Data processing method and device for determining return quantity based on inventory simulation

Abstract

The present invention discloses a data processing method and device for determining a return quantity based on inventory simulation, the method comprising obtaining data information of a target product by using a preset data mining engine; calling a preset inventory simulation engine to simulate the inventory of the target product through a preset inventory simulation model based on the data information and a specified return quantity to obtain an inventory simulation result; calculating the total cost of the target product through a preset cost calculation model based on the inventory simulation result and the data information; and calculating the future cost rate of the target product under a specified return quantity through a preset cost rate calculation model according to the total cost and inventory amount of the target product, and determining the optimal return quantity through an optimal return quantity calculation model based on an acceptable cost ratio threshold of the gross profit margin and the future cost rate.

Description

Technical Field

[0001] The present invention relates to the fields of big data analysis and warehousing logistics, and in particular to a data processing method, device, electronic equipment and medium for determining a return quantity based on inventory simulation. Background Art

[0002] Against the backdrop of increasingly mature e-commerce development, major e-commerce platforms are gradually shifting from an expansionary business model that competes in the GMV (Gross Merchandise Volume, total transaction amount within a certain period of time) field to a refined management model based on profit and loss measurement. In order to achieve refined management, it is necessary to effectively control inventory costs.

[0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0004] The existing technology mainly determines whether the product is unsalable by the unsalable turnover days threshold and / or the inventory age threshold, and then determines the unsalable inventory. However, the existing means of clearing unsalable products use the above-mentioned indirect indicators to manage inventory, which cannot directly measure the cost savings brought by the return volume, and cannot measure the changes in the cost structure after the clearance suggestion. Moreover, the turnover, which is one of the criteria for judging unsalable products, is a point-in-time calculation value, which is easily affected by factors such as recent replenishment, stockpiling, and promotions. Furthermore, the rationality of the threshold used in the calculation process of the existing unsalable inventory cannot be judged; in view of the difficulty in directly finding the "reasonable turnover days" of a product, the empirical value of 60 is used as the turnover unsalable turnover days threshold (the empirical value of the inventory age threshold is 90 days), but it is impossible to judge whether the above empirical value is correct. In addition, the existing technology does not optimize the unsalable inventory clearance plan based on the principle of cost optimization. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a data processing method, device, electronic device and medium for determining the return quantity based on inventory simulation, which can quantify the holding cost under specific inventory structure changes (returning a specific quantity under certain return nodes) by simulating future inventory, and optimize the optimal clearance strategy under the constraint of the returnable quantity, so as to find the inventory clearance plan with the best profit and loss, and guide intelligent inventory management.

[0006] To achieve the above-mentioned purpose, according to one aspect of an embodiment of the present invention, there is provided a data processing method for determining a return quantity based on inventory simulation, the method comprising: obtaining data information of a target product by using a preset data mining engine; calling a preset inventory simulation engine to simulate the inventory of the target product through a preset inventory simulation model based on the data information and a specified return quantity, thereby obtaining an inventory simulation result; calculating the total cost of the target product through a preset cost calculation model based on the inventory simulation result and the data information; and calculating the future cost rate of the target product under the specified return quantity through a preset cost rate calculation model based on the total cost and inventory amount of the target product, and determining the optimal return quantity through an optimal return quantity calculation model based on an acceptable cost ratio threshold of the gross profit margin and the future cost rate.

[0007] Further, simulating the inventory of the target product by using the preset inventory simulation model includes: determining a simulation scenario based on the commodity information included in the data information, and when determining that the simulation scenario is a conventional scenario other than the capital occupation scenario and the provision calculation scenario, performing conventional scenario simulation using a conventional scenario simulation sub-model, wherein:

[0008] Using the conventional scenario simulation sub-model to perform the conventional scenario simulation includes:

[0009] Obtain the designated return quantity and the return node corresponding to the designated return quantity, and calculate the proportion of the batch inventory to the total inventory according to the inventory information in the data information, and proportionally allocate the average daily sales in the data information to obtain the batch average daily sales of each batch; consume the inventory of each batch with the batch average daily sales until the return node, and calculate the beginning inventory of each batch at the return node: proportionally allocate the designated return quantity to obtain the return quantity of each batch: deduct the return quantity of each batch from the beginning inventory and obtain the refreshed inventory; and repeatedly execute the conventional scenario simulation with the refreshed inventory until the inventory consumption is completed or the threshold is truncated.

[0010] Furthermore, using the conventional scenario simulation sub-model to perform the conventional scenario simulation also includes: generating a data container for a natural month, the data container adopts a data structure of a two-layer nested dictionary, wherein the key of the data structure starts from the beginning of the month of the simulation start date to the beginning of the month of the truncation threshold, and the first layer value of the data structure is the remaining simulation days of the month, the actual number of days in the month, the monthly cumulative inventory, the monthly cumulative sales, and the remaining end-of-month inventory, and the inventory simulation results under the conventional scenario simulation are filled into the data container and recorded in the database as the first data set, and

[0011] Calculating the total cost using the cost calculation model includes calculating the regular storage fee and the overdue storage fee included in the total cost based on the first data set and the data information.

[0012] Furthermore, simulating the inventory of the target product by using the preset inventory simulation model further includes: when it is determined that the simulation scenario is the capital occupation scenario, using the capital occupation scenario simulation sub-model to perform capital occupation scenario simulation, wherein:

[0013] The capital occupation scenario simulation using the capital occupation scenario simulation sub-model includes: matching the account period of the target product, and determining the calculation date and the simulation start date based on the account period and the first warehousing time of each batch, and then performing the conventional scenario simulation using the conventional scenario simulation sub-model, and filling the inventory simulation result under the capital occupation scenario simulation into the data container, and recording it in the database as a second data set, and

[0014] Calculating the total cost using the cost calculation model includes calculating the capital occupation cost included in the total cost based on the second data set and the data information.

[0015] Furthermore, simulating the inventory of the target product by using the preset inventory simulation model further includes: when it is determined that the simulation scenario is the provision calculation scenario, using the provision calculation scenario simulation sub-model to perform provision calculation scenario simulation, wherein:

[0016] Using the provision calculation scenario simulation sub-model to perform the provision calculation scenario simulation includes: using the conventional scenario simulation sub-model to perform the conventional scenario simulation to calculate the end-of-period inventory of each batch as of the simulation settlement date, and filling the end-of-period inventory into the data container and recording it in the database as a third data set.

[0017] Further, calculating the total cost using the cost calculation model includes calculating the inventory impairment provision cost included in the total cost based on the third data set and the data information, and

[0018] Calculating the future cost rate by the cost rate calculation model includes dividing the total cost by the inventory amount to obtain the future cost rate under the specified return quantity.

[0019] Further, determining the optimal return quantity by using the optimal return quantity calculation model and using binary search to determine the optimal return quantity includes:

[0020] Calculate the future cost rate of zero return for the natural consumption scenario with no returns, and compare the future cost rate of zero return with the rate cap; when the future cost rate of zero return is less than or equal to the rate cap, the optimal return quantity is 0; and when the future cost rate of zero return is greater than the rate cap, calculate the future cost rate of full return, and compare the future cost rate of full return with the rate cap; when the future cost rate of full return is greater than or equal to the rate cap, the optimal return quantity is 100%; and when the future cost rate of full return is less than the rate cap, calculate the future cost rate of 50% return, and compare the size relationship of the future cost rate of 50% return with the rate cap and the rate cap; repeat the binary search until the future cost rate corresponding to the searched return quantity is between the rate cap and the rate cap or the set maximum number of searches is reached.

