In the early 1990s, the IBM Corporation decided that its microelectronics d
ivision should expand from producing parts exclusively for other IBM locati
ons to producing a range of products for diverse customers. To overhaul its
supply-chain-management applications to handle the new business, it develo
ped intelligent models to match assets with demand to determine which deman
ds it could meet when and to provide manufacturing guidelines. In 1994, the
PROFIT team began applying OR techniques to build these tools, interweavin
g linear programming with a traditional material resource planning algorith
m and a heuristic matching process based on clues established in the explos
ion algorithm. The team has deployed three core applications: a weekly divi
sion run that determines customer commitments and manufacturing requirement
s, daily manufacturing runs that identify the best use of manufacturing res
ources to meet division requirements, and a division available-to-Promise a
pplication that facilitates fast response to customers placing orders (not
described). This work has improved manufacturing utilization and customer-o
rder response time.