THE INTERACTION BETWEEN FORECASTING METHODS AND INVENTORY
MANAGEMENT POLICIES FOR INTERMITTENT DEMAND
Katrien Ramaekers and Gerrit K. Janssens
Operations Management and Logistics
Limburgs Universitair Centrum
Universitaire Campus – Building D
B-3590 Diepenbeek, Belgium
e-mail: {katrien.ramaekers,gerrit.janssens}@luc.ac.be
ABSTRACT
Inventory systems have to cope with uncertainty in
demand. The inventory control literature mostly makes
use of the Normal or Gamma distributions for describing
the demand during the lead-time. The Poisson
distribution has been found to provide a reasonable fit
when demand is very low (only a few pieces per year).
Less attention has been paid to irregular demand. This
type of demand is characterised by a high level of
variability, but may be also of the intermittent type, i.e.
demand peaks follow several periods of zero or low
demands.
In such a situation forecasting demand is considered
difficult. This research investigates the performance of
several forecasting methods for intermittent demand and
their impact on inventory management policies. A
simulation model is built in order to investigate these
effects and to serve as a guide for parameter settings of
both the forecasting models and the inventory
management policies.
The experimental design includes three forecasting
methods: simple exponential smoothing, moving average
and Croston’s method. The inventory management
policies make use of fixed review-time periods. Two
alternative policies are studied: reorder point method
with a fixed order quantity policy, and reorder point
method with an order-up-to-level policy. Intermittent
demand frequency is generated according to a Bernoulli
process or a first-order Markov process. Individual order
sizes are generated using a Gamma distribution.
INTRODUCTION
Inventory management relates to short-term decisionmaking
regarding inventories in organisations, more
specifically to decide on when to order and how much to
order. Decisions are made regarding to objectives stated
by the organisation, e.g. minimisation of costs,
maximisation of profit, or maximisation of service level.
Inventory systems have to cope with uncertainty. They
include uncertainty in demand, in costs, in lead-time and
in supplied quantity (Waters 1992, chapter 4).
In this paper we focus on the uncertainty in demand.
In re-order point models the probability distribution of
demand during the lead-time is an important
characteristic in inventory management. Most textbooks
assume that demand for an item is formed from a large
number of smaller demands from individual customers.
As a result, the authors assume that the resulting demand
is continuous and follows a Normal distribution. For fast
moving items a Normal distribution is appropriate. Silver
and Peterson (1985) recommend the Normal distribution
for items with average lead-time demand higher than 10.
Using the Normal distribution for a demand distribution
can be questioned because (1) the distribution is defined
both on the positive and negative axes; and (2) it is
symmetrical. While the Normal distribution could be
approximately correct in many cases, it is conceptually
not, and of course, cannot be used in computer
simulation as negative demand may be generated at
random. When of relevance, one rather should look for a
distribution, which is defined only for non-negative
values and allows for skewness.
In the literature on inventory control, many times
reference is made to the Gamma distribution. It is
defined only on non-negative values and, according to
the parameters of its distribution, ranges from a
monotonic decreasing function, through unimodal
distributions skewed to the right, to Normal distributions.
The Gamma distribution is attractive because of the ease
it can deal with fixed lead times and how the situations
can be extended to probabilistic distributions of lead
times (Burgin 1975). For items with lower demand,
Silver and Peterson (1985) propose the Laplace or
Poisson distributions. The Poisson distribution has been
found to provide a reasonable fit when the demand is
very low (only a few pieces per year).
Less attention has been paid to irregular demand. This
type of demand is characterised by a high level of
variability, but may be also of the intermittent type, i.e.
demand peaks follow several periods of zero or low
demand. Items with intermittent demand include service.....