[0021] According to another aspect of an embodiment of the present invention, there is provided a data processing device for determining a return quantity based on inventory simulation, the device comprising:

[0022] A data preprocessing module, which is used to obtain data information of target goods;

[0023] An inventory simulation module, the inventory simulation module is used to simulate the inventory of the target product based on the data information and the specified return quantity to obtain an inventory simulation result;

[0024] a cost calculation module, the cost calculation module being used to calculate the total cost of the target product based on the inventory simulation result and the data information; and

[0025] A profit and loss optimization module is used to calculate the future cost rate of the target product under the specified return quantity based on the total cost and inventory amount of the target product, and determine the optimal return quantity based on an acceptable cost ratio threshold of the gross profit margin and the future cost rate.

[0026] According to yet another aspect of an embodiment of the present invention, there is provided an electronic device for determining a return quantity based on inventory simulation, comprising:

[0027] one or more processors;

[0028] a storage device for storing one or more programs,

[0029] When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in one aspect above.

[0030] According to another aspect of an embodiment of the present invention, a computer-readable medium is provided, on which a computer program is stored. When the program is executed by a processor, the method described in the above aspect is implemented.

[0031] The embodiments of the above invention have the following advantages or beneficial effects: because the technical means described above are adopted, the technical problems such as the inability of existing clearing means to directly measure the cost savings brought by the return volume, the inability to measure the change in cost structure after the clearance suggestion, the turnover being easily affected, the inability to judge the rationality of the threshold and the lack of optimization of the clearing plan based on the principle of cost optimization are overcome, thereby achieving the technical effects of quantitatively calculating the holding cost, optimizing the optimal clearance strategy under the constraint of the return volume, finding the inventory clearance plan with the best profit and loss, and guiding intelligent inventory management.

[0032] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0034] Figure 1 is a schematic diagram of the main process of a data processing method for determining a return quantity based on inventory simulation according to an embodiment of the present invention;

[0035] Figure 2 is a schematic diagram for explaining the inventory simulation step of the data processing method for determining the return quantity based on inventory simulation according to an embodiment of the present invention;

[0036] Figure 3 is a schematic flow chart of an inventory simulation step of a data processing method for determining a return quantity based on inventory simulation according to an embodiment of the present invention;

[0037] Figure 4 is a schematic flow chart of the profit and loss optimization step of the data processing method for determining the return quantity based on inventory simulation according to an embodiment of the present invention;

[0038] Figure 5 is a schematic diagram of main modules of a data processing device for determining a return quantity based on inventory simulation according to an embodiment of the present invention;

[0039] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0040] Figure 7 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0042] Before describing the present invention, the specific scheme of the existing clearing technology for improvement is first described:

[0043] Solution 1: Determine whether the product is unsalable by inventory turnover at a certain point in time. SKUs (Stock Keeping Units) that exceed the unsalable turnover days threshold (the threshold is set manually) are considered unsalable, and the unsalable sales volume is the inventory volume that exceeds the unsalable turnover.

[0044] Solution 2: Determine whether it is unsalable by inventory age. SKUs that exceed the inventory age threshold (the threshold is set manually) are determined to be unsalable, and the unsalable volume is the total amount of batch inventory that exceeds the inventory age threshold.

[0045] Solution 3: Combine turnover and inventory age to jointly determine whether it is unsalable, that is, first determine whether it exceeds the unsalable turnover threshold, and then determine whether it exceeds the inventory age threshold. If both are determined as "yes", the SKU is determined to be unsalable. In the calculation of unsalable sales, first calculate the unsalable inventory through unsalable turnover, then determine the long-aged inventory through inventory age, and correct the unsalable inventory according to the long-aged inventory to obtain the final unsalable inventory.

[0046] Therefore, the existing clearing technology has many defects as mentioned above, such as the inability to directly measure the cost savings brought by the return volume, the inability to measure the change in cost structure after the clearance suggestion, the inability to determine a reasonable threshold, and the inability to optimize the clearing based on the principle of cost optimization.

[0047] In view of the above-mentioned defects, the present invention proposes a data processing method, device, electronic device and medium for determining the return quantity based on inventory simulation, which will be described in detail below in conjunction with the accompanying drawings.

[0048] The data processing method for determining the return quantity based on inventory simulation according to the present invention comprises a data preprocessing step, an inventory simulation step, a cost calculation step and a profit and loss optimization step.

[0049] The following will combine Figures 1 to 4 A data processing method for determining a return quantity based on inventory simulation according to an embodiment of the present invention is described.

[0050] Figure 1is a schematic diagram of the main process of a data processing method for determining a return quantity based on inventory simulation according to an embodiment of the present invention; Figure 2 is a schematic diagram for explaining the inventory simulation step of the data processing method for determining the return quantity based on inventory simulation according to an embodiment of the present invention; Figure 3 is a schematic flow chart of inventory simulation steps of a data processing method for determining return quantity based on inventory simulation according to an embodiment of the present invention; Figure 4 It is a schematic flow chart of the profit and loss optimization step of the data processing method for determining the return quantity based on inventory simulation according to an embodiment of the present invention.

[0051] like Figure 1 As shown, the data processing method for determining the return quantity based on inventory simulation according to an embodiment of the present invention includes S101, a data preprocessing step; S102, an inventory simulation step; S103, a cost calculation step and S104, a profit and loss optimization step. The above steps will be described in detail below.

[0052] Step S101, data preprocessing step

[0053] In the data preprocessing step, a preset data mining engine is used to obtain data information, which includes product information, inventory information, sales information and financial information.

[0054] Among them, product information includes basic product information, product lines (small, medium and large items), national average warehouse quotations, and whether it is purchased globally.

[0055] Inventory information includes logical batch (hereinafter referred to as "batch") inventory, regional dimension month-to-date cumulative inventory, and returnable quantity.

[0056] Sales volume includes future average daily sales forecast and cumulative sales volume from month to date in regional dimensions.

[0057] Financials include settlement periods, gross margins, and accrual parameters.

[0058] The data mining engine can obtain various information and parameters included in the above data information from the system database. The above individual parameters will be described in detail below.

[0059] Specifically, in order to carry out targeted inventory management and unsalable clearance, the target goods need to be classified first, based on the product line of the target goods (small and medium-sized items, large items), special product tags (for example, global shopping, gifts, virtual items, etc.), etc. The target goods are classified based on whether they are small and medium-sized items or large items, because the factors affecting the storage fees are different depending on whether the production line of the target goods is small and medium-sized items or large items. Specifically, for small and medium-sized items, the storage fee is mainly affected by their inventory volume and the discount rate of the storage area; for large items, the storage fee is also affected by the type of the goods, for example, the storage of special items and oversized items will be charged higher fees. On the other hand, the target goods are classified based on whether they are special goods, because special goods will have charging standards independent of other goods, for example, the corresponding provision coefficient for global shopping goods will be higher than that of conventional small and medium-sized goods.

[0060] In addition, in order to measure the cash flow corresponding to the inventory of goods, it is also necessary to obtain the average purchase price of the goods. Here, the average of the national warehouse quotations of the goods is used as an estimate of the average purchase price. Large-scale purchases of high-value goods will occupy more cash flow, thus generating more capital occupation costs.

[0061] The above data are all read through the distributed file system (HDFS), and the data warehouse tool (HIVE) is used for preliminary data cleaning and matching.

[0062] In addition, it is also necessary to obtain the current inventory of the target goods, and since the inventory is entered into the warehouse according to logical batches, it is also necessary to split the current inventory of the target goods according to the batches of the goods and obtain the inventory corresponding to each batch, so as to infer the inventory age of the remaining inventory of the batch through the first entry time of the logical batch. The longer the inventory age, the greater the provision coefficient of the goods, and thus more depreciation provision fees will be charged. At the same time, since the current calculation time is usually not the beginning of the month, and the calculation of overdue storage fees is based on the monthly cumulative inventory, the cumulative inventory of the first month needs to be filled for the cumulative inventory from the historical month to date. Inventory data is mainly obtained through the ERP (Enterprise Resource Planning) system, accessing the MYSQL (My Structured Query Language) database or HDFS, selecting the inventory data of the target goods, and processing and storing them.

