JP2009230555A - Demand forecast method, inventory plan decision method, demand forecast system and inventory plan decision system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain a demand forecast method can perform demand forecasting with higher accuracy. <P>SOLUTION: This demand forecast system includes: a demand forecast pattern DB 6 for storing a demand forecast pattern in a component unit; a result DB 2 for storing the past use results in a component unit; a component information-result information acquisition part 3 for acquiring a usage result of the preceding year of a component that is subjected to demand forecast from the result DB 2; and a demand forecast part 7 for calculating a pattern distance evaluation in each demand forecast pattern obtained from the demand forecast pattern DB 6, selecting a demand forecast pattern with the least separation, and calculating the number of monthly demand forecast by proportionally distributing the total annual demand forecast of this year calculated on the basis of the use results to each month according to the selected demand forecast pattern. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a demand prediction method and an inventory plan formulation method for a product for which there is little actual data regarding orders.

  In conventional inventory management, new products that have just been released do not have sufficient results, so it is difficult to predict the demand for each product, and arrangements have been made depending on the intuition and experience of the person in charge.

  Patent Document 1 below discloses an example of a demand prediction method. In this method, an order is predicted based on forecast information from a customer. Specifically, past forecast information and past performance are compared, and the deviation of lead time and quantity is calculated as a forecast correction parameter, and future forecast information is corrected using this deviation. To do.

JP 2005-25245 A

  However, in the technique described in Patent Document 1, a predicted value is obtained by correcting future forecast information using forecast information from the customer and past results. For this reason, there is a problem that the information without forecast information cannot be predicted. In addition, there are problems such as that seasonal fluctuations are not taken into consideration in the prediction, and there is no means for avoiding the risk when the prediction is lost.

  The present invention has been made in view of the above, and it is an object of the present invention to obtain a demand prediction method and an inventory plan formulation method capable of highly accurate demand prediction even for newly-released products with little track record regarding orders. And

  In order to solve the above-described problems and achieve the object, a demand prediction method according to the present invention includes a demand prediction pattern storage unit that stores one or more demand prediction patterns for each part, and all of the parts that perform demand prediction. A demand prediction pattern acquisition step of acquiring a demand prediction pattern; a usage record acquisition step of acquiring a previous year's use record of the part for which the demand prediction is performed; For each acquired demand prediction pattern, calculate the pattern distance evaluation indicating the degree of deviation from the previous year's use results of the parts for which the demand prediction is performed, compare the obtained pattern distance evaluation, and predict the demand with the smallest deviation Demand prediction pattern selection step to select a pattern, and a regression formula that shows the fluctuation of demand based on the previous year's usage record of the part that performs the demand prediction Calculating the annual demand forecast total for the current year based on the regression equation, and dividing the annual demand forecast total into each month according to the selected demand forecast pattern And a monthly demand forecast number calculating step for calculating the number.

  According to the present invention, there is an effect that demand prediction with high accuracy is possible.

  Embodiments of a demand prediction method and an inventory plan formulation method according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments.

  In the present embodiment, as a target for demand prediction, seasonal products such as air conditioners and refrigerators that are frequently shipped in summer are assumed. In addition, the “parts” is a product in a unit shipped by a business operator to a repair company or the like for repair or the like, or a unit used by itself.

  FIG. 1 is a diagram illustrating a configuration example of an inventory plan formulation system. The inventory plan formulation system in FIG. 1 includes a parts master DB 1, a performance DB 2, a parts information / result information acquisition unit 3, a prediction target component determination unit 4, an order point setting unit 5, a demand prediction pattern DB 6, a demand A prediction unit 7, an inventory standard calculation unit 8, an inventory plan formulation unit 9, a demand prediction DB 10, and an inventory plan DB 11 are provided.

  The component master DB 1 is a database for managing master information related to components. FIG. 2 is a diagram illustrating an example of a table structure of the component master DB. As shown in FIG. 2, “part number” and “release year” are set as data items here. The parts master DB 1 may be prepared as a master for an inventory plan formulation system, or may be a master operating in a normal core business system or the like.

  In the present embodiment, “year” refers to the “refrigeration year” defined by the Japan Refrigeration and Air Conditioning Industry Association. For example, “2007 freezing year” refers to the period from October 1, 2006 to September 30, 2007.

