Distributed Quay Crane Scheduling with Overlapping Constraints A distributed constraint optimization approach

Abstract-For sea port container terminals, a key objective is to increase the container throughput by minimizing the amount of time necessary to load into and discharge containers from a ship using quay cranes (QCs). In this paper we discuss the situation in which some tasks can be handled by multiple QCs, represented by so-called overlapping area constraints. Overlapping area constraints determing the tasks that more than one QC could take care of. We formulate a distributed QC scheduling problem with overlapping area constraints and cast this problem as a Distributed Constraint Optimization Problem (DCOP). A new negotiation algorithm called Extended Asynchronous BackTracking (E-ABT) is then proposed for solving the DCOP. Keywords-Quay Crane Scheduling Problem; Distributed Quay Crane Scheduling; Distributed Constraint Optimization Programming; Extended Asynchronous BackTracking I. INTRODUCTION Every year millions of TEU (Twenty-foot Equivalent Unit container) are handled by container terminals. Maritime container terminals all over the world are the most important part for transshipment and intermodal container transfers. By the end of 2013 ahnost 170 million TEU are expected to be handled [I]. Today container ships still become larger and can already carry up to 15000 TEUs [2]. To cope with the rapid growing traffic of international trade, the most important problem will be the reduction of the amount of dwell time and associated transaction cost. The way to do this is to ensure that load/unload processes of containers to/from ship are done as quickly as possible. In this paper we focus on improving the scheduling of the tasks of quay cranes (QCs). In particular, we consider how performance could be improved ifQCs would have overlapping task areas; i.e., if multiple quay cranes would be able to handle particular containers, instead of the situation in which only a single QC can handle a particular container. This will require a new approach to scheduling, in which quay cranes themselves can decide, in coordination with one another, which container to handle next. We recommend that in order to obtain better performance of loading/discharging containers by QCs, a different type of stowage plan is adopted. Nowadays, groups of containers with the same destination are typically considered in a same ship's bay [3]. This limits the number