JP2010086278A - Automatic technique selection device, automatic technique selection method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically select an optimal combination for every merchandise by a number of demand prediction techniques and stock simulation techniques. <P>SOLUTION: The automatic technique selection device includes: a merchandise sales achievement storage part 15 for storing sales achievement during a merchandise verification period; a predicted demand calculation part 11 for using a demand prediction technique to calculate a predicted demand during the verification period on the basis of the merchandise sales achievement; a theoretical stock calculation part 12 for using a stock simulation technique to calculate theoretical stock during the verification period on the basis of the predicted demand; an order prediction error output part 13 for calculating an order predicted error every combination of a demand prediction technique and the stock simulation technique from a difference between the theoretical stock and the merchandise sales achievement; and a technique evaluation part 14 for selecting a combination of the demand prediction technique and the stock simulation technique corresponding to the merchandise based on the order prediction error. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a method automatic selection device, a method automatic selection method, and a program for selecting a method such as demand prediction and inventory simulation.

In general, demand prediction and inventory simulation are used in retail stores and the like in order to prevent out-of-stock and excess inventory. However, there are many demand forecasting methods and inventory simulation methods, and selection of the demand forecasting method and inventory simulation method requires knowledge of experience, know-how, and characteristics of each product.
Therefore, it is considered beneficial to be able to automatically select the optimum combination from a number of demand forecasting methods and inventory simulation methods. This is because the combination of methods is enormous, and it takes time to manually select methods according to the characteristics of each product.

In the technology described in Patent Document 1, the forecast demand results are obtained by a plurality of demand forecasting methods based on the life cycle pattern and the latest sales results, and the shipment quantity rank, average absolute error rate, cumulative error rate, and seasonal stability are calculated. Based on the comprehensive evaluation of the predicted demand result, the optimum predicted demand result is selected as a sales plan.
Moreover, in the technique described in Patent Document 2, when setting values such as the planning cycle, procurement lead time, and planning lead time are updated, the safety stock is recalculated based on the updated setting values. Appropriate inventory management. The safety stock is a stock that is secured to prevent out of stock even if an error occurs in the prediction.
JP 2009-21020 A Japanese Patent No. 3991811

However, in the technique described in Patent Document 1, inventory transition is not considered because inventory simulation is not used. For this reason, there is a problem that it is not possible to grasp the shortage situation or the excess or deficiency of the inventory, and the evaluation cannot be made according to reality.
In the technique described in Patent Document 2, the lead time is taken into consideration, but the inventory simulation method cannot be automatically selected.
The present invention has been made in view of the above points, and an object of the present invention is to automatically select an optimal combination for each product from a number of demand prediction methods and inventory simulation methods. To provide an automatic selection method and program.

  The present invention has been made to solve the above problems, and one aspect of the present invention is a product sales record storage unit that stores a sales record during a product verification period, and a demand prediction method. A forecast demand calculation unit that calculates a forecast demand amount during the verification period based on the sale results of the merchandise read from a merchandise sales record storage unit, and a forecast calculated by the forecast demand calculation unit using an inventory simulation method A theoretical inventory calculation unit that calculates a theoretical inventory during the verification period based on a demand amount, and a combination of a demand prediction method and an inventory simulation method based on a difference between the theoretical inventory calculated by the theoretical inventory calculation unit and the sales performance of the product. An order prediction error output unit for calculating an order prediction error for each item, and a demand prediction method and an inventory simulation corresponding to the product based on the order prediction error calculated by the order prediction error output unit. And techniques evaluation unit for selecting a set of down approach, a method automatic selection device, characterized in that it comprises a.

  In addition, according to one aspect of the present invention, in the above-described method automatic selection device, a lead time storage unit that stores a lead time from ordering to delivery of a product is provided, and the theoretical stock calculation unit includes a lead time corresponding to the product. Is read from the lead time storage unit, the order quantity during the verification period is calculated from the lead time and the predicted demand quantity using the inventory simulation method, and the delivery quantity during the verification period is calculated from the order quantity. The theoretical inventory during the verification period is calculated based on the delivered quantity.

