JP2006350883A - Demand prediction value automatic determination system using knowledge database, demand prediction value automatic determination program used therefor, and storage medium storing its program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a demand prediction value automatic determination system using a knowledge database, a program and a storage medium used for it for automating and simplifying validity determination of a criterion without depending on an experience or a sense of a worker on the criterion for a demand prediction value in finial decision of a prediction model suitable for a product demand tendency and a component demand tendency. <P>SOLUTION: This demand prediction value automatic determination system predicting a product demand and a component demand is provided with a result value storage device 2 storing a product order receipt result value in the past, an order receipt plan value and the like, a knowledge data management device 3 accumulating the prediction model used for predicting a demand, a simulation parameter, and a demand prediction value criterion in the past, and a process and the like to the decision of the criterion as the knowledge database, and a demand prediction computer 1 performing demand prediction computing based on the result values in the past stored in the result value storage device 2 while referring to the process to the decision of the demand prediction value in the past according to the data accumulated in the knowledge data management device 3. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention is an automatic demand prediction value determination system using a knowledge database for predicting product demand and parts demand, for example, in order to procure parts necessary for manufacturing in anticipation of demand for products to be sold. The present invention also relates to a demand forecast value automatic determination program used therefor and a recording medium on which the program is recorded.

  Conventionally, as a demand forecasting method for forecasting product demand and parts demand in order to procure parts necessary for manufacturing in anticipation of demand for products to be sold, for example, a method as shown in FIG. is there.

FIG. 5 is a block diagram schematically showing a conventional demand prediction method.
In FIG. 5, the prediction value 13 is calculated using a prediction model 11 suitable for the demand trend of products and parts, and using fluctuation parameter information related to the demand fluctuation and simulation parameters 12 such as weighting by coefficients. Calculated. The calculated predicted value 13 may be used as it is for a production plan or a purchase plan. However, in order to satisfy conflicting effects such as inventory reduction and shortage prevention for strict delivery, An operation for confirming the validity of the predicted value based on the judgment 16 of the person in charge is performed using the set judgment standard 14 and past data 15. When it is determined by the confirmation that the predicted value is not valid, the simulation parameter 12 used for prediction is corrected or the prediction model is changed, and the predicted value is changed by re-simulation.

FIG. 6 is a flowchart showing a method of performing a general normal demand forecast.
In FIG. 6, in step ST21, demand prediction is performed using a plurality of demand prediction models, and the result is displayed. In step ST22, the one with the highest prediction accuracy than the latest prediction and the actual value is listed as the first candidate. In step ST23, when the operator determines that the result obtained in step ST22 is not appropriate based on environmental information or the like, the simulation parameter such as the coefficient of variation is changed, or the prediction model itself is changed. The simulation is performed again. The result is displayed as in 21. Steps ST21 to ST23 are repeated, and the final prediction result is determined and determined.

Here, the following prior arts have been proposed in relation to the demand prediction method (see, for example, Patent Document 1 and Patent Document 2).
Japanese Patent No. 2002-73746 (pages 3 to 4, FIGS. 1 and 8) Japanese Patent No. 2003-242432 (Page 4, FIGS. 2 and 5)

  Among them, Patent Document 1 is a part inventory management demand prediction method proposed for the purpose of improving the demand prediction accuracy of low frequency parts, and is based on parameters indicating demand characteristics after extraction of low frequency parts. It is a method that calculates the occurrence probability of demand and uses it to calculate the distribution of probability of occurrence by Monte Carlo simulation, and also predicts demand by regression analysis based on the calculated number of demands. Yes.

  In addition, Patent Document 2 accurately calculates a demand prediction calculation when a product composed of a plurality of parts is sold in a BTO (Build To Order) type production system that is assembled and delivered after an order is received from a customer. This is a demand forecasting device proposed for the purpose of performing decoupling points for each part, using the second demand forecasting calculation, and suddenly receiving large orders from the first demand forecasting result. An external factor or the like is added as a forecast adjustment value, and a third demand forecast result is calculated to improve forecast accuracy.

However, in the conventional technology, the specification of the demand generation interval and the demand number rate, which are parameters indicating the demand characteristics, are determined by the worker, and addition of the predicted adjustment value by an external factor is also performed by the operator. Yes. As for the judgment criteria in the judgment of workers, weighting is performed based on the situation at that time based on various numerical data such as past results, economic trends, seasonal fluctuation factors, management plans, orders received, and inventory numbers, and the demand forecast results Judging the validity.
In addition, the judgment criteria differ depending on the target product such as standard products that are short delivery products or made-to-order products, the impact on the critical path in the manufacturing process, the procurement lead time and price of the parts, etc. In general, it varies widely depending on the attributes and categories. Further, the validity judgment of the demand forecast result is also determined by a complicated thinking process based on various judgment criteria.
Based on the reasoning process of validity using such judgment criteria, if the demand forecast result is judged to be inappropriate, a forecast correction value is added to the input information of the demand forecast method, or simulation parameters and coefficients are weighted. Or changing the prediction model itself used for demand prediction.
Further, this correspondence is generally divided into a plurality of patterns depending on the attributes and categories of the object.

