Directed Research Project Details
Random Networks: Their Generation and Evaluation
| Author: | Adnan M Salman |
|---|---|
| Date: | March 07, 2005 |
| Time: | 10:00 |
| Location: | 220 Deschutes |
| Committee: | Art Farley (Chair) Andrzej Proskurowski Virginia Lo |
Abstract
In the last few years, with the increase in computational power, several statistical properties of large-scale networks have been studied. It has been emphasized that the topological structure of a network has a major impact on many of its properties. In this project we considered several algorithms to generate random simple graphs with some properties. The evolution of these graphs was studied. The properties of these graphs were then compared to each other and with available theoretical model. We mainly studied four types of random graphs that have received much attention in the literature. The classical random graph played the role of a reference to compare the properties of other topologies, and the graph that has been used most frequently in the literature. The power law graph captures many real complex networks such as the Internet and the Web. The regular random graph has a role in theoretical computer science. In this project we implemented a topology generator given the degree distribution of the nodes. In addition, to a topology generator based on the growth and preferential attachment principles. We studied numerically the evolution of these networks. We then compared the properties of these networks with the available graph-theoretic on one side and with other kinds of graphs on the other side. To evaluate the topologies we considered several metrics includes, clustering coefficient, average path length, betweenness, neighborhood function and the resilience of the network under node removal.