A simulation model is developed to evaluate the performance of a flexible manufacturing system with respect to time in system. The new rule to assign parts to machine-tools we propose, OOM (Only One Machine), designed to minimize parts movements performs poorer than WINQ (Work In Queue), a rule directed at balancing machine workload. Different numbers of automated guided vehicles (AGV) produce significantly different results with the best performance resulting with five AGVs. Three AGVs are too few to handle the transportation requirements, whereas seven may, to some extent, increase AGV blockage. The number of parts that can be entirely processed on one single machine is found to impact performance, but the impact is not consistent across the experimental conditions. Three rules to sequence parts to be processed are found to have a moderate impact when OOM assignment is employed, but have no impact under the WINQ assignment rule.