Parametrizing Random Topology Generating Functions

T. Kisner, F. Kaderali, and H. Hemmer (Germany)


Modelling, Random Graphs, Waxman Graphs, Connectiv ity, Simulation


Random graphs are a popular method to model the inter connections in different domains, ranging from social net works, IP networks up to peer-to-peer networks. For mod elling and simulation often different distinct graphs are re quired, which must have a minimum number of links. By choosing graph generation paramaters for random graphs resulting into a low number of links, the risk arises not to generate a connected graph, which is inapplicable for further simulations. After demonstrating the question of choosing convenient parameters for the graph generation process using random graphs as an example, we concen trate on a popular class of random graphs introduced by Waxman. We show that other investigations in literature do not comply with our constraints and propose simple manageable approximations of each characteristic for ran dom and Waxman graphs, which can be used practically to parametrize the graph generation process.

Important Links:

Go Back