WO2019053821A1

WO2019053821A1 - Ordering assistance system, ordering assistance program, and ordering assistance method

Info

Publication number: WO2019053821A1
Application number: PCT/JP2017/033092
Authority: WO
Inventors: 早苗中尾; 昌幸親松; 真佑子美濃部
Original assignee: 株式会社日立製作所
Priority date: 2017-09-13
Filing date: 2017-09-13
Publication date: 2019-03-21
Also published as: JPWO2019053821A1; JP6814302B2

Abstract

Provided is an ordering assistance system having a storage unit and a data processing unit. The data processing unit builds a demand prediction model for a commodity on the basis of sales performance held in the storage unit, calculates a predicted demand level of the commodity by use of the demand prediction model having been built, calculates, on the basis of the predicted demand level, a risk value that indicates the magnitude, for each order quantity of the commodity, of at least one of risk in which a sales opportunity loss for the commodity occurs and risk in which excessive inventory of the commodity occurs, outputs a recommended order quantity on the basis of the calculated risk value, determines whether or not the demand prediction model is to be updated on the basis of the sales performance and the predicted demand level, and updates the demand prediction model when it has been determined that the demand prediction model is to be updated.

Description

Order support system, order support program and order support method

The present invention relates to a technology for supporting the ordering of goods.

There is a technique described in Japanese Patent Application Laid-Open No. 8-279013 (Patent Document 1) in order to set an appropriate stock amount. Patent Document 1 includes “a data input / output unit having data input / output display and an arithmetic processing function, a demand amount result storage unit for storing the past demand amount results by model, and information of the demand amount result storage unit. A demand amount probability distribution prediction unit that predicts the probability distribution of demand amount by model and an appropriate stock that determines the setting range of the appropriate stock amount from the probability distribution of model type demand amount obtained from the demand amount probability distribution prediction unit Risk calculation to calculate the out-of-stock risk or stock risk of the product from the output result of the volume range setting unit, the cost storage unit that calculates the inventory cost that occurs at the time of inventory and the sales opportunity loss cost that occurs when out of stock And an appropriate inventory amount determination unit that determines the inventory amount that is the minimum risk result from the output result of the risk calculation unit as the appropriate inventory amount, and a data storage unit, and the demand amount of each market is determined by probability distribution It has been described factories and to set the risk minimum appropriate amount of inventory the amount of inventory in the distribution center by evaluating in the out-of-stock risk and inventory risk ", and by. In addition, "the demand amount actual storage unit stores the actual results of the past demand amount by product type, market and monthly basis, and the demand amount probability distribution prediction unit stores the target products from the demand amount actual storage unit. The monthly demand by market is read, the minimum and maximum demand for the target month are obtained from the information, and the probability distribution of each demand is calculated from the distribution (histogram) within those ranges. ", Is described.

Patent Document 1: Japanese Patent Application Laid-Open No. 8-279013

In patent document 1, probability distribution is set uniquely from the past performance. Therefore, it can not respond to the change of the daily buying tendency by diversification of the hobby taste in recent years.

Therefore, it is an object of the present invention to provide an order support device and an order support program, which presents an appropriate order amount that suppresses the risk of sales opportunity loss and excess inventory taking into consideration changes in purchasing tendency.

In order to solve the above problems, the present invention is an order support system having a storage unit and a data processing unit, wherein the storage unit holds information on sales results of goods, and the data processing unit A demand forecasting unit that builds a demand forecasting model of the product based on the sales results and calculates a forecasted demand volume of the product using the built-up demand forecasting model; and the product based on the forecasted demand volume A risk value calculation unit for calculating a risk value indicating at least one of the risk of occurrence of a sales opportunity loss of the product and the risk of occurrence of an excess inventory of the product for each ordered quantity of The evaluation unit includes: an order amount output unit that outputs a recommended order amount based on a value; and an evaluation unit that determines whether to update the demand forecast model based on the sales record and the forecasted demand amount. There when determining to update the forecast model, the demand prediction unit and updates the forecast model.

According to the present invention, it is possible to present an appropriate order quantity that suppresses the risk of sales opportunity loss and excess inventory, taking into consideration changes in purchasing tendency in daily ordering work.

Problems, configurations, and effects other than those described above will be apparent from the description of the embodiments below.

It is a block diagram showing a schematic structure of an order support system explained as Example 1 of the present invention. It is a block diagram showing an example of composition of an information processor (computer) used for realization of an order support device and an order terminal of Example 1 of the present invention. It is a block diagram which shows the main functions with which the order assistance apparatus of Example 1 of this invention is equipped, and the main data which an order assistance apparatus memorize | stores. It is explanatory drawing which shows an example of the forecasted demand amount calculated by the demand forecasting part of Example 1 of this invention. It is explanatory drawing which shows the example of the risk value calculation process by the risk value calculation part of Example 1 of this invention. It is explanatory drawing which shows the example of the result calculated by the risk value calculation part of Example 1 of this invention. It is a block diagram which shows the main functions with which the order placement terminal of Example 1 of this invention is equipped. It is explanatory drawing which shows an example of the order list screen which an order placement terminal displays on an output device, when a user determines order quantity in Example 1 of this invention. It is explanatory drawing which shows an example of the order detailed screen displayed when a specific goods is selected by the order list display part of the order list screen of Example 1 of this invention. It is explanatory drawing which shows an example of the sales information which the memory | storage part of the order assistance apparatus of Example 1 of this invention manages as a table of a database. It is explanatory drawing which shows an example of the external information which the memory | storage part of the order assistance apparatus of Example 1 of this invention manages as a table of a database. It is explanatory drawing which shows an example of the prediction model information which the memory | storage part of the order assistance apparatus of Example 1 of this invention manages as a table of a database. It is explanatory drawing which shows an example of the goods information which the memory | storage part of the order assistance apparatus of Example 1 of this invention manages as a table of a database. It is explanatory drawing which shows an example of the stock information which the memory | storage part of the order assistance apparatus of Example 1 of this invention manages as a table of a database. It is explanatory drawing which shows an example of the recommendation information which the memory | storage part of the order assistance apparatus of Example 1 of this invention manages as a table of a database. FIG. 12 is an explanatory view showing an example of an option setting screen for setting a condition for calculating a risk value and a recommended order amount output from an order amount output unit by the risk value calculation unit of the first embodiment of the present invention. is there. It is a flowchart which shows an example of the process which the demand forecasting part of Example 1 of this invention and a demand forecast integration part perform. It is a flowchart which shows an example of the process which the evaluation part of Example 1 of this invention performs. It is a block diagram which shows the main functions with which the order assistance apparatus of Example 2 of this invention is equipped, and the main data which an order assistance apparatus memorize | stores. It is explanatory drawing which shows an example of the prediction model information which the memory | storage part of the order assistance apparatus of Example 2 of this invention manages as a table of a database. It is explanatory drawing which shows an example of the stock information which the memory | storage part of the order assistance apparatus of Example 2 of this invention manages as a table of a database. It is a flowchart which shows an example of the process which the demand forecasting part of Example 2 of this invention and a demand forecast integration part perform. It is a flowchart which shows an example of the process which the evaluation part of Example 2 of this invention performs. It is a flowchart which shows an example of the process which the evaluation part of Example 2 of this invention performs.

Hereinafter, embodiments will be described with reference to the drawings. In the following description, the same or similar parts may be assigned the same reference numerals and duplicate descriptions may be omitted.

FIG. 1 is a block diagram showing a schematic configuration of an order support system 1 described as a first embodiment of the present invention.

As shown in FIG. 1, the order support system 1 includes a POS terminal 20, an order terminal 30, and an order support device 10 communicably connected via a communication network 4. The order support device 10 is configured using an information processing device (computer) or dedicated hardware. The order support apparatus 10 may be realized using a plurality of information processing apparatuses communicably connected. The communication network 4 is, for example, a public communication network, a local area network (LAN), a wide area network (WAN), or the like, regardless of whether it is wired or wireless.

The order support device 10 is realized, for example, using an information processing device installed in a system center or a data center. The order support device 10 may be virtually realized as a cloud server or the like.

The POS terminal 20 is installed, for example, at a cash register of a store. The POS terminal 20 registers sales information of a product (purchased product) purchased by a customer in a storage device. Also, the POS terminal 20 performs a settlement process for settling the price of the purchased item. The settlement process includes a settlement process for cash payment and a settlement process for credit payment. Such registration processing and settlement processing are known processing, and thus detailed description will be omitted.

