System and method for estimating business demand based on business influences

• US 5,712,985 A
• Filed: 10/13/1995
• Issued: 01/27/1998
• Est. Priority Date: 09/12/1989
• Status: Expired due to Fees

• Current Assignee: Stafanpolus KG Limited Liability Company
• Original Assignee: Park City Group Incorporated
• Inventors: Lee, Michael D., Tondevold, Barrire K., Pond, Jamie T., Fields, Randall K.
• Primary Examiner(s): Hayes, Gail O.
• Assistant Examiner(s): Shingala, Gita
• Application Number:
  US08/542,847
• Time in Patent Office: 837 Days
• Field of Search: 395/208, 395/210, 395/207, 364/468.01, 364/468.02, 364/468.03
• US Class Current: 705/7.31
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

A demand forecasting and production scheduling system and method creates production schedules for various business items describing a forecasted demand for the business items in a number of future time intervals. The system includes a computer managed database of various profiles, including a base profile for each business item, and a number of influence profiles. The profiles describe variations in demand for the business item in a number of time intervals. The base profile describes an underlying level of demand for a business item that is anticipated for the business item absent any influencing factors, such as promotional sales, holidays, weather variations, and the like. The variations in demand for the business item due to such influence factors are stored in the database as influence profiles. The influence profiles may be either standard, percentage, or seasonal. The forecasted demand for a business item in a number of future time intervals is determined by selective combination of the base profile for the business item and any number of influence profiles. The forecasted demand is stored in the database in a forecast profile. From the forecast profile a production schedule is created, and the business item provided according to the production schedule. Actual demand for the business item is monitored and stored. The variation between actual demand and the forecasted demand is used to update the base and influence profiles. From the updated base and influence profiles the forecasted demand is redetermined, and the production schedule updated accordingly.

Claims

1. A computer implemented production scheduling system for projecting future demand in a plurality of time intervals for at least one business item, s accounting for recurring variations in actual demand for the business item, and for scheduling production of the business item, the system comprising:

a memory divided into separate files including a database comprising;

a base profile defining a base demand for the business item in each of a first plurality of time intervals;

at least one influence profile defining a variation of the base demand in each of a second plurality of time intervals;

at least one forecast profile defining in each of a plurality of future time intervals a forecasted demand for the business item as a function of a combination of the base profile and the at least one influence profile;

a data file containing actual current demand for the business item in each of a plurality of past time intervals;

a processor coupled to the memory to receive the base profile, and the at

least one influence profile and to produce therefrom at least one forecast profile forecasting demand for the business item in each of a plurality of future time intervals;

the processor further updating the base profile and the at least one influence profile in each of selected past time intervals with a weighted minimum update value in each selected time interval; and

a production schedule apparatus coupled to the processor to receive a forecast profile and to produce therefrom a production schedule for the business item indicating a forecasted demand for the business item in a plurality of future time intervals.

2. The system of claim 1, wherein the weighted minimum update value is a function of:

a difference between the actual current demand and the forecasted demand, the difference bounded by a maximum and a minimum value; and

,

a ratio of the combined base profile and the at least one influence profile to a combined absolute value of the base profile and a combined absolute value of the at least one influence profile.

3. The system of claim 1, wherein there is provided at least one standard influence profile having unit values, the forecast profile is determined by the equation:

##EQU19## where m is a number of time intervals in the base profile;

n is a number of standard influence profiles;

standard_-- profile is a standard influence profile; and

,

base_-- profile is the base profile.

4. The system of claim 3, wherein there is provided at least one percentage influence profile having percentage values, the forecast profile is determined by the equation:

##EQU20## where forecast_-- profile is the forecast profile; and

,

percent_-- profile is a percentage influence profile.

5. The system of claim 4, wherein there is provided at least one seasonality influence profile, the forecast profile is determined by the equation:

##EQU21## where season_-- pro is the seasonality influence profile.

