JP3308654B2 - Shipment forecast system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for calculating a predicted shipment volume per unit period with a period such as one month as a unit period, and more particularly, a growth rate of an actual shipment volume per unit period. The present invention relates to a shipment amount prediction system for predicting a shipment amount by using a prediction method adapted to a product number from a plurality of prediction methods such as a prediction method based on a product number and a prediction method based on an annual change of an actual shipment amount.

[0002]

2. Description of the Related Art In order to prevent shortage or excess of inventory, it is necessary to predict the amount of shipment, and a system for making this prediction has been proposed in Japanese Patent Laid-Open No. 4-171577.

[0003] This system is based on a store-by-store order probability obtained by dividing a store-by-store product shipment volume for a certain period in the past by a store-by-store product order volume during the same period, and store-by-product order information. The product is estimated as the shipment volume. In order to evaluate the reliability of the shipment amount, an operation of dividing the predicted shipment amount by the average value of the actual shipment results by store and product for a certain period in the past is performed, and the operation result is set within a predetermined range. Is used to determine whether the value is equal to or not. When the calculation result exceeds the set range, the reliability of the prediction is determined to be low, and the shipment amount is manually corrected.

[0004]

However, when the above configuration is used, the product is in the early stage of sales or in the growth period,
In the expansion period in which the shipment volume starts to increase in proportion to the degree of penetration of the product into the market, the shipment volume changes greatly in different fiscal years, so that the reliability of the prediction is judged to be low. The same applies to products whose products have entered a retreat period and whose production volume has been reduced. Since the shipment volume greatly decreases in different fiscal years, the reliability of the prediction is determined to be low. Therefore, when trying to apply the conventional technique to the prediction of the shipment amount of various types of products, it is often necessary to make corrections manually, resulting in a problem that it is extremely time-consuming.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a prediction method suitable for a mature product having a stable shipment volume and a product in an expansion period or a contraction period. By automatically selecting and using two types of prediction methods, a suitable prediction method, it is possible to provide a shipment amount prediction system capable of accurately predicting the shipment amount of a wide variety of products without using manual correction. Is to do.

[0006]

In order to solve the above-mentioned problems, a shipping amount prediction system according to the present invention uses a unit period for predicting the shipping amount as a unit period, and sets a unit period to be predicted as a prediction target period. , Is applied to the shipment volume forecasting system that calculates the forecast shipment volume during this forecast period,
Prior to the forecast target period, the actual shipment amount in a period set in advance as a period longer than the unit period, the actual shipment amount of the previous year in the previous year period corresponding to this period, and the growth rate of the actual shipment amount Based on the data such as the average value , at least the estimated shipment volume based on the yearly change
A first forecasting unit for calculating a certain actual forecast amount, and a unit period
Calculate the expected growth rate, which is the predicted shipment volume based on the change for each interval
A second prediction unit for output, and an error calculator for calculating each predicted error between predicted actual predicted amount and elongation predictor and the actual shipments by the first and second prediction unit these,
A configuration including a prediction method determining unit that determines a predicted amount corresponding to the smallest prediction error among the plurality of prediction errors calculated by the error calculation unit as a predicted shipping amount during a period in which the actual shipping amount has already been obtained. And

[0007]

According to the present invention, the actual forecast quantity of a plurality of forecast quantities is a value obtained by dividing the actual shipment quantity in a preset period by the previous year actual shipment quantity in the previous year period corresponding to this period, and the forecast target period. Is the product of the previous year's actual shipment volume in the previous year's period corresponding to. That is, the first predicting unit that obtains the actual predicted amount predicts the shipping amount according to the prediction method when the product is assumed to be in the stable period.

[0008] The predicted growth rate of the plurality of predicted quantities is:
It is the product of the average growth rate, which is the average value of the growth rate of the actual shipment amount based on the actual shipment amount in each of a plurality of unit periods before the prediction target period, and the actual shipment amount in the latest unit period. That is, the second predicting unit that calculates the predicted growth rate predicts the shipment amount according to the prediction method when the product is assumed to be in the growth period or the contraction period.

Therefore, among the prediction errors calculated by the error calculation section, the actual prediction error indicates a prediction error when the product is assumed to be a product in a stable period, and the elongation rate prediction error indicates a product in a growth period or a contraction period. This shows the prediction error assuming that there is.

That is, when the growth rate prediction error is smaller than the actual prediction error, the product is a product in the expansion period or the contraction period. Therefore, the prediction method determination unit sets the predicted growth rate as the predicted shipment volume. When the actual prediction error is smaller than the elongation rate prediction error, the product is a product in the stable period, and the actual predicted amount is set as the predicted shipment amount.

That is, the predicted amount corresponding to the smallest prediction error among the plurality of prediction errors is set as the predicted shipment amount.

[0012]

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an electrical configuration of an embodiment of a shipping amount prediction system according to the present invention.