The ordering terminal 30 is an information processing apparatus operated by the user, and is, for example, a smart phone, a portable information terminal such as a portable telephone and a tablet terminal, a personal computer, and the like. The ordering terminal 30 may be, for example, a terminal device installed in a store or the like.

The order support apparatus 10 predicts future demand quantity from external information such as sales performance and weather for each product, and outputs a recommended order quantity in consideration of the expiration date of the product, lead time and current stock quantity.

FIG. 2 is a block diagram showing a configuration example of an information processing apparatus 100 (computer) used to realize the order support apparatus 10 and the order terminal 30 according to the first embodiment of the present invention.

As shown in FIG. 2, the information processing apparatus 100 includes a processor 101, a main storage device 102, an auxiliary storage device 103, an input device 104, an output device 105, and a communication device 106. These are communicably connected to each other via communication means such as a bus (not shown).

The processor 101 is configured using, for example, a central processing unit (CPU) or a micro processing unit (MPU). Various functions of the information processing apparatus 100 are realized by the processor 101 reading and executing the program stored in the main storage device 102.

The main storage device 102 is a device for storing programs and data, and is, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), an NVRAM (Non Volatile RAM), or the like. The auxiliary storage device 103 is a hard disk drive, a solid state drive (SSD), an optical storage device, a recording medium reading / writing device, or the like. Programs and data stored in the auxiliary storage device 103 are loaded into the main storage device 102 as needed.

The input device 104 is, for example, a keyboard, a mouse, and a touch panel. The output device 105 is, for example, a liquid crystal monitor, an LCD (Liquid Crystal Display), a graphic card, or the like. The communication device 106 is a communication interface that communicates with another device via the communication network 4 and is, for example, a network interface card (NIC), a wireless communication module, or the like.

FIG. 3 is a block diagram showing main functions of the order support apparatus 10 according to the first embodiment of the present invention and main data stored in the order support apparatus 10.

As shown in FIG. 3, the order support device 10 includes a data processing unit 110 and a storage unit 120. The data processing unit 110 further includes a demand prediction unit 111, a demand prediction integration unit 112, a risk value calculation unit 113, an order quantity output unit 114, an evaluation unit 115, and a recommendation result provision unit 116. These functions are realized, for example, by the processor 101 of the information processing apparatus 100 reading out and executing a program stored in the main storage device 102. That is, in the following description, the processes executed by the above-described units are actually executed by the processor 101 according to the instructions described in the program.

The storage unit 120 stores sales information 121, external information 122, prediction model information 123, prediction information 124, product information 125, inventory information 126, and recommendation information 127. The storage unit 120 manages such information (data) as, for example, a text file or a database table such as a RDBMS (Relational Data Base Management System). All or part of the information may be configured to be stored on another server via the communication network 4.

The demand prediction unit 111 constructs one or more demand prediction models from the sales information 121 and the external information 122 when the prediction model information 123 of the product is not generated, and generates the prediction model information 123. In the present embodiment, the prediction period is described on a day basis. The demand forecasting model uses, for example, a simple moving average model, an exponential smoothing model, an ARIMA model (Auto Regressive Integrated Moving Average model), or a multiple regression model. Depending on sales frequency, the usual demand forecasting model may not be suitable. For example, if the sales result is only a few days in a month, it is not desirable to apply the ARIMA model and the multiple regression model. If sales frequency is low, restrict the use of only a specific demand forecasting model that is suitable for that, change the unit of forecasting period, and select a longer period than normal forecasting period (1 day in the above example) It is desirable to make weekly forecasts. In this embodiment, when the sales frequency is low, it is assumed that a dedicated model (hereinafter, referred to as "low frequency model") is used. The demand forecasting unit 111 reconstructs the demand forecasting model when the update schedule is registered in the forecasting model information 123.

Further, the demand prediction unit 111 calculates a predicted demand amount based on the prediction model information 123, the sales information 121, and the external information 122, and generates the prediction information 124. The forecasted demand amount is calculated by a probability distribution according to the demand forecasting model.

FIG. 4: is explanatory drawing which shows an example of the predicted demand amount calculated by the demand forecasting part 111 of Example 1 of this invention.

In the case of FIG. 4, the probability that the demand amount per forecast period (for example, one day) is “20” is 1%, and the probability that the demand amount is “50” is 14%. An example of displaying the probability distribution of such forecast demand amount by a table is shown in FIG. 4 (a), and an example of displaying it by a graph is shown in FIG. 4 (b).

The demand forecast integration unit 112 integrates a plurality of predicted demand amounts calculated by the demand forecast unit 111 as one predicted demand amount when there are a plurality of demand forecast models of the same product. The demand prediction integration unit 112 may, for example, integrate predicted demand amounts calculated from a plurality of demand prediction models at an equal ratio. Alternatively, the demand prediction integration unit 112 makes the ratio of the predicted demand calculated from the demand prediction model having a small prediction error to a larger proportion according to the prediction error of each demand prediction model (ie, the integrated predicted demand amount). Integration so that the contribution of For example, the demand prediction integration unit 112 integrates at a rate proportional to the reciprocal of the prediction error. This improves the prediction accuracy. As the prediction error, mean absolute percentage error (MAPE), mean squared error (MSE), root mean squared error (RMSE), or the like is used.

The risk value calculation unit 113 determines the sales opportunity loss and the sales opportunity loss for each ordered quantity taking into consideration the product information 125 such as the expiration date, the available sale date, and the lead time, and the inventory information 126 based on the predicted demand amount after integration. Calculate the risk value due to excess inventory. The sales opportunity loss is a loss of sales opportunity caused by lack of stock of the product despite demand for the product. The risk value calculation unit 113 calculates the out-of-stock risk and the low display risk as risks due to the sales opportunity loss. The low display risk is the risk that the consumer will refrain from purchasing and the sales will decrease, as the remaining number of products arranged on the display shelf of a store such as a supermarket is decreasing. For example, the risk value calculation unit 113 calculates the out-of-stock risk using the following formula (1), and calculates the small display risk using the formula (2).

N is a unit of a prediction period, and in the present embodiment, N = 1 (day). m is the true demand in the forecast period. When the risk value is calculated based on the predicted demand calculated as described above, the predicted demand is m. s is an order quantity. “a” is the number serving as a reference for determining that the number of displays is small, and it is determined that a risk arises when the number of displays is a number or less. Assuming that a product of demand m sells at a constant speed, the interval between products sold is N / m days, and the number of days until the order quantity s is sold is N / m × s days.

The risk value calculation unit 113 also calculates the consumption expiration risk and the excess inventory risk as risks due to the excess inventory. Surplus inventory risk is the risk of storing excess inventory. For example, the amount of stock exceeding demand and safety stock can be considered as extra stock. For example, the risk value calculation unit 113 calculates the consumption expiration risk using the following equation (3), and calculates the surplus inventory risk using the equation (4).

Y is the number of safety stock. The safety stock number can use a preset value or may be calculated using the following equation (5). Instead of the standard deviation of the prediction error, the standard deviation of the number of purchases may be used. In addition, for the period in which the forecasted demand amount is calculated, the forecasted demand amount may be used instead of the average sales volume in the equations (3) and (4). By performing the calculation as described above, it is possible to appropriately calculate the magnitude of each type of risk based on the product information such as the expiration date and the lead time and the predicted demand amount.

In this embodiment, for example, if the out-of-stock risk is 0.2, each risk is calculated based on the number of days such that the out-of-stock days are 0.2 days. Each risk may be calculated on a monetary basis when the commodity price and the cost required for storage become clear. If calculated on a monetary basis, it will be easier to grasp the impact on management. It also allows you to compare risks between different products. On the other hand, since the cost required for storage is often ambiguous, and commodity prices also fluctuate, when dealing with various commodities, it is possible to reduce complicated processing by calculating each risk on a day basis.

The risk value calculation unit 113 calculates the expected value of the out-of-stock risk, the low display risk, the consumption overage risk and the surplus inventory risk from the probability distribution of the predicted demand amount for each order amount, adds the weighted expected values and adds them. Calculate the risk value by The weight for each risk may be equal or, for example, the out-of-stock risk may be emphasized and the out-of-stock risk may be increased. Also, by setting the weight to "0", it is possible to exclude the risk that does not need to be considered from the calculation of the risk value. For example, in the case where there is no display of goods in mail order or wholesale business, the weight of the small display number risk may be set to “0”.

