Method for forecasting a future inventory demand

A computer-implemented method for establishing inventory quantities based on patterns of demand variability comprising:

• receiving, by a processor associated with a computer, historical demand data associated with one or more part numbers, wherein historical demand data includes data gathered during one or more historical demand periods;
• performing, by the processor, a first statistical analysis on one or more part numbers to identify seasonal part numbers from among the one or more part numbers;
• establishing, by the processor based on the first statistical analysis, a seasonal fixed requirement associated with the seasonal part numbers, wherein the seasonal fixed requirement corresponds to one or more monthly inventory levels associated with an inventory environment;
• performing, by the processor, a second statistical analysis on any part number not identified as a seasonal part number to identify quasi-seasonal part numbers, wherein the second statistical analysis is configured to identify corresponding portions of the historical demand data that exceed predetermined demand thresholds over multiple demand periods;
• and
• establishing, by the processor based on the second statistical analysis, a quasi-seasonal fixed requirement associated with quasi-seasonal part numbers, wherein the quasi-seasonal fixed requirement includes a duration corresponding to a predetermined portion of the future demand period;
• determining, by the processor, a coefficient of variation for historical data corresponding to part numbers not identified as either seasonal or quasi-seasonal;
• identifying, by the processor, one or more of the part numbers as high variability part numbers if the coefficient of variation associated with a respective part number is greater than 1;
• and
• establishing, by the processor, a high variability fixed requirement associated with the one or more high variability part numbers, establishing a high variability fixed requirement includes setting an inventory level associated with a respective part number substantially equal to one month of estimated demand for a respective part number if a cost associated with the respective part number is less than a predetermined cost threshold or the cost associated with the respective part number is greater than a predetermined cost threshold and the estimated demand of the part number is greater than a predetermined demand threshold.

A computer-implemented method for forecasting a future inventory demand comprising:

• receiving, by a processor associated with a computer, historical demand data associated with a part number;
• statistically, by the processor, analyzing the historical demand data associated with the part number to identify a portion of a historical demand period as one of a seasonal, a quasi-seasonal, a high variability, or a non-seasonal;
• estimating, by the processor, a first inventory requirement for a first predetermined portion of a future demand period if the part number is identified as a seasonal part number;
• estimating, by the processor, a second inventory requirement for a second predetermined portion of a future demand period if the part number is identified as a quasi-seasonal part number, wherein the second predetermined portion of the future demand period is a multiple of the first predetermined portion;
• determining, by the processor, a coefficient of variation for historical data corresponding to part numbers not identified as either seasonal or quasi-seasonal;
• identifying, by the processor, one or more of the part numbers as high variability part numbers if the coefficient of variation associated with a respective part number is greater than 1;
• and
• establishing, by the processor, a high variability fixed requirement associated with the one or more high variability part numbers, wherein establishing a high variability fixed requirement includes setting an inventory level associated with a respective part number substantially equal to one month of estimated demand for a respective part number if a cost associated with the respective part number is less than a predetermined cost threshold or the cost associated with the respective part number is greater than a predetermined cost threshold and the estimated demand of the part number is greater than a predetermined demand threshold.