A dynamic inventory model using exponential smoothing.

Inventory systems may be analyzed using techniques which were originally devised to study the dynamic response of servomechanisms. This paper uses these methods to study an inventory system using exponential smoothing to predict a demand which is assumed to have a linear trend with uncorrelated noise components. The statistical properties of the estimator are investigated, and both exact and limiting expressions for its mean and variance are derived. The model is assumed to have a fixed time lag between the placing of a replenishment order and its receipt, and the system response is determined for impulse, step and ramp inputs. The choice of a smoothing constant is discussed, and the effect of the choice on the variance of the forecast equation, and the response time of the system is analyzed.http://www.archive.org/details/dynamicinventory00reilLieutenant, United States Nav