Our previous work presented a Petri net-based iterative synthesis policy for deadlock prevention in flexible manufacturing systems (FMS). Given the Petri net model of an FMS prone to deadlock, it aims to synthesize a live controlled Petri net. Its use for FMS control guarantees its deadlock-free operation and high performance in terms of resource utilization and system throughput. At each iteration, a first-met bad marking is singled out from the reachability graph of the Petri net. A well-established invariant-based control method is used to prevent it from being reached. This process is carried out until the net model becomes live. The method proposed is generally applicable, easy to use, effective, and straightforward, although its off-line computation is of exponential complexity. This paper presents two improvements: (a) using the Petri net reduction approach to simplify very large Petri net models so as to alleviate computation effort; and (2) simplifying the invariant-based control method. A number of FMS deadlock problems from the literature are used to illustrate them.