Container terminal optimization: Pareto front optimization for maximum port area utilization
Typ
Examensarbete för masterexamen
Master's Thesis
Master's Thesis
Program
Övrigt, MSc
Publicerad
2024
Författare
Hauptmann, Johannes
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
Large quantities of containers are moved in ports around the world every day,
motivating a focus on transport efficiency to optimize energy consumption, environmental impact, the cost of moving goods, utilization of scarce areas, lead times,
and customer satisfaction. The significant investment in container terminals (CTs)
and their equipment necessitates a high equipment occupancy rate as well. From a
sustainability perspective, port operation is not only about energy consumption but
also the equipment occupancy parameter, since the amount of equipment needed is
proportional to the material used to build it.
Recently built or upgraded CTs are automated, advanced container terminals
(ACTs). This work develops a model of a basic ACT, and expand it into a complex layout that is fast enough to compute yet realistic. The goal was to perform
multi-objective optimization on mooring time, equipment occupancy, and port land
utilization based on such a model. In a typical ACT process, quay cranes lift containers from ships at berth and load automated guided vehicles (AGVs), which transport
the containers to next to automated rail-mounted gantry cranes (ARMGs) at the
yard, where they lift the containers and stack them in areas for further distribution
outside the container terminal.
This research leverages discrete event simulation (DES) and multi-objective optimization using genetic algorithms (GAs). This approach enables the evaluation of
various alternatives and represents the stochastic behavior at the operational level.
A simulation model has been developed using the Salabim package in Python as
a base to evaluate the performance and optimize the operation of the ACT from
quay to container stack. The basic model can be simulated on a regular computer,
including the optimization, which is based on the Distributed Evolutionary Algorithms in Python (DEAP) library and its Non-dominated Sorting Genetic Algorithm
II (NSGA-II) implementation. Based on the initially basic model, the scope gradually expand to cover multiple piers and a hinterland central hub, optimizing further
aspects such as addressing the scarce land challenge in the pier areas.
It is recommended that container logistics stakeholders partner up and plan for
optimization and enhancement of the complete container management ecosystem.
This work lowers the threshold to expand the scope and guides in what direction
to improve. The model is built to be scalable, and more local conditions can be
added. With additions to the model, such as bi-directional container movement, it
should be useful to implement for real cases as a base for decision-making at new
development or enhancements of container ports.
Beskrivning
Ämne/nyckelord
Multi-objective Container Terminal Optimization , Port Discrete Simulation , Genetic Algorithm Utilization