In many industries, production systems consist of serial stages; often flexible servers are allowed to movebetween stages. Motivated by such industrial applications, we have analyzed the allocation of flexible serversin a two stage serial queuing system in order to maximize throughput. We have analyzed the system underthe assumption that the servers work collaboratively: the servers can work together on one part at the sametime. We have identified the optimal allocation policies for the servers and have shown the structure of thosepolicies is characterized by a switching curve. The exact curve depends on endogenous parameters, such asorder arrival rate and buffier size at each stage.
展开▼