DYNAMIC DECISION MAKING IN AN AUTOMATED CONTAINER YARD BLOCK VIA OPTIMIZATION-EMBEDDED DISCRETE-EVENT-SIMULATION

Abstract

This thesis introduces an optimization-embedded discrete-event simulation framework for dynamic decision making in an automated container terminal's yard block. This work aims at addressing two major problems in yard block operations, namely the Container Storage Location Assignment Problem and the Yard Crane Job Scheduling Problem. The first is concerned with the allocation of storage space for inbound containers, while the second aims at finding the optimal sequencing of the yard crane's storage and retrieval requests. An innovative top-down approach to a hybrid modelling formalism is adopted to provide a discrete-event simulation model of yard crane operations. This model is used to compare operational strategies addressing the storage allocation and job scheduling problems. In particular, an optimization model is embedded in the simulation model to perform dynamic scheduling. A rolling horizon heuristic is used reduce computing time. Finally, insights on the relative performance of operational strategies and optimization-embedded simulation are given.