[0063] After obtaining the inventory information of the goods, in order to simulate the inventory in the future, it is necessary to obtain the future inventory consumption rhythm, that is, the average daily sales. Here, the sales forecasting system interface is called, and the regional sales of the goods in the future predetermined time period (for example, 91 days) are obtained to obtain the average daily sales, which is used as an estimate of the regional inventory consumption rhythm. Then, the available inventory of the logical batches of the goods in each region can be consumed according to the average daily sales of the region, thereby obtaining a simulation of future inventory changes to achieve an estimate of the future cumulative inventory. At the same time, just like the need to obtain the cumulative inventory from the month to date, in the data preprocessing step, it is also necessary to obtain the cumulative sales from the month to date of the first month.

[0064] Step S102, inventory simulation step

[0065] Figure 2 The basic form of future inventory simulation under any return node is shown. Assuming that future sales (daily sales) are maintained at a fixed value, the future inventory is a triangular consumption pattern. If the future sales are 0 or very small, so that the inventory cannot be consumed to 0 for a long time (or infinitely long), it will be truncated at a certain set time threshold, and the triangular inventory consumption pattern will be transformed into a trapezoid (or rectangle). The truncation threshold can be set according to the specific situation, for example, the truncation threshold is set to 2 years.

[0066] Figure 2 The consumption patterns of inventory under natural consumption, single return and multiple returns are shown in the figure. The horizontal axis represents time, and its unit time (cycle) can be set to 1 day, half a day, 0.25 days, etc. The vertical axis represents the inventory (where 20 pieces, 50 pieces and 80 pieces represent the inventory of each batch). Under natural consumption, it is a triangular inventory consumption pattern. When under the set return node, the inventory is reduced by the value corresponding to the return amount, thus generating a trapezoidal inventory consumption pattern. Figure 2 Only an example of the basic trend of inventory changes is shown. The following will describe in detail the data processing method for determining the return quantity based on inventory simulation of the present invention. Figure 2 Inventory simulation steps.

[0067] In the inventory simulation step S102, for the inventory of each region of the target product, a preset inventory simulation engine is called to simulate the inventory of the target product through a preset inventory simulation model based on the data information described above and the specified return quantity, thereby obtaining an inventory simulation result. Figure 3 The simulation of the inventory of the target product by using the preset inventory simulation model specifically includes the following steps:

[0068] First, it is determined whether the simulation scenario is a capital occupation scenario or a provision calculation scenario (step S1021). If it is determined not to be a capital occupation scenario or a provision calculation scenario, a conventional scenario simulation sub-model is used to perform conventional scenario simulation.

[0069] When using the conventional scenario simulation sub-model to perform conventional scenario simulation, first generate a data container for a natural month (step S1022). For example, the data container can use a two-layer nested dictionary data structure (nested dictionary data structure), called database (data set), for example, {'2020-01-01': {simulation remaining days: 31, month real days: 31, month cumulative inventory: 0, month cumulative sales: 0, month end remaining end inventory: 0}}, the key of the data structure is from the beginning of the month when the simulation starts to the beginning of the month when the truncation threshold is located (when the natural consumption is completed and the truncation threshold has not been reached, then after the consumption is completed until the truncation threshold, the values ​​in the above data structure are all 0). The first layer of values ​​is the simulation remaining days of the month, the real days of the month, the month cumulative inventory, the month cumulative sales, and the month end remaining end inventory.

[0070] Then, get the return node t of the regional dimension i and each return node t i Return volume i (Calculation is performed in the regional dimension.) The average daily sales volume in the region is proportionally distributed according to the inventory of the logical batches to calculate the average daily sales volume of each logical batch in the region (step S1023).

[0071] Specifically, obtain the return node t specified by the simulation system i and each return node t i Return volume i . The average daily sales volume in the region is calculated according to the inventory of each logical batch. 0i} i∈batch Proportional allocation (ratio is ), i represents the logical batch, stock 0i represents the inventory of logical batch i in the region, 0 represents the number of returns, then the average daily sales volume of logical batch i in the region is Among them, fcst represents the average daily sales volume in the region, that is, the total average daily sales volume of each logical batch in the region.

[0072] Subsequently, the step proceeds to step S1024, where the inventory consumption is performed based on the average daily sales volume of each logical batch in the region calculated in step S1023 from the start date of the simulation, and the beginning inventory at the first return node is calculated, and the return quantity of each logical batch is allocated in proportion to the beginning inventory of the logical batches.

[0073] Specifically, from the simulation start date, the average daily sales volume of each logical batch in the region calculated in step S1023 is used to consume inventory, and the accumulated inventory information across the months is recorded in dataBase (data set). When the consumption reaches the first return node t1, the initial inventory of each logical batch is calculated. 1i} i∈batch ,

[0074] The calculation formula is Among them, Δt is the unit time from the simulation start date to the first return node t1. Then, the return quantity return1 at the first return node t1 is proportionally distributed according to the logical batch beginning inventory, and the proportion is The return quantity of each logical batch after allocation is {return 1i} i∈batch ,in:

[0075]

[0076] Then, the step proceeds to step S1025, where the return quantity of each logical batch is deducted from the initial inventory of each logical batch and the inventory is refreshed. Specifically, the refreshed inventory is:

[0077]

[0078] Then the step returns to step S1023, and the average daily sales volume of each logical batch in the future is updated according to the updated inventory The updated inventory is consumed according to the updated sales forecast (ie, average daily sales) and the accumulated inventory is recorded.

[0079] Repeat the above steps S1022 to S1025 until the inventory consumption is completed (or reaches the truncation threshold), and associate the filled dataBase with each logical batch of the product. The dataBase calculated in this scenario is recorded as dataBase1 (the first data set).

[0080] When it is determined in step S1021 that the simulation scenario is a capital occupation scenario, the step proceeds to capital occupation scenario simulation step S1026 to perform capital occupation scenario simulation using the capital occupation scenario simulation sub-model.

[0081] The capital occupation scenario simulation using the capital occupation scenario simulation sub-model includes the following steps:

[0082] First, match the settlement account period of the goods (step S1026). For each logical batch in each area, calculate the date after the first warehousing time is increased by the account period days. If the date is earlier than the calculation date, the calculation will start from the calculation date; if the date is later than the calculation date, the calculation date will be postponed to the date, that is, the simulation will start from this date.

[0083] For example, if the first entry date is the 1st and the accounting period is 3 days, the date after the first entry date plus the accounting period is the 3rd. The later day between the 3rd and the calculation date is taken as the calculation date, and the day after the calculation date is the simulation start date.

[0084] After completing the above step S1026, the step proceeds to the above-described step S1022 to step S1025 until the inventory simulation is completed. The dataBase calculated in this scenario is recorded as dataBase2 (the second data set).

[0085] When it is determined in step S1021 that the simulation scenario is a provision calculation scenario, the step enters the provision calculation simulation step S1027 to perform a provision calculation scenario simulation using the provision calculation scenario simulation sub-model.

[0086] The simulation of the provision calculation scenario using the provision calculation scenario simulation sub-model includes the following steps:

[0087] The end of the month of the calculation date is used as the simulation settlement date, and the remaining inventory of each logical batch in each region until the simulation settlement date is calculated using steps S1022 to S1025 of the conventional scenario simulation (step S1027). The dataBase in this scenario only has the month of the calculation date, and only needs to record the ending inventory (this inventory is the point-in-time inventory, not the cumulative inventory), which is recorded as dataBase3 (the third data set).

