Simulation-optimization for the management of the transshipment operations at maritime container terminals

Abstract

Maritime container terminals are complex infrastructures designed specifically to handle a large number of containers, and which play a relevant role in international freight transport. Terminal managers must deal with a wide variety of interrelated logistic problems, and the effectiveness and productivity of the terminal depends on their solution. Management strategies are therefore necessary to increase their ef- fectiveness and productivity, and thereby reducing the costs of these operations. This task is complicated by imprecise data and the need to satisfy several criteria, many of which are subjective, when evaluating the solutions. One of these logistic problems, the quay crane scheduling problem, has attracted the attention of many researchers since quay cranes are one of the most valuable resources in the port. Many proposals based on optimization algorithms have tackled this problem but the vast majority disregard the uncer- tainty inherent in this kind of systems and the impact of internal delivery vehicles. An intelligent system which integrates Artificial Intelligence techniques and simulation tools is pro- posed to aid terminal managers. The system combines an intelligent evolutionary algorithm to generate high quality schedules for the cranes with a simulation model that incorporates uncertainty and the impact of internal delivery vehicles. The joint use of these tools provides managers with enhanced infor- mation to decide on the quality and robustness of the proposed schedules, resulting in better solutions for everyday situations. Our intelligent system based on the optimization-simulation model provides clear benefits to mar- itime terminal management. The system efficiently identifies high quality schedules and can be used to evaluate its robustness. It is also flexible and can easily be adapted if other elements need to be intro- duced, which may affect the goodness of a schedule.