DTRP - An extension of the Delay Time model to include Replacement Possibility.

Forecasting demand for upcoming periods is an issue that has been thoroughly discussed and researched. It is of great importance as company strategies and schedules are made based on these forecasts every day. This thesis investigates and compares four types of existing forecasting methods and proposes a new one that is consistent with a failure mode in which items can also be replaced with newer models instead of invariably fixed. Specifically, there are two major contributions made to the existing literature. Firstly, it provides a detailed explanation of a comparison of four existing methods used in forecasting of demand. It evaluates the strength and weaknesses of these methods, and states the most efficient method given the setup of the sample. The findings indicate that the best forecasting method depends on the error function and the sample size of the items. Secondly, this paper defines two modes of failure and contributes a new forecasting method tailored to it. The first failure mode is set up in such a way that a failure of an item is always cause for demand. In the second failure mode, an item is either fixed or replaced. Only the items that are fixed are cause for demand. In accordance with the second failure mode, a new method is proposed. This method is under certain circumstances an improvement to existing methods.