The multiple lot-sizing problem with rigid demand and random yield frequently arises in make-to-order production systems with imperfect production processes. We focus on the multiple lot sizing problem whose yield is interrupted geometric distributed. We know that the structural properties of the cost function and optimal run size for the lot sizing problems depend on the cost structures and the yield distribution. Our purpose in this paper is to examine those structural properties of the multiple lot sizing problem with a rigid large demand, an interrupted geometric yield distribution and general cost structures. We present results characterizing the behavior of the cost function and optimal lot sizes. These results help us understand imperfect process described by the interrupted geometric distribution. Furthermore, we will show that for sufficiently large demands the optimal lots are precisely those that minimize the ratio of production cost to the expected number of good items. Keywords: Lot sizing problem, Interrupted geometric yield, Rigid demand.