FIG. 5 is an explanatory view showing an example of a risk value calculation process by the risk value calculation unit 113 according to the first embodiment of the present invention.

Specifically, FIG. 5 shows an example of the risk value calculation process when the order quantity is 50 pieces. For example, if 50 pieces of goods are ordered, the demand risk is 0 until the demand quantity is 50 pieces, but if the demand quantity exceeds 50 pieces, the stock out occurs, and the demand risk for the case of 55 pieces is the formula It becomes 0.091 from (1). From the probability distribution of the demand amount, the probability that the demand amount is 55 pieces is 12%, so the expected value is 0.0109 × 12% and becomes 0.0109. Expected values are calculated for all the demand amounts set by the probability distribution, and the final expected value of the out-of-stock risk is 0.138. Expected values for the low display risk, surplus inventory risk and consumption expiration risk calculated in the same way as the expected value for the out-of-stock risk are respectively 0.152, 0.115 and 0. Here, when the weight of the out-of-stock risk is set to 2 and the weight of the other risks to 1, the risk value is calculated to be 0.543, and this value is the risk value when the order quantity is 50 pieces.

FIG. 6 is an explanatory view showing an example of the result calculated by the risk value calculation unit 113 according to the first embodiment of the present invention. Specifically, FIG. 6 shows an example of graphically displaying the result when the risk value is calculated by the method shown in FIG. 5 for each order quantity.

The risk value calculated in the example described above is a risk value when the stock amount at the time the ordered product is delivered is zero. In practice, it is necessary to estimate the amount of stock at the time of delivery and to calculate the risk value for each ordered amount. For example, the stock quantity at the time of delivery is estimated using the following equation (6).

The risk value calculation unit 113 adjusts the ordered amount according to the stock amount at the delivery time point estimated by the equation (6) to obtain a final risk value. For example, in the case of the risk value described in FIG. 5, the risk value is 0.543 when the order quantity is 50, but when the stock quantity at the delivery point is estimated to be 10, the order quantity The risk value is 0.543 when (50−10 =) 40 are set.

The order amount output unit 114 outputs an appropriate order amount as a recommended order amount based on the risk value calculated by the risk value calculation unit 113. The order quantity output unit 114 outputs, for example, an order quantity that minimizes the risk value as a recommended order quantity. Alternatively, the user sets the range of acceptable risk values and which of sales opportunity loss and excess inventory are desired to be avoided, and the order quantity output unit 114 outputs the recommended order quantity according to the set conditions. The order quantity output unit 114 registers the recommended order quantity and the risk value calculated by the risk value calculation unit in the recommendation information 127.

The evaluation unit 115 confirms the level of the stock amount, and when the stock amount level is not appropriate, evaluates whether it is necessary to update the demand forecasting model. An appropriate inventory level can be set, for example, based on the latest average sales volume. For example, the evaluation unit 115 confirms whether the inventory level is appropriate using the following equation (7).

b _ki is the i-th sales quantity of the product k. n _k is a value indicating the order of the latest sales volume of the product k. For example, if the sales quantity of the product k is stored from 100 days ago, n _k is 100. T is the number of days for calculating the average sales volume, for example, when using the average sales volume of the last two weeks, T is 14. L _k is the lead time of the product k. The evaluation unit 115 obtains the upper limit of the inventory level by using the following equation (8) instead of the equation (7), for example, when dealing with a product whose order unit is large compared with the average sales volume.

U _k is an order unit of goods k. As a result, for products with an average sales volume of 30 pieces, a lead time of 1 day, and an order unit of 100 pieces, where the order unit greatly exceeds the average sales amount, the inventory level is always the upper limit of the inventory level It is possible to avoid the situation of exceeding.

The evaluation unit 115 evaluates the accuracy of the demand forecasting model (in other words, whether or not it is necessary to update the demand forecasting model) when the inventory level is not appropriate. For example, the evaluation unit 115 compares the past sales volume stored in the sales information 121 with the forecasted demand on the same day, and updates the demand forecasting model when the difference between the actual sales volume and the forecasted demand is large. It determines that it is necessary, and registers the update schedule in the prediction model information 123 of the product. After the update schedule is registered, the demand prediction unit 111 updates the demand prediction model (see FIG. 17 and the like for details).

In the evaluation unit 115, the index serving as a trigger for evaluating whether or not the demand forecasting model needs to be updated is not limited to the inventory level, but the stockout rate, inventory turnover rate, discard rate, average sales number change, or product life cycle The change point of may be used. The storage unit 120 may hold information on the life cycle of each product (for example, in the product information 125). The evaluation unit 115 updates the demand forecasting model at the time when the demand amount is estimated to change, such as the time when the life cycle stage changes, for example, the time when the introduction period shifts to the growth period, based on the life cycle of each product. It may be determined whether or not it is necessary. This enables the demand forecasting model to be updated at an appropriate time.

When the recommended result provision unit 116 receives the request for provision of the recommended order amount from the ordering terminal 30, the product information of all the products belonging to the category is obtained for the product category received together with the provision request, and the information of all the acquired products ( The list is sent to the ordering terminal 30.

FIG. 7 is a block diagram showing the main functions of the order placement terminal 30 according to the first embodiment of the present invention.

As shown in FIG. 7, the ordering terminal 30 includes an ordering support result receiving unit 311, an ordering support result display unit 312, and an order quantity determination unit 313. These functions are realized, for example, by the processor of the information processing apparatus 100 constituting the ordering terminal 30 reading out and executing a program stored in the main storage device 102.

The order support result receiving unit 311 receives the information transmitted from the recommendation result providing unit 116 of the order support apparatus 10.

FIG. 8 is an explanatory view showing an example of an ordering list screen 800 displayed on the output device 105 by the ordering terminal 30 when the user confirms the ordering amount in the first embodiment of the present invention.

As shown in FIG. 8, the order list screen 800 includes a product category input area 801, an order list display unit 802, an order amount input area 803, and an order amount determination operation unit 804.

When a product category is selected in the product category input area 801, the order support result display unit 312 of the order terminal 30 transmits to the order support apparatus 10 a request for providing a recommended order amount together with the selected product category. The input to the product category input area 801 may be performed by the user selecting from a list of product categories prepared in advance as described above, or may be performed by directly inputting a product category name.

The order support result display unit 312 also displays a list of recommended order quantities on the order list display unit 802 for the products of the product category input in the product category input area 801 based on the information received by the order support result reception unit 311. Do. The list of recommended order quantities displayed in the order list display unit 802 includes, for example, items of a product name, an order pattern, a trend, a predicted demand quantity, a recommended order quantity, and an order quantity.

In the ordering pattern, when the model type stored in the prediction model information 123 of each product is a low frequency model, “low frequency” is displayed, and in the other cases, “standard” is displayed. The trend indicates "increase" if the sales volume stored in the sales information 121 for each product is increasing, "decrease" if it is decreasing, and "no change" if not changing. Be done. As the order quantity, the same value as the value displayed in the recommended order quantity is displayed as the initial value. This supports the user's ordering operation.

The items displayed on the order list display unit 802 may include information that allows the user to determine the order amount, and the order pattern and the trend are not necessarily required. In the present embodiment, an example is presented in which these items are included as information that can be considered when the user decides the order amount.

The order quantity determination unit 313 receives an order quantity determination process by the user. The order quantity column displayed in the order list display unit 802 of the order list screen 800 also serves as the order quantity input area 803, and when the user desires to change the order quantity, the order quantity determination section 313 determines the order quantity input area. The input from the user input in 803 is accepted. When the user operates the order amount determination operation unit 804, the order amount determination unit 313 determines the order amount.

FIG. 9 is an explanatory view showing an example of the order details screen 900 displayed when a specific product is selected in the order list display unit 802 of the order list screen 800 according to the first embodiment of the present invention.

Selection of a product on the order list screen 800 is executed, for example, by the user clicking on a line of a specific product in the order list display unit 802. Alternatively, when the input to the ordering terminal 30 is performed via the touch panel, the user performs the input by touching a row of a specific product. As shown in FIG. 9, the order details screen 900 includes a product information display unit 901, a setting condition display unit 902, a result display unit 903, a change order amount input area 904, and an order amount update operation unit 905.