  The result DB 2 is a database for holding past use results for parts registered in the part master DB 1. The actual use is the actual use of the part. FIG. 3 is a diagram illustrating an example of a table structure of the performance DB 2. As shown in FIG. 3, “part number”, “year”, and usage record for each month from October to September are set as data items. The number of parts is set in the monthly usage record. Here, “pieces” are used as the unit of parts, but other appropriate units may be used depending on the properties of the parts.

  The part information / result information acquisition unit 3 acquires master information from the part master DB 1 and extracts newly released parts based on the “release year”. The parts extracted in this way are targets for demand prediction in the present embodiment. Further, the part information / result information acquisition unit 3 acquires data from the result DB 2 for the previous year of the target year in which the demand prediction of the demand prediction target part is performed.

  Based on the data acquired by the part information / result information acquisition unit 3, the prediction target part determination unit 4 selects the demand prediction target part as “prediction target product” or “non-prediction target product” depending on whether or not there is an annual usage record. Is determined. The order point setting unit 5 sets order points for non-predicted products.

  The demand prediction pattern DB 6 is a database for holding a demand prediction pattern for each product model. The demand prediction pattern is data obtained by patterning annual demand fluctuations in consideration of seasonal fluctuations, and several types are derived and registered for each product model. FIG. 4 is a diagram illustrating an example of a table structure of the demand prediction pattern DB 6. As shown in FIG. 4, “product model”, “pattern ID”, and a pattern for each month are set as data items. The product model in which each part is used can be identified from the part number.

  The demand forecasting unit 7 sets a demand forecast pattern, calculates an annual demand forecast total, and calculates a monthly demand forecast number based on these.

  The inventory standard calculation unit 8 calculates an inventory standard based on the monthly demand forecast number obtained by the demand prediction unit 7. Here, the inventory standard is the target inventory quantity at the end of the month when the safety inventory is expected.

  The inventory plan formulation unit 9 formulates an inventory plan for a prediction target product or a non-prediction target product and stores it in the inventory plan DB 11. Here, the inventory plan is to set the target inventory quantity at the end of the month for each month. In addition, the inventory plan formulation unit 9 stores the demand forecast pattern, the annual demand forecast total, and the monthly demand forecast number determined by the demand forecast unit 7 in the demand forecast DB 10.

  The demand prediction DB 10 is a database for holding a demand prediction for each type of part (part number) obtained by the demand prediction unit 7. FIG. 5 is a diagram illustrating an example of a table structure of the demand prediction DB. As shown in FIG. 5, “part number”, “year”, “demand forecast pattern”, “annual demand forecast total”, and “monthly demand forecast number” are set as data items.

  The inventory plan DB 11 is a database for holding monthly inventory for each part type (part number) obtained by the inventory plan formulation unit 9. FIG. 6 is a diagram illustrating an example of a table structure of the inventory plan DB 11. As shown in FIG. 6, “part number”, “year”, and inventory for each month are set as data items.

  Next, the operation (inventory plan formulation process) in the inventory plan formulation system configured as described above will be described using specific data.

  The part information / result information acquisition unit 3 in the inventory planning system of FIG. 1 acquires master information from the part master DB 1. Next, referring to the “release year” in the acquired master information, a part that has not passed three years since the release is extracted. “Three years” is intended to be excluded from the target of demand forecast because the demand is stable after three years from the release year. Other values may be used as long as it is appropriate to divide by other years depending on the characteristics of parts demand. In addition, the part information / result information acquisition unit 3 acquires data for the previous year of the target year (current year) in which the demand prediction is performed from the result DB 2 using the part number of the demand prediction target part as a key.

  When these pieces of information are acquired, the inventory planning system performs the process shown in FIG. 7 for each part extracted as a demand prediction target. FIG. 7 is a flowchart showing inventory plan formulation processing.

  The prediction target component determination unit 4 refers to the data on the demand prediction target component notified from the component information / result information acquisition unit 3, and if there is at least one year of actual use in the previous year, the “prediction target product” If there is none, it is determined that the product is “non-predictable product” (step S1). It should be noted that other criteria may be adopted in order to discriminate between the prediction target product and the non-prediction target product.