  Further, according to one aspect of the present invention, in the above-described method automatic selection device, the order prediction error output unit calculates a stockout rate and an inventory level during the verification period, and the method evaluation unit performs the verification period The shortage rate is selected from among a set of a demand forecasting method and an inventory simulation method in which the stockout rate does not exceed a predetermined upper limit.

  Further, one aspect of the present invention is a method automatic selection method performed by a method automatic selection device including a product sales record storage unit that stores a sales record during a product verification period, wherein the predicted demand calculation unit includes a demand Calculating a predicted demand amount during the verification period based on the sales performance of the product using a prediction method; and a theoretical inventory calculation unit using the inventory simulation method based on the predicted demand amount during the verification period A step of calculating a theoretical inventory of the order, and an order prediction error output unit calculating an order prediction error for each set of a demand prediction method and an inventory simulation method from a difference between the theoretical stock and the actual sales of the product, and a method The evaluation unit includes a step of selecting a combination of a demand prediction method and an inventory simulation method corresponding to the product based on the order prediction error. It is a technique automatic selection method.

Further, according to one aspect of the present invention, a computer is configured to calculate a predicted demand amount during the verification period based on a sales record during a verification period of a product stored in advance using a demand prediction method; Using the method, calculating the theoretical inventory during the verification period based on the predicted demand, and predicting the order for each set of the demand forecasting method and the inventory simulation method from the difference between the theoretical inventory and the sales performance of the product. Calculating an error; selecting a combination of a demand prediction method and an inventory simulation method corresponding to the product based on the order prediction error;
Is a program for executing

  According to the present invention, the theoretical inventory is calculated using the inventory simulation from the predicted demand amount calculated using the demand prediction method, and therefore, an order prediction error that is an evaluation index for each set of the demand prediction method and the inventory simulation method. Can be calculated. This makes it possible to automatically select an optimal combination for each product from a large number of demand prediction methods and inventory simulation methods.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
First, the words used below are defined as follows.
The verification period is a period in which there is an actual value (sales result) of the sales volume of the product, and is a period during which verification is performed.
The verification date is an arbitrary day during the verification period.
The predicted demand amount is a predicted value of the demand amount after the next day after the verification date, which is obtained based on the sales performance before the verification date. The predicted demand is calculated using a demand prediction method.
The lead time is the number of days from when a product is ordered until it is delivered, and is obtained from “delivery date−order date”.
The order quantity is a forecast value obtained based on the forecast demand quantity and the lead time days, and is a forecast value of the quantity for each day that a retail store orders goods. The order quantity is calculated using an inventory simulation method.
The delivery amount is a predicted value obtained based on the order quantity, and is a predicted value of the daily amount that the product is delivered to the retail store.
The theoretical stock is a predicted value of the stock amount obtained based on the delivery amount and the predicted demand amount. The theoretical stock on the verification date is a value obtained by adding the delivery amount to the theoretical stock on the day before the verification date and subtracting the predicted demand amount. The theoretical inventory on the first day of the verification period is the actual value of the inventory.
The order prediction error is an evaluation index obtained based on the difference between the theoretical inventory and the sales performance, and includes an inventory level, an inventory amount, a stockout quantity, and an opportunity loss. When the value obtained by subtracting the actual sales from the theoretical inventory is positive, the value is the inventory level, and 0 is the amount of missing items. In addition, when the value obtained by subtracting the actual sales from the theoretical inventory is negative, the absolute value of the value is the shortage quantity and 0 is the inventory level. When the value obtained by subtracting the actual sales from the theoretical inventory is zero, the inventory level and the quantity of missing items are zero. The stock amount is a value obtained by multiplying the stock level by the sales unit price of the product. The opportunity loss is a value obtained by multiplying the amount of missing items by the sales unit price of the product. The method automatic selection apparatus according to the present embodiment selects a combination of a demand prediction method and an inventory simulation method based on an order prediction error.