In this embodiment, the unit period for estimating the shipment amount is one month. Therefore database 4
Is the output 7 for each product number for each month.
, The actual forecast quantity that is the shipping quantity predicted by the first predicting unit 1, the predicted growth rate that is the shipping quantity predicted by the second predicting unit 2, and the third predicting unit 3 It is a block for storing each value of the business unit predicted quantity, which is the predicted shipment quantity. Therefore, the first to third prediction units 1 to 3 are bidirectionally connected to the database 4.

That is, in this embodiment, three prediction units 1 to 3 having different prediction methods are provided as prediction units.

The first forecasting unit 1 uses the one month to be forecast as a forecast target period and calculates the actual shipment volume for three months before this period as the actual shipment volume for the previous year during the same period of the previous year. Divide by This block calculates the product of the actual shipment amount of the previous year corresponding to the prediction target period and the division result, and outputs the calculation result as the actual prediction amount.

The second prediction unit 2 calculates an average growth rate, which is an average value of the growth rates of the actual shipment amount, based on the actual shipment amount in each month for three months before the prediction target period. A block that calculates the product of the average growth rate and the actual shipment amount of the last month, which is the latest unit period, and outputs the calculation result as a predicted growth rate.

The third forecasting unit 3 divides the actual shipment volume for three months before the forecast target period by the actual shipment volume of the previous year during the same period of the previous year, and calculates This block outputs a business unit forecast amount obtained by performing a correction based on experience manually using the product of the actual shipment volume and the division result as basic data.

The outputs of the first to third prediction units 1 to 3 are provided to a prediction method determination unit 5 and an error calculation unit 6.

The error calculating unit 6 calculates the actual shipping amount for each month and the past three months in which the predicted amounts of the first to third predicting units 1 to 3 have already been obtained. Actual forecast error indicating the error between the forecasted amount and actual shipment amount, growth rate forecast error indicating the error between the expected growth rate and the actual shipment amount, business unit forecast error indicating the error between the business unit forecast amount and the actual shipment amount This is a block for calculating each error, and calculates deviations of these errors in order to increase the accuracy of error comparison in the prediction method determination unit 5, and sends the calculated deviations to the prediction method determination unit 5.

The prediction method determining unit 5 is a block for comparing each of the actual prediction error, the elongation rate prediction error, and the business unit prediction error using the error deviation. When the actual prediction error is minimized, the actual predicted amount is transmitted as the predicted shipment amount from the output 8, and when the elongation percentage prediction error is minimized, the predicted elongation ratio is transmitted as the predicted shipment amount.
When the business unit prediction error is the smallest, the business unit prediction amount is transmitted as the predicted shipping amount.

FIG. 2 is a flowchart showing the operation of one embodiment of the present invention. FIG. 3 is an explanatory diagram showing the relationship between monthly data used for prediction and the period to be predicted. And FIG. 5 is an explanatory diagram showing various errors and their deviations each month. The operation of one embodiment of the present invention will be described below with reference to FIG.

In this embodiment, as shown in FIG. 3, for example, when the prediction target period is one month, prediction is performed based on each data from September to November, and the prediction target period is two months. In some cases, prediction is made based on data from October to December.

Now, assuming that the database 4 stores the actual shipment volume for each month up to February, and that the forecast target period is April, the first forecasting unit 1
By dividing the total of the actual shipments of January and February by the total of the actual shipments of the previous year during the same period of the previous year, an index of the annual change is obtained. Then, the product of this index and the actual shipment volume in April of the previous year is calculated, and the actual predicted volume as the calculation result is stored in the database 4 as the predicted shipment volume in April. It is sent to the arithmetic unit 6 (step S11).

Further, the second predicting unit 2 performs the operations in December, January,
The average growth rate, which is the average value of the growth rate, is calculated from the actual shipment quantity for each month of the month, and the product of the actual shipment quantity in February, the previous month, and the average growth rate is calculated. Then, the calculated growth rate predicted amount is stored in the database 4 as the predicted shipment amount in April, and is also transmitted to the prediction method determination unit 5 and the error calculation unit 6 (step S12).

The third predicting unit 3 performs the same prediction as that of the prior art, and calculates the total of the actual shipments in December, January and February, and the total of the actual shipments in the previous year during the same period in the previous year. The product of the index of the yearly change obtained by dividing by the above and the actual shipment amount in April of the previous year is calculated. The calculation result is used as basic data, and if necessary, a manual correction is given. The business unit forecast amount, which is the correction result, is stored in the database 4 as the forecast shipment amount in April, and the forecast method determination unit 5 And the error calculation unit 6 (step S
13).

As a result of the above operation, the database 4
In each month, data up to February is stored for the actual shipment amount and data up to April is stored for the predicted amount. FIG. 4 shows data relating to the month up to February among these data.