  When it is determined that the demand prediction target part is the “prediction target product” (step S1: Yes), the demand prediction unit 7 performs a demand prediction pattern setting process (step S2). FIG. 8 is a flowchart showing the operation of the demand prediction pattern setting process.

  The demand prediction unit 7 determines the product model from the demand prediction target part, and acquires all the demand prediction patterns from the demand prediction pattern DB 6 using this product model as a key (step S21). FIG. 9A is a diagram illustrating an example of a demand prediction pattern, and more specifically, shows three demand prediction patterns set for a certain product model.

  The demand forecasting unit 7 calculates the pattern distance evaluation by the least square method for the first demand forecasting pattern (for example, pattern 1) using the use record of the previous year that has been acquired from the record DB2 (step S22). . Here, the “pattern distance evaluation” is a value indicating the degree of divergence of the demand prediction pattern from the actual results, and is defined as, for example, the following equation (1).

Pattern distance evaluation =
{((Octual number used / average number of payouts per month)-October pattern) ²
((November usage results / average number of payouts per month)-11 month pattern) ² +
...
((September use results / average number of payouts per month) -September pattern) ² } (1)
The “average number of payouts per month” is a value obtained by dividing the annual total usage results in the previous year's results by 12 months.

  FIG. 9-2 is a diagram illustrating an example of the usage record of the previous year. Hereinafter, based on the usage record of FIG. 9-2, an optimal demand prediction pattern is examined for each demand prediction pattern of FIG. Since the value obtained by dividing “total” in FIG. 9-2 by 12: (786/12 = 65.5) is “average number of payouts per month”, the data of “pattern 1” in FIG. If it calculates using Formula (1), it will become the following (2) Formula.

Pattern distance evaluation =
{((40 / 65.5) −0.82) ² + ((41 / 65.5) −0.63) ² +
...
((74 / 65.5) -1.17) ² } = 0.720 (2)

  Next, the demand prediction unit 7 determines whether or not the calculated pattern distance evaluation is the minimum value (minimum distance) among all the demand prediction patterns (step S23). Here, the examination is started from the pattern 1, and other pattern distance evaluation has not been calculated yet. For this reason, the demand prediction pattern of the pattern 1 is temporarily preserve | saved (step S24). Next, it judges with demand forecasting part 7 not examining all demand forecast patterns (Step S25: No), and processes about the next demand forecast pattern (pattern 2).

  In the pattern 2 process, the pattern distance evaluation “1.817” is obtained by the same procedure as that for calculating the pattern distance evaluation for the demand prediction pattern of the pattern 1 (step S22). Here, since “Pattern 1 (0.720) <Pattern 2 (1.817)”, the demand prediction unit 7 compares “Pattern 2” with the value of “Pattern 1” temporarily stored. "Is large (step S23: No), the demand prediction pattern of pattern 2 is not temporarily stored. Then, it is determined that all demand prediction patterns have not been examined yet (step S25: No), and the next demand prediction pattern (pattern number: 3) is processed in the same procedure as described above.

  In the pattern 3 process, the pattern distance evaluation “4.646” is obtained by the same procedure as that for calculating the pattern distance evaluation for the demand prediction patterns of the patterns 1 and 2 (step S22). Here, since “pattern 1 (0.720) <pattern 3 (4.646)”, the demand prediction unit 7 compares “pattern 3” with the temporarily stored value of “pattern 1”. "Is large (step S23: No), and the demand prediction pattern of pattern 3 is not temporarily stored.

  Thus, the demand prediction part 7 which complete | finished examination of all the demand prediction patterns determines the demand prediction pattern set in step S2 to the finally saved "pattern 1" (step S25: No). (Step S26). Thus, in this embodiment, all the demand prediction patterns are examined, and the demand prediction pattern that has the smallest deviation from the usage record, that is, the closest use record is set (step S2).

  After executing the process of step S2, the demand prediction unit 7 next performs an annual demand prediction total calculation process (step S3). FIG. 10 is a flowchart showing the operation of the annual demand forecast total calculation process.

