Yannik Schädler, Universität Bremen, Institut für Theoretische Physik, Complex Systems, AG Prof. Dr. Stefan Bornholdt
Opinion- and consensus formation in society is an interesting dynamical process, with fashion and paradigm bubbles as an intrinsic feature. Sociophysics models that concentrate on that particular feature have been proposed in recent years , . We here study a variant of the paradigms model and study its dynamics on several network topologies. It contains agents with memory that can interact with neighbours, as well as an innovation and a group-pressure mechanism that generates herding-effects. So far this model has been studied mainly on grid-topologies. The motivation here is to include more realistic social network architectures, and to study the influence of topology in this particular model. We study innovation waiting times or innovation rates in different settings. We find that the network topology indeed changes the dynamics: Simulations show that the network structure strongly affects the likelihood of a consensus.
 Katarzyna Sznajd-Weron: Sznajd model and its applications , Acta Physica Polonica B, vol.36, no. 8 (2005)
 Bornholdt, S. and Jensen, M. H. and Sneppen, K.: Emergence and Decline of Scientific Paradigms, Phys. Rev. Lett. 106, 058701 (2011)
Patricia Zauchner, SOCIUM, Universität Bremen, AG PD Dr. Tanja Pritzlaff-Scheele
Studies showed that in purely aggregative forms of democratic decision making the systematic consideration of all relevant criteria can have harmful effects on citizens’ voting behaviour. However, deliberative forms of decision making have other requirements than purely aggregative ones; what can be harmful to one decision making process could be advantageous for another. Hence, in this presentation, I will present the results of a computer laboratory experiment that examined whether systematic processing can actually improve the quality of a discourse.
Stefan Hillmich, Universität Bremen, Informatik, AG Prof. Dr. Rolf Drechsler
The slime mold Physarum polycephalum is a current focus area of research. Many recent experiments emphasize network formation, for example by investigating growth in networks seeded from small fragments of the slime mold.
Researching the growth of a network is commonly apprehended by taking images of the network at certain intervals and analyzing them afterwards. In this work, a more encompassing graph-based model extending the work of the Institute for Biophysics is presented as well as a way of extracting this model.
This is done by transforming the original images to binary images. The area covered by the network is approximated by filling it with circles. These circles are interpreted as nodes in an undirected graph and edges are added between neighboring nodes. This results in one graph per image. To track changes in the network, referred to as events, nodes representing the same part of the slime mold in different images are determined and collected in so-called temporal clusters.
A low temporal resolution can cause many events to happen between two images. This may yield contradicting results if a single node is affected by multiple events at once. A mitigation is introduced by virtual events as a means of breaking down large events into smaller ones.