The product information display unit 901 displays information of the product (that is, the product selected by the user) stored in the product information 125.

The setting condition display unit 902 displays setting conditions regarding the method of calculating the risk value and the method of determining the recommended order amount. Details of setting of setting conditions will be described later with reference to FIG.

The result display unit 903 displays a risk value 903A stored in the recommendation information 127, and a demand change 903B indicating a change between the sales result and the forecasted demand amount. The demand change 903B is obtained by plotting the sales volume (sales record) stored in the sales information 121 and the forecasted demand stored in the prediction information 124 on a graph with the horizontal axis as a date, for example. . Here, as shown in FIG. 4, when the predicted demand is obtained as a probability distribution, the demand calculated from the probability distribution is plotted. For example, the demand with the highest probability may be plotted, or the expected value of the demand calculated from the probability distribution may be plotted. The same is true for the forecasted demand shown in FIG. 8 and the forecasted demand in the following description.

The result display unit 903 also displays the forecast demand amount and the planned delivery quantity 903C from the order placement date to the date when the quantity ordered on the order placement date is delivered in the form of a graph. In the case of FIG. 9, since the order date is February 19 and the lead time is a product on the second day, the delivery date is February 21 and the forecasted demand from February 19 to February 21 The planned delivery quantity is displayed. This allows the user to adjust the order quantity while taking into consideration the difference between the past sales results and the forecasted demand and the future delivery schedule. When the user wants to change the order amount, the user inputs the changed order amount in the change order amount input area 904, and operates the order amount update operation unit 905 (for example, the update button displayed on the screen is pressed) The order quantity can be changed by

When the update of the order quantity by the user is received, the order quantity displayed on the order list screen 800 is also updated.

FIG. 10 is an explanatory view showing an example of sales information 121 managed as a database table by the storage unit 120 of the order support apparatus 10 according to the first embodiment of the present invention.

The sales information 121 holds the results of product sales. As shown in FIG. 10, the sales information 121 is composed of one or more records having items of a product ID 1011, a sold date 1012, and a sales volume 1013. In the product ID 1011, an ID for identifying a product is set. For example, a JAN code may be used as the product ID 1011. The sold date 1012 and the sold volume 1013 respectively have the date the item was sold and its volume.

FIG. 11 is an explanatory diagram showing an example of the external information 122 managed by the storage unit 120 of the order support apparatus 10 according to the first embodiment of the present invention as a table of a database.

The external information 122 manages information necessary to use a demand forecasting model that uses data other than sales quantities, such as multiple regression models. As shown in FIG. 11, the external information 122 is composed of one or more records having items of date 1111, day of week 1112, holiday flag 1113, weather 1114, temperature 1115, and event 1116. Each record corresponds to the information for each date set on the date 1111.

In the holiday flag 1113, it is set whether the date is a holiday. For example, if the date is a holiday, the holiday flag 1113 is set to “1”, and if it is not a holiday, the holiday flag 1113 is set to “0”. In the event 1116, the presence or absence of an event to be held in the vicinity (for example, the vicinity of a store where the ordering terminal 30 is installed) on the date is set. For example, when an event is held, “1” is set to the event 1116, and when the event is not held, “0” is set to the event 1116. The external information 122 to be managed may include not only the day of the week, a holiday flag, weather, temperature and events, but also wind power, special sale status, trend, discount rate from list price, and the like.

FIG. 12 is an explanatory view showing an example of the prediction model information 123 managed as a table of the database by the storage unit 120 of the order support device 10 of the first embodiment of the present invention.

In the prediction model information 123, information of the demand prediction model constructed by the demand prediction unit 111 is managed. As illustrated in FIG. 12, the prediction model information 123 is configured of one or more records having items of a product ID 1211, a model type 1212, a frequency 1213, an error 1214, an update schedule 1215, and an update content 1216.

The model type 1212 is set with the type of demand forecasting model. In the frequency 1213, it is set whether the sales frequency of the product set in the product ID 1211 of each record is high or low. For example, the frequency is set to "low" when the sales record of the product is 2 days or less a week, and the frequency is set to "high" when the product is sold 3 days a week or more. In the error 1214, a prediction error is set when the demand amount of the product of each record is predicted by the demand prediction model set in the model type 1212. As the calculation of the prediction error, the error at the time of model construction in the data used when constructing the demand prediction model may be used, or the error between the prediction in the latest determined period and the sales results is used You may

In the update schedule 1215, “1” is set when it is necessary to update the information of the record as a result of evaluation by the evaluation unit 115, and “0” is set when it is not necessary to update. In the update content 1216, the update content of the demand forecasting model determined by the evaluation unit 115 is set. For example, in the case of a change in the type of demand forecasting model, “model type” is set, and in the case of a change in parameter, “parameter” is set. The change in the type of demand forecasting model means a change from a model other than the low frequency model to a low frequency model, or a change from a low frequency model to another model. If you do not consider changes in the type of demand forecasting model, such as not limiting the type of demand forecasting model to be used according to sales frequency, you may exclude the update content 1216 from the item of forecasting model information 123 .

FIG. 13 is an explanatory diagram of an example of the product information 125 managed by the storage unit 120 of the order support apparatus 10 according to the first embodiment of the present invention as a table of a database.

As shown in FIG. 13, the product information 125 includes one or more records having items of product ID 1311, product name 1312, category 1313, expiration date 1314, lead time 1315, and order unit 1316. In the category 1313, categories of goods (for example, snacks, seasonings, meats, vegetables, etc.) are set. The expiration date of the product is set as the number of days in the expiration date 1314. For example, in the case of a product whose expiration date is 60 days from manufacture, the expiration date 1314 is set to 60. In the case of a product that does not have a expiration date, an expiration date or an available sale period may be set instead of the expiration date. In lead time 1315, a period from ordering of goods to delivery is set. The order unit 1316 is set with a unit for ordering a product. For example, in the case of a product that can not be ordered in units of one, but needs to be ordered in units of ten, “10” is set.

FIG. 14 is an explanatory diagram of an example of inventory information 126 managed by the storage unit 120 of the order support apparatus 10 according to the first embodiment of the present invention as a table of a database.

Inventory information 126 holds a history of product inventory. As shown in FIG. 14, the inventory information 126 includes one or more records having items of item ID 1411, date 1412, inventory amount 1413, delivery amount 1414, and order amount 1415. In the stock amount 1413, the stock amount on the date set on the date 1412 of the product set in the product ID 1411 is set. In the delivery amount 1414, the quantity of the product set in the product ID 1411 is set on the date set in the date 1412 is set. In the order quantity 1415, the quantity of goods set by the goods ID 1411 ordered on the date set by the date 1412 is set. In the present embodiment, the delivery amount 1414 and the order amount 1415 are described as a configuration for managing the stock information 126. However, they may be managed separately from the stock information as the order information.

FIG. 15 is an explanatory view showing an example of the recommendation information 127 managed as a database table by the storage unit 120 of the order support apparatus 10 according to the first embodiment of the present invention.

As shown in FIG. 15, the recommendation information 127 is configured of one or more records having items of a product ID 1511, a date 1512, a recommended order quantity 1513, and a risk value 1514. The recommended order quantity determined by the order quantity output unit 114 is stored in the recommended order quantity 1513. The risk value calculated by the risk value calculation unit 113 is stored in the risk value 1514. The risk value 1514 may be a list of the ordered quantity and the risk value of the ordered quantity described in a list format or a data format corresponding to the language of the mounting program.

FIG. 16 shows an option setting screen 1600 for setting the condition for calculating the risk value by the risk value calculation unit 113 and the recommended order amount output from the order amount output unit 114 according to the first embodiment of the present invention. It is explanatory drawing which shows an example.

In the “Risk value calculation mode” column 1601 of the option setting screen 1600, the user selects one of the displayed options. In the example of FIG. 16, as options, "even," "prevent sales opportunity loss avoidance", "only consider sales opportunity loss avoidance", "precede excess inventory avoidance", "consider only excess inventory avoidance" and "manual" Is displayed.