[0088] Step S103, cost calculation step

[0089] Generally, inventory costs can be roughly divided into holding costs, fulfillment costs, out-of-stock costs and other costs.

[0090] Among them, there is a great deal of uncertainty in fulfillment costs. Factors such as fulfillment routes, product characteristics, warehouse network layout, etc. may become major influencing factors, and fulfillment costs are difficult to improve and optimize through effective management methods. Out-of-stock costs are difficult to measure accurately. They are cost forecasts based on sales loss forecasts and cannot be effectively verified.

[0091] In comparison, the cost of holding goods is essentially the cost generated by the accumulation of inventory, which has the characteristics of good measurement and can be effectively controlled through clearing measures such as "returns", "promotions" and "scrap". The existing holding costs mainly include "warehousing costs", "capital occupation costs" and "depreciation costs". Among them, "warehousing costs" can also be subdivided into "regular warehouse costs" and "overdue warehouse costs" as a penalty item.

[0092] For the above holding costs, after completing the inventory simulation, the total cost of the target goods is calculated through the preset cost calculation model based on dataBase1, dataBase2, dataBase3 and the data information described above, that is, the sum of regular storage fees, overdue storage fees, capital occupation costs and inventory depreciation accrual costs. The calculation formulas for the above four costs are shown in the following table, and the calculation of the above four costs will be described in detail in combination with the following table.

[0093] Table 1 List of calculation formulas for various costs

[0094]

[0095] (1) Regular storage fees

[0096] Regular storage fees are the fees charged for storing goods in the warehouse. They are directly related to the unit price of the goods and the cumulative number of pieces stored. Different regions will have different discount rates. The specific calculation logic is as follows:

[0097] Sum up the estimated future cumulative inventory of each month in dataBase1 for each logical batch in each region of the product to obtain the estimated future total cumulative inventory of the product in each region

[0098] Then multiply it by the regional discount rate (discount), summarize the calculation results nationwide and multiply it by the national average warehouse quotation (price).

[0099] Thus, the regular storage fee for the target product is:

[0100] (See Table 1).

[0101] (2) Expired storage fee

[0102] Expired storage fees are penalty fees incurred when goods are stored in the warehouse for too long or too long. They are directly related to the national turnover, standard turnover, and accumulated inventory of goods. Different regions will have different discount rates. In addition, the overdue storage fees are charged and calculated on a monthly basis. It is estimated that future overdue storage fees need to be calculated on a monthly basis and then summarized. The specific calculation logic is as follows:

[0103] Sum up the estimated future cumulative inventory of each month in dataBase1 of each logical batch in each region of the product to each month, summarize the estimated future cumulative sales of each month in dataBase1 according to the same logic, summarize the cumulative inventory and cumulative sales of each month to the whole country, calculate the national average turnover days of the month, if the national average turnover days is less than the standard turnover days of the product, then record the national overdue storage fee of the product in the month as 0, if the national average turnover days is greater than the standard turnover days of the product, calculate as follows:

[0104] a) Calculate the national cumulative inventory of the product for the month, recorded as stock, and calculate the cumulative inventory of each region for the month, recorded as stock i ;

[0105] b) Calculate the inventory ratio of each region for the product in the current month,

[0106] c) Calculate the excess inventory of the product in the current month, specifically, the national average turnover days (ito nation ) minus the standard turnover days of goods (ito std ), the difference is multiplied by the national cumulative sales volume of the product in that month (sales nation ), recorded as overdue

[0107] d) Calculate the overdue storage cost of the goods in each region for the month. The specific calculation is as follows:

[0108] overdue*ratio i *discount i *price*2;

[0109] e) Aggregate the overdue storage fees of the goods in each region for the month to the whole country;

[0110] The overdue storage fee is calculated as follows:

[0111] (See Table 1).

[0112] Use the above calculation method to calculate the overdue storage cost of each month in dataBase1 on a rolling basis.

[0113] (3) Capital occupation cost:

[0114] The existence of capital occupation cost is due to the fact that inventory will occupy cash flow, and cash flow itself can generate income. This part of income is a cost for the self-operated centralized supply chain. The specific calculation logic is as follows:

[0115] Aggregate the monthly cumulative inventory of each logical batch in each region of the product in dataBase2 to the estimated total future cumulative inventory of the product nationwide, recorded as stock capital(stock* in the formula), multiply the inventory by the national warehouse quotation of the product (wh_qtn) and the average daily rate of return on capital occupation (interest, also known as the conversion rate) to get the estimated future capital occupation cost of the product

[0116] (See Table 1).

[0117] (4) Provision for inventory impairment

[0118] Inventory will depreciate over time. In order to make the goods flow faster, a part of the inventory depreciation provision will be increased. This part of the depreciation provision is also a cost for inventory management. The depreciation provision is directly related to the inventory age and inventory volume of the goods. The specific calculation logic is as follows:

[0119] The remaining stock at the end of the month in dataBase3 for each logical batch of each product in each region is calculated. it ), matching the depreciation coefficient (ratio) corresponding to the inventory age of each batch (coef t ), the remaining inventory amount of each batch at the end of the month (stock it *wh_qtn) multiplied by the depreciation coefficient corresponding to the batch inventory is the future depreciation cost of the product. The specific formula is:

[0120] (See Table 1).

[0121] Step S104, the profit and loss optimization step, calculates the future cost rate of the target product under the specified return quantity through the preset cost rate calculation model according to the total cost and inventory amount of the target product described above, and determines the optimal return quantity through the optimal return quantity calculation model based on the acceptable cost ratio threshold of the gross profit margin and the future cost rate. Specifically, based on the acceptable cost ratio threshold α (a∈[0,1], preferably 0.6) of the gross profit margin space, the optimal return quantity is determined using a binary search.

[0122] In the present invention, a binary search is performed using the acceptable cost ratio threshold α of the gross profit margin space input by the business. Figure 4 The profit and loss optimization algorithm is executed as shown, which is mainly used to search for a return scenario in which the return node is the current calculation time and there is a returnable quantity constraint.

[0123] First, the future cost rate of zero return for the natural consumption scenario with no returns is calculated by a preset cost rate calculation model, where future cost rate = future cost / inventory amount, where future cost is the sum of various costs calculated in the above step S103.

[0124] Then, based on the acceptable cost ratio threshold of the gross profit margin and the future cost rate, the optimal return quantity is determined through the optimal return quantity calculation model, which specifically includes the following steps:

[0125] Determine whether the future cost rate of zero return is less than or equal to the rate cap (step S1041), where the rate cap = gross profit margin * α.

[0126] Among them, when the specified return quantity is set (for example, the return quantity here is 0), four holding costs are calculated through steps S102 and 103, and the future cost is obtained by summing the four holding costs. The inventory amount is equal to the current inventory after deducting the above-mentioned specified return quantity (here is 0 return quantity) multiplied by the warehouse quotation wh_qtn, and the calculated future cost is divided by the inventory amount to obtain the zero return future cost rate.

[0127] If it is determined in step S1401 that the future cost rate of zero returns is less than or equal to the rate upper limit (S1041: yes), the optimal solution is 0, that is, the optimal return quantity is zero.

[0128] In this case, the rate corresponding to the future costs incurred without refund (zero return) is less than the rate cap, and the profit margin is acceptable.

[0129] If it is determined in step S1041 that the future cost rate of zero returns is greater than the rate cap (step S1041: No), the future cost rate of full returns under the full return scenario is calculated, and it is determined whether the future cost rate of full returns is greater than or equal to the rate cap (step S1402).

