Flexible manufacturing systems (FMS) are generally set up to process a wide variety of parts with low to medium volume demand. The loading problem in FMS is a short-term decision issue, which addresses the simultaneous processing of a set of different parts in an efficient manner on capital-intensive resources. This study focuses on cyclic schedules in loading pallets carrying groups of identical parts to CNC machining centers. A predefined stream of pallets fixtured for different part types is repeatedly supplied to every CNC machine in cycles. Moreover, a limited number of pallets have to be shared among the cycles of machines to keep utilization and the overall throughput rate at acceptably high levels. The solution approach is an integrated mix of optimization and simulation methodologies. The validation of the model is based on data from past 6 months of operation. Comparisons are made retrospectively. (C) 2002 Elsevier Science Ltd. All rights reserved.