| Summary: | 碩士 === 國立臺灣海洋大學 === 航運管理學系 === 100 === Abstract
Container loading problems are important for sea transportation, air transportation and manufacturing. The way adopted to load cargoes into a container has direct impacts on the freight of a supply chain and cost calculation. In order to maximize profit, how to load cargoes into a container by minimizing the unused capacity is worthy of a study.
In this thesis, we study a container loading problem with multiple types of cargoes. The objective of this thesis is solving the container loading problem within a reasonable time by minimizing the unused capacity.
This thesis presents a genetic algorithm for a three dimensional container loading problem. The algorithm can be divided into two stages, construction and improvement stages. In the construction stage, a wall building method is used to calculate the unused space and to decide the loading priority of cargoes and stopping criterion. We keep building a layer within the current wall until no more space available. In order to mi-nimize the wasted space, we modified the wall building method by incorporating the concept of layer determine box. It is achieved by setting the width of each layer equal to layer determine box’s width.
In improvement stage, the genetic algorithm is used to load the container by taking the results of the construction stage as an input. Based on chromosome coding, we gen-erate the initial population and there calculate the fitness for further replication, cros-sover, and mutation. 15 test questions from Operations Research Library and a real world data are tested. Computational results show that our algorithm provides and ac-curate and efficient method for solving the container loading problem.
Keywords: container loading problem; genetic algorithm; heuristic algorithm
|
|---|