In this way, judgments by workers are based on judgment standards and thought processes that depend on the experience and intuition of the workers, and the judgment standards and thought processes are often not written or systematized. Even if it is culture and systematized, there is no one that is used as a judgment standard in direct cooperation with a system that stores demand forecast models and existing data.
As described above, the conventional demand forecasting method or system is used to determine the demand characteristics at the preparation stage in demand forecasting, the validity judgment of the demand forecasting result, and the re-demand forecasting method or input when judged to be inappropriate. The person in charge who has experience has made corrections to this information, and has not yet been automated until validation.

  The present invention has been made to solve the above-described problems, and in making a final determination of a prediction model suitable for the demand trend of products and parts, it is necessary to consider the experience and intuition of the operator with respect to the criterion of the demand prediction value. Demand prediction value automatic determination system using a knowledge database, demand prediction value automatic determination program used therefor, and recording medium on which the program is recorded, which can automate and simplify validity judgment by eliminating dependence The purpose is to provide.

In order to solve the above problems, the present invention is configured as follows.
According to the first aspect of the present invention, there is provided an automatic demand prediction value determination system for predicting product demand and parts demand in order to procure parts necessary for manufacturing in anticipation of demand for products to be sold. As a knowledge database, the actual value storage device that stores past actual order values or order plan values, the prediction model and simulation parameters used for demand forecasting, the past judgment criteria for demand forecast values, and the process up to the decision Based on the accumulated knowledge data management device and the past actual value stored in the actual value storage device while referring to the process from the data accumulated in the knowledge data management device to the determination of the past judgment of the demand forecast value And a demand prediction calculation device that performs demand prediction calculation using the demand prediction calculation device. It is characterized in that the the like.
According to a second aspect of the present invention, in the demand predicted value automatic determination system according to the first aspect, the past data used for determining the validity of the demand predicted value and the determination for determining the validity of the demand predicted value It is characterized in that the standard can be freely set by the operator.
According to a third aspect of the present invention, there is provided a program for automatically determining a predicted demand value, which is a program for causing a computer to operate the automatic predicted value determination system according to the first or second aspect.
According to a fourth aspect of the present invention, there is provided a recording medium, wherein the demand predicted value automatic determination program according to the third aspect is recorded.

  According to the demand prediction value automatic determination system according to claim 1, the demand prediction value automatic determination program used therefor, and the recording medium on which the program is recorded, a prediction model used when performing prediction can be set, and prediction In order to judge the validity of the predicted value, it is possible to select the past data used to judge the validity of the predicted value calculated as a result. By providing a device capable of inputting “judgment criteria” and storing such information in the knowledge data management device, it is possible to perform capitalization as knowledge information for demand prediction and to perform optimum demand prediction.

  Embodiments of the present invention will be specifically described below with reference to the drawings.

FIG. 1 shows a configuration of an automatic demand forecast value determination system according to the present invention.
Reference numeral 1 denotes a demand prediction calculation device that performs demand prediction, and usually uses a personal computer. Reference numeral 2 denotes a performance value storage device for storing data used for demand prediction, and these are used separately stored in a database system or the like. Reference numeral 3 denotes a knowledge data management device, which stores prediction values made by the operator and the process thereof.

The features of the present invention are as follows.
In other words, in order to procure parts necessary for manufacturing in anticipation of demand for products to be sold, in the automatic demand forecast value judgment system for forecasting product demand and parts demand, the past actual order value of the product Alternatively, the actual value storage device 2 for storing the planned order value and the like, and the knowledge data accumulated as a knowledge database of the prediction model and simulation parameters used for the demand forecast, the past judgment criteria of the demand forecast value and the process up to the determination, etc. Demand prediction based on past performance values stored in the performance value storage device 2 with reference to the management device 3 and the process from the data stored in the knowledge data management device 3 to the determination of the past judgment of the demand prediction value A demand prediction calculation device 1 that performs calculation, and the demand prediction calculation device 1 automatically performs optimum demand prediction. Has become a thing was.
In addition, it is possible for the operator to freely set additional judgment criteria for judging the validity of past data and the demand forecast value used for judging the validity of the demand forecast value.
Below, an example of the process of the program used for the demand forecast value automatic determination system using said knowledge database is described based on a flowchart.