For example, when “priority on avoiding sales opportunity loss” is selected, the risk value calculation unit 113 determines the risk of sales opportunity loss (out-of-stock risk, low display risk) and excess inventory risk (consumption expiration risk, surplus inventory risk When weighting and adding the expected value of), the weight of the risk of sales opportunity loss is made greater than the risk of excess inventory. When “even” is selected, the risk value calculation unit 113 sets the weight of the risk of loss of sales opportunity and the weight of the risk of excess inventory to the same size. In the example of FIG. 16, “Equal” is the default value, and “Equal” is automatically selected when the user does not select any option. When “only considering sales opportunity loss avoidance” is selected, the risk value calculation unit 113 sets the weight of the risk of excess inventory to zero. If “prevent excess stock avoidance” is selected, the risk value calculation unit 113 makes the weight of the risk of excess stock greater than the weight of the risk of sales opportunity loss. When “consider only excess stock avoidance” is selected, the risk value calculation unit 113 sets the weight of the risk of the sales opportunity loss to 0.

For example, the type of risk accepted may differ depending on the location conditions of the store, etc., but by performing weighting as described above for each store, a risk value meeting the store's conditions is calculated, and as a result, the store's conditions It is possible to present an order quantity that conforms to.

When the user selects "Manual", the user further inputs the weight of out-of-stock risk, low display risk, overdue consumption risk and surplus inventory risk in the input fields. In this case, the risk value calculation unit 113 calculates the risk value using the weight input by the user.

In addition, the setting method of the weighting option as described above and the weight corresponding to each option is an example, for example, weighting different from the above may be performed, for example, an option other than the above may be displayed.

The user selects one of the displayed options in the column 1602 of “How to determine the recommended order amount” on the option setting screen 1600. In the example of FIG. 16, “order amount for which the risk value is minimum” and “set an allowable range of the risk value” are displayed as options. If the recommended order quantity is the order quantity that minimizes the risk value, the user selects the “order quantity that minimizes the risk value”. When determining the recommended order quantity from the order quantity which is the risk value within the allowable range, the user selects "set the allowable range of the risk value". If you select "Set tolerance of risk value", the user further enters the maximum value of acceptable risk value in the "Maximum acceptable risk value" input field to avoid either sales opportunity loss or excess inventory. Choose whether you want to give priority. In addition, the user may either “provide an order quantity with a minimum risk value” or “do not provide a recommended order quantity” if the minimum value of the risk value exceeds the value set in the “maximum acceptable risk value”. Choose how to deal with it.

In the setting of FIG. 16, the order quantity output unit 114 sets a small order quantity to avoid an excess stock (for example, a consumption expiration risk or a surplus stock risk is minimized) in the order quantity range in which the risk value is 1 or less. Order quantity is output as the recommended order quantity.

Note that the setting of the method of determining the recommended order amount as described above is an example, and a setting different from the above may be performed, for example, options other than the above may be displayed. Or, in the column 1602 of “How to determine the recommended order volume”, the user selects “Set an acceptable range of risk value”, and enters the maximum allowable risk value in the “Allowable maximum risk value” input field. However, when it is not selected whether to give priority to avoiding the sales opportunity loss or the excess inventory, the order quantity output unit 114 outputs the sales opportunity loss within the range of the order quantity for which the risk value is 1 or less. Both the higher order quantity to be avoided (for example, the order quantity that minimizes the risk of out-of-stock or the number of displayed items) and the small order quantity that avoids excess inventory may be output as the recommended order quantity. At this time, the order quantity output unit 114 may output, together with each order quantity, information indicating which risk each order quantity is to be avoided. The user can select one by referring to the output information.

The setting of options may be performed via the screen (option setting screen 1600) as described above, or, for example, a file in which options are described may be provided to the order support apparatus 10 as an input.

<Description of processing>
Subsequently, processing performed in the order support system 1 configured as described above will be described.

FIG. 17 is a flowchart illustrating an example of processing performed by the demand prediction unit 111 and the demand prediction integration unit 112 according to the first embodiment of this invention.

The demand prediction unit 111 sets a product to be a target for which the demand is predicted (S1701). The demand forecasting unit 111 determines whether a demand forecasting model is built for the target product (S1702). If the demand forecasting model is built (S1702: YES), the process proceeds to S1703. When the demand forecasting model is not built (S1702: NO), the demand forecasting unit 111 builds the demand forecasting model and adds forecasting model information 123 (S1704).

When the demand forecasting model is built, the demand forecasting unit 111 determines whether there is a plan to update the built demand forecasting model (S1703). If there is no update schedule (S1703: NO), the process proceeds to S1706. If there is an update schedule (S1703: YES), the demand prediction unit 111 reconstructs the demand prediction model and updates the prediction model information 123 (S1705).

The demand prediction unit 111 reads out the demand prediction model of the target product from the prediction model information 123 (S1706). The demand forecasting unit 111 determines whether there is a demand forecasting model for which the forecasted demand amount is not calculated among the demand forecasting models read out in S1706 (S1707). If the forecasted demand amount has been calculated for each demand forecast model (S1707: NO), the process proceeds to S1708. If there is a demand forecasting model for which the forecasted demand volume is not calculated (S1707: YES), the process proceeds to S1709, and the demand forecasting unit 111 calculates the forecasted demand volume.

The demand forecasting unit 111 determines whether the forecasted demand amount has been calculated from two or more demand forecasting models (S1708). When only a single predicted demand amount is calculated from one demand prediction model (S1708: NO), the demand prediction unit 111 registers the predicted demand amount in the prediction information 124 (S1711) and ends the processing. If the predicted demand amount is calculated from each of two or more demand forecast models (S1708: YES), the demand forecast integration unit 112 integrates the calculated forecast demand amount (S1710), and then the demand forecast unit 111 registers the predicted demand amount after integration in the prediction information (S1711) and ends the processing.

FIG. 18 is a flowchart illustrating an example of processing performed by the evaluation unit 115 according to the first embodiment of this invention.

The evaluation unit 115 determines whether the current inventory level is appropriate (S1801). For example, the evaluation unit 115 may determine that the current inventory level is appropriate when the inventory level is within the range from the lower limit to the upper limit of the inventory amount calculated by Equation (7) or Equation (8). If the current inventory level is appropriate (S1801: YES), the process is ended. If the current inventory level is not appropriate (S1801: NO), it is determined whether or not the demand forecast model is to be updated. To advance to step S1802.

As already described, the fact that the current inventory level is not appropriate is an example of a trigger for determining whether or not to update the demand forecast model, and the evaluation unit 115 performs the determination based on another trigger in S1801. You may go. For example, the evaluation unit 115 may determine in S1801 whether or not the time when the amount of demand changes has arrived based on the life cycle of the product, and if it is determined that it has arrived, the process may proceed to S1802.

When the current inventory level is not appropriate, the evaluation unit 115 determines whether a shortage occurs (S1802). If no shortage occurs (S1802: NO), the evaluation unit 115 determines whether the inventory level shifts toward the appropriate range (S1803), and if a shortage occurs (S1802) : YES) Proceed to S1804.

The evaluation unit 115 ends the process when the inventory amount level shifts in the direction toward the appropriate range (S 1803: YES), and when not shifting in the direction toward the appropriate range (S 1803: NO) in S 1804 move on.

In S1804, the evaluation unit 115 reads the date and sales quantity from the sales information 121, and determines whether there is a change in the sales frequency. For example, the evaluation unit 115 calculates whether the sales frequency of the latest one month is higher or lower than a predetermined standard (for example, three days or more per week), and the calculated result is registered in the prediction model information 123. Frequency 1213 If it is different from the value of, it is determined that there is a change in sales frequency. If there is no change in the sales frequency (S1804: NO), the processing proceeds to S1805, and if there is a change in the sales frequency (S1804: YES), the processing proceeds to S1806.

When there is a change in the sales frequency, the evaluation unit 115 sets “1” to the change schedule of the prediction model information 123, and sets “model type” in the update content 1216 of the prediction model information 123 (S1806). After completing the setting in S1806, the evaluation unit 115 ends the process.

As described above, when “1” is set to the change schedule of the prediction model information 123 and “model type” is set to the update content 1216 of the prediction model information 123, the demand prediction unit 111 then selects FIG. In S1705, the type of demand forecasting model is changed. For example, when the sales frequency is changed from high frequency to low frequency in S1804, a low frequency model is newly constructed. On the other hand, when the sales frequency is changed from low to high in S1804, a normal type demand forecasting model such as a multiple regression model is constructed. This makes it possible to construct an appropriate demand forecasting model according to sales frequency.