6. The system of claim 1, wherein the updated base profile is determined by the equation:

##EQU22## where base_-- profile is the base profile prior to being updated;

profile_-- weight is a weighting value specifying a percentage of the base profile to apply to the updated base profile; and

delta_-- profile is a function of a difference between the actual current demand and the forecast profile, the difference bounded by a maximum and a minimum value for updating the base profile, and a ratio of the combined base profile and the at least one influence profile to a combined absolute value of the base profile and a combined absolute value of the at least one influence profile.

7. The system of claim 1, wherein an updated influence profile is determined by the equation:

##EQU23## where influence_-- profile is an influence profile prior to being updated;

profile_-- weight is a weighting value specifying a percentage of the influence profile to apply to the updated influence profile; and

delta_-- profile is a function of a difference between the actual current demand and the forecast profile, the difference bounded by a maximum and a minimum value for updating the influence profile and a ratio of the influence profile to a combined absolute value of the base profile and a combined absolute value of the at least one influence profile.

8. The system of claim 6, wherein the delta profile is determined by the equation:

##EQU24## where base_-- value is vector sum of the base profile and the at least one influence profile;

min_-- update is the minimum value for updating the base profile;

profile_-- sum is the combined absolute value of the base profile and the combined absolute value of the at least one influence profile; and

,

delta is a vector based on the function of the difference between the actual current demand and the forecast profile, the difference bounded by a maximum and the minimum value for updating the base profile.

9. The system of claim 8, wherein the delta vector is equal to the vector sum of a first vector determined by the equation:

##EQU25## and a second vector determined by the equation;

##EQU26## where actual_-- dem is the actual demand for the business item in vector form;

forecast is the forecast profile; and

,

max_-- update is the maximum value for updating the base profile.

10. The system of claim 9, wherein profile_-- sum is determined by the equation:

##EQU27## where std_-- pro is an influence profile expressed in unit values.

11. The system of claim 10, wherein the profile_-- sum further includes in the vector sum a value modifier vector determined by the equation:

##EQU28## where base_-- value is the combined base profile and at least one influence profile;

balance_-- factor is a specified weighting value for weighting the contribution of percentage influence profiles relative to standard influence profiles; and

,

percent_-- pro is a percentage influence profile.

12. The system of claim 7, wherein the delta profile is determined by the equation:

##EQU29## where std_-- pro is the at least one influence profile;

min_-- update is the minimum value for updating the base profile;

profile_-- sum includes the combined absolute value of the base profile and the combined absolute value of the at least one influence profile; and

,

delta is a vector based on the function of the difference between the actual current demand and the forecast profile, the difference bounded by a maximum and the minimum value for updating the base profile.

13. The system of claim 12, wherein the delta vector is equal to the

vector sum of a first vector determined by the equation:

##EQU30## and a second vector determined by the equation;

##EQU31## where actual_-- dem is the actual demand for the business item in vector

forecast is the forecast profile; and

,

max_-- update is the maximum value for updating the base profile.

14. The system of claim 1, wherein there is provided at least one percentage influence profile defining a percentage variation in the base demand in the base profile, each updated percentage influence profile determined by the equation:

##EQU32## where percent-pro is the percentage influence profile;

profile_-- weight is a weighting value specifying a percentage of the percentage influence profile to apply to the updated percentage influence profile; and

,

weight_-- per_-- pro is a combined value of the percentage influence profile and a delta vector, and the delta vector is a function of ratio of percentage influence profile to combinedvalue of all percentage influence profiles.

