Within the past 50 years, containerization and globalization have driven a change from small container terminals to large container terminals that need efficient logistic models to keep up with the significant growth in container traffic. Efficiently managing rubber-tired gantry cranes and planning container placement within the terminal are two ways to increase the overall efficiency of a terminal. In this thesis, we combine these strategies in a real-time dispatching tool using an approximate dynamic programming heuristic. The heuristic re-optimizes at the rate the quay crane handles containers, incorporating endogenous and exogenous information in each solution. We formulated and solved an Integer Linear Program (ILP) to estimate the heuristic's solution quality. The heuristic finds solutions within seconds and the absolute gap between the heuristic solution and the ILP solutions remained essentially constant as the size of the problem increased.
展开▼