FIG. 1 is a block diagram illustrating a functional configuration of the method automatic selection apparatus 1 according to the present embodiment.
The method automatic selection device 1 is a device that selects a combination corresponding to each product from a plurality of demand prediction methods and a plurality of inventory simulation methods. The method automatic selection apparatus 1 is provided with means for storing a program for executing processing based on a plurality of demand prediction methods and a plurality of inventory simulation methods in advance in association with the names of the methods. A function for executing processing based on a demand prediction method and an inventory simulation method is provided. The method automatic selection device 1 selects a combination of a stored demand prediction method name and an inventory simulation method name.
The demand prediction method is a method of calculating a predicted demand amount from the sales performance of a product. Examples of the demand prediction method include time series analysis AR (AutoRegressive model), MA (Moving Average model), ARIMA model (AutoRegressive Integrated Moving Average model), simple moving average, exponential smoothing method, multiple regression model, and a mixture thereof.
The inventory simulation method is a method of calculating the order quantity from the predicted demand amount of each product. The inventory simulation method uses, for example, a normal distribution or a Poisson distribution based on a method for calculating the order quantity by subtracting the stock quantity on the order date from the accumulation of the forecast demand quantity from the order date to the delivery date. There is a method of calculating an appropriate stock quantity and subtracting the stock quantity on the order date from the calculated appropriate stock quantity to calculate the order quantity.

For example, the inventory simulation based on the normal distribution is useful when the number of sales per day is several tens or more. If the number of sales per day is large, it is assumed that the actual results are based on a normal distribution. Therefore, the average value and variance are calculated from the past results, and an inventory quantity of 3σ (three times the square root of variance) is secured. In this case, the shortage rate can be suppressed to 1% or less. Specifically, the average and variance are obtained from the results of the past two weeks (changed according to the product characteristics), and the stock quantity is three times the square root of variance.
In addition, Poisson distribution is useful for few products with few daily sales and new products for which data is not collected. If the stock quantity at the point where the accumulated Poisson distribution with the actual value as an average is 99% is secured, the shortage rate can be suppressed to 1% or less.

  The method automatic selection apparatus 1 stores a predicted demand amount calculation unit 11, a theoretical inventory calculation unit 12, an order prediction error output unit 13, a method evaluation unit 14, and a product sales record table that is a sales record of products. A sales record storage unit 15, a verification period storage unit 16 for storing a verification period table indicating the verification period, and a lead time storage unit 17 for storing a lead time table indicating the lead time from ordering to delivery of each product. Consists of including.

The predicted demand amount calculation unit 11 calculates a predicted demand amount for each verification date from the product sales performance table using a demand prediction method, and generates a predicted demand table. The verification date is a date within the verification period stored in the verification period table. The calculated predicted demand amount is stored in the predicted demand table.
The theoretical inventory calculation unit 12 calculates the theoretical inventory for each verification date from the predicted demand amount calculated by the predicted demand amount calculation unit 11 and the lead time days stored in the lead time table using an inventory simulation method.
The demand prediction method storage unit 18 stores a program for executing processing based on the demand prediction method in association with the name of the method.
The stock simulation technique storage unit 19 stores a program for executing processing based on the stock simulation technique in association with the name of the technique.

The order prediction error output unit 13 calculates an order prediction error for each set of the demand prediction method and the inventory simulation from the theoretical inventory and the sales performance, and generates an order prediction error table. In the order prediction error table, the calculated inventory level, missing item quantity, opportunity loss, and inventory amount are stored.
The method evaluation unit 14 evaluates a combination of each demand prediction method and the inventory simulation method based on the order prediction error calculated by the order prediction error output unit 13 and selects an optimal combination. Moreover, the method evaluation part 14 produces | generates the evaluation result table which stores the evaluation rank which is an evaluation result.

