This paper describes a practical method for setting safety lead-times for p
urchased components in assembly systems with uncertainty in the supply proc
ess. The approach is specifically designed to be used in an MRP procurement
environment. A two-stage production model in which suppliers were uncapaci
tated with stochastic lead times was developed. Then, a combinatorial optim
ization method that took advantage of structural properties was developed t
o produce an optimal integer solution of safety lead-times. In addition to
yielding a practical lead-time setting tool, the model led to some interest
ing policy insights: (1) the flexibility achieved by bringing parts in earl
y may have value even in situations with ideal suppliers; (2) purchasing co
mponents in batches offers some of the same protection as does safety lead
time, but is less precise and subsequently more expensive; and (3) ignoring
the complicating issues of MRP procurement results in far from optimal res
ults.