## Visualisation of State Transition Graphs (2000)

Citations: | 7 - 0 self |

### BibTeX

@MISC{Ham00visualisationof,

author = {Frank van Ham and Advisor M. A. Reniers},

title = {Visualisation of State Transition Graphs},

year = {2000}

}

### OpenURL

### Abstract

Although graph drawing and graph visualisation have often been the subject of research, most of this research focuses on creating tidier drawings of relatively small graphs. However, in todays complex society, research into the visualisation of large graphs is becoming increasingly important. One can think of maps of the entire internet or call graphs for very large software programs, for example. This thesis presents a visualisation method for a class of large graphs called state transition graphs. State transition graphs can be used as a mathematical description for processes or programs, and their analysis gives much insight into the process or program they describe. The major problem in the visualisation of large graphs is the size of the graph. Displaying an entire large graph all at once may overload the user with information, making it more difficult to comprehend. We attempt to overcome this problem by grouping nodes based on a local structural property. This reduces the complexity of the original graph and by initially displaying groups of nodes instead of individual nodes, the overall structure of the graph is revealed to the user. Further, instead of minimising the number of line crossings in a graph, as is often done in the visualisation of small graphs, we focus our attention on a different, global aesthetic aspect of a graph, namely symmetry. A clear picture of the symmetry in a graph allows the user to mentally split it into a number of smaller parts. Another advantage is that only one of the symmetrical parts has to be analysed, considerably reducing the amount of work necessary. After a description of the concepts behind this visualisation method, we present results obtained from working with a prototype. Finally, we discuss the advantages and disadvantages of this type of visualisation.

### Citations

2152 |
Algorithms for Clustering Data
- Jain, Dubes
- 1988
(Show Context)
Citation Context ...ph looks like when confronted with thousands of information elements. A technique to drastically reduce visual complexity is the grouping of nodes based on a common property, also known as clustering =-=[8]-=-. Clustering based on the structure of the graph is especially useful here, since the structure of the resulting clusters generally resembles the overall structure of the original graph, which facilit... |

380 | The factors of graphs
- Tutte
- 1952
(Show Context)
Citation Context ...ocal structural node properties. We use a two-step heuristic to position the nodes. First, we assign initial positions, based on the positions of nodes in ancestor and descendant clusters, similar to =-=[12]-=-. In a second step we adjust these initial positions of nodes to increase the space between them. The following describes these two steps in more detail. Step 1: The position of a node is based on the... |

343 |
Method for visual understanding of hierarchical system structures
- Sugiyama, Tagawa, et al.
- 1981
(Show Context)
Citation Context ...ucture using a method related to cone trees; 4. Position individual nodes and edges. 3.1 Ranking In a first step nodes are assigned to a layer or rank, comparable to the standard Sugiyama-type layout =-=[11]-=-. That is, all nodes are assigned a positive discrete value that indicates their depth relative to the start node. We found two ranking methods the most useful, since they correspond with two types of... |

201 | Process Algebra - Baeten, Weijland - 1990 |

138 | Tidier drawing of trees - Reingold, Tilford - 1981 |

130 | Which aesthetic has the greatest effect on human understanding - Purchase - 1997 |

119 |
Finite Transition Systems
- Arnold
- 1994
(Show Context)
Citation Context ... role in science with applications ranging from biology to electronic engineering. This paper focuses on an application in computer science, where graphs are applied to describe Finite State Machines =-=[1]-=-. The concept of a finite state machine plays a central role in computer science, as it can be used to model processes. Subsequent study of these models can then reveal useful information about the pr... |

39 |
Cone Trees: Animated 3D
- Robertson, Mackinlay, et al.
- 1991
(Show Context)
Citation Context ...hat depicts clusters as circles in a horizontal plane. A plane is reserved for each rank, with the topmost plane containing clusters with rank 0.Thebackbonetreeislaidoutinamannerresembling cone trees =-=[10]-=-, with ancestor clusters positioned at the apex of an imaginary cone and their descendant clusters placed at the base of the cone. In other words, our clusters have the same status as tree nodes in a ... |

