Using the gamma distribution to model demand when lead time

When the lead time for an item and the demand per unit of time are both stochastic, the resulting demand during lead time is the convolution of a random number of time periods, each with random demand. If this "compound" demand follows the normal distribution, one can easily determine the...

Full description

Saved in:
Bibliographic Details
Published in:Journal of business logistics 1995-01, Vol.16 (1), p.107
Main Author: Keaton, Mark
Format: Article
Language:English
Subjects:
Approximation
Distribution planning
Inventory control
Logistics
Mathematical models
Normal distribution
Probability
Standard deviation
Studies
Supply & demand
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:When the lead time for an item and the demand per unit of time are both stochastic, the resulting demand during lead time is the convolution of a random number of time periods, each with random demand. If this "compound" demand follows the normal distribution, one can easily determine the reorder point to provide a desired level of customer service. However, there is little likelihood that demand over a random lead time will be normally distributed. Tyworth (1992) has recent offered a new approach to the problem of stochastic lead time, that does not require a compound distribution. The use of flexible gamma distribution for modeling demand per unit of time is discussed. In addition, it is shown how to compute the parameters of the distribution. A Pascal-language computer code to implement Tyworth's approach is provided.
ISSN:0735-3766
2158-1592