Modeling and Generating Intermittent Demand Manuel D. Rossetti1, Ph. D., P. E., Vijith Varghese, Yasin Unlu Department of Industrial Engineering University of Arkansas 4207 Bell Engineering Center Fayetteville, AR 72701

Abstract Demand series, which have a significant proportion of zeroes and non-zero values having a high coefficient of variation, are often considered as intermittent. Intermittent demand often occurs within the spare parts industry because of the complex arrival processes involved when failed parts are sent for inventory replacement. In addition, intermittent demand occurs when demand is collected at the stock keeping unit (SKU) level. Intermittent demand presents unique forecasting and inventory modeling issues. This paper presents and evaluates methodologies for modeling and generating intermittent demand. These models can be used as motivation for forecasting techniques and are especially relevant for generating input sequences for simulating inventory models under intermittent demand characteristics. Two intermittent demand generation models are presented. Their ability to represent intermittent demand is evaluated via a set of simulation experiments. Keywords: intermittent, sporadic, lumpy demand, demand generation

1 Corresponding author. Tel.: 479-575-6756; fax: 479-575-8431. Email address: [email protected]