[0130] If it is determined in step S1042 that the future cost rate of full return is greater than or equal to the rate cap (step S1402: yes), it means that there is no feasible solution under the constraint of the returnable quantity, so the closest solution is used as an approximation, and the optimization is exited and the optimal solution is recorded as a full return.

[0131] The rate corresponding to the future costs incurred in the case of a full return is still higher than the rate cap, and the profit margin is below the acceptable range.

[0132] If it is determined in step S1042 that the future cost rate of full return is less than the rate cap (step S1042: No), it is determined that there is a feasible solution, and full return is not the optimal solution. At this time, the step proceeds to step S1403, in which the rate cap is set to be gross profit margin*β (β<α), preferably, for example, β can be set to be 80%*α, and the relationship between the future cost rate of 50% return and the rate cap and rate cap is determined.

[0133] If the future cost rate of 50% returns is between the upper and lower limits of the rate (including the case where the upper and lower limits are equal), then the 50% return volume is the optimal solution.

[0134] If it is determined that the future cost rate of 50% return is less than the lower limit, the future cost rate of 25% return is calculated, and the relationship between the future cost rate of 25% return and the upper and lower limits of the rate is determined (step S1044).

[0135] On the other hand, if it is determined that the 50% return future cost rate is greater than the rate cap, the 75% return future cost rate is calculated, and the relationship between the 75% return future cost rate and the rate cap and rate cap is determined (step S1045).

[0136] Repeat the above steps S1044 or S1045 until the future cost rate after the return of the return quantity searched in a certain time falls between the upper and lower limits of the rate for the first time; or the number of iterations exceeds the maximum number of searches set in advance (for example, 5 times by default in the present invention).

[0137] If there is no refund, the rate is too high, and the rate corresponding to a full refund will be lower than the upper limit of the rate. At this time, a lower limit of the rate is set, and the optimal return ratio is sought through dichotomy, so that the corresponding rate under this return ratio will fall within the range between the upper limit and the lower limit of the rate.

[0138] It should be noted that the above-mentioned use of binary search to determine the optimal return quantity is not the only implementation method of the profit and loss optimization of the present invention. For example, the optimal solution of the return quantity can also be obtained by using algorithms such as taboo search and grid search. Existing algorithms such as taboo search and grid search for obtaining the optimal solution can be applied to the profit and loss optimization of the present invention, and will not be described in detail here.

[0139] The above describes the data processing method for determining the return quantity based on inventory simulation according to the present invention, which can quantify the holding cost under specific inventory structure changes (returning specific quantities under certain return nodes) by simulating future inventory, and optimize the optimal clearance strategy under the constraint of the returnable quantity, thereby finding the inventory clearance plan with the best profit and loss, and guiding intelligent inventory management.

[0140] According to another aspect of the present invention, a data processing device 200 for determining the quantity of returned goods based on inventory simulation is provided. Figure 5 As shown, it includes a data preprocessing module 201, an inventory simulation module 202, a cost calculation module 203 and a profit and loss optimization module 204. The above modules will be described in detail below.

[0141] Data preprocessing module 201

[0142] The data preprocessing module 201 is used to obtain data information, which includes commodity information, inventory information, sales information and financial information.

[0143] Among them, product information includes basic product information, product lines (small, medium and large items), national average warehouse quotations, and whether it is purchased globally.

[0144] Inventory information includes logical batch inventory, regional month-to-date cumulative inventory, and returnable quantity.

[0145] Sales volume includes future average daily sales forecast and cumulative sales volume from month to date in regional dimensions.

[0146] Finance includes settlement period, gross profit margin and provision factor.

[0147] The various information and parameters included in the above data information can be obtained from the system database, and the above individual parameters will be described in detail below.

[0148] Specifically, in order to carry out targeted inventory management and unsalable clearance, the data preprocessing module 201 needs to first classify the target goods, which is based on the product line of the target goods (small and medium-sized items, large items), special product tags (for example, global shopping, gifts, virtual items, etc.), etc. The target goods are classified as small and medium-sized items or large items because the factors affecting the storage fees are different for the production line of the target goods, whether it is small and medium-sized items or large items. Specifically, for small and medium-sized items, the storage fee is mainly affected by its inventory and the discount rate of the storage area; while for large items, the storage fee is also affected by the type of the goods, for example, the storage of special items and oversized items will be charged higher fees. On the other hand, the target goods are classified as special goods because special goods will have charging standards independent of other goods, for example, the corresponding provision coefficient for global shopping goods will be higher than that of conventional small and medium-sized items.

[0149] In addition, in order to measure the capital flow corresponding to the inventory of goods, the data preprocessing module 201 also needs to obtain the average purchase price of the goods. Here, the average of the national warehouse quotations of the goods is used as an estimate of the average purchase price. Large-scale purchases of high-value goods will occupy more cash flow, thus generating more capital occupation costs.

[0150] The above data are all read through the distributed file system (HDFS), and the data warehouse tool (HIVE) is used for preliminary data cleaning and matching.

[0151] In addition, the data preprocessing module 201 also needs to obtain the inventory of the target goods at the current time point, and because the inventory is put into the warehouse according to logical batches, it is also necessary to split the current time point inventory of the target goods according to the batches of the goods and obtain the inventory corresponding to each batch, so as to infer the inventory age of the remaining inventory of the batch through the first entry time of the logical batch. The longer the age of the goods, the greater the provision coefficient, and thus more depreciation provision fees will be charged. At the same time, since the current calculation time point is usually not the beginning of the month, and the calculation of overdue storage fees is based on the monthly cumulative inventory, the cumulative inventory of the first month needs to be filled for the cumulative inventory from the historical month to the present. The inventory data is mainly obtained through the ERP (Enterprise Resource Planning) system, accessing the MYSQL (My Structured Query Language) database or HDFS, selecting the inventory data of the target goods and processing and storing them.

[0152] After obtaining the inventory information of the goods, in order to simulate the inventory in the future, the data preprocessing module 201 needs to obtain the future inventory consumption rhythm, that is, the average daily sales. Here, the sales forecasting system interface is called, and the regional sales of the goods in the future predetermined time period (for example, 91 days) are obtained to obtain the average daily sales, which is used as an estimate of the regional inventory consumption rhythm. Then, the available inventory of the logical batches of the goods in each region can be consumed according to the average daily sales of the region, thereby obtaining a simulation of future inventory changes to achieve an estimate of the future cumulative inventory. At the same time, just as it is necessary to obtain the cumulative inventory from the month to date, in the data preprocessing step, it is also necessary to obtain the cumulative sales from the month to date of the first month.

[0153] Inventory Simulation Module 202

[0154] Figure 2 The basic form of future inventory simulation under any return node is shown. Assuming that future sales (daily sales) are maintained at a fixed value, the future inventory is a triangular consumption pattern. If the future sales are 0 or very small, so that the inventory cannot be consumed to 0 for a long time (or infinitely long), it will be truncated at a certain set time threshold, and the triangular inventory consumption pattern will be transformed into a trapezoid (or rectangle). The truncation threshold can be set according to the specific situation, for example, the truncation threshold is set to 2 years.

[0155] Figure 2The consumption patterns of inventory under natural consumption, single return and multiple returns are shown in the figure. The horizontal axis represents time, and its unit time (cycle) can be set to 1 day, half a day, 0.25 days, etc. The vertical axis represents the inventory (where 20 pieces, 50 pieces and 80 pieces represent the inventory of each batch). Under natural consumption, it is a triangular inventory consumption pattern. When under the set return node, the inventory is reduced by the value corresponding to the return amount, thus generating a trapezoidal inventory consumption pattern. Figure 2 Only examples of basic trends in inventory changes are shown.

