JPH05113983A - Merchandise ordering system - Google Patents

Abstract

PURPOSE:To easily decide an exact ordering amount by predicting a sales amount while considering the degree of influences exerted upon the sales amount of merchandise by respective factors. CONSTITUTION:At an inferring means A to output a sales amount B to an ordering amount decision part C, inferred results 1-(n) obtained by fuzzy inference with respective factors 1-(n) as input parameters are adjusted by adjusting values 1-(n) based on the correlation of the respective factors 1-(n) with the respective sales amounts, and the fixed value of the sales amount is calculated from the obtained inferred values 1-(n).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a product ordering system for predicting a sales amount of each product by fuzzy inference from a plurality of factors and determining an ordering amount of the product based on the result of the prediction and the stock amount.

[0002]

2. Description of the Related Art The order quantity of a product in a store is determined by the difference between the sales volume and the inventory quantity on the ordering date, and in the conventional store, an ordering person determines the order quantity of each product based on experience. It was

[0003]

However, there are various factors that determine the order quantity of a product, and the degree of influence of each of these many factors on the sales amount varies depending on each product. Therefore, based on the value of each factor. There is a problem that it is difficult to predict the sales volume of each product individually, and ordering work becomes difficult. Therefore, a skilled ordering person is indispensable for ordering products, which causes a problem that the personnel cost in the store rises sharply.

The object of the present invention is to infer the sales amount of each product by fuzzy inference using a large number of factors as input variables, and adjust the inference result based on the correlation with the sales amount of each factor. Enables to obtain inference results according to the degree of influence of each factor on the amount of sales, facilitates accurate forecasting of the amount of sales, and eliminates the need for a person in charge of ordering and reduces labor costs It is to provide a merchandise ordering system.

[0005]

As shown in FIG. 1, the product ordering system of the present invention comprises an inference means, an inventory management section, and an order quantity determination section, and the sales amount output from the inference means is the inventory management section. It is input to the order quantity determination unit together with the stock quantity output from. The order quantity determination unit outputs the difference between the sales quantity and the stock quantity as the order quantity of the product. The inference means executes fuzzy inference for each of the factors 1 to n that are input and obtains inference results 1 to n. The inference result 1 to the inference result n are adjusted by the adjustment value 1 to the adjustment value n based on the correlation of the sales amount of each of the factors 1 to n, and the inference value 1
~ Get the inference value n. A definite value is obtained from these inferred values 1 to n and is output to the order quantity determination unit as the sales amount.

[0006]

In the present invention, the inference result of fuzzy inference using each of a plurality of factors as input variables is adjusted based on the correlation of each factor with the sales amount. Therefore, the sales amount of the product is accurately predicted in consideration of the degree of influence of each of the plurality of factors on the sales amount.

[0007]

FIG. 2 is a block diagram showing the configuration of a commodity ordering system according to an embodiment of the present invention. The product ordering system 1 includes a sales forecasting unit 2, an order quantity determination unit 3, an inventory management unit 4, an information processing unit 5, and an EOS (electronic ordering / ordering system) 6.
And POS (Point of Sale Information Management System) 7. Order information is input to the inventory management unit 4 from the EOS 6 and sales information is input from the POS 7. The inventory quantity of the product is calculated according to the ordering information and the sales information, and is stored in the inventory management unit 4. Information processing unit 5
Acquires and processes information on various factors that affect the sales of products. A keyboard and a sensor (not shown) are connected to the information processing unit 5. Factors such as presence / absence of advertisement of another store and weather are input from the keyboard, and factors such as temperature are detected by the sensor. Entered in. The information processing unit 5 classifies the inputted factor information by date and time and outputs it to the sales forecasting unit 2.

The sales forecasting unit 2 forecasts the sales amount of each product based on the information about factors affecting sales collected by the information processing unit 5. The forecast result of the sales forecasting unit 2 is output to the order quantity determining unit 3. The order quantity determination unit 3
Inventory information output from the inventory management unit 4 and the sales forecasting unit 2
The order quantity of each product is determined based on the prediction information output from. The order for ordering each product is sent to the EOS 6, and the product is ordered through the ordering terminal and the telephone line according to the ordering data coded for each product.

