JP2021103377A - Ordering recommendation system, ordering recommendation method, and program - Google Patents

Abstract

To provide an ordering recommendation system, an ordering recommendation method, and a program that can increase accuracy in calculation of a recommended amount of ordering.SOLUTION: An ordering recommendation system of an embodiment has an ordering recommended amount calculation unit. At the time of ordering products, the ordering recommended amount calculation unit subtracts a predicted quantity of products in stock at the timing of delivery from a standard quantity of stock at the timing of delivery to calculate an insufficient number of products for the standard quantity of stock at the timing of delivery. The ordering recommended amount calculation unit adds, to the insufficient number of products for the standard quantity of stock at the time of delivery, the predicted number of products demanded in a period from the timing of delivery to earlier timing of the timing of delivery for the next ordering and the timing at which a sales expiration date of products to be ordered this time arrives, to thereby calculate the recommended amount of ordering the products.SELECTED DRAWING: Figure 1

Description

Embodiments of the present invention relate to ordering recommendation systems, ordering recommendation methods, and programs.

Conventionally, there has been known a technique for calculating a recommended order quantity of a product on the order date so that the product ordered on the order date does not run out in the store by the day when the product arrives at the store.

However, in the conventional technique, the method of calculating the recommended order quantity has not been sufficiently examined in the case where the sales period of the product may be shorter than the ordering interval for the reason of maintaining the quality of the product.

Japanese Unexamined Patent Publication No. 2007-200185

An object to be solved by the present invention is to provide an ordering recommendation system, an ordering recommendation method, and a program capable of improving the accuracy of calculating the ordering recommended quantity.

The ordering recommendation system of the embodiment has an ordering recommended quantity calculation unit. At the time of ordering a product, the order recommended quantity calculation unit calculates the shortage with respect to the standard stock quantity at the delivery timing by subtracting the estimated stock quantity of the product at the delivery timing from the standard stock quantity at the delivery timing. In addition, the ordering recommended quantity calculation unit responds to the shortage of the standard stock quantity at the delivery timing from the delivery timing to the delivery timing of the next order and the earlier timing of the expiration of the sales deadline of the product ordered this time. The recommended quantity for ordering the product is calculated by adding the estimated demand for the product during the period between them.

The figure which shows an example of the inventory optimization system. The figure for demonstrating the calculation process of the theoretical stock quantity. The figure for demonstrating the calculation process of the theoretical stock quantity. The figure which shows an example of the time series data of the correction theoretical stock quantity. The figure for demonstrating the calculation process of the correction theory stock quantity. The figure for demonstrating the calculation process of discount aggregate data. The figure for demonstrating the calculation process of the waste aggregate data. The figure for demonstrating the calculation process of the theoretical inventory data for each individual item. Diagram to explain the calculation process of theoretical inventory data by category. The figure for demonstrating the calculation process of the purchase customer number data. The figure for demonstrating the calculation process of the data for calculating the number of missed sales. The figure which shows an example of the data for calculating the number of missed sales. The figure for demonstrating the calculation process of the estimated purchase rate and the number of missed sales by individual item. The figure for demonstrating the calculation process of the oversold amount by individual item in the shortage period considering the purchase of a substitute product. The figure for demonstrating the preprocessing of training data. The figure for demonstrating the preprocessing of training data. The figure which shows an example of the learning data. A flowchart showing an example of demand forecast processing. The figure which shows an example of the learning data. A flowchart showing an example of demand forecast processing. The flowchart which shows an example of the calculation process of a reference stock amount. The figure which shows an example of the calculation method of a unit unsold loss. The figure which shows an example of the demand forecast residual distribution of a product. The figure which shows an example of the demand forecast residual distribution of a product. The figure which superimposes the demand forecast residual distribution of a product, the amount of a loss on sale of a product, and the amount of a loss on sale of a product. The figure which superimposes the demand forecast residual distribution of a product, the amount of a loss on sale of a product, and the amount of a loss on sale of a product. The figure which superimposes the demand forecast residual distribution of a product, the amount of a loss on sale of a product, and the amount of a loss on sale of a product. The figure which superimposes the demand forecast residual distribution of a product, the amount of a loss on sale of a product, and the amount of a loss on sale of a product. The figure for demonstrating an example of the calculation process of a reference stock amount. The figure for demonstrating an example of the calculation process of a reference stock amount. The figure for demonstrating the calculation process of the recommended order quantity. The figure for demonstrating the calculation process of the recommended order quantity. The figure for demonstrating the calculation process which corrects the recommended order quantity based on the sales performance for each content of an uncertain product. The flowchart which shows an example of the calculation process which corrects the recommended order quantity based on the sales performance for each content of an indefinite product. The figure for demonstrating the calculation process of the recommended order quantity considering the sale deadline of a product. The figure for demonstrating the calculation process of the recommended order quantity considering the sale deadline of a product. The figure for demonstrating the calculation process of the recommended order quantity considering the sale deadline of a product. The figure for demonstrating the calculation process of the recommended order quantity considering the sale deadline of a product. The figure for demonstrating the calculation process of the recommended order quantity considering the sale deadline of a product.

Hereinafter, the inventory optimization system, the inventory optimization method, and the program of the embodiment will be described with reference to the drawings.

FIG. 1 is a configuration diagram of the inventory optimization system 10 of the embodiment. The inventory optimization system 10 includes, for example, a loss estimation system 100 and an ordering recommendation system 200. The ordering recommendation system 200 predicts the demand for the product based on the missed sale data of the product estimated by the loss estimation system 100 and the sales performance data of the product, and the inventory of the product is based on the predicted demand for the product. It is a system that determines the recommended quantity for ordering products so as to optimize. The products include products produced in the store and products delivered based on orders placed outside the store. Therefore, the word "ordering" used in the embodiment may be better read as "production" for products produced in the store. Similarly, it may be better to read "ordering recommendation" as "production recommendation" and "ordering recommended amount" as "production recommended amount". In the embodiment, the explanatory text generated by these replacements is omitted.

[Loss estimation system]
The loss estimation system 100 is realized by one or more processors. The loss estimation system 100 is a system for estimating discount loss, disposal loss, and oversold loss of goods. The loss estimation system 100 includes, for example, a control unit 110 and a storage unit 140. The control unit 110 includes, for example, an inventory quantity calculation unit 112, an inventory quantity correction unit 114, a discount loss calculation unit 116, a disposal loss calculation unit 118, a shortage determination unit 120, a missed sale calculation unit 122, and a single item. It includes a separate demand amount calculation unit 124, a category-specific demand amount calculation unit 126, an evaluation unit 128, a missed sale correction unit 130, and a missed sale output unit 132. These components are realized, for example, by a hardware processor such as a CPU executing a program (software). Some or all of these components are hardware (circuits) such as LSI (Large-scale integrated circuit), ASIC (Application Specific integrated circuit), FPGA (Field-Programmable gate array), and GPU (Graphics Processing Unit). It may be realized by the part; including circuitry), or it may be realized by the cooperation of software and hardware. The program may be stored in advance in a storage device such as an HDD or a flash memory (a storage device including a non-transient storage medium), or a removable storage medium such as a DVD or a CD-ROM (non-transient). It is stored in a sex storage medium) and may be installed by attaching the storage medium to a drive device.

The inventory quantity calculation unit 112 calculates the inventory quantity of the product. The inventory quantity calculation unit 112 calculates the inventory quantity of the product based on, for example, the theoretical inventory data 142 read from the storage unit 140. The theoretical inventory data 142 is data obtained by calculating the inventory quantity of a product in time series based on information related to an increase or decrease in the inventory quantity, and a data group related thereto.

When the inventory quantity of the product calculated by the inventory quantity calculation unit 112 is a negative value, the inventory quantity correction unit 114 corrects the inventory quantity of the product to a positive value, and the calculation result of the inventory quantity of the corrected product. Is stored in the storage unit 140 as theoretical inventory data 142. The inventory quantity correction unit 114 may, for example, correct the inventory quantity of the product to a non-negative value and store information indicating that the product has been corrected. The inventory quantity correction unit 114 may indicate that the product has been corrected, for example, by switching the correction flag of the product from off to on.

The discount loss calculation unit 116 calculates the discount loss of the product, and stores the calculation result of the discount loss of the product as the product data 144 in the storage unit 140. The discount loss calculation unit 116 calculates the discount loss of a product by, for example, integrating the discount amount of the product with the number of discounted sales of the product.

The disposal loss calculation unit 118 calculates the disposal loss of the product, and stores the calculation result of the disposal loss of the product as the product data 144 in the storage unit 140. The disposal loss calculation unit 118 calculates the disposal loss of the product by, for example, integrating the cost price of the product with the number of discarded products.

The shortage determination unit 120 determines whether or not the product is out of stock based on the number of products in stock. The out-of-stock determination unit 120 does not determine the out-of-stock of the product when the in-stock quantity of the product is a positive value in the time-series data of the in-stock quantity of the product, and the in-stock quantity of the product is "0 (zero)". In that case, it is determined that the product is out of stock. At this time, the shortage determination unit 120 may not determine the shortage of the product, for example, when the correction flag of the product is switched on by the inventory quantity correction unit 114.

The oversold calculation unit 122 calculates the oversold amount of the product, and stores the calculation result of the oversold amount as the unsold data 146 in the storage unit 140. The oversold amount is the quantity of products that are considered to have missed a sales opportunity due to a shortage, and the oversold loss is a loss calculated by integrating the number of products with the price per unit of the product. The missed-sale calculation unit 122 calculates the missed-sale amount of the product based on, for example, the sales performance of the product for each time zone.

The demand amount calculation unit 124 for each individual item calculates the unsold amount based on the unsold amount of the product calculated by the unsold calculation unit 122, totals the unsold amount of each individual item by category, and divides the unsold amount of each item into categories. The demand amount of the category is calculated by adding the sales results.

The category-specific demand amount calculation unit 126 calculates the demand amount of a category including a plurality of products based on the category sales of the day of the week / time zone in which the product is determined not to be out of stock by the shortage determination unit 120.

The evaluation unit 128 evaluates the degree of influence of the sale of alternative products included in the same category on the sales performance of the product. In the evaluation unit 128, for example, the larger the ratio of the total value of the demand amount for each product calculated by the demand amount calculation unit 124 for each item to the demand amount for the category calculated by the demand amount calculation unit 126 for each category, the greater the ratio. Greatly evaluate the degree of influence of alternative products. Increasing the degree of influence of alternative products means that many alternative products are sold.

The oversold correction unit 130 corrects the oversold amount of the product calculated by the oversold calculation unit 122 based on the degree of influence evaluated by the evaluation unit 128.

The missed sale output unit 132 reads the missed sale data 146 from the storage unit 140, and outputs the missed sale amount of the product indicated by the missed sale data 146 to the external device.

[Ordering recommendation system]
The ordering recommendation system 200 is realized by one or more processors. The ordering recommendation system 200 predicts the demand for a product based on the sales performance data of the product and the missed sales data of the product. The missed-selling data of a product is data showing the quantity and the estimated loss amount estimated to be missed-sold when the product misses a sales opportunity due to a shortage. In addition, the ordering recommendation system 200 considers the unsold loss of the product and the lost sales of the product, and sets the inventory amount with respect to the predicted sales amount based on the demand forecast of the product so that the expected loss amount of the product is minimized. Determined as the standard inventory amount. Then, the ordering recommendation system 200 determines the difference between the standard inventory amount and the predicted inventory amount based on the demand forecast of the product as the ordering recommended amount.

The ordering recommendation system 200 includes, for example, a control unit 210 and a storage unit 230. The control unit 210 includes, for example, a demand forecasting unit 212, a reference inventory calculation unit 214, an ordering recommended quantity calculation unit 216, and an ordering recommended quantity output unit 218. These components are realized, for example, by a hardware processor such as a CPU executing a program (software). Some or all of these components may be realized by hardware such as LSI, ASIC, FPGA, GPU (including circuit part: circuitry), or realized by collaboration between software and hardware. May be good. The program may be stored in advance in a storage device such as an HDD or a flash memory (a storage device including a non-transient storage medium), or a removable storage medium such as a DVD or a CD-ROM (non-transient). It is stored in a sex storage medium) and may be installed by attaching the storage medium to a drive device.

The demand forecast unit 212 generates a demand forecast model M1 based on the sales record data 232 of the product stored in the storage unit 230, and inputs a forecast time zone, weather forecast information, etc. into the demand forecast model M1 to produce the product. Demand forecast. For example, the demand forecasting unit 212 adds up the number of missed sales for each sales time zone of the product that missed the sales opportunity of the product due to the shortage to the number of sales based on the sales performance data 232 for each sales time zone of the product. Then, the number of products sold for each sales time zone is corrected. Then, the demand forecast unit 212 inputs the forecast target sales time zone and weather forecast information into the demand forecast model M1 learned using the learning data including the number of sales of the corrected product for each sales time zone. By doing so, the demand forecast of the product is made.

The demand forecasting unit 212 determines, for example, whether or not there is an assortment of products for each sales time zone based on the number of sales of the corrected product for each sales time zone, and determines the sales time zone for which it is determined that there is no assortment. The demand forecast model M1 is trained using the training data including the number of products sold in the excluded sales time zone. For example, when the number of products sold for each sales time zone of the corrected product is "0 (zero)" in all the time zones of the day, the demand forecasting unit 212 has an assortment of products in all the time zones of the day. Assuming that there is no product, the judgment result of the product assortment is corrected. Alternatively, the demand forecasting unit 212 sets the sales period of the product to a predetermined length, for example, when the number of sales of the corrected product for each sales time zone is "0 (zero)" in all the time zones of the day. If it is less than, it is assumed that there is no product assortment in all time zones of the day, and the judgment result of the product assortment is corrected. The determination of no assortment referred to here is a process different from the assortment target determination of the theoretical inventory data 142 of the loss estimation system 100 described later. Products whose sales period is less than a predetermined length include, for example, meat, fish, vegetables, and prepared foods. The demand forecasting unit 212 stores the forecasting result of the demand of the product in the storage unit 230 as the demand forecasting data 234 in association with the weather information and the calendar information at the time of selling the product. The weather information may be not only the weather but also, for example, average temperature, precipitation, sunshine duration, and the like.