[0156] The inventory simulation module 202 is used to simulate the inventory of each area of ​​the target product. The operations performed by the inventory simulation module 202 are as follows:

[0157] Determine whether the simulation scenario is a capital occupation scenario or a provision calculation scenario. If it is determined not to be a capital occupation scenario or a provision calculation scenario, perform a regular scenario simulation.

[0158] When performing a conventional scenario simulation, a data container for a natural month is first generated. For example, the data container can use a two-layer nested dictionary data structure (nested dictionary data structure), called a database, for example, {'2020-01-01':{simulation remaining days: 31, month real days: 31, month cumulative inventory: 0, month cumulative sales: 0, month end remaining end inventory: 0}}, the key of the data structure is from the beginning of the month when the simulation starts to the beginning of the month when the truncation threshold is located (when the natural consumption is completed and the truncation threshold has not been reached, then all the values ​​in the above data structure after the consumption is completed until the truncation threshold are 0). The first layer of values ​​is the simulation remaining days of the month, the real days of the month, the month cumulative inventory, the month cumulative sales, and the month end remaining end inventory.

[0159] Then, get the return node t of the regional dimension i and each return node t i Return volume i (Calculated in the regional dimension.) The average daily sales volume in the region is proportionally distributed according to the inventory of the logical batches to calculate the average daily sales volume of each logical batch in the region.

[0160] Specifically, obtain the return node t specified by the simulation system (for example, the profit and loss optimization module described later) i and each return node t i Return volume i . The average daily sales volume in the region is calculated according to the inventory of each logical batch. 0i} i∈batch Proportional allocation (ratio is ), i represents the logical batch, stock0i represents the inventory of logical batch i in the region, 0 represents the number of returns, then the average daily sales volume of logical batch i in the region is Among them, fcst represents the average daily sales volume in the region, that is, the total average daily sales volume of each logical batch in the region.

[0161] Subsequently, starting from the simulation start date, the inventory consumption is performed based on the average daily sales volume of each logical batch in the calculated area, and the beginning inventory at the first return node is calculated. The return quantity of each logical batch is allocated in proportion to the beginning inventory of the logical batch.

[0162] Specifically, from the simulation start date, the average daily sales volume of each logical batch in the calculated area is used to consume inventory, and the cumulative inventory and other information across the months are recorded in dataBase. When the consumption reaches the first return node t1, the initial inventory of each logical batch is calculated. 1i} i∈batch , the calculation formula is Among them, Δt is the unit time from the simulation start date to the first return node t1. Then, the return quantity return1 at the first return node t1 is proportionally distributed according to the logical batch beginning inventory, and the proportion is The logical return quantity after allocation is {return 1i} i∈batch ,

[0163] in,

[0164] Then, the return quantity of each logical batch is deducted from the initial inventory of each logical batch and the inventory is refreshed. Specifically, the refreshed inventory is:

[0165]

[0166] Then, based on the updated inventory, refresh the average daily sales of each logical batch in the future

[0167] The updated inventory is consumed according to the updated sales forecast (ie, average daily sales) and the accumulated inventory is recorded.

[0168] Repeat the above operation until the inventory is consumed (or reaches the interception threshold), and associate the filled dataBase with each logical batch of the product. The dataBase calculated in this scenario is recorded as dataBase1.

[0169] When it is determined that the simulation scenario is a capital occupation scenario, the capital occupation scenario simulation is entered.

[0170] First, match the product account period. For each logical batch in each region, calculate the date after the first warehousing time is increased by the account period days. If this date is earlier than the calculation date, the calculation will start from the calculation date; if this date is later than the calculation date, the calculation date will be postponed to this date, that is, the simulation will start from this date.

[0171] For example, if the first entry date is the 1st and the accounting period is 3 days, the date after the first entry date plus the accounting period is the 3rd. The later day between the 3rd and the calculation date is taken as the calculation date, and the day after the calculation date is the simulation start date.

[0172] Then, the same operations as the conventional scenario simulation are performed until the inventory simulation is completed. The dataBase calculated in this scenario is recorded as dataBase2.

[0173] When it is determined that the simulation scenario is an accrual calculation scenario, the accrual calculation simulation is entered.

[0174] First, the end of the month of the calculation date is used as the simulation settlement date, and then the same operations as the conventional scenario simulation are performed to calculate the remaining inventory of each logical batch in each region until the simulation settlement date. The dataBase in this scenario only has the month of the calculation date, and only the ending inventory needs to be recorded, which is recorded as dataBase3.

[0175] Cost Calculation Module 203

[0176] Generally, inventory costs can be roughly divided into holding costs, fulfillment costs, out-of-stock costs and other costs.

[0177] Among them, there is a great deal of uncertainty in fulfillment costs. Factors such as fulfillment routes, product characteristics, warehouse network layout, etc. may become major influencing factors, and fulfillment costs are difficult to improve and optimize through effective management methods. Out-of-stock costs are difficult to measure accurately. They are cost forecasts based on sales loss forecasts and cannot be effectively verified.

[0178] In comparison, the cost of holding goods is essentially the cost generated by the accumulation of inventory, which has the characteristics of good measurement and can be effectively controlled through clearing measures such as "returns", "promotions" and "scrap". The existing holding costs mainly include "warehousing costs", "capital occupation costs" and "depreciation costs". Among them, "warehousing costs" can also be subdivided into "regular warehouse costs" and "overdue warehouse costs" as a penalty item.

[0179] For the above-mentioned holding costs, the cost calculation module 203 calculates the regular storage fee, overdue storage fee, capital occupation cost and inventory depreciation provision cost based on dataBase1, dataBase2 and dataBase3. The calculation formulas of the above-mentioned four costs are shown in Table 1. The calculation operations of the above-mentioned four costs performed by the cost calculation module 203 will be described in detail below in combination with the above-mentioned Table 1.

[0180] (1) Regular storage fees

[0181] Regular storage fees are the fees charged for storing goods in the warehouse. They are directly related to the unit price of the goods and the cumulative number of pieces stored. Different regions will have different discount rates. The specific calculation logic is as follows:

[0182] Sum up the estimated future cumulative inventory of each month in dataBase1 for each logical batch in each region of the product to obtain the estimated future total cumulative inventory of the product in each region

[0183] Then multiply it by the regional discount rate (discount), summarize the calculation results nationwide and multiply it by the national average warehouse quotation (price).

[0184] Thus, the regular storage fee for the target product is:

[0185] (See Table 1).

[0186] (2) Expired storage fee

[0187] Expired storage fees are penalty fees incurred when goods are stored in the warehouse for too long or too long. They are directly related to the national turnover, standard turnover, and accumulated inventory of goods. Different regions will have different discount rates. In addition, the overdue storage fees are charged and calculated on a monthly basis. It is estimated that future overdue storage fees need to be calculated on a monthly basis and then summarized. The specific calculation logic is as follows:

[0188] Sum up the estimated future cumulative inventory of each month in dataBase1 of each logical batch in each region of the product to each month, summarize the estimated future cumulative sales of each month in dataBase1 according to the same logic, summarize the cumulative inventory and cumulative sales of each month to the whole country, calculate the national average turnover days of the month, if the national average turnover days is less than the standard turnover days of the product, then record the national overdue storage fee of the product in the month as 0, if the national average turnover days is greater than the standard turnover days of the product, calculate as follows:

[0189] a) Calculate the national cumulative inventory of the product for the month, recorded as stock, and calculate the cumulative inventory of each region for the month, recorded as stock i ;

[0190] b) Calculate the inventory ratio of each region for the product in the current month,

[0191] c) Calculate the excess inventory of the product in the current month, specifically, the national average turnover days (ito nation ) minus the standard turnover days of goods (ito std ), the difference is multiplied by the national cumulative sales volume of the product in that month (sales nation ), recorded as overdue;

[0192] d) Calculate the overdue storage cost of the product in each region in the current month. The specific calculation is as follows: overdue*ratio i *discount i *price*2;

[0193] e) Aggregate the overdue storage fees of the goods in each region for the month to the whole country;

[0194] The overdue storage fee is calculated as follows:

[0195] (See Table 1).