For example, in the case of daily foods ordered daily, the sales data and the order data up to yesterday are set to POS 7 and EOS, respectively, in order to determine the quantity ordered today.
6 to the inventory management unit 4. The stock management unit 4 outputs the current stock quantity of each product to the order quantity determination unit 3 based on the sales data and the order data. On the other hand, the information processing unit 5 inputs external and internal factors that affect today's sales from the input information to the sales prediction unit 2. Based on this information, the sales forecasting unit 2 forecasts the sales amount for today for each product and outputs it to the order quantity determining unit 3. Subsequently, the information processing unit 5 sends information about tomorrow's sales to the sales forecasting unit 2. The sales forecasting unit 2 forecasts the sales volume tomorrow according to this information. The prediction result is output to the order quantity determination unit 3 in the same manner as today's prediction result, and the order quantity determination unit 3 obtains the stock quantity at the time of closing today based on the current inventory information and the sales forecast information of today, Tomorrow's sales forecast information is added to this value to determine today's order quantity.

FIG. 3 is a diagram showing in detail the structure of the sales forecasting unit of the above-mentioned product ordering system. The sales forecasting unit 2 includes an inference control unit 11 connected to the order quantity determining unit 3 and the information processing unit 5 via an interface 18. A first inference unit 12, a second inference unit 13, a third inference unit 14, and a correlation storage unit 19 are connected to the inference control unit 11. The first inference unit 12 to the third inference unit 14 are rule bases 1.
This rule base 1 has each of 5 to 17
Perform fuzzy inference according to the fuzzy rules stored in 5-17. The inference control unit 11 uses the information input from the information processing unit 5 and the POS 7 through the interface 18 to determine the factors that affect the number of customers, the factors that affect the sales amount of each unit, and the sales amount of each individual product. First reasoning section 1
2, input to the second inference unit 13 and the third inference unit 14.
The first inference unit 12 executes fuzzy inference based on the factors selectively input from the inference control unit 11, and outputs the inference result of the number of customers. Further, the second inference unit 13 is the inference control unit 11
Fuzzy inference is executed according to the contents of the factor data selectively input from, and the sales amount of each unit is inferred. Similarly, the third inference unit 14 infers the sales amount of each individual product based on the content of the factor data selectively input from the inference control unit 11. The inference control unit 11 outputs the number of customers, which is the inference result of the first inference unit 12, to the second inference unit 13 together with other factor data. The inference result of the second inference unit 13 is input to the third inference unit 14.

All rules used for fuzzy inference in each inference unit are expressed by fuzzy production rules. For example, factors input to the first inference unit that infers the number of customers include the day of the week, the presence / absence of advertisements at other stores, the presence / absence of special sale items, the weather, and the temperature.

The second reason for inferring the sales amount of each unit is
Factors input to the inference unit include the day of the week, the weather, the weather change, the temperature, the presence or absence of an advertisement, and the result of inference of the number of customers in the first inference unit 12. The factors that are input to the third inference unit 14 that infers the sales amount of each individual product include the day of the week, the weather, the weather change, the presence or absence of an advertisement, and the temperature, and the number of customers that is the inference result of the first inference unit 12. , The sales amount for each unit, which is the result of the inference in the second inference unit, the price of the product, the discount amount, and the like.

The rules stored in the rule bases 15 to 17 are represented by if x = A then y = B, and the relationship "if x is A, y is B" is stored. For example, when two types of udon a and b having different brands are handled in the store, the rule base 17 provided in the third inference unit 14 includes rules related to udon units including both udon a and b. I remember that if udon unit sells very well then sells very well if udon unit sells well then sells well, and as a rule related to temperature, if if cool, then if sells a little if then sells too much I remember not being.

The correlation storage unit 19 stores the correlation between the factor data input to the third inference unit 14 and the sales amount (actual sales amount) of each product which is the inference result.
This correlation is quantified by, for example, a partial correlation coefficient of a known quantification theory. For example, for the product a shown in FIG. 4, the partial correlation coefficient of the unit sales is 0.702, and The partial correlation coefficient is 0.489. this is,
When the degree of influence of the sales amount of the unit including the commodity a on the sales amount of the commodity a is 0.702, the influence degree of the temperature is 0.489. Which factor has the largest influence on the sales amount of the product, and the value of the influence of each factor naturally varies from product to product, and the correlation Kikube 19 shows the partial correlation coefficient of each factor for each product. Remembered

FIG. 8 is a flow chart showing the processing procedure of the product ordering system when determining the order quantity of a product. When the operator selects a product whose order quantity is to be determined (n1), the factor data (input variable data) input from the information processing unit 5 is read (n2), and the partial correlation coefficient of the factor data for the product is read. From the correlation storage unit 19 (n3). After that, the fuzzy rule for the factor data is read from the rule base, and fuzzy inference is executed in the inference unit (n4). For example, brand a
When inferring the sales amount of udon noodles, if the temperature input from the information processing unit 5 is at the position indicated by the arrow in the membership function of the antecedent part shown in FIG. 5, then if it is cool then then a little fuzzy A rule is selected, and the label “Slightly sell” of the consequent part membership function representing the sales amount of the product is truncated at the degree of belonging 0.8 of the antecedent part. Further, if the unit sales amount input from the inventory management unit 4 is at the position of the arrow of the antecedent membership function shown in FIG. 6, the rule that if the unit of if udon sells well sells well is selected. , The label of "sell well" in the consequent part membership function is truncated at the degree of belonging 0.8 in the antecedent part membership function.

