Modeling framework
In order to derive the benefits and distribution of
benefits from VMI, we construct a two-level supply
chain consisting of a single supplier and a single
buyer and examine inventory management practices
before and after the implementation of VMI. A
simple supply chain is used for computational
ease, though many of the results can be generalized
to more complex supply chains. We assume that a
single stock-keeping unit is transacted between the
supplier and the buyer and that the buyer faces
external demand from consumers. In a supply
chain without VMI, the supplier observes consumer
demand only indirectly through the buyer’s ordering
policy. With VMI, the supplier’s information system
directly receives consumer demand data. Consumer
demand is assumed to be deterministic and known
to the buyer under normal (non-VMI) operating
procedures and known to both the buyer and supplier
under VMI. The supplier’s and buyer’s inventory
carrying charges per unit are denoted as H and
h, respectively (throughout the paper, parameters in
upper case and lower case are associated with the
supplier and buyer, respectively). The supplier’s and
buyer’s cost of placing an order are denoted as C,
c (without VMI), and cV (with VMI)—there is no
need to define CV since it is assumed that VMI does
Y. Yao et al. / Decision Support Systems 43 (2007) 663–674 665
not change the supplier’s order cost. Without loss
of generality, we assume that order lead times, the
time between when an order is placed and when
the shipment is received, are negligible for both the
supplier and the buyer. That is, once a replenishment
order is placed by either the supplier or the
buyer, the replenishment shipment is received instantaneously.
The type of supply chain described above
approximates those in a number of industries like
automobiles where the buyer’s demand is reasonably
stable and predictable. Toyota, for example, faces
fairly predictable demand, at least in the short-tomedium
run, for many of its models. Certain consumer
goods (e.g., staples such as sugar, flour, and
canned soups) may also face fairly stable and predicable
demands.
The ordering process considered is an inventory
review system where orders are placed at pre-determined
reorder points. Since demands are known, the
main cost minimization parameter available to the
buyer in a non-VMI system is its order size. In
other words, the buyer must determine its order quantity.
Once the buyer reaches its reorder point, a replenishment
request is passed to the supplier, and the
order quantity is immediately shipped to the buyer.
The supplier then reviews its inventory and plans its
own ordering or production processes. The major
difference between not using and using VMI is that
the buyer’s order quantity is determined by the supplier
in a VMI system. Fig. 1 presents the modeling
framework.