35 | Graphs: An Introductory Approach - Wilson, J - 1990 |

29 | Graph visualisation and navigation in information visualisation - Herman, Melançon, et al. - 1999 |

28 |
Graph Drawing: Algorithms for the Visualisation of Graphs
- Battista, Eades, et al.
- 1999
(Show Context)
Citation Context ...ements. Fig 1. 2D Visualization of a state transition graph with 191 states 2. Approach Much research has been done in the area of two dimensional (2D) graph drawing. For an overview see for instance =-=[2]-=-. The general approach for 2D visualizations of directed graphs is to produce a pleasant looking picture by optimizing one or two aspects of the visualization. In the case of larger graphs however, th... |

22 |
Visualizing the structure of the World Wide Web
- Munzner, Burchard
- 1995
(Show Context)
Citation Context ...tiveness when dealing with a large number of nodes. Several attempts to overcome these problems have been made by using 3D visualizations and by applying different techniques such as hyperbolic space =-=[9]-=-, hierarchical node clustering [13],[4] or self organizingsnetworks [7]. Since finding a general solution for visualizing large directed graphs can be difficult, most existing methods are forced to us... |

21 |
Research report: Interacting with huge hierarchies: beyond cone trees
- Carriere, Kazman
- 1995
(Show Context)
Citation Context .... • N > 1: We space the clusters evenly over the base of a cone with its apex at the center of A. The base diameter of this cone can be computed by using a recursive method similar to the one used by =-=[3]-=-. However, since positioning all N descendant clusters over the base may not always be a symmetrical solution we make the following three exceptions: • If there is a unique largest cluster among the d... |

8 | A balancing act: analyzing a distributed lift system in
- Groote, Pang, et al.
- 2001
(Show Context)
Citation Context ...the human perceptual system with respect to pattern recognition are indispensable. Figure 11 shows a graph consisting of 2860 nodes, depicting the communication protocol for a modular car lift system =-=[5]-=- where three hydraulic jacks support one platform. If we compare the visualization of this graph to the one depicting the same protocol for two jacks we can expect to find some similarities, since bot... |

2 |
Visualization of State Transition Graphs. Masters thesis available from http://www.win.tue.nl/˜vanwijk
- Ham
- 2000
(Show Context)
Citation Context ...e two nodes x and y to be equivalent iff a row (x=z1,z2,...,zN=y) of nodes with equal rank exists, such that for all 1 ≤ i<N D(zi ) ∩ D( zi+1 ) is not empty (see Fig 3). It can be fairly easily shown =-=[6]-=- that this is an equivalence relation, so by definition its equivalence classes are non-empty and disjoint, which makes them very suitable to use as clusters. Since all nodes in a cluster have the sam... |

1 |
Multilevel Visualization of Clustered Graphs”, Graph Drawing 96
- Eades, Feng
- 1996
(Show Context)
Citation Context ...er of nodes. Several attempts to overcome these problems have been made by using 3D visualizations and by applying different techniques such as hyperbolic space [9], hierarchical node clustering [13],=-=[4]-=- or self organizingsnetworks [7]. Since finding a general solution for visualizing large directed graphs can be difficult, most existing methods are forced to use domain-specific information. In this ... |

1 |
et al., "Narcissus: Visualising Information
- Hendley
- 1995
(Show Context)
Citation Context ... overcome these problems have been made by using 3D visualizations and by applying different techniques such as hyperbolic space [9], hierarchical node clustering [13],[4] or self organizingsnetworks =-=[7]-=-. Since finding a general solution for visualizing large directed graphs can be difficult, most existing methods are forced to use domain-specific information. In this case we focus on large state tra... |

1 |
et al., “Layout for visualizing Large
- Ware
(Show Context)
Citation Context ... number of nodes. Several attempts to overcome these problems have been made by using 3D visualizations and by applying different techniques such as hyperbolic space [9], hierarchical node clustering =-=[13]-=-,[4] or self organizingsnetworks [7]. Since finding a general solution for visualizing large directed graphs can be difficult, most existing methods are forced to use domain-specific information. In t... |