System and method for forecasting intermittent demand

• US 6,205,431 B1
• Filed: 10/29/1998
• Issued: 03/20/2001
• Est. Priority Date: 10/29/1998
• Status: Expired due to Term

• Current Assignee: Smart Software, Inc.
• Original Assignee: Smart Software, Inc.
• Inventors: Smart, Charles Nelson, Willemain, Thomas Reed
• Primary Examiner(s): Stamber, Eric W.
• Assistant Examiner(s): MEINECKE DIAZ, SUSANNA M
• Application Number:
  US09/182,249
• Time in Patent Office: 873 Days
• Field of Search: 705/1, 705/7, 705/10, 705/28, 705/29, 700/90, 700/91
• US Class Current: 705/7.25
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 10/06315 Needs-based resource requirements planning or analysis
• CPC Class Codes: G06Q 10/06375 Prediction of business process outcome or impact based on a proposed change
• CPC Class Codes: G06Q 10/0639 Performance analysis of employees; Performance analysis of enterprise or organisation operations
• CPC Class Codes: G06Q 10/087 Inventory or stock management, e.g. order filling, procurement or balancing against orders
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

A system and method for forecasting intermittent demand. The forecasting technique utilizes sample reuse techniques to build a distribution of predicted cumulative lead time demand values that can be analyzed using statistical methods, be used as input for an inventory control system, be used as input for a sales planning system, etc.

Claims

1. A computerized method for forecasting intermittent demand for a lead time, comprising the steps of:

providing a data set of intermittent data comprising a predetermined number of historical demand values;

calculating lead time demand values for the lead time, wherein the calculating step comprises the step of sampling values from the historical demand values;

summing the lead time demand values to provide a lead time demand sum; and

repeating the calculating and summing steps a predetermined number of times to provide a distribution of lead time demand sums that forecast intermittent demand for inventory requirements.

2. The method of claim 1, wherein at least ten percent of the historical demand values are zero.

3. The method of claim 1, wherein the calculating step further includes the step of determining if a next lead time demand value is either zero or non-zero.

4. The method of claim 3, wherein the step of determining if the next lead time demand value is zero is accomplished with a Markov model.

5. The method of claim 3, wherein if the next lead time demand value is non-zero, the next lead time demand value is randomly chosen from among the non-zero historical demand values.

6. The method of claim 5, wherein a jittered lead time demand value is calculated as a random selection of an integer that is in a neighborhood of the lead time demand value.

7. The method of claim 6, wherein the jittered lead time demand value (S) is calculated as follows:

8. The method of claim 7, wherein if S≦

0, then S is forced to LTV.

9. The method of claim 1, further comprising the step of inputting the distribution into an inventory control system.

10. A program product, having a computer usable medium having computer readable code embodied therein that, when executed, comprises:

means for examining intermittent data comprised of a predetermined number of historical demand values made up of zero and non-zero values;

means for calculating lead time demand values for the lead time, wherein the calculating means samples values from the historical demand values;

means for summing the lead time demand values to provide a lead time demand sum;

looping means for generating a plurality of lead time demand sums;

storing means for storing the plurality of lead time demand sums; and

means for providing the plurality of lead time demand sums that forecast inventory requirements.

11. The program product of claim 10, wherein the calculating means sequentially calculates the lead time values, and comprises:

means for determining if each lead time value to be calculated is zero or non-zero;

means for randomly assigning one of the non-zero historical demand values to each non-zero lead time value; and

means for jittering each non-zero lead time value.

12. The program product of claim 11, wherein the means for determining if each lead time value to be calculated is zero or non-zero utilizes a Markov model.

13. The program product of claim 11, wherein the jittering means chooses a randomly selected integer that neighbors each non-zero lead time value.

14. The program product of claim 10, further comprising an inventory control means that receives as input the plurality of lead time demand sums.

15. The program product of claim 10, wherein the looping means loops N times, wherein N is a user defined value.

16. A computer system, comprising:

a processing unit;

a computer system memory accessible to the processing unit;

a data set comprising intermittent data stored within the computer system memory, wherein the intermittent data is comprised of a plurality of time series values;

a mechanism for examining the intermittent data; and

a mechanism for generating a plurality of series of lead time demand values, wherein each series of lead time demand values is based upon at least one sampling selection from among the time series values.

17. The computer system of claim 16, wherein the program further comprises:

a mechanism for summing the lead time demand values within each series of lead time demand values to provide a plurality of lead time demand sums; and

a mechanism for statistically analyzing the plurality of lead time demand sums.