When there is no change in the sales frequency, the evaluation unit 115 compares the sales volume of the sales information 121 with the predicted demand volume of the prediction information 124 and determines whether there is a difference between the two (S1805). At this time, the evaluation unit 115 determines that there is a difference between the two when the difference between the sales quantity and the forecasted demand amount is larger than a predetermined reference. For example, the evaluation unit 115 may determine that there is a difference between the sales volume and the forecasted demand amount when the difference between the sales volume and the forecasted demand amount is 10% or more of the average sales volume in the last two weeks. . The period for comparing the sales volume and the forecasted demand amount is not limited to the last two weeks, and may be ten days or one month. Also, the determination of the difference is not limited to 10% or more of the average sales volume, and may be performed based on a different ratio or a specific numerical value. If there is no divergence between the sales volume and the predicted demand amount (S1805: NO), the processing is ended, and if there is a divergence (S1805: YES), the processing proceeds to S1807.

In S1807, the evaluation unit 115 sets “1” to the change schedule of the prediction model information 123, and sets “updated content” 1216 of the prediction model information 123 as “parameter”. After completing the setting in step S1807, the evaluation unit 115 ends the process. In this case, the demand prediction unit 111 then changes the parameters of the demand prediction model in S1705 of FIG. This improves the forecasting accuracy of the demand forecasting model.

In the first embodiment described above, the prediction period is described as a unit of one day, but a plurality of days may be put together as one prediction period as necessary. For example, the prediction period may be on a weekly basis. Alternatively, it may be a unit shorter than one day, such as one hour.

In addition, before long holidays such as New Year holidays and Bon holidays, the sales volume of various products may increase, and there may be a shortage of resources required for delivery and delivery work. Therefore, in the long run (two weeks, three weeks, etc.), the future demand quantity may be forecasted, and the order quantity may be increased ahead of time for products with a long consumption term.

As described above, the ordering support system 1 of this embodiment calculates the risk of sales opportunity loss and excess inventory taking into consideration the product information (expiration date, lead time, etc.) and the predicted demand amount in the future, Present an appropriate order quantity to curb Further, the order support system 1 of the present embodiment updates the demand forecasting model when the stock level is not appropriate. For this reason, it is possible to present an appropriate order quantity while coping with changes in purchasing trends.

The order support system 1 of the first embodiment is set to be updated immediately when the demand forecasting model needs to be updated. On the other hand, in the order support system 1 of the second embodiment described below, the period of information (sales information 121 and external information 122) used for rebuilding the demand forecasting model when the demand forecasting model needs to be updated. Set up and update the demand forecasting model after a sufficient amount of information has been accumulated for reconstruction. In addition, the order support system 1 of the second embodiment does not perform demand forecasting (i.e., does not construct a demand forecasting model) for products whose sales frequency is extremely low (i.e., does not construct a demand forecasting model), and sets order support. Here, ordering means that the store places an order according to the order received from the customer. That is, the store does not have the stock of the product set to correspond in the order placement order (unless receiving the order of the product from the customer).

The same reference numerals as in the first embodiment shown in FIGS. 1 to 18 denote the same parts as in the first embodiment, so the description will be omitted and the second embodiment will be described focusing on the differences with the first embodiment. The order support system 1 will be described.

FIG. 19 is a block diagram showing the main functions of the order support apparatus 10A of the second embodiment of the present invention and the main data stored by the order support apparatus 10A.

As shown in FIG. 19, the data processing unit 110 of the order support device 10A of the second embodiment includes a demand prediction unit 111A instead of the demand prediction unit 111, and an evaluation unit 115A instead of the evaluation unit 115. Further, the storage unit 120 stores prediction model information 123A instead of the prediction model information 123, and stores inventory information 126A instead of the inventory information 126.

FIG. 20 is an explanatory diagram of an example of prediction model information 123A managed by the storage unit 120 of the order support apparatus 10A according to the second embodiment of the present invention as a table of a database.

In addition to the items of the record of the prediction model information 123 in the first embodiment, the record of the prediction model information 123A of the second embodiment further includes the data start date 1217 and the level adjustment 1218. Further, the record of the prediction model information 123A has items of a model type 1212A in place of the model type 1212 and an update schedule 1215A in place of the update schedule 1215 and an update content 1216A in place of the update content 1216.

In the model type 1212A, when each product identified by the value of the product ID 1211 is a corresponding product in order placement, "order placement" is set. If each product is not a corresponding product in order placement, the type of demand forecasting model is set as in the first embodiment.

In the update schedule 1215A, a schedule date for updating the demand forecasting model is set. In the update content 1216A, in addition to the “model type” and “parameter” set as the update content 1216 of the prediction model information 123 in the first embodiment, the products supported by the order acceptance change from the order acceptance to the prediction target It is registered as "new" when it becomes.

In the data start date 1217, when the demand forecasting model needs to be updated, the start date of data used when rebuilding the demand forecasting model is set. For example, when information on the product is accumulated in the sales information 121 from May 17, 2015, the data from May 17, 2015 is used when the demand forecasting model is first constructed. Thereafter, when the sales trend changes, such as a change in sales volume level or a change in sales frequency, the data used for rebuilding the demand forecast model uses the data after the change in the sales trend. If the latest sales trend has changed on October 1, 2016, the data start date 1217 is set to "2016/10/1." This can prevent deterioration in prediction accuracy that may occur when constructing a demand prediction model using data with different sales tendencies.

In level adjustment 1218, when updating the demand forecasting model, the amount of data used to build the demand forecasting model is not sufficient, and until the sufficient amount of data is accumulated, it is calculated by the existing demand forecasting model When dealing with by adjusting the level of the forecast demand amount, “1” is set, and otherwise “0”. For example, if the data start date is "2016/10/10" and the amount of data required to build the demand forecast model is set to two weeks, the demand forecast model is updated to 2016/10/15 or later Ru. Until then, when performing level adjustment, the level adjustment 1218 is set to “1”. The amount of data required to construct a demand forecasting model may be set for each type of demand forecasting model, or the same value may be set for all types of demand forecasting models.

FIG. 21 is an explanatory diagram showing an example of inventory information 126A managed by the storage unit 120 of the order support apparatus 10A of the second embodiment of the present invention as a table of a database.

In addition to the items of the record of the inventory information 126 in the first embodiment, the record of the inventory information 126A of the second embodiment further includes an item of the discard amount 1416. In the discard amount 1416, when the product is discarded due to consumption expiration or the like, the quantity is set.

The demand prediction unit 111A constructs one or more demand prediction models from the sales information 121 and the external information 122 when the prediction model information 123A of the product is not generated, and generates the prediction model information 123A. The demand forecasting unit 111A does not construct the demand forecasting model of the product when the sale frequency of the product is extremely low, or the occurrence of discarding due to consumption expiration is large, and the product is regarded as a product corresponding to the order acceptance order. It registers in prediction model information 123A. For example, if the demand prediction unit 111A has a sales frequency equal to or less than a predetermined frequency (for example, once a month or lower than the sales frequency serving as the determination criterion of the low frequency model in the first embodiment), the sales interval is If it is longer than the expiration date registered in the product information 125 of the product, it is determined that the sales frequency is extremely low. In addition, the demand forecasting unit 111A determines that there is a large amount of discarding when the ratio of the discard amount to the delivery (purchasing) amount of the product (hereinafter referred to as a discard rate) is higher than a value set in advance. .

The demand forecasting unit 111A reconstructs the demand forecasting model when the update schedule is registered in the forecasting model information 123A. Further, when the product is not a corresponding product in order placement, the demand prediction unit 111A calculates the predicted demand amount based on the prediction model information 123A, the sales information 121, and the external information 122, and generates the prediction information 124.

The evaluation unit 115A compares the past sales volume stored in the sales information 121 with the predicted demand amount on the same day when the inventory level is not appropriate for the product not supported by the order placement, and the actual sales volume When the difference between the forecast demand amount and the demand amount is large, it is determined that the demand forecast model needs to be updated, and the update schedule is registered in the forecast model information 123A of the product. The evaluation unit 115A also determines whether or not the product corresponding to the received order is to be supported by the received order, and registers so as to cancel the received order as necessary.

<Description of processing>
FIG. 22 is a flowchart illustrating an example of processing performed by the demand prediction unit 111A and the demand prediction integration unit 112 according to the second embodiment of this invention.

