A genetic algorithm for solving the quay crane scheduling and allocation problem

Abstract

International sea-freight container transportation has grown dramatically over the last years and container terminals represent nowadays a key actor in the global shipping network. The management of a container terminal is a complex process that involves many decisions. Among the problems to be solved, there are the spatial allocation of containers on the terminal yard, allocation of ships to berths and cranes, scheduling priorities and operations in order to minimize ship’s turnaround time, one of the main indicators of the terminal performance for the shipping companies. An efficient use of quay cranes is crucial, since quay cranes are highly expensive and represent one of the most scarce resources in the terminal. The quay crane allocation problem aims to efficiently assign quay cranes to vessels that must be operated over a given time horizon. The allocated cranes must be sufficient to complete the workload within the given time window, although many configurations are possible. In this paper, we will mainly focus on the quay crane scheduling problem, considering other logistic activities as given. This problem can be split into two sub-problems. First, specific quay cranes must be assigned to specific tasks. Second, a detailed schedule of the loading and unloading moves for each quay crane should be constructed. In this case, the number of quay cranes assigned to the vessel is assumed to be known in advance. The paper presents a solution for solving the quay crane scheduling and allocation problem that is based on working with modern optimization techniques such as genetic algorithms. The genetic algorithms are soft computing techniques that are used for optimization and search. Thus, the solution is obtained through a C++ built on simulation model that, over the analytical modelling of container’s terminals activity, has the main advantage of providing a greater level of detail and avoids too many simplifications. Also, three different scenarios that describe operational situations have been taken into account.