The truncated normal–gamma mixture as a distribution for lead time demand

A central problem in inventory control concerns the statistical description of total demand during the lead time between placement and receipt of an order. Since both the demand per unit time and the lead time are random variables, the distribution of lead time demand is formulated as a mixture. In application, one or both components of the mixture are often assumed to be normally distributed despite their nonnegativity. In practice, this assumption is taken to imply that a normal curve, truncated at zero, is a valid description. Several approximations to the true tail area probabilities (given truncation) are examined. In turn, this leads to an improved and simpler expression for the normal–gamma mixture. Some numerical results are also presented.