A genetic algorithm heuristic for solving the quay crane scheduling problem with time windows

Abstract

One of the most important operations in marine container terminals is quay crane scheduling. The quay crane scheduling problem (QCSP) involves scheduling groups of containers to be loaded and unloaded by each quay crane. It also requires addressing practical issues such as minimum spacing between quay cranes and precedence relationships between container groups. This study addresses the QCSP with one additional consideration: time availability of quay cranes. This problem is referred to as QCSP with time windows (QCSPTW) in the literature. This article discusses the genetic algorithm (GA) developed to solve the QCSPTW. It builds on a previously developed GA to solve the QCSP by the authors. The results of a large set of numerical experiments using benchmark instances highlight several key characteristics of the proposed solution approach: (i) the developed GA can provide near optimal solutions in a faster time for medium and large-sized instances (overall average gap is less than 3 per cent), and (ii) the developed GA leads to an improvement in the solution quality (lower vessel turnaround time) for instances with fragmented time windows (time windows that are broken up into two or more non-contiguous segments).