This paper presents a model of a bottleneck facility that performs two
distinct types of operations: ''regular'' and ''rework.'' Each job is
subjected to a test after completing the regular operation at the bot
tleneck. If the job passes the test, then it continues its process dow
nstream. Otherwise, the job will cycle back to the bottleneck stage fo
r rework operation. Upon the completion of a batch of regular jobs, Me
decision maker observes the amount of rework and decides on whether t
o switch over to process the reworks or continue to process another ba
tch of regular jobs. It is assumed that both switch-over time and cost
are incurred when the facility switches from performing one type of o
peration to a different type. The goal of the analysis is to character
ize Me optimal operating policy for the bottleneck so that the average
operating cost is minimized. In order to characterize the optimal ope
rating policy, we first formulate the problem as a semi-Markov decisio
n process. Then we show that there exists an optimal ''threshold'' ope
rating policy that can be described as follows: upon completion of a b
atch of regular jobs, switch over to process the reworks only if the n
umber of reworks exceeds a critical value. In addition, we develop a s
imple procedure to compute the critical value that specifies the optim
al threshold policy. Moreover, we evaluate the impact of batch sizes,
yield, and switch-over time on the optimal threshold policy.