Comparitive Study Experiment Of Graph Drawing Algorithms Based On Spring Model

Network data becomes more and more with the advent of the information age. People's ability of collecting data is far stronger than the ability of understanding and analyzing data. The problem how to analyse the network data collected conveniently and rapidly is to be handled difficultly. Scholars gradually pay more attention to it.Graph visualization technology transforms the network data and the relationship of the network data into graph on screen. It is convenient for people to watch the network data. So people can easily understand more clearly implied information of network data. The important problem is how to display clearly the relationship between each node on screen. So graph layout algorithm is the core method of graph visualization.As a result of advantage and disadvantage of every graph layout drawing algorithm, the problem how to present the ideal layout is difficult. People should use appropriate algorithm to present the ideal layout.Also, People should use relevent graph algorithm according to different network data and different layout requirements. The LinLog algorithm of T.M.J. Fruchterman and E.M.Reingold, DH algorithm of Davidson and David Hareland FR algorithm of T.M.J.Fruchterman and E.M.Reingold based on the mechanical models are used large amounts of data to test them. This paper does comparative study of the three graph drawing algorithms.Those major works of this paper are as follows.(1) This paper introducing the basic concept of graph and graph layout also introduces many common aesthetic criterias about graph layout. This paper also descripts detailedly several graph drawing algorithms based on the spring model.(2) This paper introduces the strategy of data set selection. It use symmetrical graph data collection selected to test the symmetry effecttion of graph.It uses randomly generation data collection to test area,edge cross, the ratio of the longest edge and the shortest edge, the longest edge and algorithm efficiency.Also,clustering feature is tested in three algorithms with the real complex network data collection and computer-generated network.(3) FR algorithm, LinLog algorithm, DH algorithm are implemented by Java language.(4) The three algorithms are compared in aesthetic criterias, clustering features and efficiency detailedly. The experimental result is analyzed in advantages and disadvantages of three algorithms.