FIG. 2 is a schematic diagram illustrating a data structure of a product sales record table stored in the product sales record storage unit 15 in the present embodiment.
As shown in the figure, the product sales record table is a two-dimensional tabular data composed of rows and columns, and each of the product JAN (Japan Article Number) code, date, sales quantity, unit price, and inventory quantity. Has a column of items. Each row of this table exists for each set of product JAN code and date. The product JAN code is a code for identifying each product. The sales quantity (sales result) is the sales quantity at the end of the day. The unit price is the unit sales price of the product. The stock quantity is the stock quantity at the end of the day.

FIG. 3 is a schematic diagram illustrating a data structure of a verification period table stored in the verification period storage unit 16 according to the present embodiment.
As shown in the figure, the verification period table has a column for each item of the verification start date and the verification end date. The date is stored in the verification start date. The date after the verification start date is stored in the verification end date. The verification period is a period from the verification start date to the verification end date.

FIG. 4 is a schematic diagram showing a data structure of a lead time table stored in the lead time storage unit 17 in the present embodiment.
As shown in the figure, the lead time table is two-dimensional tabular data composed of rows and columns, and has columns for each item of product JAN code and lead time days. Each row of this table exists for each product JAN code.

  FIG. 5 is a schematic diagram illustrating a data structure of a predicted demand table generated by the predicted demand amount calculation unit 11 in the present embodiment. The forecast demand table is a two-dimensional tabular data consisting of rows and columns, and has columns for each item of product JAN code, demand forecast method, verification date, forecast date, and forecast demand. Yes. Each row of this table exists for each set of product JAN code, demand prediction method, verification date, and prediction date. In the verification date, any date in the verification period set in the verification period table is stored. In the prediction date, any date from the day after the verification date to the verification end date is stored. The predicted demand amount stores the predicted demand amount when the prediction date is verified on the verification date using the demand prediction method. That is, the predicted demand amount calculation unit 11 calculates the predicted demand amount on the prediction date using the product sales record table before the verification date.

  FIG. 6 is a schematic diagram illustrating a data structure of an order prediction error table generated by the order prediction error output unit 13 according to the present embodiment. The order forecast error table is a two-dimensional tabular data consisting of rows and columns. The product JAN code, demand forecast method, verification date, theoretical inventory, forecast demand, inventory simulation method, It has columns for each item of quantity, inventory level, opportunity loss, and inventory value. Each row of this table exists for each set of product JAN code, demand prediction method, verification date, and inventory simulation method.

  FIG. 7 is a schematic diagram illustrating a data structure of an evaluation result table generated by the method evaluation unit 14 in the present embodiment. The evaluation result table is two-dimensional tabular data composed of rows and columns, and includes columns for each item of a product JAN code, an evaluation rank, a demand prediction method, and an inventory simulation method. The evaluation result table exists for each set of the product JAN code, the demand prediction method, and the inventory simulation method. A positive integer is stored in the evaluation rank. The smaller the number of evaluation ranks, the higher the evaluation.

  Next, with reference to FIG. 8, the method selection processing by the method automatic selection device 1 will be described. FIG. 8 is a flowchart showing the procedure of the method selection process. In the method selection process, a combination of a demand prediction method and an inventory simulation method that are optimal for a certain product is selected. The method automatic selection device 1 performs a method selection process on all the products stored in the product sales record table, and selects a combination of a demand prediction method and an inventory simulation method that are optimal for each product.

