124] See also Figure 4 , is a schematic diagram of the structure of a warehouse order quantity prediction device provided by an embodiment of the present invention, Figure 4 The warehouse order quantity prediction device 40 shown mainly includes:

[0125] The acquisition module 41 is used to obtain the actual order quantity of multiple historical periods from the warehouse historical data;

[0126] A processing module 42 is used to obtain a predicted fitting value of the next period according to the actual order quantities of the multiple historical periods;

[0127] The output module 43 is used to obtain the single quantity prediction value of the next period according to the predicted fitting value and the preset expected strength information corresponding to the next period.

[0128] Figure 4 The warehouse order quantity prediction device 40 of the embodiment shown can be used to perform Figure 2 The implementation principles and technical effects of the steps in the method embodiment shown are similar and will not be repeated here.

[0129] Optionally, the processing module 42 is used to determine the optimized fitting values ​​of the multiple historical periods based on the actual order quantities of the multiple historical periods; and to perform fitting processing for the next period on the optimized fitting values ​​of the multiple historical periods using a preset timing prediction algorithm to obtain predicted fitting values.

[0130] Optionally, the processing module 42 is used to perform fitting processing on the actual order quantities of the multiple historical periods for each historical period using a preset timing fitting algorithm to obtain initial fitting values ​​corresponding to the multiple historical periods; and determine optimized fitting values ​​of the multiple historical periods based on the initial fitting values ​​corresponding to the multiple historical periods.

[0131] Optionally, the processing module 42 is used to obtain the trend fitting factor corresponding to each of the historical periods; and determine the optimized fitting value corresponding to each of the historical periods based on the initial fitting value corresponding to each of the historical periods, the trend fitting factor and the actual order quantity.

[0132] Optionally, the processing module 42 is used to use the product of the initial fitting value corresponding to each of the historical periods and the trend fitting factor as the adjusted fitting value; obtain the relative error between the adjusted fitting value corresponding to each of the historical periods and the actual single quantity; if the preset stop condition is not met, update the initial fitting value and the trend fitting factor, and return to execute the product of the initial fitting value corresponding to each of the historical periods and the trend fitting factor as the adjusted fitting value; if the preset stop condition is met, use the current initial fitting value as the optimized fitting value.

[0133] Optionally, the preset stop condition includes at least one of the following:

[0134] When the trend fitting factor is equal to 1, the absolute value of the relative error is less than 0.15;

[0135] When the trend fitting factor is not equal to 1, the absolute value of the relative error is less than 0.1;

[0136] The update execution times reaches a preset upper limit.

[0137] Optionally, the processing module 42 is used to reduce the trend fitting factor by a preset step size to obtain a current trend fitting factor if the relative error is greater than or equal to 0; if the relative error is less than 0, increase the trend fitting factor by a preset step size to obtain a current trend fitting factor; for each historical period, take the quotient of the actual order quantity and the current trend fitting factor as the current initial fitting value.

[0138] Optionally, the processing module 42 is used to obtain pre-stored historical order quantity forecast values ​​and historical expected strength information of each of the historical periods; obtain the historical forecast fitting values ​​of each of the historical periods based on the historical order quantity forecast values ​​and the historical expected strength information of each of the historical periods; and obtain the trend fitting factor corresponding to each of the historical periods based on the actual order quantity of each of the historical periods and the historical forecast fitting values.

[0139] Optionally, the output module 43 is further used to obtain business attribute information corresponding to the next cycle before obtaining the single quantity forecast value of the next cycle based on the predicted fitting value and the preset expected strength information corresponding to the next cycle, the business attribute information indicating the promotion information of the next cycle, and the business attribute information corresponding to the preset expected strength information; and use the expected strength information corresponding to the business attribute information as the expected strength information corresponding to the next cycle.

[0140] Optionally, the output module 43 is used to use the product of the predicted fitting value and the expected intensity information corresponding to the next period as the single quantity prediction value of the next period.

[0141] The warehouse order quantity prediction device provided in this embodiment is used to obtain the actual order quantity of multiple historical periods from the warehouse historical data through the acquisition module; the processing module is used to obtain the predicted fitting value of the next period according to the actual order quantity of the multiple historical periods; the output module is used to obtain the order quantity prediction value of the next period according to the predicted fitting value and the preset expected strength information corresponding to the next period. On the basis of predicting the warehouse order quantity with the actual order quantity of multiple historical periods, the expected strength information corresponding to the next period is also introduced, the influence of human control factors on the warehouse order quantity is taken into account, and the accuracy and reliability of the warehouse order quantity prediction are improved.

[0142] See also Figure 5 , is a schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention, the electronic device 50 includes: a processor 51, a memory 52 and a computer program; wherein

[0143] The memory 52 is used to store the computer program, and the memory may also be a flash memory. The computer program is, for example, an application program, a functional module, etc. for implementing the above method.

[0144] The processor 51 is used to execute the computer program stored in the memory to implement the various steps performed by the electronic device in the above method. For details, please refer to the relevant description in the above method embodiment.

[0145] Optionally, the memory 52 may be independent or integrated with the processor 51 .

[0146] When the memory 52 is a device independent of the processor 51, the electronic device may further include:

[0147] The bus 53 is used to connect the memory 52 and the processor 51 .

[0148] The present invention also provides a readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it is used to implement the methods provided by the various embodiments described above.