  First, the demand prediction part 7 calculates | requires a linear regression type about the performance of the last year (step S31). Specifically, each month of the previous year is set as an x value, and October, November,..., September is set as x = 1, 2,. In addition, the usage record of each month of the previous year is defined as a y value. Then, a linear regression line “Y = aX + b” is calculated.

  In calculating the linear regression line, if the following equation (3) is defined, a and b are the following equations (4). Note that n is the number of data samples.

_{sigma_x = x i (i = 1,2} , ..., n)
sigma_y = y _i (i = 1, 2,..., n)
sigma_xx = x _i x _i (i = 1, 2,..., n)
sigma_xy = x _i y _i (i = 1, 2,..., n) (3)

a = (n × sigma_xy−sigma_x × sigma_y)
/ (N × sigma_xx−sigma_x × sigma_x)
b =
(Sigma_xx x sigma_y-sigma_x x sigma_xy)
/ (N × sigma_xx−sigma_x × sigma_x) (4)

  Hereinafter, when the primary regression line is calculated using the actual use values of FIG. 9-2, the following equation (5) is obtained, and a and b are the following equations (6).

sigma_x = 1 + 2 + ... + 12 = 78
sigma_y = 40 + 41 + ... + 74 = 786
sigma_xx = 1 × 1 + 2 × 2 +... + 12 × 12 = 650
sigma_xy = 1 × 40 + 2 × 41 +... + 12 × 74 = 6395 (5)

a = (12 × 6395−78 × 786) / (12 × 650−78 × 78) = 8.993
b =
(650 * 786-78 * 6395) / (12 * 650-78 * 78) = 7.045
(6)

As a result, the linear regression line becomes the following equation (7).
Y = 8.993X + 7.045 (7)

  Next, since the slope of the obtained regression line is positive (step S32: Yes), the demand prediction unit 7 adds up the current year region of the regression line (step S33). That is, assuming that the annual demand in the current year fluctuates along the obtained linear regression line, the values of Y obtained using the values of X = 13 to 24 are totaled. When applied to the above equation (7), the following equation (8) is obtained.

  (Annual demand forecast total) = (8.993 × 13 + 7.045) + (8.993 × 14 + 7.045) +... + (8.993 × 24 + 7.045) = 2081 (8)

  On the other hand, when the slope of the obtained regression line is negative (step S32: No), the demand prediction unit 7 sums up the usage results for each month in the previous year (step S34).

  And the demand prediction part 7 determines the total calculated in this way as an annual demand prediction total (step S35). Here, if the slope of the regression line is negative, the total annual demand forecast for the current year will be less than or equal to the previous year when the current year areas of the regression line are summed. However, analysis of past performance has revealed that there is a high probability that demand for newly set parts will increase. For this reason, in this embodiment, the total of the actual results of the previous year is set as the total annual demand forecast of the current year, and the correction is made so as to anticipate at least the demand of the previous year. As described above, when the annual demand forecast total is calculated, the annual demand forecast total calculation processing in step S3 ends.

  Next, the demand forecasting unit 7 calculates the monthly demand forecast number by distributing the annual demand forecast sum calculated in step S3 to each month using the demand forecast pattern determined in step S2 (step S4). In addition, when the annual total of the distributed monthly demand forecast number does not reach the annual demand forecast total, the shortage is allocated to the demand forecast number in October. For example, using the above-described demand forecast pattern 1 in FIG. 9-1 and the annual demand forecast total: 2081 obtained in step S34, the following (9) is obtained, and the monthly demand forecast number shown in FIG. 11 is obtained. . FIG. 11 is a diagram illustrating an example of the demand forecast number for each month.

Expected number of October = 2081/12 × 0.82
Expected number of November = 2081/12 × 0.63
...
Expected number of September = 2081/12 × 1.17 (9)

  If the total annual demand forecast calculated in step S3 is 12 or less, the total number is set as the demand forecast number for October.

Next, the stock standard calculation unit 8 calculates the stock standard based on the monthly demand forecast number obtained by the demand prediction unit 7 (step S5). The stock standard when an arbitrary month is m is the following equation (10). In addition, m is a natural number and repeats 1-12.
m month stock standard = (m + 1) monthly demand forecast number + (m + 2) monthly demand forecast number (10)

  When the formula (10) is applied to the monthly demand forecast number shown in FIG. 11, the inventory standard for each month becomes the following formula (11).

