2006
DOI: 10.1007/11618058_22
|View full text |Cite
|
Sign up to set email alerts
|

An Experimental Comparison of Fast Algorithms for Drawing General Large Graphs

Abstract: In the last decade several algorithms that generate straightline drawings of general large graphs have been invented. In this paper we investigate some of these methods that are based on force-directed or algebraic approaches in terms of running time and drawing quality on a big variety of artificial and real-world graphs. Our experiments indicate that there exist significant differences in drawing qualities and running times depending on the classes of tested graphs and algorithms.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

1
67
0
2

Year Published

2009
2009
2016
2016

Publication Types

Select...
7
1
1

Relationship

0
9

Authors

Journals

citations
Cited by 71 publications
(73 citation statements)
references
References 20 publications
1
67
0
2
Order By: Relevance
“…Even though small graphs are im-possible to embed in Euclidean space such that the Euclidean distance between vertices is proportional to the graph distance (that would be ideal for a direct correspondence between the real graph-the terrain-and the visualization-the map), still one can make visualizations that capture much of the network structure. Typically they organize the vertices such that there is a positive correlation between the graph and Euclidean distances [51]. This is enough to see differences between networks of different degree-degree correlations [112] or to identify dense clusters by the eye [42].…”
Section: Future Outlookmentioning
confidence: 99%
“…Even though small graphs are im-possible to embed in Euclidean space such that the Euclidean distance between vertices is proportional to the graph distance (that would be ideal for a direct correspondence between the real graph-the terrain-and the visualization-the map), still one can make visualizations that capture much of the network structure. Typically they organize the vertices such that there is a positive correlation between the graph and Euclidean distances [51]. This is enough to see differences between networks of different degree-degree correlations [112] or to identify dense clusters by the eye [42].…”
Section: Future Outlookmentioning
confidence: 99%
“…We measured the CPU running times for distance computation and the layout algorithm. Descriptions of the test graphs are given in [14,15]. Figure 2 shows for both Pivot and Landmark MDS that the running time for the breadth-first searches for matrix C in O(km) time indeed dominates over the computation times for spectral decomposition of C T C and the final coordinates, which together are in O(k 3 + k 2 n) time.…”
Section: Discussionmentioning
confidence: 99%
“…This suggests the use of our method for efficient generation of initial placements for further processing, which is crucial for many algorithms. In the experimental study of [14] these graphs posed serious difficulties for most methods.…”
Section: Qualitymentioning
confidence: 99%
“…Các giải thuật phổ biến nhất đều dựa trên giải pháp tương đối đơn giản đó là mô hình lực có hướng (forced-directed model) [5,6] và cho kết quả tốt (ít sự giao cắt giữa các cung, cấu trúc cân đối) đối với các đồ thị có kích thước nhỏ (vài trăm đỉnh). Một số giải thuật khác [7,8] dựa trên quá trình tính toán nhiều pha đã chứng tỏ khả năng thích ứng với các đồ thị có kích thước lớn (vài nghìn đỉnh). Các giải thuật này khá thành công trong việc hiển thị cấu trúc nhóm của đồ thị khi mà các nhóm này được sinh ra một cách "tự nhiên" dựa trên cấu trúc nội tại của đồ thị.…”
Section: Giới Thiệuunclassified