[0149] Among them, the readable storage medium can be a computer storage medium or a communication medium. The communication medium includes any medium that facilitates the transmission of a computer program from one place to another. The computer storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, a readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application-specific integrated circuit (Application Specific Integrated Circuits, abbreviated as: ASIC). In addition, the ASIC can be located in a user device. Of course, the processor and the readable storage medium can also exist in a communication device as discrete components. The readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0150] The present invention also provides a program product, which includes an execution instruction, which is stored in a readable storage medium. At least one processor of a device can read the execution instruction from the readable storage medium, and at least one processor executes the execution instruction so that the device implements the methods provided in the above various embodiments.

[0151] In the above-mentioned embodiment of the electronic device, it should be understood that the processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention can be directly embodied as being executed by a hardware processor, or can be executed by a combination of hardware and software modules in the processor.

[0152] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A warehouse order quantity prediction method, characterized in that: include: Obtain the actual order quantity of multiple historical periods from the warehouse historical data; Using a preset time series fitting algorithm, the real order quantities of the multiple historical periods are fitted for each historical period to obtain initial fitting values ​​corresponding to the multiple historical periods; According to the historical single quantity forecast value and the historical predicted strength information of each of the historical periods, the historical forecast fitting value of each of the historical periods is obtained; Obtaining a trend fitting factor corresponding to each of the historical periods according to the actual order quantity of each of the historical periods and the historical forecast fitting value; The product of the initial fitting value and the trend fitting factor corresponding to multiple historical periods is used as the adjusted fitting value; Obtaining the relative error between the adjusted fitting value and the actual single quantity corresponding to each of the historical periods; If the preset stop condition is not met, the initial fitting value and the trend fitting factor are updated, and the process returns to execute the product of the initial fitting value and the trend fitting factor corresponding to each of the historical periods as the adjusted fitting value; If the preset stop condition is met, the current initial fitting value is used as the optimized fitting value; Using the preset time series prediction algorithm, the optimized fitting values ​​of the multiple historical periods are subjected to fitting processing for the next period to obtain the predicted fitting value for the next period; According to the predicted fitting value and the preset estimated strength information corresponding to the next period, the single quantity prediction value of the next period is obtained.

2. The method according to claim 1, characterized in that The preset stop condition includes at least one of the following: When the trend fitting factor is equal to 1, the absolute value of the relative error is less than 0.15; When the trend fitting factor is not equal to 1, the absolute value of the relative error is less than 0.1; The update execution times reaches a preset upper limit.

3. The method according to claim 1, characterized in that The updating of the initial fitting value and the trend fitting factor comprises: If the relative error is greater than or equal to 0, the trend fitting factor is reduced by a preset step size to obtain a current trend fitting factor; If the relative error is less than 0, the trend fitting factor is increased by a preset step size to obtain a current trend fitting factor; For each of the historical periods, the quotient of the actual order volume and the current trend fitting factor is used as the current initial fitting value.

4. The method according to claim 1, characterized in that: Before obtaining the historical forecast fitting value of each historical period according to the historical single quantity forecast value and the historical expected strength information of each historical period, the method includes: Obtain the pre-stored historical order quantity forecast values ​​and historical estimated strength information for each of the historical periods.

5. The method according to claim 1, characterized in that The preset time series fitting algorithm and the preset time series prediction algorithm both include the ETS algorithm.

6. The method according to claim 1, characterized in that Before obtaining the predicted value of the order quantity of the next period according to the predicted fitting value and the preset predicted strength information corresponding to the next period, the method further includes: Acquire service attribute information corresponding to the next period, where the service attribute information corresponds to preset estimated strength information; The estimated strength information corresponding to the service attribute information is used as the estimated strength information corresponding to the next period.

7. The method according to any one of claims 1 to 6, characterized in that: The obtaining, according to the predicted fitting value and the preset predicted strength information corresponding to the next period, a predicted value of the order quantity of the next period includes: The product of the predicted fitting value and the expected intensity information corresponding to the next period is used as the single quantity prediction value of the next period.

8. A warehouse order quantity prediction device, characterized in that: include: The acquisition module is used to obtain the actual order quantity of multiple historical periods from the warehouse historical data; A processing module, used for obtaining a predicted fitting value of the next period according to the actual order quantities of the multiple historical periods; An output module, used to obtain a predicted value of the order quantity of the next period according to the predicted fitting value and preset predicted strength information corresponding to the next period; The processing module is specifically used to: perform fitting processing of each historical period on the real order quantity of the multiple historical periods using a preset time series fitting algorithm to obtain initial fitting values ​​corresponding to the multiple historical periods; According to the historical single quantity forecast value and the historical predicted strength information of each of the historical periods, the historical forecast fitting value of each of the historical periods is obtained; Obtaining a trend fitting factor corresponding to each of the historical periods according to the actual order quantity of each of the historical periods and the historical forecast fitting value; The product of the initial fitting value and the trend fitting factor corresponding to multiple historical periods is used as the adjusted fitting value; Obtaining the relative error between the adjusted fitting value and the actual single quantity corresponding to each of the historical periods; If the preset stop condition is not met, the initial fitting value and the trend fitting factor are updated, and the process returns to execute the product of the initial fitting value and the trend fitting factor corresponding to each of the historical periods as the adjusted fitting value; If the preset stop condition is met, the current initial fitting value is used as the optimized fitting value; The preset time series prediction algorithm is used to perform fitting processing for the next cycle on the optimized fitting values ​​of the multiple historical cycles to obtain the predicted fitting values ​​for the next cycle.

9. An electronic device, characterized in that: include: A memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the warehouse order quantity prediction method according to any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, it is used to implement the warehouse order quantity prediction method described in any one of claims 1 to 7.

Images