Combining simulation and multi-objective optimisation for equipment quantity optimisation in container terminals

  • Zhougeng Lin

Student thesis: Doctoral Thesis


This thesis proposes a combination framework to integrate simulation and multi-objective optimisation (MOO) for container terminal equipment optimisation. It addresses how the strengths of simulation and multi-objective optimisation can be integrated to find high quality solutions for multiple objectives with low computational cost. Three structures for the combination framework are proposed respectively: pre-MOO structure, integrated MOO structure and post-MOO structure. The applications of the three structures under the combination framework for following three problems are discussed in the thesis: internal truck quantity optimisation based on post-MOO structure, multiple equipment quantity optimisation based on post-MOO structure and multiple equipment quantity optimisation based on integrated MOO structure.

The truck quantity optimisation problem in modern container terminals, which aims to improve operational efficiency and reduce cost, is discussed in the thesis. This is a multi-objective problem because multiple factors need to be considered in order to guarantee owner's service quality and profitability. A simulation model and a multi-objective optimisation model are built under the combination framework. According to the combination framework and structures proposed, the "Data Processing" is defined as data fitting which generates a set of fitting coefficients and base functions. Solutions provide a series of choices for container terminal operators.

As a further study based on the truck quantity optimisation problem, a multiple equipment (including trucks) quantity optimisation problem is raised. The problem is discussed and a series of optimisation models based on post-MOO structure for multiple equipment deployment are built for the container terminal daily decision making in the consideration of multiple variables and objectives. Simulation and multi-objective optimisation are combined to build integrated optimisation models under the combination framework. The problem is solved by a genetic algorithm.

Based on the multiple equipment quantity optimisation problem raised above, another combination structure, namely MOO leading integrated structure, is employed to solve the same problem in order to find good enough solutions with less computational cost. The "Data Processing" in the combination framework is defined as data fitting and the "Searching Techniques" is defined as dynamic MOO search. The data fitting generates a set of fitting coefficients and base functions and the dynamic MOO search is a technique to explore the next searching positions based on the Pareto solutions. The results demonstrate that the integrated MOO structure finds better or close to best solutions comparing to the post-MOO structure and the computational cost is likely to be less.
Date of AwardJan 2013
Original languageEnglish
Awarding Institution
  • University of Portsmouth
SupervisorDylan Jones (Supervisor) & Xiang Song (Supervisor)

Cite this