期刊刊名：明道學術論壇 　 卷期：6卷1期

篇名出版日期：2010年3月1日

作者：柯志忠,楊志隆,Chih-Chung Ke, Chih-Lung Yang

語言：English

關鍵字：regular graph,vertex,trail,degree,正則圖,頂點,追蹤,度數

被點閱次數：4次

閱讀時間：1414sec

摘要：
How to generate regular graphs effectively has become an important issue in different applications regarding spatial relationship description, such as the construction of a neighborhood system for comparing similarities among blocks of pixels in image processing. Usually, this was developed by graph theorists who concentrated on generating regular graphs uniformly at random, with a relatively high probability of construction failure. In this study, three efcient algorithms, Vertex Scan, Random Trail, and Structural Recursion, are proposed to generate regular graphs. The performance of each algorithm depends on the number of vertices and the degree of each vertex. In addition to comparative analysis, computer simulations are also performed to evaluate these algorithms in terms of probability of success. A conclusion is drawn about how to choose the best algorithm under various scenarios.

　　　　　[全文下載]