Dynamic product routing in a hybrid flow shop

In a hybrid flow shop (HFS), products can often follow multiple possible routes. Routing rules are used to determine per job which route is the best. We design a routing solution where we model the HFS as a network of nodes, where every node represents an individual part of the HFS. From this network of nodes, we derive every possible route a job can take. Every route indicates which nodes the job will follow and which operation will be performed at which node. For every job, we determine its best route on the basis of three attributes: its throughput time, the processing costs and the machine utilization. With the AHP method, we can apply different weights to every attribute. We analyze if the routing rules behave as we expect and if they can outperform a set of standard routing rules.