Managing resource allocation, scheduling and simulation for an intermodal container terminal

An intermodal container terminal is a complex system that manages the process of container loading and unloading exchanging transport modes. In particular, an intermodal sea-port moves containers from the yard to the ships and vice versa. A set of resources, such as yard and quay cranes, and straddle carriers, direct the flows of containers. Minimizing the resources needed to serve the workload is the most critical task in a port, because resources determine the terminal costs. It is also a difficult task, partly because it must be addressed for a relatively long time horizon. Here, the resource allocation problem is formulated as the problem of optimally designing a network of flow. Approximate solutions are obtained by formalizing the problem as a mixed-integer linear program, and by partially searching the solutions space by the method of branch and bound. This method was applied to a set of real world instances from the La Spezia Container Terminal, in Italy. On average, the solution was 6.6% far from the continuous relaxation bound; moreover, the cost of the computed allocation plans was 67% of the costs originally planned by the terminal managers. The actual sustainability of the computed plans was verified by designing and implementing both a scheduling algorithm, that optimizes the ships' loading/unloading list, and a detailed discrete-event simulator of the terminal. The test system executes a loop: the computed allocations are passed to the scheduler that produces the list of container moves; on this basis, the simulation model updates the terminal state that is then returned to the resource allocation module. The system was able to serve the real workload for one week, respecting all the ships' deadlines, thus validating the proposed approach.