Performance evaluation of container handling in a transshipment hub

Queueing networks are often adopted as conceptual paradigms to support decision making in container terminals. Modeling container storage/retrieval operations on the yard of a transshipment hub is no exception to the rule. Here a detailed and tuned discrete-event simulation model versus a simplified and approximate analytical approach is investigated. Using both real data and company operating practices requires particular care when modeling human-operated vehicles that circle between the quay and yard for container handling. Numerical results show that vehicle-centric closed queueing networks should be preferred to decomposed networks under Poisson-type renewal flows. The finite source effects of the the finite fleet of vehicles cannot be disregarded. The take-away for queueing theoreticians is the need for renewed efforts on closed models, while simulationists should shift towards digital-twins to model complex subsystems such as the yard area.