October inventory standard = 109 + 111 = 220
November stock standard = 111 + 92 = 203
...
August stock standard = 203 + 142 = 345
September stock standard = 142 + 109 = 251 (11)

  In the present embodiment, since the demand forecast is performed only for one year, the forecast value of the current year is used for calculating the inventory standard for the month spanning the next year. For example, the August inventory standard requires monthly demand forecasts for September of the current year and October of the following year (next year of the current year), but no data exists for October of the following year. Therefore, the data for October of the current year is used instead of the data for October of the following year. Similarly, the September inventory standard requires monthly demand forecasts for October and November in the following year, but since neither data exists, instead of the data for October and November in the following year The data for October and November of the current year are used.

  As described above, an inventory standard as shown in FIG. 12 is obtained. FIG. 12 is a diagram illustrating an example of inventory standards for each month of the current year.

  On the other hand, when it is determined that the demand prediction target part is “non-prediction target product” (step S1: No), the order point setting unit 5 sets an order point for the non-prediction target product (step S6). First, the order point setting unit 5 determines whether the part is an “electric part” or a “non-electric part” based on the part number. This is because the scale of use is different between electric parts and non-electric parts. The background is that electrical components are fragile and the procurement time is long, while non-electrical components are difficult to break and the procurement time is short. In the present embodiment, components are classified only into electrical components and non-electrical components. However, the components may be further hierarchized, and in that case, more detailed demand prediction is possible. Further, another classification method may be adopted.

  Based on the determination result, as an example, “20” is set as an ordering point for a part designated as an electrical part, and “12” is set as an ordering point for a part designated as a non-electrical part. These constants are values that clear the target of the stockout rate (for example, the stockout rate of 2% or less, etc.) after statistical analysis of past performance and carry out inventory transition simulation. Has been. Hereinafter, these constants are referred to as “minimum holding number”.

  Next, the inventory plan formulation unit 9 sets each month's inventory according to the inventory standard calculated by the inventory standard calculation unit 8 in step S5 and stores it in the inventory plan DB 11 (step S7). Here, the “inventory plan” refers to a process of determining the inventory for each month based on the inventory standard. In setting the inventory for each month, follow the policies such as “Secure parts ahead of schedule” and “Keep the minimum necessary number”. For example, (1) if the inventory standard in October is “0”, the larger of the annual demand forecast total or the minimum number of holdings is taken as the inventory for that month. (2) If the inventory standard for months other than October is “0”, the inventory standard for the previous month is used as the inventory for that month. (3) If the monthly inventory standard is below the minimum number of inventory, the minimum number of inventory is taken as inventory for that month.

  For example, when performing inventory planning based on the inventory standard shown in FIG. 12, assuming that the part shown in FIG. 12 is a “non-electric part”, the minimum number of possessions is “12”. Looking at the inventory standards for each month, the inventory standards for any month are not "0". Also, the inventory standard for any month is not less than “12”. Therefore, the inventory plan formulation unit 9 sets the values shown in FIG. 12 as the planned inventory for each month as it is and stores it in the inventory plan DB 11. Here, “planned inventory” refers to inventory that is determined to be secured at the end of each month.

  FIG. 13A is a diagram illustrating an example of monthly demand prediction calculated for a certain electrical component. When inventory planning is performed based on the monthly demand forecast as shown in FIG. 13A, the minimum holding number is “20”. FIG. 13B is a diagram illustrating the monthly demand forecast of FIG. 13A, the inventory standard obtained from the monthly demand forecast, and the planned inventory which is a determined value.

  When the stock standard is calculated, for example, in October, the value of November and the value of December are totaled to “0”. Therefore, from the policy (1), the total annual demand forecast is set as the planned inventory for October. In November and December, the stock standard is “0”. Therefore, according to the policy (3), the minimum inventory number “20” is set as the planned inventory for November and December. In January, the value in February and the value in March are summed to “0”. Therefore, from the policy (2), the inventory standard for December, which is the previous month, is set as the planned inventory. Similarly, from March to June, the inventory standard is inventory, and from July to September, the inventory standard is below the minimum number of stocks, so “20” is the planned inventory. In this way, the “inventory plan (determined value)” of FIG. 13-2 is obtained.