[Step S1: Predicted demand calculation]
First, in step S1, the predicted demand calculation unit 11 calculates the predicted demand for each verification date from the product sales performance table using each demand prediction method. A specific procedure of the predicted demand amount calculation process will be described with reference to FIG. FIG. 9 is a flowchart showing the procedure of the predicted demand amount calculation process.
First, in step S <b> 101, the predicted demand calculation unit 11 selects one demand prediction method that can be executed by the method automatic selection device 1.
Next, in step S102, the predicted demand calculation unit 11 sets the verification start date set in the verification period table as the verification date.
Next, in step S103, the predicted demand calculation unit 11 sets the next day of the verification date as the predicted date.
Next, in step S104, the predicted demand amount calculation unit 11 extracts the sales quantity corresponding to the product from the product sales performance table and the date before the verification date, and extracts it using the selected demand prediction method. Calculate the forecast demand on the forecast date from the sales quantity. Then, the predicted demand amount calculation unit 11 stores the calculated predicted demand amount in the predicted demand table.
Next, in step S105, the predicted demand calculation unit 11 determines whether or not the predicted date is the verification date + the maximum lead time. The maximum lead time days is the maximum value of the lead time days stored in the lead time table. If the predicted date is the verification date + the maximum lead time, the process proceeds to step S107. On the other hand, if the predicted date is before the verification date + the maximum lead time, the process proceeds to step S106.
In step S106, the predicted demand calculation unit 11 sets the next day of the prediction date as the prediction date, and returns to step S104.

On the other hand, in step S107, the predicted demand amount calculation unit 11 sets the next day of the verification date as the verification date.
Next, in step S108, the predicted demand calculation unit 11 determines whether the verification date is a verification end date. If the verification date is the verification end date, the process proceeds to step S109. On the other hand, if the verification date is before the verification end date, the process returns to step S103.
In step S109, the predicted demand calculation unit 11 determines whether or not the processing shown in steps S102 to S108 has been performed for all the demand prediction methods. When all the demand prediction methods are performed, the process is terminated. On the other hand, if not, the process proceeds to step S110.
In step S110, the next demand prediction method is selected and the process returns to step S102.

[Step S2: Order Quantity Calculation]
Next, in step S2, the theoretical inventory calculation unit 12 calculates the order quantity from the forecast demand table and the lead time table for each verification date using each inventory simulation method. A specific procedure of the order quantity calculation process will be described with reference to FIG. FIG. 10 is a flowchart showing the order quantity calculation process.
First, in step S <b> 201, the theoretical inventory calculation unit 12 selects one demand prediction method that can be executed by the method automatic selection device 1. And the forecast demand amount corresponding to the selected demand forecast method is extracted from the forecast demand table produced | generated by step S1.
Next, in step S <b> 202, the theoretical stock calculation unit 12 selects one stock simulation method that can be executed by the method automatic selection device 1.
Next, in step S203, the theoretical inventory calculation unit 12 sets the verification start date set in the verification period table as the verification date.
Next, in step S204, the theoretical inventory calculation unit 12 extracts the one corresponding to the verification date from the predicted demand amount extracted in step S201. Then, the theoretical inventory calculation unit 12 reads the lead time days corresponding to the product from the lead time table, and uses the inventory simulation method selected in step S202 to verify the verification date from the read lead time days and the extracted predicted demand amount. Calculate the order quantity.

Next, in step S205, the theoretical inventory calculation unit 12 determines whether or not the verification date is the verification end date set in the verification period table. If the verification date is the verification end date, the process proceeds to step S207. On the other hand, if the verification date is before the verification end date, the process proceeds to step S206.
In step S206, the theoretical stock calculation unit 12 returns to step S204 with the next day of the verification date as the verification date.
On the other hand, in step S207, the theoretical inventory calculation unit 12 determines whether or not the processes shown in steps S203 to S206 have been performed for all the inventory simulation methods. If all inventory simulation methods have been performed, the process proceeds to step S209. On the other hand, if not all inventory simulation methods have been performed, the process proceeds to step S208.
In step S208, the theoretical inventory calculation unit 12 selects the next inventory simulation method and returns to step S203.
On the other hand, in step S209, the theoretical inventory calculation unit 12 determines whether or not the processes shown in steps S202 to S208 have been performed for all the demand prediction methods. When all the demand prediction methods are performed, the process is terminated. On the other hand, when not carrying out about all the forecast demand methods, it progresses to Step S210.
In step S210, the theoretical inventory calculation unit 12 selects the next forecast demand method and returns to step S202.