18. The computer system of claim 16, wherein the generating mechanism determines if a next lead time demand value to be calculated is zero or non-zero.

19. The computer system of claim 18, wherein the generating mechanism randomly assigns a non-zero one of the plurality of time series values to the next lead time demand value if the next lead time demand value is calculated as non-zero.

20. The computer system of claim 16, wherein the generating mechanism comprises a jittering mechanism for altering lead time demand values to neighboring integer values.

21. The computer system of claim 16, wherein the at least one sampling selection randomly selects a single non-zero value.

22. The computer system of claim 16, wherein the at least one sampling selection selects a consecutive series of values.

23. A computerized method for generating a lead time demand (LTD) distribution for a lead time, comprising the steps of:

providing a data set of intermittent data comprising a predetermined number of historical demand values;

calculating subseries sums for the lead time based on the data set of intermittent data;

providing a cumulative distribution function (CDF);

calculating the LTD distribution based on the CDF and subseries sums; and

providing the LTD distribution to manage product requirements.

24. The method of claim 23, wherein the CDF is a non-parametric function.

25. The method of claim 23, wherein the CDF is a parametric function.

26. The method of claim 25, wherein the CDF utilizes a lognormal probability density function.

27. The method of claim 25, wherein the CDF utilizes a normal probability density function.

28. The method of claim 23, further comprising the step of inputting the LTD distribution into an inventory control system.

29. A program product embodied on a computer usable medium, for generating a lead time demand (LTD) distribution, wherein said program product, when executed, comprises:

means for examining a data set of intermittent data comprising a predetermined number of historical demand values;

means for calculating subseries sums based on the data set of intermittent data;

means for calculating the LTD distribution based on a cumulative distribution function (CDF) and the subseries sums; and

means for providing the LTD distribution to manage product requirements.

30. The program product of claim 29 further comprising an analysis means for reporting statistical summaries of the LTD distribution.

31. The program product of claim 29, further comprising an inventory control means for managing inventory based on the LTD distribution.

32. The program product of claim 29, wherein the CDF is nonparametric and is given as:

33. The program product of claim 29, wherein the CDF is parametric and is given as:

34. The program product of claim 33, wherein G{circumflex over ( )}(X) is estimated by integrating g(X) from −

∞

to X, where

35. The program product of claim 33, wherein G{circumflex over ( )}(X) is estimated by integrating g(X) from −

∞

to X, where

36. An inventory management program product embodied on a computer usable medium, that, when executed, comprises:

a forecast mechanism, the forecast mechanism including;

,

means for examining intermittent data comprised of a predetermined number of historical demand values made up of zero and non-zero values;

means for calculating lead time demand values for a lead time, wherein the calculating means samples values from the historical demand values;

means for summing the lead time demand values to provide a lead time demand sum;

looping means for generating a plurality of lead time demand sums; and

storing means for storing the plurality of lead time demand sums on a computer system memory; and

an inventory control mechanism, including;

means for reading the plurality of lead time demand sums; and

means for outputting inventory reorder information that specifies when a product order should be placed.

37. The program product of claim 36, wherein the calculating means sequentially calculates the lead time values, and comprises:

means for determining if each lead time value to be calculated is zero or non-zero;

means for randomly assigning one of the non-zero historical demand values to each non-zero lead time value; and

means for jittering each non-zero lead value.

38. The program product of claim 36, wherein the inventory control mechanism further comprises means for outputting performance measures.

39. The program product of claim 36, wherein the inventory reorder information comprise a reorder point and an order quantity.

40. The program product of claim 36, wherein the inventory reorder information comprises a review interval and an order-up-to value.

41. The program product of claim 38, wherein the performance measures include inventory/order costs and service level percentages.

42. An inventory management program product embodied on a computer usable medium, that, when executed, comprises:

a forecast mechanism, the forecast mechanism including;

means for examining a data set of intermittent data comprising a predetermined number of historical demand values;

means for calculating subseries sums based on the data set of intermittent data;

means for calculating a lead time demand (LTD) distribution based on a cumulative distribution function (CDF) and the subseries sums;

an inventory control mechanism, including;

means for reading the plurality of lead time demand sums; and

means for outputting inventory reorder information that specifies when an order should be placed.

43. The program product of claim 42, wherein the inventory control mechanism further comprises means for outputting performance measures.

44. The program product of claim 42, wherein the inventory reorder information comprises a reorder point and an order quantity.

45. The program product of claim 42, wherein the inventory reorder information comprises a review interval and an order-up-to value.

46. The program product of claim 43, wherein the performance measures include inventory/order costs and service level percentages.