Zusammenfassung für den Vortrag am 12.01.2009 (14:00 Uhr)Complex Networks Seminar
Claudius Gros (Institut für Theoretische Physik, Universität Frankfurt)
Cognition-inspired graph theory: Vertex routing models and recursive clique graphs
We report work in progress on a new class of vertex routing models and on recursive clique graphs, which are both motivated by our long-term research on cognitive system theory. Information arriving to a vertex B arriving from a vertex A is routed to vertex C in a vertex routing model. We define a class of random vertex routing models with and without a memory trace and the notion of information centrality based on the long-term information flow in terms of cyclic attractors. We find that the information centrality is extensive/sub-extensive for models with/without a memory trace. A clique is a maximal fully connected subgraph and a clique graph is the meta-network of cliques. We study numerically the possible occurrence of a phase-transition in recursive clique graphs with respect to their growth properties, starting from Erdös-Renyi graphs, as a function of the initial connectivity.