Optimized routing in flexible manufacturing systems with blocking

Flexible manufacturing systems (FMSs) with limited storage capacity can be modeled as queuing networks with finite buffers and general service times. An approach is presented for improving the performance of such systems by distributing the routing probabilities across workstations in an optimal way. The proposed approach is suitable for online application in situations where analytic models are not available. Case studies have shown that reasonably good estimated optima can be obtained with very short computer run lengths in comparison with conventional simulation methods