Next, the degree of belonging is adjusted in consideration of the degree of influence of the factors that determine the sales amount of the udon of brand a. For example, in the example shown in FIG. 4, the partial correlation coefficient of unit sales is 0.702 and the partial correlation coefficient of air temperature is 0.489. The effect of temperature is about 70%. Therefore,
This 70% is multiplied by the degree of belonging of the antecedent part shown in FIG. 5, and a predetermined label of the consequent part membership function is truncated by this value (0.8 * 0.7 = 0.56). By adjusting the degree of belonging based on the partial correlation coefficient, the degree of influence of each factor on the sales amount of the product is considered.

The above fuzzy inference is executed for all factor data (n5 → n2), and the membership functions of the product sales amounts obtained in all fuzzy rules are superposed as shown in FIG. The position of the arrow in FIG. 7) is output as the sales amount inference result (n6). Then, the inventory quantity of the selected product is read from the inventory management unit 4 (n7), and the difference between the sales quantity and the inventory quantity is output as the order quantity of the product (n8, n9).

The partial correlation coefficient stored in the correlation storage unit 19 is regularly calculated by the inference control unit 11 of the sales forecasting unit 2. In this calculation, the inference control unit 11 calculates using the quantification type 1 of the known quantification theory based on the actual sales amount of the product input from the POS 7 as shown in FIG. 9 (n11, n12). .. The obtained partial correlation coefficient is updated and stored in the correlation storage unit 19 (n13).

As described above, according to this embodiment, the inferred value of the sales amount can be adjusted by using the partial correlation coefficient of each factor that is regularly updated, and the influence degree of the factor on the sales amount can be considered. The sales volume can be accurately predicted.

[0020]

According to the present invention, the sales amount of a product can be accurately predicted in consideration of the degree of influence of each of a plurality of factors on the sales amount, facilitating the ordering work of the product,
There is an advantage that the person in charge of ordering is unnecessary and the labor cost can be reduced.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of the present invention.

FIG. 2 is a diagram showing a configuration of a product ordering system that is an embodiment of the present invention.

FIG. 3 is a block diagram of a sales forecasting unit forming a part of the product ordering system.

FIG. 4 is a diagram showing a part of a partial correlation coefficient stored in a correlation storage unit that constitutes a sales prediction unit of the product ordering system.

FIG. 5 is a diagram showing membership functions of an antecedent part and a consequent part of a fuzzy rule that uses temperature as an input variable in the fuzzy inference in the sales forecasting unit of the product ordering system.

FIG. 6 is a diagram showing membership functions of an antecedent part and a consequent part of a fuzzy rule in which a unit sales amount in a sales forecasting unit of the product ordering system is an input variable.

FIG. 7 is a diagram showing a calculation method for obtaining a definite value of an inference result in the sales prediction unit of the product ordering system.

FIG. 8 is a flowchart showing a processing procedure in the product ordering system.

FIG. 9 is a flowchart showing a part of a processing procedure of an inference control section in the sales prediction section of the product ordering system.

[Explanation of symbols]

 1-Commodity Ordering System 2-Sales Prediction Unit 11-Inference Control Unit 12-First Inference Unit 13-Second Inference Unit 14-Third Inference Unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazuhiko Ueki 1-2-1, Miyanishi, Matsuyama City Headquarters No. 1 Building Fuji Co., Ltd. (72) Hideki Suzumura 1-2-1, Miyanishi, Matsuyama City Headquarters No. 1 Building Fuji, Inc. (72) Inventor, Hajime Takebayashi, No. 10 Hanazono Todocho, Ukyo-ku, Kyoto Omron Co., Ltd. (72) Inventor, Chinami Yoshida, No. 10 Hanazono Todo-cho, Ukyo-ku, Kyoto Omron Corporation

Claims (1)

[Claims] 1. In a product ordering system for predicting a sales amount of each product and determining an order amount based on the result of the prediction and an inventory amount, each of a plurality of factors affecting the sales amount of the product is used as an input variable. The product ordering is characterized in that the inference result of the fuzzy rule inference is adjusted based on the correlation of each factor to the sales amount to obtain the inference value of the sales amount, and the inference means for obtaining the definite value from the inference value is provided. system.