A simulation study comparing alternative inventory buffering techniques in the presence of forecast bias

Forecasting and forecast error have been frequent research paper topics. The effect this error has on a production system has been shown to be a critical point in determining supply chain costs. These costs come in the form of increased inventory, lost sales and production instability. A critical component of this error is the degree of forecast bias. Bias affects the forecast and resulting statistical safety stock by over or under compensating for this form of error. This paper evaluates different methods for adjusting the safety stock in the presence of bias using a simulation model. The simulation model included several different characteristics to match various environments. The results have been presented as a guide to the master scheduler. This guide allows a master scheduler to determine the most appropriate safety stock method using characteristics that match a particular environment. The traditional safety stock methods, as well as several alternatives, have been evaluated to provide a robust comparison and set of recommendations.