  As described above, in the present embodiment, the demand forecast pattern is set in step S2, the annual demand forecast total is calculated in step S3, and the monthly demand forecast is obtained based on these results. In this way, since the annual demand forecast total and the demand forecast pattern are handled independently, it is possible to finely adjust the number of monthly demand forecasts by operating one of the parameters. This leads to the recent JIT (Just In Time) idea of stocking only the necessary amount when necessary.

  For example, even in the middle of the fiscal year, by correcting the annual demand forecast total, the value of the monthly demand forecast can be changed while maintaining the demand forecast pattern. As a result, the monthly demand trend is in line with the actual situation, but if a shortage in volume is expected, the total annual demand forecast can be set to a larger value. In addition, by selecting a new demand prediction pattern or creating a new one, it is possible to perform an optimal demand forecast for each month while maintaining the total annual demand forecast. For example, when a part-specific demand occurs, a demand prediction pattern can be individually created and applied. Further, for example, when there is a part having a large difference between the monthly forecast and the actual result, it is possible to make a prediction according to the actual situation by creating a new demand forecast pattern.

  As described above, in the present embodiment, for newly set parts with a small track record, a demand forecast pattern and a total annual demand forecast are obtained separately from past usage records, and the demand forecast is performed using these. did. Thereby, demand prediction is possible only from the past results. Moreover, since the demand forecast pattern and the annual demand forecast total are calculated separately, it is possible to accurately perform demand forecast. Moreover, since a demand prediction pattern is derived, seasonal variations can be taken into consideration.

  Moreover, since a demand forecast can be changed by adjusting any one of a demand forecast pattern and an annual demand forecast total, the adjustment according to the actual condition can be performed easily.

  In addition, since the minimum number of holdings is set separately from the demand prediction and the planned inventory for each month is adjusted, it is possible to prepare for the case where the demand prediction is lost, and the risk of missing parts can be avoided.

  In the present embodiment, assuming that the finished product is a summer product, the refrigeration year was adopted and the demand forecast was performed. However, even if the year assuming other seasonal variation characteristics is adopted, the above Similar effects can be obtained.

  As described above, the demand forecasting method according to the present invention is useful when formulating an inventory plan, and is particularly suitable for demand forecasting of products with little past performance such as newly set parts.

It is a figure which shows the structural example of an inventory plan formulation system. It is a figure which shows an example of the table structure of components master DB. It is a figure which shows an example of the table structure of track record DB. It is a figure which shows an example of the table structure of demand prediction pattern DB. It is a figure which shows an example of the table structure of demand forecast DB. It is a figure which shows an example of the table structure of inventory plan DB. It is a flowchart which shows an inventory plan formulation process. It is a flowchart which shows the operation | movement of a demand prediction pattern setting process. It is a figure which shows the example of a demand prediction pattern. It is a figure which shows the example of the use performance of the last year. It is a flowchart which shows the operation | movement of an annual demand forecast total calculation process. It is a figure which shows an example of the demand forecast number of each month. It is a figure which shows an example of the stock standard of each month of this year. It is a figure which shows the example of a monthly demand forecast. It is a figure which shows the monthly demand forecast of Drawing 13-1, the stock standard obtained from this monthly demand forecast, and the planned stock which is a decision value.

Explanation of symbols

1 Parts master DB
2 results DB
3 Part information / result information acquisition part 4 Prediction target part determination part 5 Order point setting part 6 Demand prediction pattern DB
7 Demand Forecasting Unit 8 Inventory Standard Calculation Unit 9 Inventory Plan Formulating Unit 10 Demand Forecast DB
11 Inventory plan DB

Claims (16)

