For service provider companies, it is important to repair the faulty goods returned from the market as soon as possible. To do so the faulty parts need to be replaced at earliest possible time. If there are no spare parts available on stock, their order and delivery will increase the throughput time of repair process. Having a reliable demand prediction method that is able to forecast the spare parts demand in a precise manner would allow the service provider to keep the appropriate quantity of different spare parts on its stock. The goal of my work was to implement a method that can be considered as a supplementary one to the statistical forecasting methods.
The theoretical background of a method that is based upon the deterministic modeling of time series as well as its practical implementation in OpenERP is introduced in this work. The necessary calculations, application of fuzzy methods, the algorithm for historical demand data analysis, the database to store the data, the OpenERP system and the calculations done in MATLAB are discussed here.
The developed module is able to provide long term forecasts for demand of active parts based on historical demand information gained form end of life parts. The application can show the forecast both in text and graphical format.