The error calculator 6 calculates the actual prediction error, the growth rate prediction error, and the business unit prediction error for December, January, and February when the actual shipment amount and each predicted amount are obtained. Then, based on the calculated error, from the data for March of each prediction error, the following formula

[0029]

(Equation 1)

By using, the deviation of each prediction error is calculated. FIG. 5 shows prediction amounts, prediction errors, and deviations of the prediction errors in December, January, and February, and these deviations are given to the prediction method determination unit 5 (step S14).

The prediction method determination unit 5 determines a prediction method that minimizes the deviation from the three types of deviations provided from the error calculation unit 6. The data shown in FIG. 5 indicates that the deviation of the performance prediction error is the minimum, and indicates that the product number for which the shipment amount is to be predicted is a product whose shipment amount is stable. . Therefore, the prediction method determination unit 5 determines that the actual predicted amount is the predicted shipping amount with the highest accuracy and outputs
Is used to send the actual predicted amount to the outside.

Further, the prediction method determining unit 5 determines that the product number to be predicted is a product in the expansion period of the shipment amount, such as a product in the early stage of development, or a product in the reduction period of the shipment amount. When the deviation of the prediction error is minimized, the predicted amount of elongation is determined to be the predicted shipping amount with the highest accuracy, and the predicted amount of elongation is sent to the outside. When the deviation of the business unit prediction error is minimized, the business unit prediction amount is sent to the outside as the predicted shipment amount.

As described above, in the present embodiment, the third
By providing the prediction unit 3, the prediction corrected manually is also performed, so that it is possible to improve the prediction accuracy even for a product in which the change in the shipment amount is special.

Note that the present invention is not limited to the above-described embodiment, and the error calculation section 6 is not limited to the past shipment quantity for which the monthly actual shipment quantity and the first to third prediction sections 1 to 3 have already been obtained. For the three months, the description has been given of the case where each error of the actual forecast error, the growth rate forecast error, and the business unit forecast error is calculated for each month, and the deviation is calculated from the calculated result.
For example, it is possible to adopt a configuration in which an error is calculated only for the previous month (in this case, the configuration of the prediction method determination unit is configured to compare the error itself).

Further, the configuration of the error calculation section can be configured to calculate an error for a certain period in the past.

[0036]

The shipping quantity forecasting system according to the present invention allows, for example, a first forecasting section of a plurality of forecasting sections to calculate an actual forecasting quantity which is a forecasted shipping quantity based on a yearly change. Then, a predicted growth rate, which is a predicted shipment volume based on a change per unit period such as one month, is calculated. Further, during the period in which the actual shipment volume has already been obtained, the error between the predicted actual volume and the actual shipment volume and the error between the predicted growth rate and the actual shipment volume are calculated by the error calculation unit. Based on these errors, by using the prediction method determining unit, when the error of the actual predicted amount is small, the actual predicted amount is used as the predicted shipment amount, and when the error of the predicted elongation ratio is small, the predicted elongation amount is predicted. Shipment volume. As a result, two types of forecasting methods are automatically selected: a forecasting method suitable for a product in a mature period in which shipment volume is stable, and a forecasting method suitable for a product in an expansion period or a contraction period. It is possible to improve the accuracy in predicting the shipment amount of various types of products without using correction.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an electrical configuration of an embodiment of a shipping amount prediction system according to the present invention.

FIG. 2 is a flowchart showing the operation of one embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a relationship between data for each unit period used for prediction and a prediction target period.

FIG. 4 is an explanatory diagram showing a relationship between each predicted quantity and an actual shipment quantity.

FIG. 5 is an explanatory diagram showing various errors and their deviations for each unit period.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 1st prediction part 2 2nd prediction part 4 database 5 prediction method determination part 6 error calculation part

Claims (1)

(57) [Claims] 1. A shipping volume prediction system that calculates a predicted shipping volume in a prediction target period by setting a period serving as a unit for predicting a shipping volume as a unit period and a unit period serving as a prediction target as a prediction target period. The actual shipment volume in a period before the prediction target period and set in advance as a period longer than the unit period,
Based on data such as the actual shipment volume of the previous fiscal year and the average growth rate of the actual shipment quantity in the previous fiscal year corresponding to this period , at least the estimated shipment quantity based on the annual change
A first prediction unit for calculating an actual predicted amount, which is a unit, and a unit
Estimated growth rate, which is the estimated shipment volume based on the
A second prediction unit for calculating the actual績予 predicted by the first and second prediction unit these
An error calculation unit for calculating a prediction error between the surveying and elongation rate prediction amount and the actual shipment amount, and a smallest one of a plurality of prediction errors calculated by the error calculation unit in a period in which the actual shipment amount has already been obtained. A shipping amount prediction system comprising: a prediction method determining unit that determines a predicted amount corresponding to a prediction error as a predicted shipping amount.