[0196] Use the above calculation method to calculate the overdue storage cost of each month in dataBase1 on a rolling basis.

[0197] (3) Capital occupation cost:

[0198] The existence of capital occupation cost is due to the fact that inventory will occupy cash flow, and cash flow itself can generate income. This part of income is a cost for the self-operated centralized supply chain. The specific calculation logic is as follows:

[0199] Aggregate the monthly cumulative inventory of each logical batch in each region of the product in dataBase2 to the estimated total future cumulative inventory of the product nationwide, recorded as stock capital (stock* in the formula), multiply the inventory by the national warehouse quotation of the product (wh_qtn) and the average daily rate of return on capital occupation (interest, also known as the conversion rate) to get the estimated future capital occupation cost of the product

[0200] (See Table 1).

[0201] (4) Provision for inventory impairment

[0202] Inventory will depreciate over time. In order to make the goods flow faster, a part of the inventory depreciation provision will be increased. This part of the depreciation provision is also a cost for inventory management. The depreciation provision is directly related to the inventory age and inventory volume of the goods. The specific calculation logic is as follows:

[0203] The remaining stock at the end of the month in dataBase3 for each logical batch of each product in each region is calculated. it ), matching the depreciation coefficient (ratio) corresponding to the inventory age of each batch (coef t ), the remaining inventory amount of each batch at the end of the month (stock it *wh_qtn) multiplied by the depreciation coefficient corresponding to the batch inventory is the future depreciation cost of the product. The specific formula is:

[0204] (See Table 1).

[0205] Profit and Loss Optimization Module 204

[0206] The profit and loss optimization module 204 is used to determine the optimal return quantity by using a binary search based on the acceptable cost ratio threshold α of the gross profit margin space.

[0207] In the present invention, the profit and loss optimization module 204 uses the acceptable cost ratio threshold α (a∈[0,1], preferably 0.6) of the gross profit margin space input by the business to perform a binary search. Figure 4 The profit and loss optimization algorithm is shown, which is mainly used to search for a return scenario where the return node is the current calculation time and there is a returnable quantity constraint.

[0208] The specific optimization operation performed by the profit and loss optimization module 204 is as follows:

[0209] First, calculate the future cost rate of zero return for the natural consumption scenario with no returns, where the future cost rate = future cost / inventory amount, where the future cost is the sum of the various costs calculated by the above-mentioned cost calculation module 203, and determine whether the future cost rate of zero return is less than or equal to the rate cap, where the rate cap = gross profit margin*α.

[0210] Among them, when the return quantity is set (for example, the return quantity here is 0), four holding costs are obtained through simulation and calculation of the inventory simulation module 202 and the cost calculation module 203, and the future cost is obtained by summing up the four holding costs. The inventory amount is equal to the current inventory after deducting the above-mentioned specified return quantity (here is 0 return quantity) multiplied by the warehouse quotation wh_qtn, and the calculated future cost is divided by the inventory amount to obtain the zero return future cost rate.

[0211] If it is determined that the future cost rate of zero returns is less than or equal to the rate upper limit, the optimal solution is 0, that is, the optimal return quantity is zero.

[0212] In this case, the rate corresponding to the future costs incurred without refund (zero return) is less than the rate cap, and the profit margin is acceptable.

[0213] If it is determined that the future cost rate of zero returns is greater than the rate cap, the future cost rate of full returns under the full return scenario is calculated, and it is determined whether the future cost rate of full returns is greater than or equal to the rate cap.

[0214] If it is determined that the future cost rate of full return is greater than or equal to the rate cap, it means that there is no feasible solution under the constraint of the returnable quantity, so the closest solution is used as an approximation, and the optimization is exited and the optimal solution is recorded as a full return.

[0215] The rate corresponding to the future costs incurred in the case of a full return is still higher than the rate cap, and the profit margin is below the acceptable range.

[0216] If it is determined that the future cost rate of full return is less than the rate cap, it is determined that there is a feasible solution, but full return is not the optimal solution. At this time, the rate cap is set to be gross profit margin*β (β<α), preferably, for example, β=80%*α, and the relationship between the future cost rate of 50% return and the rate cap and rate cap is determined.

[0217] If the future cost rate of 50% returns is between the upper and lower limits of the rate (including the case where the upper and lower limits are equal), then the 50% return volume is the optimal solution.

[0218] If it is determined that the future cost rate of 50% returns is less than the lower limit of the rate, the future cost rate of 25% returns is calculated, and the relationship between the future cost rate of 25% returns and the upper and lower limits of the rate is determined.

[0219] On the other hand, if it is determined that the 50% return future cost rate is greater than the rate cap, the 75% return future cost rate is calculated, and the relationship between the 75% return future cost rate and the rate cap and rate cap is determined.

[0220] Repeat the above operation until the future cost rate after a certain return volume falls between the upper and lower limits of the rate for the first time; or the number of iterations exceeds the maximum number of searches set in advance (for example, 5 times by default in the present invention).

[0221] If there is no refund, the rate is too high, and the rate corresponding to a full refund will be lower than the upper limit of the rate. At this time, a lower limit of the rate is set, and the optimal return ratio is sought through dichotomy, so that the corresponding rate under this return ratio will fall within the range between the upper limit and the lower limit of the rate.

[0222] It should be noted that the above-mentioned profit and loss optimization module using binary search to determine the optimal return quantity is not the only implementation method of the profit and loss optimization of the present invention. For example, algorithms such as taboo search and grid search can also be used to obtain the optimal solution for the return quantity. Existing algorithms such as taboo search and grid search for obtaining the optimal solution can be applied to the profit and loss optimization of the present invention, and will not be described in detail here.

[0223] The above describes a data processing device for determining return quantities based on inventory simulation according to the present invention, which can quantify the holding costs under specific inventory structure changes (returning specific quantities under certain return nodes) by simulating future inventory, and optimize the optimal clearance strategy under the constraint of the returnable quantity, thereby finding the inventory clearance plan with the best profit and loss, and guiding intelligent inventory management.

[0224] Figure 6 An exemplary system architecture 600 to which the clearing method or clearing device according to the embodiment of the present invention can be applied is shown.

[0225] like Figure 6 As shown, system architecture 600 may include terminal devices 601, 602, 603, network 604 and server 605. Network 604 is used to provide a medium for communication links between terminal devices 601, 602, 603 and server 605. Network 604 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0226] Users can use terminal devices 601, 602, and 603 to interact with server 605 through network 604 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 601, 602, and 603, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0227] The terminal devices 601 , 602 , and 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0228] The server 605 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 601, 602, and 603. The backend management server may analyze and process the received data such as product information query requests, and feed back the processing results (such as target push information, product information - only an example) to the terminal device.

[0229] It should be noted that the clearing method provided in the embodiment of the present invention is generally executed by the server 605 , and accordingly, the clearing device is generally arranged in the server 605 .

[0230] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0231] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system 700 of a terminal device suitable for implementing an embodiment of the present invention. Figure 7 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0232] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0233] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0234] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the system of the present invention are executed.

[0235] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0236] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0237] The modules involved in the embodiments of the present invention may be implemented in software or hardware. The modules described may also be set in a processor, for example, it may be described as: a processor includes a data preprocessing module, an inventory simulation module, a cost calculation module and a profit and loss optimization module. The names of these modules do not constitute a limitation on the modules themselves in some cases. For example, the profit and loss optimization module may also describe a module that uses a binary search to determine the optimal return quantity.