The reference inventory calculation unit 214 calculates the inventory amount that minimizes the expected loss amount of the product as the reference inventory amount, and stores the calculation result of the reference inventory amount as the reference inventory data 236 in the storage unit 230.

The ordering recommended quantity calculation unit 216 calculates the ordering recommended quantity of the product, and stores the calculation result of the ordering recommended quantity as the ordering recommended data 238 in the storage unit 230. The ordering recommended quantity calculation unit 216 calculates, for example, the estimated stock quantity of products at a future point in time when the ordered products are lined up in the store, and calculates the difference from the standard stock quantity at the same future time point as the quantity that is insufficient at the future time point. In addition, the ordering recommended quantity calculation unit 216 sets the total number of predicted demands after the time when the products to be ordered at the present time are lined up in the store and before the time when the products to be ordered next time are lined up in the store as the insufficient quantity. By adding to the calculated number, the recommended order quantity of the product is calculated.

The ordering recommended quantity output unit 218 reads the ordering recommended data 238 from the storage unit 230, and outputs the ordering recommended quantity of the product indicated by the ordering recommended data 238 to the external device.

[Calculation of theoretical inventory]
Next, the calculation process of the theoretical inventory will be described.

The inventory quantity calculation unit 112 calculates the theoretical inventory data of the product. The inventory quantity calculation unit 112 calculates the theoretical inventory quantity of a product by using, for example, the inventory fluctuation number from the start time to the end time of the time zone. The evaluation items of the inventory fluctuation number include, for example, the number of purchases, the number of moving goods received, the number of moving goods issued, the number of returned goods, the number of productions, the number of wastes, and the number of sales. The number of purchases is data indicating the number of products purchased by the store from the seller. The number of moving goods is data indicating the number of products received from another store and products received from another department in the same store (for example, products received from the meat department to the side dish department). The number of moving shipments is data indicating the number of products shipped to another store and products shipped to another department in the same store (for example, products shipped from the meat department to the side dish department). The number of returned goods is data indicating the number of goods returned from the store to the seller and from the store to the company's distribution warehouse. The number of products produced is data indicating the number of products produced in the store, such as prepared foods. The number of discarded items is data indicating the number of items discarded by the store. The number of sales is data indicating the number of products sold by the store.

The inventory quantity calculation unit 112 changes the calculation method of the theoretical inventory data of the product between the product evaluated by the cost method and the product evaluated by the selling price reduction method. A product evaluated by the cost method is a product that has a relatively short sales period and does not take inventory into consideration. A product evaluated by the selling price reduction method is a product that has a relatively long sales period and is inventoried.

As shown in FIG. 2, when calculating the theoretical inventory data of the product to be evaluated by the cost method, the inventory quantity calculation unit 112 first sets the time zone t1, which is a time point past the current time point, as the calculation start time point. , The theoretical stock quantity at the start time of the time zone t1 is set to "0 (zero)". This is because it is difficult to calculate and manage the inventory amount of products evaluated by the cost method at a certain point in time. Then, the inventory quantity calculation unit 112 calculates the theoretical inventory quantity "X1" at the end time of the time zone t1 by adding the inventory fluctuation number of the time zone t1 to the theoretical inventory quantity of the start time of the time zone t1. To do. Here, the number of inventory fluctuations may be a negative value. Further, the inventory quantity calculation unit 112 applies the theoretical inventory quantity “X1” at the end time of the time zone t1 as the theoretical inventory quantity at the start time of the time zone t2. After that, the inventory quantity calculation unit 112 sets the time zone t6, which is the current time point, as the calculation end time point, and sequentially calculates the theoretical inventory quantity from the time zone t1 to the time zone t6.

As shown in FIG. 3, when calculating the theoretical inventory data of the product to be evaluated by the selling price reduction method, the inventory quantity calculation unit 112 first calculates the theoretical inventory quantity “Y” at the end time of the time zone t4, which is the inventory timing. calculate. Further, the inventory quantity calculation unit 112 subtracts the inventory fluctuation number of the time zone t4 from the theoretical inventory quantity "Y" at the end time of the time zone t4, so that the theoretical inventory quantity "Y1" at the start time of the time zone t4 Is calculated. Further, the inventory quantity calculation unit 112 applies the theoretical inventory quantity “Y1” at the start time of the time zone t4 as the theoretical inventory quantity “Y1” at the end time of the time zone t3. After that, the inventory quantity calculation unit 112 sequentially calculates the theoretical inventory quantity from the time zone t4 to the time zone t1 with the time zone t1 as the calculation start time in the time zone before the inventory timing. Further, the inventory quantity calculation unit 112 applies the theoretical inventory quantity “Y” at the end time of the time zone t4 to the theoretical inventory quantity at the start time of the time zone t5. Then, the inventory quantity calculation unit 112 adds the inventory fluctuation number of the time zone t5 to the theoretical inventory quantity "Y" at the start time of the time zone t5, so that the theoretical inventory quantity "Y5" at the end time of the time zone t5 Is calculated. Further, the inventory quantity calculation unit 112 applies the theoretical inventory quantity “Y5” at the end time of the time zone t5 as the theoretical inventory quantity “Y5” at the start time of the time zone t6. After that, the inventory quantity calculation unit 112 sequentially calculates the theoretical inventory quantity from the time zone t5 to the time zone t6, with the time zone t6 as the current time point as the calculation end time in the time zone after the inventory timing.

When the theoretical inventory quantity of the product becomes inconsistent due to some data and becomes a negative value, the inventory quantity correction unit 114 corrects the theoretical inventory quantity of the product and calculates it as the corrected theoretical inventory quantity. The corrected theoretical inventory quantity is, for example, a positive value. When the inventory quantity correction unit 114 calculates the correction theoretical inventory quantity of a product, the product is excluded from the target of the out-of-stock determination. As a result, it is possible to prevent the product from being erroneously determined to be out of stock.

[Calculation of corrected theoretical inventory quantity]
Next, the calculation process of the correction theoretical inventory quantity of the product will be described.

When the inventory quantity correction unit 114 corrects the time series data of the inventory quantity calculated by the inventory quantity calculation unit 112, the inventory quantity correction unit 114 determines a period to be corrected with respect to the time series data of the inventory quantity and traces back the time axis of the period. The time series data of the number of stocks is corrected in order.

When the inventory quantity to be corrected is a negative value, the inventory quantity correction unit 114 corrects it to a positive value (for example, "1"). The inventory quantity correction unit 114 sets the inventory quantity to a positive value (for example, when the inventory quantity to be corrected is "0 (zero)" and the inventory quantity at the previous time is corrected. 1 ”) is corrected, and if the stock quantity at the previous time is not corrected, the stock quantity is not corrected. The inventory quantity correction unit 114 does not correct the inventory quantity when the inventory quantity to be corrected is a positive value.

FIG. 4 is a graph showing time-series data before and after the theoretical inventory quantity of commodities is corrected. The inventory quantity correction unit 114 calculates the time-series data of the corrected theoretical inventory quantity by correcting the time-series data of the theoretical inventory quantity for each period of estimation of the loss of the product, for example. The time-series data of the theoretical inventory quantity is, for example, data including the theoretical inventory quantity calculated at regular time intervals from the time when the store first receives the target product. In the illustrated example, the period of product loss estimation is the period from the time zone t1 to the time zone t10.

When the inventory quantity correction unit 114 corrects the time-series data of the theoretical inventory quantity, first, the theoretical inventory quantity becomes a negative value from the data of the theoretical inventory quantity included in the period of estimating the loss of the product. Search for data. For example, the inventory quantity correction unit 114 sets the final time point included in the period of estimation of the loss of the product as the calculation start time point, and searches for data in which the theoretical inventory quantity has a negative value while tracing back the time axis from that time point. In the illustrated example, the inventory quantity correction unit 114 searches for the theoretical inventory quantity data at the end time of the time zone t10 as data having a negative value.

The inventory quantity correction unit 114 corrects the theoretical inventory quantity to be calculated next. In the illustrated example, the inventory number correction unit 114 traces the time axis back from the theoretical inventory number at the end time of the time zone t10, and sets the theoretical inventory quantity at the start time of the time zone t10 as the next calculation target. Then, the inventory quantity correction unit 114 searches for the data of the theoretical inventory quantity at the start time of the time zone t10 as data having a negative value, following the data of the theoretical inventory quantity at the end time of the time zone t10.

For example, when the theoretical stock quantity in a certain time zone is corrected retroactively on the time axis, the stock quantity correction unit 114 calculates the theoretical stock quantity at the start time of the time zone in the time zone (time) to be calculated next. It is applied as the theoretical stock quantity at the end time (the time zone immediately before going back the axis). In the illustrated example, since the stock quantity correction unit 114 corrects the theoretical stock quantity at the start time of the time zone t10, the theoretical stock quantity is calculated by calculating the theoretical stock quantity at the end time of the time zone t9 to be calculated next. Apply as a number.

When the stock quantity correction unit 114 corrects the theoretical stock quantity at the end time of a certain time zone and the theoretical stock quantity at the start time of the time zone is "0 (zero)", the stock quantity correction unit 114 of the time zone When the theoretical stock quantity at the start time is corrected and the theoretical stock quantity at the end time of a certain time zone is not corrected, and the theoretical stock quantity at the start time of the time zone is "0 (zero)", The theoretical stock quantity at the start time of that time zone is not corrected. In the illustrated example, the inventory number correction unit 114 corrects the theoretical inventory quantity at the end time of the time zone t4, and the theoretical inventory quantity at the start time of the time zone t4 is “0 (zero)”. The theoretical stock quantity at the start time of the band t4 is corrected. Further, since the inventory quantity correction unit 114 does not correct the theoretical inventory quantity at the end time of the time zone t2 and the theoretical inventory quantity at the start time of the time zone t2 is "0 (zero)", the time zone t2 Do not correct the theoretical stock quantity at the start time of.

FIG. 5 is a diagram for explaining the calculation process of the corrected theoretical inventory quantity for each period of the loss estimation of the product. In the illustrated example, the inventory quantity calculation unit 112 calculates the inventory along the time axis in the forward direction for each predetermined period. In FIG. 5, the inventory quantity calculation unit 112 calculates the initial value of the time series data of the inventory quantity included in the first predetermined period as “0 (zero)”. Then, the inventory quantity correction unit 114 performs correction calculation along the time axis in units of loss estimation periods. In each loss estimation period, the inventory quantity correction unit 114 performs the correction calculation so as to go back in time as described above. Since the theoretical inventory quantity in the first loss estimation period includes negative values, the inventory quantity correction unit 114 corrects these values to positive values. As a result, the inventory quantity calculation unit 112 calculates the inventory with the initial value of the theoretical inventory quantity in the second loss estimation period as a positive value. Further, since the inventory quantity correction unit 114 includes negative values in the time series data of the theoretical inventory quantity in the second loss estimation period, these values are corrected to positive values. The inventory quantity correction unit 114 does not correct the theoretical inventory quantity because the theoretical inventory quantity in the third loss estimation period does not include a negative value.

[Calculation of missed sales]
Next, the calculation process of the missed sale of the product will be described.

First, as shown in FIG. 6, the discount loss calculation unit 116 creates the discount total data 154 based on the purchase history data 150 and the discount history data 152.

The discount history data 152 is data relating to the sales history of the discounted product. The discount history data 152 is data created in advance by the discount loss calculation unit 116 from the purchase history data 150, and is created by extracting a record to be a discount from the sales code or the sales code name of the purchase history data 150. The data items of the discount history data 152 are, for example, a store code, a product code, a transaction date and time, a discount sales number, and a discount amount. The discounted sales quantity is data on the quantity of the discounted product sold. The discount amount is data on the amount of the discounted goods.

The discount aggregated data 154 is aggregated data for each time zone of the sales history of the discounted product, for example, with the hourly unit as the minimum record. The data items of the discount aggregate data 154 include, for example, a store code, a product code, a start time of a date / time zone, an end time of a date / time zone, a number of discounted sales, and a discount amount. The discounted number of sales is data showing the total value of the number of discounted products sold for each time zone. The discount amount is data showing the total value of the discount amount of the product for each time zone.

The discount loss calculation unit 116 refers to the discount history data 152, and uses the store code, the product code, the start time of the date / time zone, and the end time of the date / time zone as keys to discount the products included in the purchase history data 150. The discount total data 154 is created by totaling the number of sales and the discount amount of the product by store, product, date / time zone.

Next, as shown in FIG. 7, the disposal loss calculation unit 118 creates the disposal total data 158 based on the disposal history data 156.

The disposal history data 156 is data related to the disposal history of the product. The data items of the disposal history data 156 include, for example, a store code, a product code, a disposal date and time, a disposal quantity, and a disposal amount. The store code is identification data of the store where the product is discarded. Here, the disposal amount is, for example, the cost price of the discarded product.

