Design and performance analysis of a heavily loaded material handling system

We consider the problem of designing a 2-stations Automated GuidedVehicle System (AGVS). The AGV System consists of a pool of vehicles that transports products from one station to the other station through an unidirectional guidepath. We seek a model to estimate the minimal number of vehicles needed to guarantee some target service level expressed in terms of mean waiting time. To achieve this we also need to design dispatching rules in order to utilize the vehicles in the best way. Our solution procedure starts by computing the necessary fill rate in order to respect the maximum mean waiting time. We use a Reorder Point Inventory Policy and Markov Chain Theory to determine the dispatching rules and to estimate the minimum number of vehicles required to guarantee the fill rate. Our computer simulationsindicate that the model offers a good degree of approximation.