Forecast Sensitivity

Abstract

Along the supply chain, in distribution centers, stores, dealers, so forth, forecasts are in continual need for inventory decisions to project the flow of demands over the future months for each item stocked. The more accurate the forecasts, the better the inventory decisions and the more profitable the entity. A 10 % decrease in the measure of the forecast error will result in approximately a 10+ percent decrease in the amount of safety stock needed.

Summary

The more accurate the forecasts , the better the inventory decisions on when and how much stock to carry for each item in a stocking location. These decisions are vital to the profitability of the entity at all of its inventory locations. The coefficient of variation of the 1 month ahead forecast error is used to measure of the forecast accuracy. A series of simulations are run to determine how the number of months of demand history affects the cov when the demand history is from a horizontal demand pattern and from a trend demand pattern. Another series of simulations are aimed at measuring the forecast accuracy depending on the parameters and the forecast model chosen for an item when the demand pattern is horizontal, trend or seasonal. Another series of simulations are run to measure how an outlier demand in the demand history influences the accuracy of the forecast.