The processes given the same reference numerals as the processes performed by the demand prediction unit 111 and the demand prediction integration unit 112 in the first embodiment of FIG. 17 are the same as the processes in FIG. 17 and thus detailed description will be omitted.

If the demand forecasting model has not been built (S1702: NO), the demand forecasting unit 111A should respond to the order by ordering the product based on the sales frequency, sales interval, and discard rate of the target product set in S1701. It is determined whether it is a product (S2212). If it is a product to be dealt with by order placement (S2212: YES), the processing proceeds to S2213, and if no order placement is required (S2212: NO), the processing proceeds to S1704. The demand prediction unit 111A registers “order acceptance order” as the “order acceptance order” in the model type 1212A of the prediction model information 123A when the commodity is a commodity to be dealt with by order acceptance (S2213).

If the demand forecasting model that has already been built has an update schedule (S1703: YES), the demand forecasting unit 111A determines whether the scheduled update date registered in the update schedule 1215A of the forecast model information 123A has passed (S2214). If the scheduled update date has passed (S2214: YES), the processing proceeds to S1705, and if the scheduled update date has not come (S2214: NO), the processing proceeds to S2215.

If the demand forecasting unit 111A determines that the scheduled update date has not come (S2214: NO), if it executes S1705, if it determines that there is no scheduled update (S1703: NO), if it executes S2213, When S1704 is executed, it is determined whether the demand forecast of the product is necessary based on whether or not the model type 1212A set in the prediction model information 123A of the target product is "order acceptance" (S2215). If the model type 1212A is not "order acceptance", the demand prediction unit 111A determines that a demand forecast is necessary (S2215: YES), and proceeds to S1706. When the model type 1212A is “order acceptance order”, the demand prediction unit 111A determines that the demand prediction is not necessary (S2215: NO) and ends the processing.

If there is a demand forecasting model for which the forecasted demand volume is not calculated (S1707: YES), the demand forecasting unit 111A computes the forecasted demand volume (S1709A). In the calculation of the forecast demand amount in S1709A, the demand forecasting unit 111A adjusts the level of the forecast demand amount when the level adjustment 1218 of the forecast model information 123A of the product is set to “1”. For example, in the predetermined period (for example, the last two weeks), the demand prediction unit 111A continues the day when the predicted demand amount is larger than the sales volume for a predetermined length (for example, 10 days) and the median of the error is the latest 2 When it becomes more than the predetermined ratio (for example, 10%) of the average sales volume of a week, you may adjust so that a predicted demand amount may become small by integrating | accumulating the numerical value less than 1 with the calculated predicted demand amount. Also, for example, in the last two weeks, the demand prediction unit 111A continues the case where the predicted demand amount is less than the sales volume for 10 days or more, and the median of the error is 10% or more of the average sales volume for the last two weeks. Alternatively, the predicted demand amount may be adjusted to be large by integrating a numerical value larger than 1 with the calculated predicted demand amount.

23 and 24 are flowcharts showing an example of processing performed by the evaluation unit 115A according to the second embodiment of this invention.

About the process which attached the same code | symbol as the process which the evaluation part 115 in Example 1 of FIG. 18 performs, since it is the same as the process of FIG. 18, detailed description is abbreviate | omitted.

The evaluation unit 115A determines whether the target product is a product that is supported by the order placement (S2308). If the target product is a product that supports order acceptance (S2308: YES), the process advances to S2309, and if the target product is not a product that supports order acceptance (S2308: NO), the process advances to S1801.

The evaluation unit 115A determines whether or not the target product is to be used as the received order when the target product is a product corresponding to the received order (S2309). Here, the evaluation unit 115A performs the continuation determination based on the sales frequency, the sales interval, and the discard rate of the product. If it is determined that the order is to be made continuously (S2309: YES), the process is ended, and if the order is released (S2309: NO), the process proceeds to S2310.

In S2310, the evaluation unit 115A registers "new" in the updated content 1216A of the predicted model information 123A, and starts data on the date when the change in sales frequency, sales interval, and discard rate caused the cancellation of the received order. Register as day 1217. Further, a date obtained by adding the number of days until data of an amount necessary for constructing a demand forecasting model is added to the data start date 1217 is registered in the update schedule 1215A as a scheduled update date.

If there is a change in the sales frequency (S1804: YES), the evaluation unit 115A proceeds to S2311. In S2311, when the evaluation unit 115A changes the sales frequency to a decreasing direction, the evaluation unit 115A determines, based on the sales frequency, the sales interval, and the discard rate of the product, whether to switch to the correspondence in order placement. When switching to the correspondence in the order placement (S2311: YES), the processing proceeds to S2312, and when not switching to the correspondence in the placement ordering (S2311: NO), the processing proceeds to S2313.

When switching to the order acceptance order, the evaluation unit 115A deletes the record in which the product ID 1211 of the prediction model information 123A is the ID of the product, sets the model type 1212A as "order acceptance order", and the frequency 1213 as "low". Information is newly added as prediction model information 123A of the product (S2312).

When not switching to the order acceptance order, the evaluation unit 115A identifies the day on which the change in sales frequency occurred (S2313), and is necessary to change the type of demand forecasting model to the forecasting model information 123A of the product. Information is registered (S1806A), and the process proceeds to S2315.

In S1806A, the evaluation unit 115A registers the date specified in S2313 in the data start date 1217 of the prediction model information 123A, and sets the number of days until data of the amount necessary to construct the demand prediction model is accumulated on the data start date. The added date is registered in the update schedule 1215A of the prediction model information 123A, and is registered as "model information" in the update content 1216A of the prediction model information 123A.

If there is a discrepancy between the sales volume and the level of the predicted demand (S1805: YES), the evaluation unit 115A specifies the day when the discrepancy between the sales volume and the level of the forecasted demand occurs (S2314). For example, the evaluation unit 115A determines that the average sales volume for a predetermined period (for example, the immediately preceding week) is predetermined based on the average sales amount for a predetermined period (for example, the immediately preceding week) at a predetermined time (for example, seven days before that). The time when a day that is larger than the ratio (for example, 1.3 times or more) continues for a predetermined number of days (for example, 3 days) or more, or a day that is smaller than the predetermined ratio (for example, 0.7 times or less) It may be determined that the time when the number of days (for example, three days) or more has continued is the day when the level of the sales volume and the forecasted demand amount deviates. The evaluation unit 115A registers information necessary for updating the parameters of the demand prediction model in the prediction model information 123A of the product (S1807A), and proceeds to S2315.

In S1807A, the evaluation unit 115A registers the date specified in S2314 in the data start date 1217 of the prediction model information 123A, and sets the number of days until data of the amount necessary for constructing the demand prediction model is accumulated on the data start date. The added date is registered in the update schedule 1215A of the prediction model information 123A, and is registered as "parameter" in the update content 1216A of the prediction model information 123A.

In S2315, the evaluation unit 115A determines whether to cope with by adjusting the level of the forecast demand amount calculated by the existing demand forecast model until the scheduled update date registered in the forecast model information 123A. For example, in a predetermined period (for example, the last two weeks), the evaluation unit 115A continues the day when the predicted demand amount is larger than the sales volume for a predetermined number of days (for example, 10 days), and the median of the error is for the last two weeks. It may be determined that the level adjustment is to be performed when the average sales amount is equal to or more than a predetermined ratio (for example, 10%). The evaluation unit 115A also sets the level when the predicted demand amount falls below the sales volume for 10 days or more and the median of the error is 10% or more of the average sales volume for the last 2 weeks in the last 2 weeks. It may be determined that the adjustment is to be carried out.

If the level adjustment is to be performed (S2315: YES), the evaluation unit 115A sets “1” to the level adjustment 1218 of the prediction model information 123A (S2316). If the level adjustment is not to be performed (S2315: NO) The level adjustment 1218 of the model information 123A is set to “0” (S2317), and the process is ended.

As described above, according to the order support system 1 of the second embodiment, when updating the demand forecasting model according to the change in the sales trend, the demand forecasting model is used by using the data after the change in the sales trend. Build Therefore, it is possible to perform demand forecasting with higher accuracy, and to present more appropriate order quantity. In addition, for products with extremely low sales frequency, the demand forecast and the recommended order volume are not presented, and it is set so as to cope with order placement. Therefore, it is possible to reduce the disposal of products generated due to low sales frequency, and to reduce the number of products for which the recommended order quantity is to be calculated, thereby reducing resources required for processing. it can.

