Summary: | 碩士 === 國立臺灣海洋大學 === 資訊工程學系 === 99 === Abstract
In this thesis study, we propose a new evolutionary algorithm to solve a class of constrainted minimum spanning tree problems, which constrains on vertice degrees and path delays are considered. Different varieties of constrainted minimum spaning tree problems have been discussed in literature. We compare our proposed algorithm, ECMST (evolutionary algorithm for constrainted minimum spanning tree problems), with two other schemes which based on Ant Colony Optimization (ACO) and Genetic Algorithm (GA) strategies individually. Experimental results show that the proposed algorithm, ECMST, outperforms the other two approaches, especially in dense graphs.
Key Words: Constrained Minimum Spanning Tree, Evolutionary Algorithm, Genetic Algorithm, Ant Colony Optimization Algorithm.
|