The disposal aggregated data 158 is aggregated data for each time zone of the disposal history of the product, for example, with the hourly unit as the minimum record. The data items of the disposal total data 158 include, for example, a store code, a product code, a start time of a date / time zone, an end time of a date / time zone, a disposal quantity, and a disposal amount. The disposal amount is data showing the total value of the disposal amount of the discarded products for each time zone.

The disposal loss calculation unit 118 refers to the disposal history data 156 and discards the products included in the disposal history data 156 by using the store code, the product code, the start time of the date / time zone, and the end time of the date / time zone as keys. Disposal total data 158 is created by totaling the amount of money by store, product, date / time zone.

Next, as shown in FIG. 8, the miss-selling calculation unit 122 creates the theoretical inventory data 162 for each individual item by combining the theoretical inventory data 142 and the miss-selling calculation category master 160.

The theoretical inventory data 142 is data in which the number of inventories of each product existing in the store is, for example, one hour as the minimum record. The data items of the theoretical inventory data 142 are, for example, the department code, the subcategory code, the store code, the product code, the start time of the date / time zone, the end time of the date / time zone, the theoretical inventory quantity of the start time, and the end time. Includes theoretical stock quantity, assortment target determination, and out-of-stock flag. Although not shown, the data items of the theoretical inventory data 142 further include out-of-stock time, selling unit price, original unit price, and putting unit price. Here, as the theoretical inventory quantity, it is preferable to use the processing result of the inventory quantity correction unit 114. The department code is identification data indicating the type of department to which the product belongs. The sub-classification code is identification data indicating the type of sub-classification, which is the classification of the lower layer of the product. The store code is identification data indicating the store where the product is placed. The product code is identification data indicating the type of product. The start time of the date / time zone is data indicating the start time of each record. The end time of the date / time zone is data indicating the end time of each record. The theoretical inventory quantity at the start time is data indicating the inventory quantity of the products existing in the store at the start time. The theoretical inventory quantity at the end time is data indicating the inventory quantity of products existing in the store at the end time. The assortment target determination is data indicating whether or not the products are in stock as store inventory (whether or not the products are in stock that can be sold at the store). It is determined that the product is out of the assortment. In the assortment target determination, "0" is given when the product is not the target of the assortment, and "1" is given when the product is the target of the assortment. The data to which "0" is given as the assortment target judgment is not used for the calculation of the missed sale of the product. The out-of-stock flag is data indicating whether or not the product is out of stock. For example, when the theoretical inventory of the product becomes "0 (zero)", it is determined that the product is out of stock. The out-of-stock flag is given "0" when the product is not out of stock, and "1" is given when the product is out of stock. The out-of-stock time is data indicating the length of time that a product is out of stock. The selling unit price is data indicating the selling price of a product. The original unit price is data indicating the cost of a product. The price-added unit price is data showing the difference between the selling price of a product and the cost price.

The sales loss calculation category master 160 is data that defines information about products required for calculating sales loss for each category. The data items of the sales loss calculation category master 160 include, for example, a department code, a subclassification code, a calculation particle size, and a calculation category. The department code is identification data indicating the type of department to which the product belongs. The sub-classification code is identification data indicating the type of sub-classification, which is the classification of the lower layer of the product. The calculated particle size defines whether the product category corresponds to "product designated purchase" or "category purchase". "Product designated purchase" means an irreplaceable product that a customer purchases by designating the product when the customer purchases the product at the store. "Category designated purchase" means a substitutable product in which a customer purchases a substitute product in the case of a shortage when the customer purchases the product at the store. The calculation category is given a sub-category code for each product as a category name for calculating sales loss.

The theoretical inventory data 162 for each item is, for example, data for a single item with an hourly unit as the minimum record. The missed sale calculation unit 122 creates the theoretical inventory data 162 for each individual item by combining the theoretical inventory data 142 and the missed sale calculation category master 160 using the department code and the subclassification code as keys. The data items of the theoretical inventory data 162 for each item are, for example, the department code, the subcategory code, the store code, the product code, the start time of the date / time zone, the end time of the date / time zone, the theoretical inventory quantity of the start time, and the end. Includes theoretical stock quantity at time, assortment target judgment, out-of-stock flag, calculated grain size, and calculated category. Although not shown, the data items of the theoretical inventory data 162 for each item further include the out-of-stock time, the selling unit price, the original unit price, and the pricing unit price.

Next, as shown in FIG. 9, the oversold calculation unit 122 creates the theoretical inventory data 164 for each category based on the theoretical inventory data 162 for each item. The theoretical inventory data 164 by category is data of product categories grouped by using the hourly unit as the minimum record and the calculation category and the calculation particle size as keys. In the illustrated example, the oversold calculation unit 122 has a specified time zone in the range of "2018/10/1 12:00" to "2018/10/1 13:00", and the calculation particle size is "category purchase". In addition, the theoretical inventory data 164 for each category is created by grouping the data whose calculation category is "CS2".

Next, as shown in FIG. 10, the missed sale calculation unit 122 creates the purchase customer number data 166 based on the purchase history data 150. The purchase history data 150 is, for example, data for a single product with an hourly unit as the minimum record.

The purchase history data 150 is data relating to the purchase date and time of the product, as well as the presence or absence of return processing and correction processing at the time of purchase of the product. The purchase history data 150 is created based on, for example, sales data of products acquired from a POS (Point of sale) register installed in each store. The data items of the purchase history data 150 are, for example, a store code, a product code, a sales code, a sales code name, a sales quantity, a sales amount, a selling unit price, an original unit price, a transaction date and time, a POS number, a transaction number, a return flag, and a correction flag. including. The store code is identification data indicating the store where the product was purchased. The product code is identification data indicating the purchased product. The sales code is identification data indicating the type of transaction made through the POS register. The selling unit price is data showing the selling price per product. The original unit price is data showing the cost per product. The transaction date and time is data related to the date and time when the transaction was made through the POS register, such as the purchase of a product, the return of a product, and the correction of purchase data of a product. The POS number is identification data of the POS register. The transaction number is identification data relating to a transaction given each time a transaction is made through the POS register. The return flag is data indicating whether or not the product has been returned, the initial value is "0", and the product is switched to "1" when the product is returned. The correction flag is data indicating whether or not the product purchase history has been corrected, the initial value is "0", and the value is switched to "1" when the product purchase history is corrected.

The purchase customer number data 166 is data on the number of purchase customers for each time zone, for example, with the hourly unit as the minimum record. The data items of the purchase customer number data 166 include, for example, the store code, the start time of the date / time zone, the end time of the date / time zone, the number of sales customers, the number of returned customers, the number of correction customers, and the number of purchase customers. The number of sales customers is data indicating the number of customers who have accounted for goods through the POS register, and corresponds to the number of data records of the transaction number in the purchase history data 150. The number of returned customers is data indicating the number of customers who have returned the product, and corresponds to the number of data records in which the return flag is "1" in the purchase history data 150. The number of corrected customers is data indicating the number of customers who requested the clerk to correct the purchase history of the product at the time of purchasing the product, and corresponds to the number of data records in which the correction flag is "1" in the purchase history data 150. To do. The number of purchasing customers is data indicating the number of customers who purchased the product for each time zone, and corresponds to the number of customers obtained by subtracting the number of corrected customers from the number of selling customers. That is, in the correction process of the purchase history of the product, the transaction history is once canceled and the sales of the product are re-recorded. In this case, it is not possible to distinguish between the recording of product sales and the re-recording of product sales. Therefore, when the product purchase history is corrected, the number of data records of the transaction number is larger than the number of purchase customers. Become. Therefore, the number of purchasing customers is calculated by subtracting the number of corrected customers from the number of selling customers. Regarding the return processing, the number of customers is counted in the original transaction that is the target of the return, so the number of returns is not deducted.

Next, as shown in FIG. 11, the oversold calculation unit 122 has the theoretical inventory data 162 for each individual item.
By combining the theoretical inventory data 164 by category, the purchase history data 150, the number of purchase customers data 166, and the day / holiday master data 168, the data 170 for calculating the number of missed sales is created. The day / holiday master data 168 is data for defining the day of the week and the holiday for each date. The data items of the day / holiday master data 168 include, for example, dates, holidays, days of the week, and weekday holidays. A date is data about a date that is the subject of a definition. A holiday is data indicating whether or not a date to be defined corresponds to a holiday. The day of the week is data indicating the day of the week of the date to be defined. Weekday holidays are data indicating whether the date to be defined is classified into "weekdays" or "holidays". In this case, "weekday" is given when the date to be defined is not a national holiday and is between Monday and Friday. In addition, "holiday" is given when the date to be defined is a national holiday, or when the day of the week to be defined is Saturday or Sunday.

The data 170 for calculating the number of missed sales is data that defines information used for calculating the number of missed sales, for example, with an hour unit as the minimum record. The data items of the data 170 for calculating the number of missed sales are, for example, department code, subcategory code, store code, product code, start time of date / time zone, end time of date / time zone, assortment target judgment, shortage. Includes flags, calculated granularity, calculated categories, days of the week, weekday holidays, number of customers purchased, number of sales, and sales amount. The number of customers sold is the same as the number of purchased customers data 166 described in [0060]. The number of sales is data indicating the number of products sold in one hour from the start time of the date / time zone to the end time of the date / time zone. The number of sales is acquired based on, for example, the sales quantity of the product included in the data obtained by combining the theoretical inventory data 162 for each item and the theoretical inventory data 164 for each category.

The missed sale calculation unit 122 combines the records of the theoretical inventory data 162 for each item, the master data 168 for the day of the week and holidays, the number of purchase customers data 166, and the purchase history data 150, using the store code as a key, to increase the number of missed sales. The calculation data 170 is created.

Next, the missed sale calculation unit 122 calculates the assumed purchase rate and the number of missed sales by individual item with reference to the data 170 for calculating the number of missed sales. The estimated purchase rate is calculated for the combination of the product and the time zone, and is the probability that the customer purchases the product when the customer visits the store. The assumed purchase rate is calculated by, for example, dividing the number of sales of a product for each time zone (hourly unit) by the number of purchase customers, which is the number of customers who have accounted for the product at the store in the same time zone. Here, the same time zone is defined by, for example, a combination of a day of the week and a time zone. That is, the same day of the week and the same time zone are set as the same time zone. The number of missed sales by individual item is the number of missed sales for each time zone of the product. The number of missed items sold separately is also calculated for the combination of the product and the time zone, but is calculated especially for the time zone when there is a shortage. The number of missed sales by individual item is the number of missed sales by individual item based on the purchase rate. For example, it is calculated by integrating the number of purchase customers, the assumed purchase rate, and the out-of-stock time ratio. The out-of-stock time ratio indicates the ratio of the time when the product was out of stock when the number of products in stock was "0 (zero)" in the time zone when the product was out of stock.

In the calculation of the estimated purchase rate, the missed-sale calculation unit 122 starts with the data 170 for calculating the missed-sale number according to the following priority order when there is no sample number that requires the same day of the week and the same time zone without any shortage. Complement the data by gradually relaxing the conditions for extracting the data.

Priority 1: No shortages on the same day and time.
Priority 2: Same day of the week (weekdays / holidays) ・ No shortages at the same time.
Priority 3: Regardless of the day of the week ・ No shortages at the same time.
Priority 4: Including shortages on the same day and time.
Priority 5: Same day of the week (weekdays / holidays) ・ Including shortages in the same time zone.
Priority 6: Regardless of the day of the day ・ Including shortages at the same time.

If there is a time zone when the product is out of stock, the missed sales calculation unit 122 determines the amount of missed sales in the time zone when the product was out of stock and the sales record in the time zone when the product was not out of stock. Estimate using. When the missed sale calculation unit 122 estimates the amount of missed sales during the time period when the product was out of stock using the sales results during the time zone when the product was not out of stock, the time when the product was out of stock. Priority is given to the sales performance of products in the time zone close to the zone. Further, when the missed sale calculation unit 122 estimates the amount of missed sales during the time period when the product was out of stock using the sales record during the time zone when the product was not out of stock, the product is out of stock. Compared to the sales performance of the product in the time zone after the time zone, the sales performance of the product in the time zone before the time zone when the product was out of stock is preferentially used.

FIG. 12 is a diagram showing an example of data 170 for calculating the number of missed sales classified according to the presence or absence of a product shortage. In the illustrated example, the data 170 for calculating the number of missed sales only during the time period when there is a shortage is shown as “Table A”. Further, in the illustrated example, the data 170 for calculating the number of missed sales only during the time period when there is no shortage is shown as “Table B”.

