The application of time series models in MRP simulation

The primary objective of this research is to investigate MRP under the uncertainty demand and the change of lot size rules using three major time series methods; Intervention analysis, transfer function and State Space models. A significant part of the research work is devoted to the development and implementation of a MRP simulation program which is used to measure the MRP performance under various conditions. More than one hundred simulation experiments were conducted and analyzed, twenty four simulations were further investigated and illustrated using the time series methods, they consisted of (i) five demand intervention analysis (ii) nine lot size intervention analysis (iii) seven transfer functions (iv) three State Space models. The simulations are analyzed through three important stages of Identification, Estimation and Diagnostics. It is found that the demand variation and lot size rules have significant effects on the MRP performance. Their dynamic relationships can be adequately represented using the time series models. The integrated simulation and time series approach is more useful in the study of the dynamic and transient behavior of MRP than the conventional steady state analysis. It is anticipated that by slight modification of the simulation program, similar studies can be performed on other major MRP policies. Hence alternative MRP designs can be evaluated. The time series methods have been successfully adapted to MRP simulation and the models serve as an important aid to the MRP analyst. This integrated modelling approach appears to be a powerful methodology for the design and analysis of complex manufacturing systems. Furthermore, this research suggests that the modelling approach may find application in other aspects of manufacturing.