[Step S3: Calculate theoretical inventory]
Next, in step S3, the theoretical inventory calculation unit 12 calculates the theoretical inventory for each verification date for each set of the demand prediction method and the inventory simulation method based on the order quantity calculated in step S2. A specific procedure of the theoretical inventory calculation process will be described with reference to FIG. FIG. 11 is a flowchart showing the procedure of theoretical inventory calculation processing.
First, in step S301, the theoretical inventory calculation unit 12 selects a combination of a demand prediction method and an inventory simulation.
Next, in step S302, the theoretical inventory calculation unit 12 sets the verification start date set in the verification period table as the verification date.
Next, in step S303, the theoretical inventory calculation unit 12 reads the lead time days corresponding to the product from the lead time table, and sets the order quantity in (verification date-lead time days) as the delivery quantity on the verification date.
Next, in step S304, the theoretical stock calculation unit 12 adds the theoretical stock on the day before the verification date to the calculated delivery amount, and subtracts the predicted demand amount on the verification date to obtain the theoretical stock on the verification date. The theoretical inventory calculation unit 12 stores the calculated theoretical inventory in the order prediction error table.
Next, in step S305, the theoretical inventory calculation unit 12 determines whether the verification date is a verification end date in the verification period table. If the verification date is the verification end date, the process proceeds to step S307. On the other hand, if the verification date is before the verification end date, the process proceeds to step S306.
In step S306, the theoretical inventory calculation unit 12 sets the next day of the verification date as the verification date and returns to step S303.
On the other hand, in step S307, the theoretical inventory calculation unit 12 determines whether or not the processing shown in steps S302 to S306 has been performed for all sets of demand prediction methods and inventory simulations. If it is performed for all the sets, the process is terminated. On the other hand, if not done for all the sets, the process proceeds to step S308.
In step S <b> 308, the theoretical inventory calculation unit 12 selects the next set of demand prediction method and inventory simulation and returns to 302.

[Step S4: Order Prediction Error Calculation]
Next, in step S4, the order prediction error output unit 13 calculates an order prediction error for each verification date for each set of the demand prediction method and the inventory simulation method. A specific procedure of the order prediction error calculation process will be described with reference to FIG. FIG. 12 is a flowchart illustrating a procedure of order prediction error calculation processing.
First, in step S401, the order prediction error output unit 13 selects a combination of a demand prediction method and an inventory simulation.
Next, in step S402, the order prediction error output unit 13 sets the verification start date set in the verification period table as the verification date.
Next, in step S403, the order prediction error output unit 13 reads the sales quantity and unit price corresponding to the product and the verification date from the sales performance table. Then, the order prediction error output unit 13 calculates a value obtained by subtracting the sales quantity from the theoretical inventory corresponding to the previous day of the verification demand set and the simulation combination of the demand prediction method. However, if delivery is carried out before the start of business on that day, or if the product delivered on that day is available for sale, the quantity delivered on the day of verification is added to the theoretical inventory corresponding to the day before the date of verification, and the Calculate the value after subtracting the quantity. If the calculated value is positive, that value is used as the inventory level. On the other hand, if the calculated value is negative, the absolute value is used as the missing item quantity. Further, the order prediction error output unit 13 sets the opportunity loss as a value obtained by multiplying the shortage quantity by the read unit price. The order prediction error output unit 13 sets a value obtained by multiplying the inventory level by the read unit price as the inventory amount. Then, the order prediction error output unit 13 stores the calculated inventory level, shortage quantity, opportunity loss, and inventory amount in the order prediction error table.