As shown in FIG. 13, the missed-sale calculation unit 122 sets the amount of missed sales during the time period when the product to be assorted is out of stock during the time zone when the product to be assorted is not out of stock. When estimating using the data of, the oversold amount estimation data 172 is referred to. The oversold amount estimation data 172 is data in which "table A" and "table B" are associated with a store code, a product code, a time zone, and a day of the week as keys. The data item of the oversold amount estimation data 172 includes a data item extracted from "Table A", a data item extracted from "Table B", and a data item as "calculation data". The data items extracted from "Table A" include, for example, a store code, a product code, a date, a time zone, a day of the week, and the number of customers. The data items extracted from "Table B" include, for example, the reference date, the number of reference customers, and the number of reference sales. The reference date is data corresponding to the date in "Table B". The number of reference customers is data corresponding to the number of sales customers in "Table B". The reference sales number is data corresponding to the sales number in "Table B". The data item as "calculation data" includes, for example, a date difference, a date difference for rank calculation, and a date difference rank. The date difference is data indicating the value of the number of days obtained by subtracting the reference date from the date. The date difference for rank calculation is a parameter that serves as a reference when calculating the date difference rank. For the date difference for rank calculation, for example, when the date difference is a positive value, the value of the date difference is given, and when the date difference is a negative value, a predetermined value is given to the absolute value of the date difference (this example). Then, a value obtained by adding "56") is given. That is, in the embodiment, when the reference date is before the date, it is used with priority over the case where the reference date is after the date. In other words, if the date difference is negative (data of a date in the future than the out-of-stock time is used), the priority of the data of the past date is increased by making the date difference larger than the past date difference. .. If the date difference is negative, the number of 56 days to be added to make the date difference larger than the past data uses the data from the shortage date to the ± 8th week (that is, ± 56th day). This is an example of the case where 56 days are added in order to increase the priority of the past data. In other words, even if you use sales data (purchase rate base, customer unit price base) for a period that is too far away, the estimation result is not credible, so the data used is for the last 8 weeks (both past and future from the time of shortage). There is. There is no particular limit on the ± 8th week, but it must be a number of weeks that is not lacking in credibility. The date difference rank is data that defines the priority of each record of the missed sale amount estimation data 172 when calculating the assumed purchase rate and the number of missed sells by individual item. In this example, the missed-sale calculation unit 122 extracts data having a date difference rank of "4" or less, and by aggregating and using the extracted data, it corresponds to a time zone when the product is out of stock. Calculate the estimated purchase rate and the number of missed items sold separately.

Next, the demand amount calculation unit 124 for each individual item calculates the estimated number of sales based on the purchase rate by integrating the number of purchasers with respect to the estimated purchase rate. In addition, the demand amount calculation unit 124 for each individual item calculates the number of missed sales based on the purchase rate by integrating the out-of-stock time ratio with respect to the estimated number of sales based on the purchase rate. In addition, the demand amount calculation unit 124 for each individual item calculates the amount of missed sales based on the purchase rate by integrating the unit selling price of the product with respect to the number of missed sales based on the purchase rate. Then, the demand amount calculation unit 124 for each individual item totals the total demand amount for each individual item, which is the sum of the total value of the oversold amount based on the purchase rate and the total sales amount of the products for a plurality of products belonging to the category. Calculate the amount of missed sales by category (based on customer unit price).

The demand amount calculation unit 126 for each category calculates the estimated sales amount based on the customer unit price by integrating the number of purchasing customers with the estimated customer unit price. The customer unit price is calculated, for example, by dividing the sales amount of the product by the number of purchasing customers. Further, when the estimated sales amount based on the customer unit price is larger than the sales amount, the demand amount calculation unit 126 for each category sets the amount obtained by subtracting the sales amount from the estimated sales amount based on the customer unit price as the estimated sales amount based on the customer unit price. Calculate the amount of missed sales (based on the unit price per customer). On the other hand, when the estimated sales amount based on the customer unit price is less than or equal to the sales amount, the demand amount calculation unit 126 for each category sets the estimated sales amount based on the customer unit price as "0 (zero)" and sets the sales amount by category (customer unit price basis). Is calculated.

The demand amount calculation unit 126 for each category calculates the oversold amount based on the customer unit price by integrating the out-of-stock time ratio with the estimated sales amount based on the customer unit price. Then, the demand amount calculation unit 126 for each category totals the demand amount for each category by adding the total value of the oversold amount based on the customer unit price and the total sales amount of the products for the plurality of products belonging to the category. Calculate as.

The evaluation unit 128 corrects the oversold amount for each item based on the total demand amount for each item calculated by the demand amount calculation unit 124 for each item and the total demand amount for each category calculated by the demand amount calculation unit 126 for each category. To do. In the example shown in FIG. 14, the evaluation unit 128 calculates the shortage period category sales by adding the CLS (Category Lost Sales), which is the amount of missed sales for each item in the shortage period, to the category sales CAS (Category Actual sales) in the shortage period. Out of stock period estimated from category sales on the same day of the week and the same time zone that is not out of stock If the amount of category demand CD (Category Demand) is exceeded, sold separately during the out of stock period so as to suppress the excess The missed amount CLS is compressed and corrected. For example, the evaluation unit 128 calculates the oversold amount CLS for each item in the out-of-stock period after correction by multiplying the oversold amount CLS by individual item in the out-of-stock period before amendment by a correction coefficient (CD / (CAS + CLS)). To do. The evaluation unit 128 compresses and corrects the oversold amount CLS for each item during the shortage period in consideration of the influence of the purchase of alternative products for the products to which "category purchase" is given as the calculation particle size. It should be noted that the individual item-specific oversold amount CLS compressed and corrected by the evaluation unit 128 may or may not be used for the demand forecast of the product by the demand forecast unit 212.

[Pre-processing of product demand forecast]
Next, the preprocessing of the demand forecast of the product will be described.

As shown in FIG. 15, the demand forecasting unit 212 first totals the number of sales recorded for each sales time of the product for each sales time zone and calculates the number of sales for each sales time zone. The sales time zone may be, for example, an hour unit, a daily unit, or a week unit. For example, the demand forecasting unit 212 totals the number of sales of products having a common product code for each sales time zone, and calculates the number of sales of the products for each sales time zone.

Next, as shown in FIG. 16, the demand forecasting unit 212 adds up the number of missed sales for each sales time zone of the product to the number of sales for each sales time zone of the product, so that each sales time zone of the product is added. Correct the number of sales.

The demand forecast unit 212 trains the demand forecast model by machine learning using the learning data including the number of sales of the corrected product for each sales time zone. Machine learning techniques include, for example, multiple regression analysis, random forest, and support vector machines. When the demand forecasting unit 212 learns the demand forecasting model, the demand forecasting unit 212 may or may not use the data of the sales time zone in which the corrected number of sales is "0 (zero)" as the learning data.

When learning the demand forecast model, the demand forecasting unit 212 totals the number of sales recorded for each sales time of the product for each predetermined period and calculates the number of sales for each predetermined period. The predetermined period is a period longer than the length of the sales time zone described above. For example, when the sales time zone is one hour unit, the predetermined period is one day unit, and when the sales time zone is one day unit, the predetermined period is one week unit. When the number of sales in a predetermined period is "0 (zero)", the demand forecast unit 212 does not use the number of sales for each sales time included in the predetermined period as learning data.

More specifically, as shown in FIG. 17, when the amount of data of the sales record of the product stored in the storage unit 230 is smaller than the reference value set by the user of the system, the demand forecast unit 212 may use the demand forecast unit 212. By machine learning using the sales performance of products for each product group as learning data, a machine learning model for each product group is trained as an example of the demand forecast model M1. The machine learning model for each product group is a model in which the sales performance of products for each product group is used as learning data and trained by machine learning. In the illustrated example, the data items of the sales performance of the products for each product group include the major classification, the middle classification, the minor classification, the product code, the weather, the date, the day of the week, the time zone, the selling price, and the number of sales. The major classification, middle classification, and minor classification are data items that indicate the classification of products by hierarchy, and are used to narrow down the products included in the product group. The weather, date, day of the week, time zone, and selling price are data items used for input to the demand forecast model and are used as learning data. The number of sales is a data item output from the demand forecast model M1 and is used as teacher data.

When the amount of product sales record data stored in the storage unit 230 is larger than the reference value set by the system user, the demand forecast unit 212 uses the product sales record for each product as learning data. By the machine learning used, the machine learning model for each product is trained as an example of the demand forecast model. The product-specific machine learning model is a model in which the sales performance of a product for each product is used as learning data and trained by machine learning. In the illustrated example, the data items of the sales performance for each product include the product code, the weather, the date, the day of the week, the time zone, the selling price, and the number of sales. The weather, date, day of the week, time zone, and selling price are data items used for input to the demand forecast model and are used as learning data. The number of sales is a data item output from the demand forecast model and is used as teacher data.

[Product demand forecast]
Next, the first embodiment of the demand forecast of the product will be described.

The demand forecasting unit 212 inputs the forecasted sales time zone to the demand forecast model M1 learned using the learning data including the number of sales of the corrected product for each sales time zone. Make a demand forecast.

When forecasting the demand for a product, the demand forecasting unit 212 changes the demand forecasting model M1 according to the amount of data of the sales record data 232 stored in the storage unit 230. For example, the demand forecasting unit 212 predicts the demand for a product by using the demand forecasting model M1 with high forecasting accuracy as the amount of data of the sales record data 232 stored in the storage unit 230 increases. For example, when the amount of data of the sales record data 232 stored in the storage unit 230 is smaller than the reference value set by the user of the system, the demand forecast unit 212 demands for each product group including a plurality of products. When the demand forecast of the product is performed using the forecast model M1 and the amount of data of the sales record data 232 stored in the storage unit 230 is larger than the reference value set by the system user, the demand forecast for each product is performed. The demand forecast of the product is performed using the model M1. Further, for example, when the amount of data of the sales record data 232 stored in the storage unit 230 is smaller than the reference value set by the user of the system, the demand forecast unit 212 statistics the sales record data 232 of the product. When the demand forecast of the product is made using the statistical model analyzed in the above manner and the amount of data of the sales record data 232 stored in the storage unit 230 is larger than the reference value set by the system user, the product Demand forecast of products is performed using a machine learning model that uses sales performance data 232 as learning data.

When the amount of sales performance data is less than the first threshold value, the demand forecasting unit 212 forecasts the demand for products using the average model for each product group.

When the amount of sales performance data is equal to or greater than the first threshold value and less than the second threshold value, the demand forecasting unit 212 forecasts the demand for the product using the product-specific average model. The average model for each product is a model that uses the average value of the sales performance of each product. The average model by product is a model in which the target of data is narrow and the prediction method is simple.

When the amount of sales performance data is equal to or greater than the second threshold value and less than or equal to the third threshold value, the demand forecasting unit 212 forecasts the demand for the product using the machine learning model for each product group. The machine learning model for each product group is a model in which the target of data is wide and the prediction method is complicated.

When the amount of sales performance data is equal to or greater than the third threshold value, the demand forecasting unit 212 forecasts the demand for the product using the machine learning model for each product. The product-specific machine learning model is a model in which the target of data is narrow and the prediction method is complicated.

That is, as the amount of sales performance data increases, the demand forecasting unit 212 uses the demand forecasting model for product demand forecasting as an average model for each product group, an average model for each product group, a machine learning model for each product group, and a product. Switch in the order of another machine learning model. However, the demand forecasting unit 212 may switch the priority order of the product-specific average model and the product group-specific machine learning model according to the amount of sales performance data. Then, when the amount of data of the sales record data 232 stored in the storage unit 230 is small, the demand forecast unit 212 predicts the demand of the product using the demand forecast model M1 for each product group including a plurality of products. When the amount of data of the sales record data 232 stored in the storage unit 230 is large, the demand forecast model M1 for each product is used to forecast the demand for the product. Further, when the amount of data of the sales record data 232 stored in the storage unit 230 is small, the demand forecast unit 212 predicts the demand of the product using a statistical model that statistically analyzes the sales record of the product. When the amount of sales record data stored in the storage unit 230 is large, the demand forecast of the product is performed using the machine learning model trained by using the product sales record data 232 as learning data.

Next, the processing flow of the first embodiment of the demand forecast of the product will be described.

As shown in FIG. 18, the demand forecasting unit 212 first acquires weather information (step S10).

Next, the demand forecasting unit 212 acquires calendar information (step S12).

Next, the demand forecasting unit 212 acquires the selling price information (step S14).

Next, the demand forecasting unit 212 determines whether or not the amount of data of the product sales record data 232 stored in the storage unit 230 is equal to or greater than the first threshold value (step S16). When the demand forecasting unit 212 determines that the amount of data of the sales performance of the product is less than the first threshold value, the demand forecasting unit 212 forecasts the demand of the product using the average model for each product group (step S18). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the data amount of the product sales record data 232 is equal to or greater than the first threshold value, the demand forecast unit 212 determines whether or not the data amount of the product sales record data 232 is equal to or greater than the second threshold value. Step S20). When the demand forecasting unit 212 determines that the amount of data of the product sales performance data 232 is less than the second threshold value, the demand forecasting unit 212 forecasts the demand for the product using the product-specific average model (step S22). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the data amount of the product sales record data 232 is equal to or greater than the second threshold value, the demand forecast unit 212 determines whether or not the data amount of the product sales record data 232 is equal to or greater than the third threshold value. Step S24). When the demand forecasting unit 212 determines that the amount of data of the product sales performance data 232 is less than the third threshold value, the demand forecasting unit 212 predicts the demand of the product using the machine learning model for each product group (step S26). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the amount of data of the sales performance of the product is equal to or greater than the third threshold value, the demand forecasting unit 212 predicts the demand of the product using the machine learning model for each product (step S28). This completes the processing of this flowchart.

Next, a second embodiment of product demand forecast will be described.

When the demand forecasting unit 212 uses the product sales record data 232 as learning data to forecast the demand for the product using the machine learning model, the amount of data of the sales record data 232 stored in the storage unit 230 is a predetermined amount. If it is less than, the demand forecast of the product is performed using the first machine learning model, and if the amount of data of the sales record data 232 stored in the storage unit 230 is equal to or more than a predetermined value, the second machine learning is performed. Forecast product demand using a model. The second machine learning model is a model trained by machine learning using learning data having a larger amount of data than the first machine learning model.

More specifically, as shown in FIG. 19, when the demand forecasting unit 212 does not accumulate the product sales record data and the data amount of the product sales record data 232 is less than the fourth threshold value, , The first machine learning model is learned as a machine learning model for each product by machine learning using the product sales performance data 232 for each product. In the example of FIG. 19, as the data items of the sales performance data 232 for each product, the product code, the weather, the date, the day of the week, the time zone, the selling price, and the number of sales are included. The weather, date, day of the week, time zone, and selling price are data items used for input to the demand forecast model and are used as learning data. The number of sales is a data item output from the demand forecast model and is used as teacher data.

