Simulating Container Terminal Performances: Microscopic vs. Macroscopic Modelling Approaches

Most of the existing contributions for container terminal analysis follow approaches based on optimization models, that are mainly useful to support strategic decisions about terminal container configuration. Many others contributions follow the simulation approach, which allows a detailed analysis but may lead to computational problems and is rather computer demanding, especially when resulting models are used to support optimization. This paper investigated the prediction reliability of two different approaches to container terminal simulation: microscopic and macroscopic. The former simulating single container movement, the latter simulating container flows movement. The microscopic model was a discrete event simulation model, the macroscopic model was a dynamic discrete time based (space-time) network assignment model. Both modelling approaches were implemented and compared taking advantage of some significant investment made by the Salerno Container Terminal (Italy) between 2005 and 2011. In particular, disaggregate (microscopic) and an aggregate (macroscopic) simulation models implemented in 2005 were validated with a large set of data acquired in 2011 after some structural and functional terminal modifications. Through this analysis it was possible to analyze the prediction reliability of both simulation approaches and to draw some operational guidelines.