Next, in step S404, the order prediction error output unit 13 determines whether the verification date is a verification end date in the verification period table. If the verification date is the verification end date, the process proceeds to step S406. On the other hand, if the verification date is before the verification end date, the process proceeds to step S405.
In step S405, the order prediction error output unit 13 sets the next day of the verification date as the verification date and returns to step S403.
On the other hand, in step S406, the order prediction error output unit 13 determines whether or not the processing shown in steps S402 to S405 has been performed for all sets of demand prediction methods and inventory simulations. If it is performed for all the sets, the process is terminated. On the other hand, if not done for all the sets, the process proceeds to step S407.
In step S407, the order prediction error output unit 13 selects the next set of demand prediction method and inventory simulation, and returns to step S402.
In addition, the note prediction error output unit 13 sets the ratio of the number of days in which the amount of missing items is 1 or more with respect to the entire verification period as the missing item rate during the verification period.

[Step S5: Method selection]
Next, in step S5, the technique evaluation unit 14 selects a combination of a demand prediction technique and an inventory simulation technique that is optimal for the product based on the order prediction error. Specifically, the technique evaluation unit 14 assigns a smaller evaluation rank in order from the smallest stock level average among the sets in which the stockout rate during the verification period does not exceed a predetermined threshold. And the method evaluation part 14 selects the pair of the demand prediction method and inventory simulation method with the smallest evaluation rank. Further, the technique evaluation unit 14 stores the assigned evaluation rank in the evaluation result table.
In this embodiment, the combination of the demand forecasting method and the inventory simulation method is selected from the shortage rate and the inventory level. For example, among the sets in which the average value of the opportunity loss does not exceed a predetermined reference value, You may select the thing with the smallest average value of money. Alternatively, for each product, criteria for missing item quantity, missing item date, inventory level, inventory amount, and opportunity loss may be set, and a set that does not exceed the criterion may be selected.

  As described above, according to the present embodiment, the theoretical inventory is calculated using the inventory simulation from the predicted demand amount calculated using the demand prediction method, and therefore, an evaluation index for each set of the demand prediction method and the inventory simulation method. It is possible to calculate an order prediction error that is This makes it possible to automatically select an optimal combination for each product from a large number of demand prediction methods and inventory simulation methods.

  In the present embodiment, the demand prediction method based only on the sales performance information is taken as an example, but information other than the sales performance may be used in accordance with each demand prediction method. Specifically, the category, size, weight, price, release date, manufacturer, sales method, presence / absence of sales, presence / absence of advertisement or flyer, weather, temperature, humidity, etc. Can be used as input. In that case, the product sales record storage unit 15 separately records product information such as product category, size, weight, price, release date, manufacturer, and sales method in association with the product JAN code. An information table, sales promotion information, presence / absence of advertisements / flyers, sales day information such as weather, humidity, temperature, etc. of the day may be provided with a sales date information table recorded in association with each sales date. .

Further, the method selection processing may be performed by recording a program for realizing each step shown in FIG. 8 on a computer-readable recording medium. Here, the “computer system” may include an OS and hardware such as peripheral devices.
Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” means a flexible disk, a magneto-optical disk, a ROM, a writable nonvolatile memory such as a flash memory, a portable medium such as a CD-ROM, a hard disk built in a computer system, etc. This is a storage device.

Further, the “computer-readable recording medium” means a volatile memory (for example, DRAM (Dynamic DRAM) in a computer system that becomes a server or a client when a program is transmitted through a network such as the Internet or a communication line such as a telephone line. Random Access Memory)), etc., which hold programs for a certain period of time.
The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.
The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

  As described above, the embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to the above, and various design changes and the like can be made without departing from the scope of the present invention. It is possible to