When the product sales record data 232 is accumulated and the data amount of the product sales record data 232 exceeds the fourth threshold value, the demand forecasting unit 212 uses machine learning using the product sales record data 232 for each product. , Train the second machine learning model as a product-specific machine learning model. In the illustrated example, the data items of the sales performance data 232 for each product include the product code, the weather, the date, the day of the week, the time zone, the selling price, the number of sales, and the statistical value of the selling price. The statistical value of the selling price is a value obtained by statistically processing the selling price of a plurality of data having the same product code, weather, day of the week, and time zone, and is, for example, an average value of the selling price. The selling price statistical value may be a variance value of the selling price. The product code, weather, date, day of the week, time zone, selling price, and average value of selling price are data items used for input to the demand forecast model and are used as learning data. The number of sales is a data item output from the demand forecast model and is used as teacher data.

Next, the processing flow of the second embodiment of the demand forecast of the product will be described.

As shown in FIG. 20, the demand forecasting unit 212 first acquires weather information (step S30).

Next, the demand forecasting unit 212 acquires calendar information (step S32).

Next, the demand forecasting unit 212 acquires the selling price information (step S34).

Next, the demand forecasting unit 212 determines whether or not the amount of data of the product sales record data 232 stored in the storage unit 230 is equal to or greater than the first threshold value (step S36). When the demand forecasting unit 212 determines that the amount of data of the product sales performance data 232 is less than the first threshold value, the demand forecasting unit 212 forecasts the demand for the product using the average model for each product group (step S38). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the data amount of the product sales record data 232 is equal to or greater than the first threshold value, the demand forecast unit 212 determines whether or not the data amount of the product sales record data 232 is equal to or greater than the second threshold value. Step S40). When the demand forecasting unit 212 determines that the amount of data of the product sales performance data 232 is less than the second threshold value, the demand forecasting unit 212 forecasts the demand for the product using the product-specific average model (step S42). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the data amount of the product sales record data 232 is equal to or greater than the second threshold value, the demand forecast unit 212 determines whether or not the data amount of the product sales record data 232 is equal to or greater than the third threshold value. Step S44). When the demand forecasting unit 212 determines that the amount of data of the product sales performance data 232 is less than the third threshold value, the demand forecasting unit 212 predicts the demand of the product using the machine learning model for each product group (step S46). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the data amount of the product sales record data 232 is equal to or greater than the third threshold value, the demand forecast unit 212 determines whether or not the data amount of the product sales record data 232 is equal to or greater than the fourth threshold value. Step S48). When the demand forecasting unit 212 determines that the amount of data of the product sales performance data 232 is less than the fourth threshold value, the demand forecasting unit 212 forecasts the demand for the product using the first machine learning model as the machine learning model for each product (step S50). ). This completes the processing of this flowchart.

When the demand forecasting unit 212 determines that the amount of data of the sales performance of the product is equal to or greater than the fourth threshold value, the demand forecasting unit 212 predicts the demand of the product using the second machine learning model as the machine learning model for each product (step S52). This completes the processing of this flowchart.

[Calculation of standard inventory amount]

Next, the processing flow of the reference inventory calculation unit 214 will be described with reference to the flowchart of FIG.

First, the reference inventory calculation unit 214 calculates the unit unsold loss L1 of the product based on the sales performance of the product (step S60).

Next, the reference inventory calculation unit 214 calculates the unit sales loss L2 of the product by subtracting the purchase cost of the product from the regular selling price per unit of the product (step S62).

Next, the reference inventory calculation unit 214 calculates the demand forecast residual distribution of the product by subtracting the demand forecast value for each time zone from the actual sales value for each time zone and calculating the residual for each time zone. (Step S64).

Next, the reference inventory calculation unit 214 calculates the expected loss amount L of the product based on the demand forecast residual distribution of the product, the unit unsold loss L1 of the product, and the unit unsold loss L2 of the product. The inventory amount that minimizes the expected loss amount L is calculated as the reference inventory amount of the product (step S66).

Next, the reference inventory calculation unit 214 determines whether or not all the products have been searched (step S68). When the reference inventory calculation unit 214 determines that all the products have not been searched, the target product is changed (step S70), and the process is returned to step S60. Then, the reference inventory calculation unit 214 repeats the processes of steps S60 to S70 until the reference inventory amount of the products is calculated for all the products. When the reference inventory calculation unit 214 determines that all the products have been searched, the processing of this flowchart ends.

Next, a first example of the calculation process of the standard inventory amount will be described.

The unit unsold loss L1 indicates the average loss amount when one unit of the product is unsold during the selling time unit such as one day or one hour. The unit unsold loss L1 is determined, for example, based on the sales performance of the product. The unit unsold loss L1 is the total number of sales and the total discount loss when the product is sold at the discounted selling price, the total number of disposal and the total disposal loss when the product is discarded, and the product carried over to the next sales time unit. Determined based on numbers.

FIG. 22 is a diagram showing an example of a method for calculating the unit unsold loss L1. In FIG. 22, the sales time unit is the time from the opening time to the closing time of the store, and the inventory transition of the product is distinguished for each time zone in which the product is manufactured. In addition, the selling price of the product is gradually discounted as the selling time elapses from the time when the product is manufactured. The unit unsold loss L1 in such a case is calculated using the following formula (1). That is, the total loss due to discount and the total loss due to disposal are divided by the total number of products sold at a discount in units of sales time, the number of products discarded, and the number of products carried over to the next sales time unit, and the average unsold loss amount is calculated. I'm calculating. In addition, in the formula (1), the average may be calculated for the right side in a period consisting of a plurality of sales time units.

The unit unsold loss L1 is calculated by using, for example, the following formula (2) when the product is a product that has not been discarded or sold at a discount, or when data regarding the disposal or discount sales of the product is not stored. , The cost of the product may be calculated by dividing by the number of units of the sales period of the product.

The unit sales loss L2 indicates the average loss amount when a sales opportunity of one product is missed due to a shortage. The unit sales loss L2 is calculated by subtracting the purchase cost of the product from the regular selling price per unit of the product.

Next, the reference inventory calculation unit 214 creates a demand forecast residual distribution. The demand forecast residual distribution is the distribution of residuals for each sales time unit. The reference inventory calculation unit 214 calculates the difference data between the actual sales value of the product in the sales time unit and the demand forecast value as the residual, and creates the distribution of the residual in each sales time unit as the demand forecast residual distribution. The demand forecast value is a value based on the demand forecast of the product by the demand forecast unit 212. Here, the actual sales value may be a value corrected by adding the missed sales in the demand forecasting unit 212.

FIG. 23 is a diagram showing a frequency distribution which is an example of a demand forecast residual distribution. In the illustrated example, the horizontal axis is the residual k and the vertical axis is the frequency h (k). The frequency h (k) indicates the appearance frequency of the residual k for each time zone. In the illustrated example, the frequency h (0) is the largest when the residual is “0 (zero)”, and the larger the absolute value of the residual, the smaller the frequency h (k) tends to be.

FIG. 24 is a diagram showing a probability distribution which is an example of a demand forecast residual distribution. In the illustrated example, the horizontal axis is the residual r and the vertical axis is the probability p (r). The probability p (r) indicates the appearance probability of the residual r. The probability p (r) is calculated from the average value μr of the residuals and the standard deviation value σr, assuming that the residual r is a normal distribution, using the following equation (3).

Next, the reference inventory calculation unit 214 calculates the expected loss amount L based on the unit unsold loss L1 of the product, the unit unsold loss L2 of the product, and the demand forecast residual distribution of the product. The expected loss amount L indicates the expected value of the product loss amount based on the demand forecast of the product.

FIG. 25 shows a frequency distribution h (k) which is an example of a demand forecast residual distribution of a product, a loss on sale of the product kL2, and a sale of the product when the inventory amount of the product is the same as the demand forecast amount. It is the figure which superposed the residual loss amount-kL1. In this example, the expected loss amount L is calculated using the following formula (4).

That is, when the residual k is a positive value, the amount of lost sales is kL2, and the probability of occurrence is h (k) / Σh (n). When the residual k is a negative value, the unsold loss amount is −kL1, and the probability of occurrence is h (k) / Σh (n). Therefore, the expected loss amount L is calculated by adding up the value obtained by multiplying the unsold loss amount or the unsold loss amount for each residual k by the probability of occurrence.

FIG. 26 shows a frequency distribution h (k), which is an example of a demand forecast residual distribution of goods, and a loss on sale of goods (ky) when the inventory amount of goods is y more than the demand forecast amount. ) L2 and the unsold loss amount − (ky) L1 are shown in an overlapping manner. In this example, the expected loss amount L is calculated using the following formula (5).

That is, when the residual k is larger than y, the oversold loss amount is (ky) L2, and the probability of occurrence is h (k) / Σh (n). When the residual k is smaller than y, the unsold loss amount is − (ky) L1, and the probability of occurrence is h (k) / Σh (n). Therefore, the expected loss amount L (y) is calculated by adding up the value obtained by multiplying the unsold loss amount or the unsold loss amount for each residual k by the probability of occurrence.

FIG. 27 shows the probability distribution p (r), which is an example of the demand forecast residual distribution of the product, the unsold loss amount rL2 of the product, and the unsold loss amount-when the inventory amount of the product is the same as the demand forecast amount. It is the figure which showed by superimposing rL1. In this example, the expected loss amount L is calculated using the following formula (6).

That is, when the residual r is a positive value, the oversold loss amount is rL2, and the probability of occurrence is p (r). When the residual r is a negative value, the unsold loss amount is −rL1 and the probability of occurrence is p (r). Therefore, the expected loss amount L is calculated by integrating the value obtained by multiplying the unsold loss amount or the unsold loss amount of the residual r and the occurrence probability.

FIG. 28 shows a probability distribution p (r), which is an example of a demand forecast residual distribution of a product, and a lost sale amount (ry) of the product when the inventory amount of the product is y more than the demand forecast amount. ) L2 and the unsold loss amount − (ry) L1 are shown in an overlapping manner. In this example, the expected loss amount L is calculated using the following formula (7).

That is, when the residual r is larger than y, the oversold loss amount is (ry) L2, and the probability of occurrence is p (r). When the residual r is smaller than y, the unsold loss amount is − (ry) L1, and the probability of occurrence is p (r). Therefore, the expected loss amount L (y) is calculated by integrating the value obtained by multiplying the unsold loss amount or the unsold loss amount of the residual r and the occurrence probability.

Next, the reference inventory calculation unit 214 searches for an addition amount y that minimizes the expected loss amount L. As a specific example, the addition amount y is changed by adding 1 from the value obtained by multiplying the demand forecast amount of the product predicted by the demand forecast unit 212 by -1 to the value of the demand forecast amount, and the expected loss amount. L is calculated to find the addition amount y that minimizes the expected loss amount L. Then, the reference inventory calculation unit 214 calculates the reference inventory amount of the product by adding the searched addition amount y to the demand forecast amount of the product predicted by the demand forecast unit 212.

Next, as a second example of the calculation process of the reference inventory amount, a method of calculating the addition amount y that minimizes the expected loss amount will be described. The calculation of the unit unsold loss L1 of the product, the calculation of the unit unsold loss L2 of the product, and the calculation of the demand forecast residual distribution of the product are common to the first example.

The standard inventory calculation unit 214 missed the sales opportunity of one product due to the unit unsold loss L1 which is the average loss amount when one unit of the same product is left unsold during the sales time unit and the shortage. The ratio of the unit sales loss L2 to the total value with the unit sales loss L2, which is the average loss amount in the case, is calculated. In addition, the reference inventory calculation unit 214 arranges the residuals in a plurality of sales time units in ascending order. Then, when the residuals in the plurality of sales time units are arranged in ascending order, the reference inventory calculation unit 214 is located in the order of the values obtained by integrating the above ratios with respect to the number of residual data in the plurality of sales time units. The value obtained by adding the demand forecast value to the residual amount to be calculated is calculated as the standard inventory amount of the product.

FIG. 29 is a diagram for explaining a second example of the calculation process of the reference inventory amount. In the illustrated example, the frequency distribution is calculated as the demand forecast residual distribution.

First, the reference inventory calculation unit 214 calculates the difference data between the sales actual value and the demand forecast value of the product in the sales time unit as the residual k, and sorts the residual k in ascending order (ascending order).

Next, the reference inventory calculation unit 214 calculates the ratio (= L2 / (L1 + L2)) of the unit unsold loss to the total value of the unit unsold loss L1 and the unit unsold loss L2 for the same product.

Next, the reference inventory calculation unit 214 determines the residual k at the [ratio rate x number of residual data N] th-th residual k, counting from the smallest, as the addition amount y that minimizes the expected loss amount L.

Then, the reference inventory calculation unit 214 calculates the reference inventory amount of the product by adding the addition amount y determined as described above to the demand forecast amount of the product predicted by the demand forecast unit 212.

When an outlier is included in the residuals in a plurality of sales time zones, the reference inventory calculation unit 214 calculates the reference inventory amount of the product based on the residual excluding the outliers. For example, the reference inventory calculation unit 214 sets a constant multiple of the interquartile range of the residual k as the permissible range of the residual k, and determines that the residual k outside the permissible range is an abnormal value and excludes it.

The calculation method of the standard inventory amount when the demand forecast residual distribution is expressed by the probability distribution will be described.