15. A computer implemented method of forecasting future demand for a business item, the method operable in a computer system including a memory for storing a plurality of files and procedures, comprising the steps

storing in a memory actual demand for the business item in each of a first plurality of time intervals;

storing in the memory a base profile defining a historical actual demand for the business item in each of a second plurality of time intervals;

storing in the memory at least one influence profile defining a variation in demand for the business item due to an influence factor in each of a plurality of time periods, wherein the influence profiles include at least one of a set comprised of;

a standard influence profile defining a variation in demand in a same unit value as the base profile;

a percentage influence profile describing a variation in demand in as a percent change;

combining the base profile and selected influence profiles to produce a production schedule including a forecast profile projecting demand for the business item in a selected plurality of future time intervals;

monitoring actual demand for the business item in each of the future time intervals as such future time intervals become present or past time intervals, to create an actual demand profile; and

,

updating the base profile and the selected influence profiles with respect to the actual demand profile, wherein updating comprises;

determining a total error between the actual demand and the forecast profile;

determining for the base profile a percent contribution of the base profile to the total error;

determining for each influence profile a percent contribution of the influence profile to the total error;

updating the base profile by its percent contribution to the total error; and

,

updating each influence profile by its percent contribution to the total error.

16. The method of claim 15, wherein the step of updating the base profile and the selected influence profiles, comprises, for each past or present time interval, the steps of:

combining the base profile with selected influence profiles;

determining a weighted delta profile for the base profile describing the impact of the base profile on the forecast profile;

determining a weighted delta profile for each influence profile describing the impact of the influence profile on the forecast profile;

updating the base profile by its weighted delta profile, and a weighting factor associated with the base profile; and

,

updating each influence by its weighted delta profile, and a weighting factor associated with the influence profile.

17. The method of claim 16, wherein the step of combining the base profile and selected influence profiles to create a forecast profile further comprises the steps of:

summing the base profile with selected standard influence profiles;

normalizing percentage influence profiles to define demand in a same unit as the base profile;

summing the normalized percentage influence profiles; and

,

multiplying summed percentage influence profiles by the summed base profile and standard influence profile to produce a forecast profile.

18. The method of claim 17, further comprising the step of:

summing the forecast profile with a seasonality influence profile describing a variation in demand resulting from a seasonal influence to produce an adjusted forecast profile.

19. The method of claim 15, wherein there is provided at least one standard influence profile having unit values, the step of combining the base profile and selected influence profiles comprises determining the forecast profile by the equation:

##EQU33## where m is a number of time intervals in the base profile;

n is a number of standard influence profiles;

standard_-- profile is a standard influence profile; and

,

base_-- profile is the base profile.

20. The method of claim 19, wherein there is provided at least one percentage influence profile having percentage values, the step of combining the base profile with selected influence profiles further comprises determining the forecast profile with the equation:

##EQU34## where forecast_-- profile is the forecast profile; and

,

percent_-- profile is a percentage influence profile.

21. The method of claim 19, wherein there is provided at least one seasonality influence profile, the step of combining the base profile with selected influence profiles further comprises determining the forecast profile with the equation:

##EQU35## where season_-- pro is the seasonality influence profile.

22. The method of claim 15, the step of updating the base profile updating the base profile by its percent contribution to the total error comprises determining the updated base profile with the equation:

##EQU36## where base_-- profile is the base profile prior to being updated;

profile_-- weight is a weighting value specifying a percentage of the base profile to apply to the updated influence profile; and

delta_-- profile is the percent contribution to the total error as a function of a difference between the actual current demand and the forecast profile, the difference bounded by a maximum and a minimum value for updating the base profile, and a ratio of the combined base profile and the at least one influence profile to a combined absolute value of the base profile and a combined absolute value of the at least one influence profile.

23. The method of claim 15, the step of updating each influence profile by its percent contribution to the total error comprises each influence profile with the equation:

##EQU37## where influence_-- profile is an influence profile prior to being updated;

profile_-- weight is a weighting value specifying a percentage of the influence profile to apply to the updated influence profile; and

,

delta_-- profile is the percent contribution to the total error as a function of a difference between the actual current demand and the forecast profile, the difference bounded by a maximum and a minimum value for updating the influence profile and a ratio of the influence profile to a combined absolute value of the base profile and a combined absolute value of the at least one influence profile.