[0238] As another aspect, the present invention further provides a computer-readable medium, which may be included in the device described in the above embodiment; or may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes:

[0239] Use the preset data mining engine to obtain data information of target goods;

[0240] Calling a preset inventory simulation engine to simulate the inventory of the target product through a preset inventory simulation model based on the data information and the specified return quantity, thereby obtaining an inventory simulation result;

[0241] Based on the inventory simulation result and the data information, calculating the total cost of the target product by a preset cost calculation model; and

[0242] According to the total cost and inventory amount of the target product, the future cost rate of the target product under the specified return quantity is calculated through a preset cost rate calculation model, and based on the acceptable cost ratio threshold of the gross profit margin and the future cost rate, the optimal return quantity is determined through an optimal return quantity calculation model.

[0243] According to the technical solution of the embodiment of the present invention, it is possible to simulate future inventory and thus quantify the holding cost under specific inventory structure changes (returning a specific amount under certain return nodes), and to search for the optimal clearance strategy under the constraint of the returnable quantity, thereby finding the inventory clearance plan with the best profit and loss and guiding intelligent inventory management.

[0244] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data processing method for determining return quantity based on inventory simulation, characterized in that: include: Use the preset data mining engine to obtain data information of target goods; Calling a preset inventory simulation engine to simulate the inventory of the target product through a preset inventory simulation model based on the data information and the specified return quantity, including: determining a simulation scenario based on the commodity information included in the data information, and when determining that the simulation scenario is a regular scenario other than a capital occupation scenario and a provision calculation scenario, performing a regular scenario simulation using a regular scenario simulation submodel; thereby obtaining an inventory simulation result; wherein performing the regular scenario simulation using the regular scenario simulation submodel includes: obtaining the specified return quantity and the return node corresponding to the specified return quantity, and According to the inventory information in the data information, calculate the ratio of the batch inventory to the total inventory, and proportionally distribute the average daily sales in the data information to obtain the batch average daily sales of each batch; use the batch average daily sales to consume the inventory of each batch until the return node, and calculate the beginning inventory of each batch at the return node; proportionally distribute the designated return quantity to obtain the return quantity of each batch; deduct the return quantity of each batch from the beginning inventory to obtain the refreshed inventory; and repeatedly perform the conventional scenario simulation with the refreshed inventory until the inventory consumption is completed or the threshold is truncated; Based on the inventory simulation result and the data information, calculating the total cost of the target product by a preset cost calculation model; and According to the total cost and inventory amount of the target product, the future cost rate of the target product under the specified return quantity is calculated through a preset cost rate calculation model, and based on the acceptable cost ratio threshold of the gross profit margin and the future cost rate, the optimal return quantity is determined through an optimal return quantity calculation model.

2. The method according to claim 1, characterized in that: in, Using the conventional scenario simulation sub-model to perform the conventional scenario simulation also includes: Generate a data container for a natural month, the data container adopts a two-layer nested dictionary data structure, wherein the key of the data structure starts from the beginning of the month when the simulation starts to the beginning of the month when the cutoff threshold is located, and the first layer value of the data structure is the remaining simulation days of the month, the actual number of days in the month, the monthly cumulative inventory, the monthly cumulative sales, and the remaining ending inventory at the end of the month, and Fill the inventory simulation results under the conventional scenario simulation into the data container and record them as the first data set, and Calculating the total cost using the cost calculation model includes calculating the regular storage fee and the overdue storage fee included in the total cost based on the first data set and the data information.

3. The method according to claim 2, characterized in that in, The step of simulating the inventory of the target product by using the preset inventory simulation model further includes: When it is determined that the simulation scenario is the capital occupation scenario, a capital occupation scenario simulation sub-model is used to perform capital occupation scenario simulation, wherein: Using the capital occupation scenario simulation sub-model to simulate the capital occupation scenario includes: matching the account period of the target product, and determining the calculation date and the simulation start date based on the account period and the first warehousing time of each batch, and then performing the conventional scenario simulation using the conventional scenario simulation sub-model, and Fill the inventory simulation results under the capital occupation scenario simulation into the data container and record them as a second data set, and Calculating the total cost using the cost calculation model includes calculating the capital occupation cost included in the total cost based on the second data set and the data information.

4. The method according to claim 3, characterized in that: in, The step of simulating the inventory of the target product by using the preset inventory simulation model further includes: When it is determined that the simulation scenario is the provision calculation scenario, the provision calculation scenario simulation sub-model is used to perform provision calculation scenario simulation, wherein: Using the provision calculation scenario simulation sub-model to perform the provision calculation scenario simulation includes: Using the conventional scenario simulation sub-model to perform the conventional scenario simulation to calculate the ending inventory of each batch to the simulation settlement date, and The ending inventory is filled into the data container and recorded as a third data set.

5. The method according to claim 4, characterized in that in, Calculating the total cost using the cost calculation model includes calculating the inventory impairment provision cost included in the total cost based on the third data set and the data information, and Calculating the future cost rate by the cost rate calculation model includes dividing the total cost by the inventory amount to obtain the future cost rate under the specified return quantity.

6. The method according to any one of claims 1 to 5, characterized in that: in, The optimal return quantity is determined by the optimal return quantity calculation model, and the optimal return quantity is determined by binary search, and includes: Calculate a zero-return future cost rate for a natural consumption scenario with no returns, and compare the zero-return future cost rate with a rate cap; When the zero return future cost rate is less than or equal to the rate cap, the optimal return amount is 0; and when the zero return future cost rate is greater than the rate cap, the full return future cost rate is calculated and compared with the rate cap; When the future cost rate of full return is greater than or equal to the rate cap, the optimal return amount is 100%; and when the future cost rate of full return is less than the rate cap, the future cost rate of 50% return is calculated, and the relationship between the future cost rate of 50% return and the rate cap and rate cap is compared; The binary search is repeated until the future cost rate corresponding to the searched return quantity is between the rate upper limit and the rate lower limit or the set maximum number of searches is reached.

7. A data processing device for determining the quantity of returned goods based on inventory simulation, characterized in that: include: A data preprocessing module, which is used to obtain data information of target goods; An inventory simulation module is used to simulate the inventory of the target product based on the data information and the specified return quantity, including: determining a simulation scenario based on the commodity information included in the data information, and when determining that the simulation scenario is a conventional scenario other than the capital occupation scenario and the provision calculation scenario, using a conventional scenario simulation submodel to perform conventional scenario simulation; obtaining an inventory simulation result; wherein using the conventional scenario simulation submodel to perform the conventional scenario simulation includes: obtaining the specified return quantity and the return node corresponding to the specified return quantity, and according to the The inventory information in the data information is used to calculate the ratio of the batch inventory to the total inventory, and the average daily sales volume in the data information is proportionally allocated to obtain the batch average daily sales volume of each batch; the inventory of each batch is consumed by the batch average daily sales volume until the return node, and the beginning inventory of each batch at the return node is calculated; the designated return quantity is proportionally allocated to obtain the return quantity of each batch; the return quantity of each batch is deducted from the beginning inventory to obtain the refreshed inventory; and the conventional scenario simulation is repeatedly performed with the refreshed inventory until the inventory consumption is completed or the threshold is truncated; a cost calculation module, the cost calculation module being used to calculate the total cost of the target product based on the inventory simulation result and the data information; and A profit and loss optimization module is used to calculate the future cost rate of the target product under the specified return quantity based on the total cost and inventory amount of the target product, and determine the optimal return quantity based on an acceptable cost ratio threshold of the gross profit margin and the future cost rate.

8. An electronic device for determining return quantity based on inventory simulation, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

9. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images