Period Batch Control is a production planning system that decomposes the cellular manufacturing system into N stages and gives each stage the same amount of time P to complete the required operations. Applying the PBC system in combination with cellular manufacturing is said to result in short and reliable throughput times, and low inventory costs. However, the design of the PBC system, i.e. the size of N and P, does not follow directly from the cellular organization. We show that varying the values of both the number of stages N and the period length P greatly influences the performance of PBC. Furthermore, if both N and P have been determined, system performance is still very sensitive for different allocations of operations to the stages. We present a mathematical programming model that determines an optimal allocation of operations to stages, using a longest path orientation accompanied by a bottleneck orientation.