24. The method of claim 22, wherein the delta profile is determined by the equation:

##EQU38## where base_-- value is vector sum of the base profile and the at least one influence profile;

min_-- update is the minimum value for updating the base profile;

profile_-- sum is the combined absolute value of the base profile and the combined absolute value of the at least one influence profile; and

,

delta is a vector based on the function of the difference between the actual current demand and the forecast profile, the difference bounded by a maximum and the minimum value for updating the base profile.

25. The method of claim 24, wherein the delta vector is set equal to the vector sum of a first vector determined by the equation:

##EQU39## and a second vector determined by the equation;

##EQU40## where actual_-- dem is the actual demand for the business item in vector form;

forecast is the forecast profile; and

,

max_-- update is the maximum value for updating the base profile.

26. The method of claim 25, wherein profile_-- sum is determined by the equation:

##EQU41## where std_-- pro is an influence profile expressed in unit values.

27. The method of claim 26, wherein the profile_-- sum further includes in the vector sum a value modifier vector set equal to:

##EQU42## where base_-- value is the combined base profile and at least one influence profile;

balance_-- factor is a specified weighting value for weighting the contribution of percentage influence profiles relative to standard influence profiles; and

,

percent_-- pro is a percentage influence profile.

28. The method of claim 23, wherein the delta profile is determined by the equation:

##EQU43## where std_-- pro is the at least one influence profile;

min_-- update is the minimum value for updating the base profile;

profile_-- sum includes the combined absolute value of the base profile and the combined absolute value of the at least one influence profile; and

,

delta is a vector based on the function of the difference between the actual current demand and the forecast profile, the difference bounded by a maximum and the minimum value for updating the base profile.

29. The method of claim 28, wherein the delta vector is set equal to the vector sum of a first vector equal to:

##EQU44## and a second vector equal to;

##EQU45## where actual_-- dem is the actual demand for the business item in vector form;

forecast is the forecast profile; and

,

max_-- update is the maximum value for updating the base profile.

30. The method of claim 15, wherein there is provided at least one percentage influence profile defining a percentage variation in the base demand in the base profile, the step of updating the base profile and the selected influence profiles comprises setting each updated percentage influence profile equal to:

##EQU46## where percent-pro is the percentage influence profile;

profile_-- weight is a weighting value specifying a percentage of the percentage influence profile to apply to the updated percentage influence profile; and

,

weight_-- per_-- pro is a combined value of the percentage influence profiles and a delta vector, the delta vector being a function of ratio of percentage influence profile to combinedvalue of all percentage influence profiles.

31. The method of claim 16, wherein there is provided at least one standard influence profile having unit values, and at least one percentage influence profile having percentage values, the base profile is combined with selected influence profiles using the equation:

##EQU47## where;

percent_-- profile is a percentage influence profile;

standard_-- profile is a standard influence profile;

m is a number of time intervals in the base profile;

n is a number of standard influence profiles; and

base_-- value is determined by the equation;

##EQU48## where;

standard_-- profile is a standard influence profile; and

,

base_-- profile is the base profile.

32. The method of claim 31, wherein a weighted delta profile for the base profile is determined by the equation:

##EQU49## where;

min_-- update is the minimum value for updating the base profile;

profile_-- sum includes the combined absolute value of the base profile and the combined absolute value of the at least one influence profile; and

,

delta is a vector based on the function of the difference between the actual current demand and the forecast profile, the difference bounded by a maximum and the minimum value for updating the base profile.

33. The method of claim 32, wherein the delta vector is set equal to the vector sum of a first vector equal to:

##EQU50## and a second vector equal to;

##EQU51## where;

actual_-- dem is the actual demand for the business item in vector form;

forecast is the forecast profile; and

,

max_-- update is a maximum value for updating the base profile.

34. The method of claim 33, wherein the updated based profile is determined by the equation:

##EQU52## where;

base_-- profile is the base profile prior to being updated;

profile_-- weight is a decimal weighting value specifying a percentage of the base profile to apply to the updated base profile; and

delta_-- profile is the weighted delta profile for the base profile.