The standard inventory calculation unit 214 missed the sales opportunity of one product due to the unit unsold loss L1 which is the average loss amount when one unit of the same product is left unsold during the sales time unit and the shortage. The ratio of the unit sales loss L2 to the total value with the unit sales loss L2, which is the average loss amount in the case, is calculated. Further, the reference inventory calculation unit 214 sets the value obtained by adding the demand forecast value to the residual value in which the value of the cumulative distribution function obtained by integrating the probability distribution representing the demand forecast residual distribution matches the ratio as the reference of the product. Calculated as inventory.

FIG. 30 is a diagram for explaining a third example of the calculation process of the reference inventory amount. In the illustrated example, a cumulative distribution function that integrates the probability distributions that represent the demand forecast residual distribution is shown.

First, the reference inventory calculation unit 214 calculates the ratio rate (= L2 / (L1 + L2)) of the unit unsold loss to the total value of the unit unsold loss L1 and the unit unsold loss L2 for the same product.

Next, the reference inventory calculation unit 214 generates a cumulative distribution function that integrates the probability distribution representing the demand forecast residual distribution, and the expected loss amount L determines the residual r whose value of the cumulative distribution function matches the ratio rate. It is determined as the minimum addition amount y.

Then, the reference inventory calculation unit 214 calculates the reference inventory amount of the product by adding the addition amount y determined as described above to the demand forecast amount of the product predicted by the demand forecast unit 212.

In reality, it is difficult to calculate the residual r in which the value of the cumulative distribution function matches the ratio rate, so the standard normal distribution table is used and the probability (area) of the standard normal distribution matches the ratio rate. The value x of the random variable can be obtained, and the residual r can be obtained from the average value ave of the demand prediction residual distribution and the standard deviation value sigma. The residual r is calculated using, for example, the following equation (8).

[Calculation of recommended order quantity]
Next, a first calculation example of the recommended order quantity of the product will be described.

The ordering recommended quantity calculation unit 216 calculates the ordering recommended quantity of the product using the following formula (9.1).

ORR (X) is a function that outputs the recommended order quantity of the date with the date X as a parameter. OIV (X) is a function that outputs the reference inventory amount at the time of opening, which minimizes the loss on the date X. LT (X) is a function that outputs the lead time when an order is placed on the date X. Lead time means the time required from ordering to delivery. DLV is an integer representing the number of days from delivery to the state where it can meet the demand of consumers. DLV is determined, for example, depending on the flight from the distribution center or process center to the store. For example, "0" is given for early courier services that can be used for demand on the day of delivery, and "1" is given for flights that are available for demand on the day after delivery. It is a parameter to be given. IVF (X) is a function that outputs the estimated inventory amount at the time of opening the store on the date X. INT (X) is a parameter indicating the interval from the day of ordering to the next ordering. DF (T) is a function that outputs the predicted number of demands for each date. That is, the recommended order quantity on date X is the difference between the standard inventory quantity and the predicted inventory quantity after the date LT (X) + DLV from the time the order is placed to the time when the product is lined up at the store. Calculated as the sum of the estimated demand numbers for each date before they are lined up in stores. Hereinafter, for the sake of brevity, the time difference between delivery and lining up at the store may be omitted, and the time after the date LT (X) + DLV may be expressed as the delivery timing.

Here, the first to second terms on the right side of the equation (9.1) are the differences between the standard inventory amount and the reserve inventory amount at the delivery timing, and the first and second terms on the right side for obtaining this difference. The term is calculated to obtain the amount for which the reserve stock amount is insufficient with respect to the standard stock amount. Therefore, although the method explanation is omitted in the equation (9.1) so as not to complicate the method explanation, in the actual implementation of the program code, when the calculation results of the first and second terms on the right side are negative numbers. You may add the process of replacing it with zero.

Although X in the formula (9.1) has been described as a date, X may represent a time zone when used as a calculation of the recommended number of products to be produced in units of time. In that case, T, LT, and DLV are also used as time zones instead of dates. That is, the recommended production number ORR (X) in the time zone X is the time from the production instruction to the time when the product is lined up at the store. It is calculated as the quantity obtained by adding the estimated number of demands for each time zone before the next production product is lined up in the store.

In addition, the equation (9.1) may include an operation corresponding to consideration of the production in the store. For example, if you want to keep at least P inventories to be lined up in the store, compare P with the calculation result of the first term OIV on the right side of equation (9.1), and make the first term on the right side larger. You may replace it with.

In addition, for example, the number of orders or the number of production may need to be a multiple of an integer determined by the product for reasons such as production lots. In that case, the number calculated by the formula (9.1) is the integer. The recommended quantity for ordering or the recommended quantity for production may be rounded up to a multiple of.

FIG. 31 is a diagram showing a process of calculating the recommended order quantity when the DLV is “0 (zero)”. In the illustrated example, the goods are ordered on date X. Then, the ordered product is delivered on the date "X + LT (X)" on which the lead time LT (X) has elapsed from the date X. Since the delivered product has a DLV of "0 (zero)", the product is displayed in the store on the date "X + LT (X)" which is the day of delivery. Further, on the date "X + INT (X)" when INT (X) has passed from the date "X", the next product is ordered. Then, on the date "X + INT (X) + LT (X + INT (X))" on which the lead time LT (X + INT (X)) has elapsed from the date "X + INT (X)", the next ordered product is delivered. In addition, the product is displayed in the store on the date "X + INT (X) + LT (X + INT (X))" which is the day of the next delivery. The ordering recommended quantity calculation unit 216 subtracts the predicted inventory quantity at the time of opening from the standard inventory quantity at the time of opening on the date “X + LT (X)”. In addition, the order recommendation quantity calculation unit 216 sets the deducted value from the date "X + LT (X) +1" which is the day after the first delivery day to the date "X + INT (X) + LT" which is the day before the next product display date. The recommended quantity for ordering the product is calculated by adding the forecast quantity up to X + INT (X) -1 "and the inventory quantity that should be left after the product is sold according to the demand forecast for the purpose of effect effect.

FIG. 32 is a diagram showing a process of calculating the recommended order quantity when the DLV is “1”. In the illustrated example, the goods are ordered on date X. Then, the ordered product is delivered on the date "X + LT (X)" on which the lead time LT (X) has elapsed from the date X. Since the delivered product has a DLV of "1", the product is displayed in the store on the date "X + LT (X) + 1", which is the day after the delivery. Further, on the date "X + INT (X)" when INT (X) has passed from the date "X", the next product is ordered. Then, on the date "X + INT (X) + LT (X + INT (X))" on which the lead time LT (X + INT (X)) has elapsed from the date "X + INT (X)", the next ordered product is delivered. In addition, the product is displayed in the store on the date "X + INT (X) + LT (X + INT (X)) + 1", which is the day after the next delivery. The ordering recommended quantity calculation unit 216 subtracts the predicted inventory quantity at the time of opening from the standard inventory quantity at the time of opening on the date “X + LT (X) +1”. In addition, the order recommendation quantity calculation unit 216 sets the deducted value from the date "X + LT (X) + 2", which is the day after the product is displayed in the store, to the date "X + INT (X)", which is the day before the next product display date. ) + LT (X + INT (X)) ”and the inventory amount that should be left after the product is sold according to the demand forecast for the purpose of effect effect, the recommended quantity for ordering the product is calculated.

The ordering recommended quantity calculation unit 216 calculates the IVF of the second term on the right side of the equation (9.1), that is, the predicted inventory quantity, using the following equation (9.2). Equation (9.2) represents a method of calculating the future inventory forecast amount IVF (X + P) P days after the date X on the date X.

Here, IV (X) of the first term on the right side of the equation (9.2) is the inventory calculation result at the time of opening on the date X, and the calculation method is calculated according to the theoretical inventory calculation method already described. The RVF (T) in the second term on the right side of the equation (9.2) is the estimated number of deliveries on the date T, and is calculated based on, for example, the order record before the date X. That is, the formula (9.2) sets the estimated number of stocks at the time of opening of the store after P days have passed from the date X, the number of stocks at the time of opening of the date X, and the planned number of deliveries in the period from the date X to the day before the date X + P. In addition, it means that it is calculated by further subtracting the number of demand forecasts from the date X to the day before the date X + P.

Further, here, the formula (9.2) is a formula for obtaining the estimated number of stocks at the time of opening the store in the future after the lapse of P days from the date X. In equation (9.2), the method explanation is omitted so as not to complicate it, but in the actual implementation of the program code, if the calculation result on the right side is a negative number, it may be replaced with zero. ..

[Calculation of recommended quantity for ordering indefinite products]
When calculating the recommended order quantity of indefinite product, which is a product whose contents of each pack are not the same, the recommended order quantity calculation unit 216 describes the unit corresponding to the contents, that is, the weight or the like as a unit. It is necessary to correct the recommended order quantity given by the procedure in consideration of packing. Therefore, the ordering recommended quantity calculation unit 216 first converts the number of sales packs for each content of the product into an integer value closest to the numerical value based on the ratio of the sales performance for each content of the product. Next, the ordering recommended quantity calculation unit 216 calculates the converted value of the sales volume for each content content of the product by integrating the content content of the product with the obtained number of sales packs. Then, the ordering recommended quantity calculation unit 216 is obtained in advance from the converted value of the number of sales packs and the sales amount for each content of the product so that the difference from the ordering recommended quantity obtained in advance in units of weight becomes small. The recommended order quantity is corrected, and the corrected value is output as the recommended order quantity after correction of the product.

FIG. 33 is data showing an example of the ordering recommended amount calculation data 390 for calculating the ordering recommended amount in units of the weight of the indefinite product. The data items of the order recommended quantity calculation data 390 are, for example, the content V, the actual sales volume ratio P, the product of the content V and the actual sales volume ratio P, and the product of the content V and the actual sales volume ratio P. Includes ratio, recommended order quantity M before correction, number of packs N, number of rounded packs N, conversion value of sales volume, error, and remarks. The content V is data indicating the type of the content of the product. The actual sales volume ratio P is data showing the ratio of the actual number of sales for each content of the product based on the actual sales of the product. The ratio of the product of the content V and the actual sales volume ratio P is the ratio of the product of the content V and the actual sales volume ratio P normalized in the numerical range of 0 to 1 so that the total is 1. is there. The recommended order quantity M before correction of each content is a value obtained by multiplying the given recommended order quantity by the ratio of V × P. The number of packs N is data obtained by dividing the recommended order quantity M before correction by the content volume V. The converted value of the sales amount is a value obtained by integrating the rounded number of packs N and the content V. The error is a value obtained by subtracting the recommended order quantity M before correction from the converted value of the sales volume. The remarks are data showing additional information regarding the error value. In the illustrated example, the remarks are labeled "plus maximum" for the data with the largest absolute value of the positive error and "minus" for the data with the largest absolute value of the negative error. The "maximum" label is given.

When the total value of the converted values of the sales volume for each product content is smaller than the pre-obtained ordering recommended quantity, the ordering recommended quantity calculation unit 216 targets the content of the product with a smaller amount than the difference. As a result, the conversion value of the number of packs N and the sales volume is increased. In the example shown in FIG. 33, when the total error is negative, the ordering recommended quantity calculation unit 216 adds the number of packs N and the converted value of the sales volume for the content that does not exceed the absolute value of the total error. To do. Further, in the example shown in FIG. 33, the ordering recommended quantity calculation unit 216 gives priority to the content having a high actual sales volume ratio among the contents that do not exceed the absolute value of the error, and gives priority to the number of packs N and the sales volume. Add the converted value of. In this example, the ordering recommended quantity calculation unit 216 has a negative absolute value of "200", so that the content "200", which has a high actual sales volume ratio among the contents "100" and "200" that do not exceed the remainder. Priority is given to the converted value of the sales volume. Then, the ordering recommended amount calculation unit 216 sets the total value of the converted values of the sales amount as the ordering recommended amount after correction with respect to the previously obtained ordering recommended amount.

When the total value of the converted values of the sales volume for each content of the product is smaller than the given recommended quantity of the order, the difference is smaller than the minimum value of the content of the product. The total value of the converted values of the sales volume is used as the recommended order quantity after correction with respect to the recommended order quantity obtained in advance, without increasing the number of sales packs and the conversion value of the sales volume.

When the total value of the converted values of the sales volume for each content of the product is larger than the pre-obtained order-recommended quantity, the order-recommended quantity calculation unit 216 calculates the pre-obtained order-recommended quantity by the difference. Increase. Further, the ordering recommended amount calculation unit 216 sets the increased ordering recommended amount as the corrected ordering recommended amount with respect to the previously obtained ordering recommended amount. In the example shown in FIG. 33, when the total of the errors is positive, the ordering recommended amount calculation unit 216 adds the total of the errors to the pre-obtained ordering recommended amount to obtain the corrected ordering recommended amount.

In this way, for indefinite products, the recommended order quantity given in units corresponding to the content capacity is calculated based on the sales performance ratio for each content capacity, and the recommended order quantity after correction corresponding to each content capacity and the recommended order quantity are calculated. Obtain the adjusted recommended order quantity that matches the total value.

Next, the processing flow of the first calculation example of the recommended quantity for ordering the product will be described.

As shown in FIG. 34, the ordering recommended quantity calculation unit 216 first updates the demand forecast quantity and the reference inventory quantity for a predetermined period after the ordering day (step S80).

Next, the ordering recommended quantity calculation unit 216 calculates each starting inventory quantity on the day of ordering by using the theoretical inventory calculation method (step S82).

Next, the ordering recommended quantity calculation unit 216 calculates each start-time predicted inventory quantity for a predetermined period after the day of ordering by using the formulas (9.2) to (10.2) (step S84).

Next, the ordering recommended quantity calculation unit 216 calculates the ordering recommended quantity on the day of ordering using the formulas (9.1) to (10.1) (step S86).