FIG. 2 shows a method for determining a predicted value while referring to past determination information using the automatic demand prediction determination system of the present invention.
First, in step ST1, the demand prediction calculation device 1 calculates a demand prediction value by simulation and displays the result. When the result of the demand forecast value is displayed, if it is determined in step ST2 whether or not the judgment criterion is met, whether or not the same judgment has been made in the past is searched and the same judgment is made in the past In step ST3, past data and determination criteria used at the time of determination are displayed. Finally, in step ST4, a prediction process is added and updated. Based on this information, the worker can make a determination similar to that performed in the past.

Next, FIG. 3 shows a method for performing automatic prediction by a computer and automatic determination of validity using a demand prediction automatic determination system of the present invention.
First, in step ST11, the demand prediction calculation device 1 searches the knowledge data management device for whether or not there has been a thought process for determining validity of the same item or object in the past with respect to the demand prediction value calculated by the simulation. . Next, if there is knowledge data (thinking process) for determining validity, in step ST12, the same determination as that determination process is automatically executed. As a result, it is compared in step ST13 whether or not the demand predicted value matches the reference. If the demand predicted value matches the reference, the predicted value is displayed as the final predicted value in step ST14. If not, the knowledge database is referred to in step ST15 to check whether there is any other prediction method. If there is any other prediction model, the simulation is executed again in step ST16 using another prediction model method. Based on the predicted value in the result, the process returns to step ST12 again, and is repeated until the judgment criteria are met.

FIG. 4 is a schematic diagram showing a screen input method according to the operation of the demand forecast value automatic determination program.
At 31, a demand prediction simulation is executed. At that time, it is possible to select a “prediction model” (311) to be used and to set a “simulation parameter” (312) to be used for prediction, and execute “simulation” (313) according to the setting contents, and the result As a demand forecast value (314).

  With regard to the demand prediction result calculated at 31, the validity is determined at 32. Select “Past data” (321) to be used for determination, select “Judgment criteria” (322) for judging validity, and if the actual predicted value does not match this judgment criteria, repeat, Another “past data” (321) or the like is designated and a determination is made. If the criterion is met, the predicted value is adopted. If the predicted value does not match the criterion, or if it is determined that the predicted value is not valid due to some use, re-prediction is executed at 31 again.

  In this way, by repeating 31 and 32, a reasonable predicted value is finally obtained, but each information input and set in the process is stored in the knowledge data management device as a knowledge database.

Since the embodiment of the present invention has the above-described configuration, it is possible to know the determination criteria determined by the worker, information used for the determination, determination of a re-demand prediction method when it is determined to be inappropriate, correction to input information, and the like. By accumulating in the data management device, it is possible to capitalize the judgment criteria and thought processes that depend on the worker as knowledge information.
In addition, by using the knowledge information that has been capitalized to visualize the validity judgment process, information is shared and used for reference, eliminating dependence on the experience and intuition of workers and automating validity judgment. Simplify and re-demand forecast method determination can be simplified.

  The present invention can be applied as a demand prediction system in the manufacturing process of industrial products.

1 shows a configuration of an automatic demand prediction value determination system according to the present invention. The method of determining a predicted value while referring to past determination information using the automatic demand prediction determination system of the present invention is shown. 1 shows a method for performing automatic prediction by a computer and automatic determination of validity using a demand prediction automatic determination system of the present invention. It is the schematic diagram which showed the screen input method along operation | movement of a demand predicted value automatic determination program. It is a block diagram of the conventional demand prediction method. It is a flowchart in the case of performing general normal demand prediction.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Demand prediction calculating device 2 Actual value storage device 3 Knowledge data management device

Claims (4)

In the demand forecast value automatic judgment system for forecasting product demand and parts demand in order to procure parts necessary for manufacturing in anticipation of demand for products to be sold,
An actual value storage device for storing the past order actual value or order planned value of the product;
A knowledge data management device that stores, as a knowledge database, prediction models and simulation parameters used for demand prediction, past judgment criteria of demand prediction values, processes up to the determination, and the like;
Demand prediction that performs a demand prediction calculation based on the past actual value stored in the actual value storage device while referring to the process from the data accumulated in the knowledge data management device to the determination of the past judgment of the demand predicted value An arithmetic unit;
And a demand forecast value automatic determination system, wherein the demand forecast calculation device automatically performs optimum demand forecast.   The past data used for judging the validity of the demand forecast value and the judgment standard for judging the validity of the demand forecast value can be additionally set by an operator freely. The demand prediction automatic determination system according to Item 1.   An automatic demand prediction value determination program, which is a program for causing a computer to operate the automatic demand prediction value determination system according to claim 1.   A recording medium in which the demand prediction value automatic determination program according to claim 2 is recorded.

Images