A demand prediction pattern acquisition step of acquiring all demand prediction patterns of parts for which demand prediction is performed, from demand prediction pattern storage means for storing one or more demand prediction patterns for each part;
A usage record acquisition step of acquiring a previous usage record of a part for which the demand prediction is performed, from a record storage unit that stores a past use record in a part unit;
For each of the acquired demand prediction patterns, calculate a pattern distance evaluation indicating the degree of deviation from the actual usage of the part for which the demand prediction is performed in the previous year, compare the obtained pattern distance evaluation, and the demand with the smallest deviation A demand forecast pattern selection step for selecting a forecast pattern;
An annual demand forecast total calculating step for deriving a regression equation indicating a change in demand based on the previous year's use results of the parts for which the demand forecast is performed, and obtaining an annual demand forecast total for the current year based on the regression equation;
A monthly demand forecast number calculating step of calculating a monthly demand forecast number by allocating the annual demand forecast total to each month according to the selected demand forecast pattern;
Demand forecasting method characterized by including.   2. The demand forecast method according to claim 1, wherein when the total annual demand forecast based on the regression formula is lower than the use result of the previous year, the use result of the previous year is set as the total demand forecast of the current year. .   The demand prediction method according to claim 1, wherein the pattern distance evaluation is obtained by a least square method.   The demand prediction method according to claim 1, wherein the regression equation is a linear regression line. A demand forecast correction step for correcting a demand forecast by adjusting at least one of the annual demand forecast total and the selected demand forecast pattern;
The demand prediction method according to claim 1, further comprising: An inventory plan development method including the demand prediction method according to any one of claims 1 to 5,
An inventory standard determination step for determining an inventory standard that is a target at the end of each month based on the monthly demand forecast number calculated in the process of calculating the monthly demand forecast number;
An inventory planning method characterized by including: A prediction target determination step for determining whether a part that performs demand prediction is a prediction target part or a non-prediction part;
An order point setting step for setting a minimum holding number that is a value that clears a predetermined shortage rate as the monthly demand prediction number when it is determined that the part is a non-prediction part in the prediction target determination step;
The inventory plan development method according to claim 6, further comprising: An inventory planning step for determining an annual inventory plan based on the inventory criteria;
The inventory plan formulation method according to claim 6 or 7, further comprising: Demand forecast pattern storage means for storing one or more demand forecast patterns per component;
A record storage means for storing past use records in parts;
From the results storage means, parts information results information acquisition means for acquiring the use results of the previous year of the parts for which demand prediction is performed,
A pattern distance indicating the degree of deviation from the previous year's use record of the part for which the demand prediction is obtained for each demand prediction pattern, by acquiring all demand prediction patterns of the part for which the demand prediction is performed from the demand prediction pattern storage unit Calculate the evaluation, compare the obtained pattern distance evaluation, select the demand forecast pattern with the smallest discrepancy, and derive the regression equation that shows the fluctuation of the demand based on the actual usage in the previous year. A demand forecasting means for calculating a monthly demand forecast number by calculating an annual demand forecast total for the current year based on an equation and distributing the annual demand forecast total to each month according to the selected demand forecast pattern;
A demand forecasting system characterized by comprising: The demand forecasting means is
10. The demand forecasting system according to claim 9, wherein when the total annual demand forecast based on the regression formula is lower than the use record of the previous year, the use record of the previous year is set as the total demand forecast of the current year. . The demand forecasting means is
The demand prediction system according to claim 9 or 10, wherein the pattern distance evaluation is obtained by a least square method.   The demand prediction system according to claim 9, wherein the regression equation is a linear regression line. The demand prediction means further includes:
The demand forecast according to any one of claims 9 to 12, wherein the demand forecast is corrected by adjusting at least one of the annual demand forecast sum and the selected demand forecast pattern. system. An inventory plan development system comprising the demand prediction system according to any one of claims 9 to 13,
Inventory standard calculation means for determining the inventory standard that is the target at the end of each month based on the monthly demand forecast number calculated by the demand prediction means,
An inventory planning system characterized by comprising: About the part for which the demand is predicted, a prediction target part determination means for determining whether the part is a prediction target part or a non-prediction part;
When it is determined that the prediction target part determination unit is a non-prediction part, an order point setting unit that sets a minimum possession number that is a value that clears a predetermined shortage rate as the monthly demand forecast number;
The inventory planning system according to claim 14, further comprising: Inventory planning means for determining an annual inventory plan based on the inventory criteria;
The inventory plan development system according to claim 14 or 15, further comprising:

Images