The present invention is not limited to the embodiments described above, but includes various modifications. For example, the embodiments described above have been described in detail for better understanding of the present invention, and are not necessarily limited to those having all the configurations of the description. In addition, part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. In addition, with respect to a part of the configuration of each embodiment, it is possible to add, delete, and replace other configurations.

Further, each of the configurations, functions, processing units, processing means, etc. described above may be realized by hardware, for example, by designing part or all of them with an integrated circuit. Further, each configuration, function, and the like described above may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files for realizing each function can be stored in a nonvolatile semiconductor memory, a hard disk drive, a storage device such as a solid state drive (SSD), or a computer readable non-volatile memory such as an IC card, an SD card, or a DVD. It can be stored on a temporary data storage medium.

Further, the control lines and the information lines indicate what is considered to be necessary for the explanation, and not all the control lines and the information lines in the product are necessarily shown. In practice, almost all configurations may be considered to be mutually connected.

Claims

An order support system having a storage unit and a data processing unit, the order support system comprising:
The storage unit holds information on sales results of products.
The data processing unit
A demand forecasting unit that constructs a demand forecasting model of the product based on the sales record and calculates a forecasted demand amount of the product using the constructed demand forecasting model;
Based on the predicted demand amount, the risk of calculating a risk value indicating at least one of the risk of occurrence of sales opportunity loss of the product and the risk of occurrence of excess inventory of the product for each ordered quantity of the product A value calculation unit,
An order quantity output unit that outputs a recommended order quantity based on the calculated risk value;
An evaluation unit that determines whether to update the demand forecasting model based on the sales record and the forecasted demand amount,
The order support system characterized in that the demand forecasting unit updates the demand forecasting model when it is determined that the evaluation unit updates the demand forecasting model.
The order support system according to claim 1, wherein
The storage unit further holds product information including a expiration date of the product and a lead time which is a time from ordering to receiving the product.
The order support system, wherein the risk value calculation unit calculates the risk value based on the predicted demand amount, the ordered amount of the product, and the product information.
The order support system according to claim 1, wherein
The storage unit further holds information on the life cycle of the product,
The ordering support system, wherein the evaluation unit determines whether or not the demand forecast model is to be updated at a time when it is estimated that the demand amount of the product changes from the life cycle of the product.
The order support system according to claim 1, wherein
The storage unit further holds information on the stock amount of the product,
The ordering support system, wherein the evaluation unit determines whether or not the demand forecast model is to be updated, when it is determined that the stock amount of the product is not appropriate.
The order support system according to claim 1, wherein
When the evaluation unit determines that the demand forecasting model is to be updated, updating of the demand forecasting model is performed based on the sales frequency of the product and the difference between the forecasted demand amount of the product and the sales performance in the past. Determine the content,
The order support system, wherein the demand forecasting unit updates the determined contents with respect to the demand forecasting model.
The ordering support system according to claim 5, wherein
The demand forecasting unit calculates the forecast demand amount for each first period of a predetermined length based on the sales record when the sales frequency of the product is higher than a predetermined first reference. Build a demand forecasting model
The evaluation unit
The demand forecasting unit calculates a predicted demand amount for each second period longer than the first period based on the sales result when the sales frequency of the product becomes lower than the first standard. Determine the updated contents of the demand forecasting model so as to construct the demand forecasting model for the purpose or to construct a specific type of demand forecasting model suitable for low frequency of sales;
When the frequency of sales of the product is higher than the first standard, and the difference between the predicted demand amount of the product and the sales performance in the past is larger than a predetermined second standard, the demand forecasting unit determines the sales The order support system is characterized by determining the update content of the demand forecasting model so as to change the parameter of the demand forecasting model for calculating the forecasted demand for each first period based on the actual results. .
The order support system according to claim 1, wherein
The demand forecasting unit does not construct a demand forecasting model of the product if it determines that the sale frequency of the product is lower than a predetermined third standard, or if it determines that the amount of waste of the product is large. An order support system that outputs information indicating that the product is ordered according to an order received.
The order support system according to claim 7, wherein
The storage unit holds information indicating the expiration date of the product, the past purchase amount of the product, and the past disposal amount of the product.
The demand forecasting unit determines that the sales frequency of the product is lower than the third reference when the sales interval of the product is longer than the expiration date of the product, and the product for the past purchase amount of the product is determined It is determined that the amount of discarding of the product is large when the ratio of the amount of discarding in the past exceeds a predetermined standard.
The order support system according to claim 1, wherein
The demand forecasting unit constructs a plurality of demand forecasting models of the product based on the sales record, and calculates a plurality of forecast demand quantities of the product using the constructed plurality of demand forecasting models.
The order support system, wherein the data processing unit further includes a demand prediction integration unit that integrates the plurality of predicted demand amounts.
The ordering support system according to claim 9, wherein
The demand forecast integration unit calculates forecast errors of the demand forecast models by comparing the forecast demand amount calculated using each of the plurality of demand forecast models in the past with the past sales results. An order support system characterized by integrating the plurality of forecast demand amounts so that the contribution of the forecast demand amount calculated using a demand forecast model with a small forecast error becomes large.
The order support system according to claim 1, wherein
The risk value calculation unit calculates a risk that a sales opportunity loss of the product occurs and a risk that an excess inventory of the product occurs, and calculates the risk value by integrating predetermined weights to both and integrating them.
The order quantity output unit outputs the order quantity of the product for which the risk value is minimized as the recommended order quantity, or the risk value falls within a predetermined allowable range, and the sales opportunity loss for the product At least one of the order quantity of the product which minimizes the risk of occurrence of the product, and the order quantity of the product whose risk value is within a predetermined allowable range and the risk of excess inventory of the product is minimal An order support system, which outputs as the recommended order quantity.
The order support system according to claim 1, wherein
The store selling the item further comprises an ordering terminal used to order the item;
The data processing unit further includes a recommendation result providing unit that transmits the recommended order amount to the ordering terminal,
The ordering terminal is
Display a screen for accepting an order for the product,
An order support system, wherein the recommended order amount is displayed as an initial value of the order amount of the product on a screen for receiving an order for the product.
An order support program for causing a computer system having a processor and a storage device connected to the processor to execute the following procedure:
A demand forecasting procedure of constructing a demand forecasting model of the product based on sales results of the product held in the storage device and calculating a forecasted demand amount of the product using the constructed demand forecasting model;
Based on the predicted demand amount, the risk of calculating a risk value indicating at least one of the risk of occurrence of sales opportunity loss of the product and the risk of occurrence of excess inventory of the product for each ordered quantity of the product Value calculation procedure,
An order quantity output procedure for outputting a recommended order quantity based on the calculated risk value,
An evaluation procedure for determining whether to update the demand forecasting model based on the sales record and the forecasted demand amount;
An ordering support program that causes the processor to execute the demand forecasting procedure to update the demand forecasting model when it is determined that the demand forecasting model is updated in the evaluation procedure.
The ordering support program according to claim 13, wherein
The demand forecasting step includes a step of constructing a plurality of demand forecasting models of the product based on the sales record, and calculating a plurality of forecasted demand amounts of the commodity using the constructed plurality of demand forecasting models.
The ordering support program according to claim 1, wherein the ordering support program further causes the processor to execute a demand prediction integration procedure for integrating the plurality of forecast demand amounts.
An ordering support method executed by an ordering support system having a storage unit and a data processing unit,
The storage unit holds information on sales results of products.
The order support method is
A demand forecasting procedure in which the data processing unit constructs a demand forecasting model of the product based on the sales record and calculates a forecasted demand amount of the product using the constructed demand forecasting model;
The data processing unit indicates, based on the predicted demand amount, at least one of a risk of occurrence of a sales opportunity loss of the product and a risk of an excess inventory of the product, for each ordered quantity of the product. Risk value calculation procedure for calculating the risk value;
An order quantity output procedure in which the data processing unit outputs a recommended order quantity based on the calculated risk value;
An evaluation procedure of determining whether the data processing unit updates the demand forecast model based on the sales record and the forecasted demand amount;
The order support method, wherein the data processing unit executes the demand forecasting procedure to update the demand forecasting model when it is determined that the demand forecasting model is updated in the evaluation procedure.