Transfer batch and buffer management are two key elements in successfully implementing the Drum-Buffer-Rope technique, which is a well-known shop-floor control method based on the theory of constraints. A production planner often encounters difficulties in determining the transfer-batch size and the schedule for entrance stations. In this study a systematic framework is developed to search the sizes of transfer batches and the sequencing of the bottleneck and gate stations. A genetic algorithm is employed to search for the transfer-batch sizes and sequences of the sublots on the constraint resources. Then, the backward method is used to schedule the non-constraint resources preceding the constraint, and the forward method is used to schedule the non-constraint resources succeeding the constraint. The simulation package Arena is used to accomplish the detailed schedule. The proposed approach is compared with three other methods that employ the forward method. The effect of equal and non-equal transfer batches and the effect of the location of the sub-constraint relative to the constraint are also observed. The results show that the proposed framework can achieve significant improvement with regard to the objectives of makespan and total-flow time.