Next, the ordering recommended quantity calculation unit 216 determines whether or not the product is an indefinite product that is packed and sold in a plurality of content types (step S88). When the order recommended quantity calculation unit 216 determines that the product is an indefinite product, the order recommended quantity calculation unit 216 corrects and calculates the recommended order quantity based on the pack number division calculation based on the actual sales volume ratio of each content (step S90), and performs the processing. The process proceeds to step S92. On the other hand, when the ordering recommended quantity calculation unit 216 determines that the product is not an indefinite product, the process proceeds to step S92 without going through the process of step S90.

After that, the ordering recommended quantity output unit 218 outputs the ordering recommended quantity calculated by the ordering recommended quantity calculation unit 216 (step S92). This completes the processing of this flowchart.

Although the daily calculation is shown as an example in the explanation of the flow of the calculation process of the recommended order quantity, the same processing step is calculated in the time zone unit in correspondence with the explanation of the calculation formula of the recommended order quantity described above. It may be executed as a process. The same applies to the second calculation example described later.

Further, in the case of daily calculation, the flow of the calculation process of the recommended order quantity may be executed once a day at the same time zone, for example, on a day when the product for which the calculation target of the recommended order quantity can be ordered can be ordered. However, it may be executed multiple times a day. In the case of hourly calculation, for example, it may be executed once an hour, a plurality of times an hour, or once every several hours during a time zone in which the product for which the recommended order quantity is to be calculated can be ordered. The same applies to the second calculation example described later.

Next, a second calculation example of the recommended quantity for ordering the product will be described.

The ordering recommended quantity calculation unit 216 calculates the estimated stock quantity of the ordered products at the future point in time when they are lined up in the store, and calculates the difference from the standard stock quantity at the same future time point as the quantity that is insufficient at the future time point. In addition, the ordering recommended quantity calculation unit 216 calculates the total estimated demand number after the time when the products to be ordered at the present time are lined up in the store and before the time when the products to be ordered next time are lined up in the store as the above-mentioned insufficient quantity. The recommended quantity for ordering the product is calculated by adding it to the number of products. The ordering recommended quantity calculation unit 216 calculates the ordering recommended quantity of the product using, for example, the following formula (10.1).

EX is a parameter representing the time from when a product is displayed in a store until the sales deadline expires. The difference from the first calculation example shown by the formula (9.1) is that the total period of the third term on the right side is calculated in consideration of the sales period of the product. The third term on the right side of equation (10.1) calculates the forecast demand after the future time when the ordered products are lined up in the store, and the difference between the forecast stock quantity of the product at the future time and the standard stock quantity. Is added to. The third term on the right-hand side of equation (10.1) is the total period from the timing when the next ordered product is lined up at the store or the timing when the sales deadline of the product ordered this time expires on date X until the earlier timing. It is supposed to be. The third item on the right side of the formula (10.1) is that if the sales deadline of the product to be ordered this time expires earlier than the timing when the product to be ordered next time is lined up at the store, the quantity to be ordered this time is placed on date X. The amount of demand after the product sales deadline has expired is not added.

The first and second terms on the right side of equation (10.1) perform calculations to obtain the difference between the standard inventory amount and the estimated inventory amount at the delivery timing, and the estimated inventory amount is relative to the standard inventory amount. An operation is performed to find the quantity for which the quantity is insufficient. Therefore, although the method description is omitted in the equation (10.1) so as not to complicate the method explanation, in the actual implementation of the program code, when the calculation results of the first and second terms on the right side are negative numbers. You may add the process of replacing it with zero.

Further, the ordering recommended quantity calculation unit 216 calculates the IVF of the second term on the right side of the equation (10.1), that is, the predicted stock quantity, using the following equation (10.2). Equation (10.2) represents a method of calculating the future inventory forecast amount IVF (X + P) after P days have passed from the date X on the date X.

The fourth and fifth terms on the right side of equation (10.2) are the terms added to equation (9.2). The fourth term on the right side of the formula (10.2) is the number of products delivered as of the date X that satisfy the condition that the sales deadline expires by the time the date X + P arrives. The fifth item on the right side of 2) is the number of products scheduled to be delivered as of the date X, which satisfy the condition that the sales deadline expires by the time the date X + P arrives. Equation (10.2) is a condition for which the sales deadline expires at the future time point to be predicted among the products that have been delivered and the products that are scheduled to be delivered in the prediction calculation of the future inventory quantity with respect to the equation (9.2). Consideration is added to deduct the quantity of products that meet the requirements.

Here, the formula (10.2) is a formula for obtaining the estimated number of stocks at the time of opening the store in the future after the lapse of P days from the date X. Therefore, in Eq. (10.2), the method explanation is omitted so as not to complicate it, but in the actual implementation of the program, if the calculation result on the right side is a negative number, it may be replaced with zero. Good.

Hereinafter, the ordering recommended quantity calculation unit 216 calculates the number of stocks after the lead time in the calculation of the formula (10.2) for the products delivered at the time of calculating the ordering recommended quantity, and also the formula ( How to set the demand addition period after the lead time in the calculation of 10.1) will be described with reference to the drawings.

The example of "Case A1" shown in FIG. 35 shows a case where the sales period is smaller than the lead time. In this case, the sales deadline of all the delivered products expires before the lead time of the products ordered on the day of ordering elapses. Therefore, in the ordering recommended quantity calculation unit 216, the fourth term on the right side of the equation (10.2) cancels the first term. Further, in the example shown in FIG. 35, the sales deadline of the ordered product expires before the lead time of the next ordered product elapses. Therefore, the demand addition period after the lead time of the third item on the right side of the formula (10.1) is until the time when the sales deadline of the product ordered on the day of order expires, and after that, the lead time of the product ordered next time. It will be a planned shortage period until the elapse of.

“Case A2” shown in FIG. 36 does not correspond to “Case A1”, but past deliveries in which the product sales deadline expires earlier than the time when the product lead time LT (X) elapses from the day of ordering Indicates the case where it exists. In the illustrated example, there is a product delivered in the past whose sales deadline expires before the lead time of the product ordered on the day of ordering elapses. Therefore, the ordering recommended quantity calculation unit 216 subtracts the quantity of the product satisfying the condition from the first term on the right side by the fourth term on the right side of the equation (10.2). Further, in the example of FIG. 36, the sales deadline of the ordered product expires before the lead time of the next ordered product elapses. Therefore, the demand addition period after the lead time of the third item on the right side of the formula (10.1) is until the time when the sales deadline of the product ordered on the day of order expires, and after that, the lead time of the product ordered next time. It will be a planned shortage period until the elapse of.

“Case A3” shown in FIG. 37 indicates a case in which neither “Case A1” nor “Case A2” is applicable in the relationship between the sales deadline of the delivered product and the lead time of the product ordered on the day of ordering. ing. In the relationship shown in the figure, the timing at which the sales deadline of the product delivered in the past expires is later than the time when the lead time LT (X) of the product elapses from the day of ordering. In such a case, the order recommended quantity calculation unit 216 does not need to subtract the quantity of the delivered product according to the fourth item on the right side of the equation (10.2). Further, in the example of FIG. 37, the sales deadline of the ordered product expires after the lead time of the next ordered product elapses. Therefore, the demand addition period after the lead time of the third item on the right side of the equation (10.1) is until after the lead time of the product ordered next time, and the planned shortage period of the product does not occur.

Next, the ordering recommended quantity calculation unit 216 calculates the number of stocks after the lead time in the calculation of the formula (10.2) for the product scheduled to be delivered at the time of calculating the ordering recommended quantity, and also the formula. How to set the demand addition period after the lead time in the calculation of (10.2) will be described with reference to the drawings.

“Case B1” shown in FIG. 38 shows a case where there is a product waiting for delivery whose sales deadline expires earlier than the time when the lead time LT (X) of the product elapses from the day of ordering. In this case, the ordering recommended quantity calculation unit 216 subtracts the quantity of the product under the condition from the second term on the right side according to the fourth term on the right side of the equation (10.2). Further, in the example of FIG. 38, the sales deadline of the ordered product expires before the lead time of the next ordered product elapses. Therefore, the demand addition period after the lead time of the third item on the right side of the formula (10.1) is until the time when the sales deadline of the product ordered on the day of order expires, and after that, the lead time of the product ordered next time. It will be a planned shortage period until the elapse of.

“Case B2” shown in FIG. 39 shows a case that does not correspond to “Case B1”. In the relationship shown in the figure, the timing at which the sales deadline of the product waiting for delivery expires is later than the time when the lead time LT (X) of the product elapses from the day of ordering. In such a case, the order recommended quantity calculation unit 216 does not need to subtract the quantity of the product to be delivered by the fifth item on the right side of the equation (10.2). Further, in the example of FIG. 39, the sales deadline of the product ordered on the day of the order expires after the lead time of the product ordered next time elapses. Therefore, the demand addition period after the lead time of the third item on the right side of the equation (10.1) is until after the lead time of the product ordered next time, and the planned shortage period of the product does not occur.

The above formula (10.2) calculates the number of products whose sales deadline has not expired after the lapse of P days from the date X by subtraction, but the number of similar products is calculated by the formula (10.2). ) Does not have to be calculated by faithful subtraction. In formula (10.2), the number of products scheduled to be delivered by the delivery timing is added to the inventory quantity of the delivered products whose sales deadline has not expired at the delivery timing, and the products up to the delivery timing are added. By subtracting the planned demand quantity of, the estimated inventory quantity of the product at the delivery timing is calculated. Equation (10.2) is an equation showing an embodiment of forecast inventory quantity calculation in consideration of the sales deadline so that comparison with equation (9.2) can be easily explained. Therefore, for example, the number of delivered products whose sales deadline has not expired after the lapse of P days from the date X is set as the first term on the right side of the formula (10.2), and the sales deadline has expired after the lapse of P days from the date X. The number of products scheduled to be delivered within the lapse of P days from the date X that has not been set may be the second term on the right side of the same formula, and the fourth and fifth terms on the right side of the same formula may be omitted.

According to the loss estimation system of the embodiment described above, the inventory quantity of the product is calculated, and if the calculated inventory quantity of the product is a negative value, the inventory quantity of the product is corrected and the inventory quantity of the product is corrected. The product is not judged to be out of stock for the specified time zone. As a result, the accuracy of determining the product shortage can be improved.

In addition, the loss estimation system corrects the inventory quantity of the product to a positive value when the calculated inventory quantity of the product is a negative value, and when the inventory quantity of the product is "0 (zero)", the product Is determined to be out of stock. As a result, the accuracy of determining the product shortage can be improved.

In addition, when the calculated inventory quantity of the product is a negative value, the loss estimation system corrects the inventory quantity of the product to a non-negative value, and also indicates that the product has been corrected for the relevant time zone. If the information indicating that the time zone of the product has been corrected is saved, the product is not judged to be out of stock for the time zone. As a result, the accuracy of determining the product shortage can be improved.

In addition, the loss estimation system corrects the time-series data of the number of stocks in order so as to go back in time with respect to the time-series data of the number of stocks, and the number of stocks to be corrected is "0 (zero)" and When the stock quantity at the previous time is corrected, the stock quantity is corrected to a positive value. As a result, the accuracy of determining the product shortage can be further improved.

Further, the loss estimation system does not correct the inventory quantity when the inventory quantity to be corrected is "0 (zero)" and the inventory quantity at the previous time is not corrected. As a result, the accuracy of determining the product shortage can be further improved.

In addition, when the loss estimation system corrects the time-series data of the number of stocks along the time axis in units of a predetermined period, the initial value of the time-series data of the number of stocks included in the first predetermined period is set to "0 ( Calculate as "zero)". As a result, the time-series data of the number of inventories is repeatedly corrected in units of predetermined periods as the data is accumulated, and easily becomes a positive value without the correction, so that the accuracy of the inventory data of the product can be improved.

In addition, according to the ordering recommendation system of the embodiment, the difference data between the actual sales value and the predicted demand value of the product in units of sales time is calculated as the residual, and the same product is sold when one product is unsold. The ratio of the unit unsold loss to the total value of the unit unsold loss, which is the hourly loss, and the unit unsold loss, which is the hourly loss when one product misses a sales opportunity due to a shortage. Is calculated, and when the difference data in the plurality of sales time zones calculated by the difference calculation is arranged in ascending order, the position is in the order of the value obtained by integrating the ratio to the number of difference data data in the plurality of sales time zones. The value obtained by adding the demand forecast value to the residual amount is calculated as the reference inventory amount, which is the inventory amount that minimizes the expected loss amount of the product. As a result, the standard inventory amount can be calculated without requiring complicated calculation processing, so that the processing load of the computer when performing the numerical calculation related to inventory optimization can be reduced.

Further, according to the ordering recommendation system of the embodiment, the difference data between the actual sales value of the product in the sales time unit and the demand forecast value is calculated as the residual, and the distribution of the residual in each sales time unit is the demand forecast residual. Created as a distribution, generate a cumulative distribution function that integrates the probability distribution that represents the demand forecast residual distribution, and for the same product, the unit unsold loss, which is the amount of loss in the sales time unit when one product is unsold. , Calculate the ratio of the unit sales loss to the total value of the unit sales loss, which is the loss amount of the sales time unit when one product misses the sales opportunity due to the shortage, and the value of the cumulative distribution function is The residual that matches the ratio is calculated as the reference inventory amount, which is the inventory amount that minimizes the expected loss amount of the product. As a result, the standard inventory amount can be calculated without requiring complicated calculation processing, so that the processing load of the computer when performing the numerical calculation related to inventory optimization can be reduced.