35. The method of claim 33, wherein there is provided at least one standard influence profile, the weighted delta profile for each standard influence profile is determined by the equation:

##EQU53## where std_-- pro is the at least one standard influence profile,

min_-- update is the minimum value for updating the base profile,

profile_-- sum includes the combined absolute value of the base profile and the combined absolute value of the at least one influence profile, and,

delta is the delta vector.

36. The method of claim 15, wherein the updated influence profile is determined by the equation:

##EQU54## where;

influence_-- profile is an influence profile prior to being updated;

profile_-- weight is a decimal weighting value specifying a percentage of the influence profile to apply to the updated influence profile; and

delta_-- profile is the weighted delta profile for the influence profile.

37. The method of claim 15, wherein there is provided at least one percentage influence profile defining a percentage variation in the base demand in the base profile, the weighted delta profile for each percentage influence profile is determined by the equation:

##EQU55## where percent-pro is the percentage influence profile;

profile_-- weight is a decimal weighting value specifying a percentage of the percentage influence profile to apply to the updated percentage influence profile; and

,

weight_-- per_-- pro is weighted percentage profile determined by a function of a combined value of the percentage influence profiles, and the delta vector.

38. The method of claim 37, wherein the weighted percentage profile is determined by the equation:

##EQU56## where percent_-- pro is a percentage profile; and

,

delta_-- percent_-- pro is a weighted delta percentage profile determined by the equation;

##EQU57## where % profile_-- sum is the summation of the percentage influence profiles; and

,

δ

percent is a profile describing the overall impact of the percentage influence profiles on a combined base profile and influence profiles.

39. The method of claim 38, wherein δ

percent is determined by the equation;

##EQU58## where;

delta is the delta vector;

base_-- value is the combined base profile and standard influence profiles;

tot_-- delta is the weighted delta profile of the base profile combined with the weighted delta profiles of the standard influence profiles; and

,

percent_-- value is normalized combined percentage influence profiles.

40. The method of claim 32, wherein profile_-- sum is determined by the equation:

##EQU59## where std_-- pro is an influence profile expressed in unit values.

41. The method of claim 40, wherein the profile_-- sum further includes in the vector sum a value modifier vector set equal to:

##EQU60## where balance_-- factor is a specified weighting value for weighting the

contribution of percentage influence profiles relative to standard influence profiles; and

,

percent_-- pro is a percentage influence profile.

42. The method of claim 19, wherein the forecast profile is further determined by the equation:

##EQU61## where season_-- pro is a seasonality influence profile defining variations in demand on a seasonal basis; and

,

forecast_-- pro is the forecast profile.

43. The method of claim 15, further comprising the steps of:

selecting influence profiles;

associating the selected influence profiles as a group; and

,

combining the group of selected influence profiles with the base profile to create a forecast profile.

44. The method of claim 15, wherein the step of storing in the memory at least one influence profile defining a variation in demand for the business item comprises the steps of:

specifying a business item;

specifying a plurality of time intervals for the influence profile;

specifying a weighting factor to be used to update the influence profile; and

,

providing a data value for each time interval specified in the influence profile.

45. The method of claim 44, wherein the step of specifying a plurality of time intervals further includes:

specifying a beginning of a first time interval; and

,

specifying a length of time for subsequent time intervals.

46. The method of claim 44, wherein the step of specifying a plurality of time intervals further includes:

specifying a beginning time of a first time interval;

specifying an ending time of a last time interval; and

,

specifying a length of each time interval between the beginning time and ending time.

47. The method of claim 44, further comprising the steps of:

specifying a unit type for the influence profile, the unit type being one of a numerical count, a currency value, or a physical quantity unit.

48. The method of claim 44, further comprising the step of:

specifying the type of influence profile being one of a standard influence profile, a percentage influence profile, or a seasonal influence profile.