It is a block diagram which shows the function structure of the method automatic selection apparatus in this embodiment. It is the schematic which shows the data structure of the goods sales performance table which the goods sales performance memory | storage part in this embodiment memorize | stores. It is the schematic which shows the data structure of the verification period table which the verification period memory | storage part in this embodiment memorize | stores. It is the schematic which shows the data structure of the lead time table which the lead time memory | storage part in this embodiment memorize | stores. It is the schematic which shows the data structure of the forecast demand table which the forecast demand amount calculation part in this embodiment produces | generates. It is the schematic which shows the data structure of the order prediction error table which the order prediction error output part in this embodiment produces | generates. It is the schematic which shows the data structure of the evaluation result table which the method evaluation part in this embodiment produces | generates. It is a flowchart which shows the procedure of the method selection process in this embodiment. It is a flowchart which shows the procedure of the predicted demand amount calculation process in this embodiment. It is a flowchart which shows the procedure of the order amount calculation process in this embodiment. It is a flowchart which shows the procedure of the theoretical inventory calculation process in this embodiment. It is a flowchart which shows the procedure of the order prediction error calculation process in this embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Method automatic selection apparatus 11 ... Forecast demand amount calculation part 12 ... Theoretical inventory calculation part 13 ... Order prediction error output part 14 ... Method evaluation part 15 ... Merchandise sales performance memory | storage part 16 ... Verification period memory | storage part 17 ... Lead time memory | storage part

Claims (5)

A product sales record storage unit for storing the sales results during the product verification period;
Using a demand prediction method, a predicted demand amount calculation unit that calculates a predicted demand amount during the verification period based on the sales result of the product read from the product sales result storage unit;
A theoretical inventory calculation unit that calculates a theoretical inventory during the verification period based on the predicted demand calculated by the predicted demand calculation unit using an inventory simulation method;
An order prediction error output unit for calculating an order prediction error for each set of a demand prediction method and an inventory simulation method from the difference between the theoretical stock calculated by the theoretical stock calculation unit and the actual sales of the product;
A method evaluation unit that selects a combination of a demand prediction method and an inventory simulation method corresponding to the product based on the order prediction error calculated by the order prediction error output unit;
A method automatic selection device comprising: A lead time storage unit that stores the lead time from ordering to delivery of products,
The theoretical inventory calculation unit reads a lead time corresponding to the product from a lead time storage unit, calculates an order quantity during the verification period from the lead time and the predicted demand amount using an inventory simulation method, and 2. The method automatic selection device according to claim 1, wherein a delivery amount during the verification period is calculated from an amount, and a theoretical inventory during the verification period is calculated based on the calculated delivery amount. The order prediction error output unit calculates a stockout rate and an inventory level during the verification period,
The method evaluation unit selects a combination of a demand forecasting method and an inventory simulation method in which the stockout rate does not exceed a predetermined upper limit during the verification period, and the average value of the inventory level is minimum. The method automatic selection device according to claim 1 or 2. It is a method automatic selection method performed by a method automatic selection device provided with a product sales result storage unit that stores the sales results during the product verification period,
A step of calculating a predicted demand amount during the verification period based on a sales record of the product, using a demand prediction method, a predicted demand amount calculation unit;
A theoretical inventory calculating unit calculating a theoretical inventory during the verification period based on the predicted demand using an inventory simulation method;
An order prediction error output unit calculating an order prediction error for each set of a demand prediction method and an inventory simulation method from a difference between the theoretical inventory and the sales performance of the product;
A method evaluation unit selecting a set of a demand prediction method and an inventory simulation method corresponding to the product based on the order prediction error;
A method automatic selection method characterized by comprising: On the computer,
Calculating a predicted demand amount during the verification period based on the sales performance during the verification period of the product stored in advance using a demand prediction method;
Calculating a theoretical inventory during the verification period based on the forecast demand using an inventory simulation method;
Calculating an order prediction error for each set of a demand prediction method and an inventory simulation method from the difference between the theoretical inventory and the actual sales of the product;
Selecting a combination of a demand prediction method and an inventory simulation method corresponding to the product based on the order prediction error;
A program for running

Images