In the ordering recommendation system, the unit unsold loss is determined based on the sales performance data of the product. As a result, the accuracy of calculation of the standard inventory amount can be improved.

In addition, the order recommendation system determines the unit unsold loss based on at least the total discount loss when the product is sold at the discounted selling price and the total disposal loss when the product is discarded. As a result, the accuracy of calculating the standard inventory amount can be further improved.

In addition, the ordering recommendation system calculates the standard inventory amount based on the difference data excluding the abnormal value when the difference data in a plurality of sales time zones contains an abnormal value. As a result, the accuracy of calculating the standard inventory amount can be further improved.

Further, according to the demand forecasting device of the embodiment, the demand forecast of the product is performed by inputting the sales record of the product stored in the storage unit into the demand forecast model, and when the demand forecast of the product is performed, the product is stored in the storage unit. The demand forecast model is changed according to the amount of sales performance data. As a result, the accuracy of product demand forecasting can be improved.

Further, the demand forecasting device predicts the demand of the product by using the demand forecasting model having high prediction accuracy as the amount of sales record data stored in the storage unit increases. As a result, the accuracy of product demand forecasting can be further improved.

Further, when the amount of sales record data stored in the storage unit is relatively small, the demand forecasting device predicts the demand of the product by using the demand forecast model for each product group including a plurality of products. When the amount of sales record data stored in the storage unit is relatively large, the demand forecast model for each product is used to forecast the demand for the product. As a result, the accuracy of product demand forecasting can be further improved.

In addition, when the amount of sales record data stored in the storage unit is relatively small, the demand forecaster predicts the demand of the product using a statistical model that statistically analyzes the sales record of the product and stores it. When the amount of sales record data stored in the unit is relatively large, the demand forecast of the product is performed using the machine learning model trained by using the product sales record as learning data. As a result, the accuracy of product demand forecasting can be further improved.

Further, when the demand forecasting device predicts the demand of a product using a machine learning model trained by using the sales performance of the product as learning data, the amount of sales performance data stored in the storage unit is less than a predetermined amount. In the case of, the demand forecast of the product is performed using the first machine learning model, and when the amount of sales record data stored in the storage unit is equal to or more than the predetermined amount, it is compared with the first machine learning model. The demand for products is forecast using the second machine learning model, which has a large amount of training data. As a result, the accuracy of product demand forecasting can be further improved.

In addition, the demand forecasting device of the embodiment adds up the number of missed sales for each sales time zone of the product that missed the sales opportunity of the product due to the shortage of the product to the number of sales based on the sales performance for each sales time zone of the product. Then, the number of sales for each sales time zone of the product is corrected, and the sales time zone for which the forecast is made for the demand forecast model learned using the learning data including the number of sales for each sales time zone of the corrected product. By entering, the demand for the product is predicted. As a result, the accuracy of product demand forecasting can be improved.

In addition, the demand forecasting device determines whether or not there is an assortment of products for each sales time zone based on the number of products sold for each sales time zone of the corrected product, and excludes the sales time zone for which it is determined that there is no product assortment. By inputting the sales time zone to be forecasted into the demand forecast model learned using the learning data including the number of products sold in the sales time zone, the demand forecast of the product is performed. As a result, the accuracy of product demand forecasting can be further improved.

Further, when the number of sales corrected for a certain period is "0 (zero)", the demand forecasting device determines that there is no assortment of products in that period. As a result, the accuracy of product demand forecasting can be further improved.

Further, in the demand forecasting device, when the number of sales corrected for a certain period is "0 (zero)" and the sales period of the product is less than the predetermined length, there is no assortment of products in that period. Is determined. As a result, the accuracy of product demand forecasting can be further improved.

In addition, the loss estimation system of the embodiment calculates the amount of missed sales of the product, which is the loss when the sales opportunity is missed due to the shortage, based on the sales performance of the product for each time zone, and calculates the amount of missed sales of the product. Calculate the amount of missed sales based on, and add the sales of the product to the amount of missed sales to calculate the demand amount for each item of the product, and the demand of the category based on the sales of the category including multiple products. Calculate the amount of money, evaluate the degree of influence on the sales performance of the product by selling the alternative product included in the same category based on the demand amount of each item of the product and the demand amount of the category, and the evaluated influence Correct the amount of missed sales of goods based on the degree. As a result, it is possible to improve the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed.

In addition, the loss estimation system evaluates the degree of influence as the ratio of the total value of the demand amount for each product to the demand amount of the category increases. As a result, the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed can be further improved.

In addition, the loss estimation system uses the sales record of the time zone when the product was out of stock and the sales record of the time zone when the product was not out of stock when there is a time zone when the product was out of stock. By complementing with, the amount of missed sales of the product is calculated. As a result, the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed can be further improved.

In addition, if there is a time zone when the product is out of stock, the loss estimation system will display the sales results of the time zone when the product was out of stock, and the sales of the day classification and time zone when the product was not out of stock. By supplementing with actual results, the amount of missed sales of products is calculated. As a result, the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed can be further improved.

In addition, the loss estimation system can be used to classify the days of the week when the product was not out of stock if the product was out of stock for a certain period of time and the product was not out of stock in other periods. By supplementing with the sales performance including the time zone, the amount of missed sales of the product is calculated. As a result, the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed can be further improved.

In addition, when the loss estimation system supplements the sales performance of the time zone when the product was out of stock by using the sales performance of the time zone when the product was not out of stock, the time zone when the product was out of stock. The closer it is, the more priority is given to the sales performance of the product. As a result, the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed can be further improved.

In addition, when the loss estimation system supplements the sales performance of the time zone when the product was out of stock by using the sales performance of the time zone when the product was not out of stock, the time zone when the product was out of stock. Compared to the sales performance of the product in the later time zone, the sales performance of the product in the time zone before the time zone when the product was out of stock is preferentially used. As a result, the accuracy of estimating the loss when the sales opportunity of the product due to the shortage is missed can be further improved.

In addition, the ordering recommendation system of the embodiment converts the number of sales based on the sales performance for each content of the product into the integer value closest to the numerical value, and integrates the content of the product with the converted integer value. The converted value of the sales volume for each content of the product is calculated, and the recommended order quantity is corrected so that the difference from the total value of the converted value of the sales volume based on the sales performance for each content of the product is small. As a result, the accuracy of calculating the recommended order quantity can be improved.

In addition, the ordering recommendation system targets the content of products that is smaller than the difference when the total value of the converted values of the sales volume for each content of the product is smaller than the recommended order quantity to be corrected. The converted value of the sales volume is increased, and the total value of the converted value of the sales volume for each content of the product after the increase is corrected to obtain the recommended order quantity. As a result, the accuracy of calculating the recommended order quantity can be improved.

In addition, the ordering recommendation system is when the total value of the converted values of the sales volume for each product content is smaller than the ordering recommended quantity to be corrected, and the difference is smaller than the minimum value of the product content. Is the recommended order quantity after correcting the total value of the converted value of the sales volume for each content of the product without increasing the conversion value of the sales volume. As a result, the accuracy of calculating the recommended order quantity can be further improved.

In addition, the ordering recommendation system increases the recommended ordering amount to be corrected when the total value of the converted values of the sales volume for each content of the product is larger than the recommended ordering amount to be corrected. The recommended quantity for ordering is adjusted to the recommended quantity. As a result, the accuracy of calculating the recommended order quantity can be further improved.

Further, the ordering recommendation system of the embodiment calculates the shortage number with respect to the standard stock quantity at the delivery timing by subtracting the estimated stock quantity of the product at the delivery timing from the standard stock quantity at the delivery timing at the time of ordering the product. Furthermore, with respect to the shortage of the standard stock quantity at the delivery timing, the product in the period from the delivery timing to the delivery timing of the next order and the earlier timing of the expiration date of the product ordered this time. The recommended quantity for ordering is calculated by adding the expected number of demands. As a result, the accuracy of calculating the recommended order quantity can be improved.

In addition, the ordering recommendation system is scheduled to deliver to the number of delivered products in stock under the condition that the sales deadline has not expired at the delivery timing, and by the delivery timing under the condition that the sales deadline has not expired at the delivery timing. By adding the number of products and subtracting the estimated demand for the products up to the delivery timing, the estimated inventory quantity of the products at the delivery timing is calculated. As a result, the accuracy of calculating the recommended order quantity can be improved.

Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, as well as in the scope of the invention described in the claims and the equivalent scope thereof.

10 ... Inventory optimization system, 100 ... Loss estimation system, 110 ... Control unit, 112 ... Inventory quantity calculation unit, 114 ... Inventory quantity correction unit, 116 ... Discount loss calculation unit, 118 ... Disposal loss calculation unit, 120 ... Out of stock Judgment unit, 122 ... Missed sales calculation unit, 124 ... Demand amount calculation unit for individual items, 126 ... Demand amount calculation unit for each category, 128 ... Evaluation unit, 130 ... Missed sales correction unit, 132 ... Missed sales output unit, 140 ... Memory Department, 142 ... theoretical inventory data, 144 ... product data, 146 ... missed sale data, 200 ... ordering recommendation system, 210 ... control unit, 212 ... demand forecasting unit, 214 ... standard inventory calculation unit, 216 ... ordering recommended quantity calculation unit , 218 ... Recommended order quantity output unit, 230 ... Storage unit, 232 ... Sales record data, 234 ... Demand forecast data, 236 ... Standard inventory data, 238 ... Order recommended data, M1 ... Demand forecast model.

Next, as shown in FIG. 9, the oversold calculation unit 122 creates the theoretical inventory data 164 for each category based on the theoretical inventory data 162 for each item. The theoretical inventory data 164 by category is data of product categories grouped by using the hourly unit as the minimum record and the calculation category and the calculation particle size as keys. In the illustrated example, the oversold calculation unit 122 has a specified time zone in the range of "2018/10/1 12:00" to "2018/10/1 13:00", and the calculation particle size is "category purchase". In addition, the theoretical inventory data 164 for each category is created by grouping the data whose calculation category is "CS 1".

The data 170 for calculating the number of missed sales is data that defines information used for calculating the number of missed sales, for example, with an hour unit as the minimum record. The data items of the data 170 for calculating the number of missed sales are, for example, department code, subcategory code, store code, product code, start time of date / time zone, end time of date / time zone, assortment target judgment, shortage. Includes flags, calculated granularity, calculated categories, days of the week, weekday holidays, number of customers purchased, number of sales, and sales amount. The number of purchased customers is the same as the number of purchased customers data 166 described in [0060]. The number of sales is data indicating the number of products sold in one hour from the start time of the date / time zone to the end time of the date / time zone. The number of sales is acquired based on, for example, the sales quantity of the product included in the data obtained by combining the theoretical inventory data 162 for each item and the theoretical inventory data 164 for each category.

Here, the second term from the first term on the right side of equation (9.1) is, but is the difference between the reference inventory and forecast inventory in delivery timing, the first term on the right side for determining the difference first two paragraphs, prediction inventory is performing an operation for obtaining the quantity short of the standard inventory quantity. Therefore, although the method explanation is omitted in the equation (9.1) so as not to complicate the method explanation, in the actual implementation of the program code, when the calculation results of the first and second terms on the right side are negative numbers. You may add the process of replacing it with zero.

Although X in the formula (9.1) has been described as a date, X may represent a time zone when used as a calculation of the recommended number of products to be produced in units of time. In that case, T, LT, and DLV are also used as time zones instead of dates. In other words, production recommended number ORR (X) in the time zone X is, to the difference between the time from the production instruction to line up in the over-the-counter LT (X) + standard inventory amount after the DLV and the predicted amount of inventory, arranged in over-the-counter It is calculated as the sum of the estimated demand numbers for each time period before the next production product is lined up in the store.

Claims (4)

At the time of ordering the product, the estimated stock quantity of the product at the delivery timing is subtracted from the reference stock quantity at the delivery timing to calculate the shortage number with respect to the standard stock quantity at the delivery timing, and further, the said at the delivery timing. For the shortage of the standard stock quantity, the estimated demand quantity of the product in the period from the above delivery timing to the delivery timing of the next order and the earlier timing of the expiration date of the product ordered this time. An ordering recommendation system equipped with an ordering recommended number calculation unit that calculates the recommended ordering quantity of products by adding. The ordering recommended number calculation unit plans to deliver to the number of delivered products under the condition that the sales deadline has not expired at the delivery timing by the delivery timing under the condition that the sales deadline has not expired at the delivery timing. The estimated number of products in stock at the delivery timing is calculated by adding the number of the products and subtracting the estimated demand number of the products up to the delivery timing.
The ordering recommendation system according to claim 1. The computer
At the time of ordering the product, the number of shortages with respect to the standard stock quantity at the delivery timing is calculated by subtracting the estimated stock quantity of the product at the delivery timing from the standard stock quantity at the delivery timing.
With respect to the shortage of the standard inventory quantity at the delivery timing, the period from the delivery timing to the delivery timing of the next order and the earlier timing of the expiration of the sales deadline of the product ordered this time. Calculate the recommended quantity for ordering products by adding the estimated demand for products.
Recommended ordering method. On the computer
At the time of ordering the product, the process of calculating the shortage with respect to the standard stock quantity at the delivery timing by subtracting the estimated stock quantity of the product at the delivery timing from the standard stock quantity at the delivery timing.
With respect to the shortage of the standard inventory quantity at the delivery timing, the period from the delivery timing to the delivery timing of the next order and the earlier timing of the expiration of the sales deadline of the product ordered this time. The process of adding the estimated demand for products to calculate the recommended quantity for ordering products, and
A program that executes.

Images