SC13 Social Simulation Research: Modeling and Visualization of Opinion Dynamics
In this series of lectures I would like to introduce main concepts of social simulation research using examples from opinion dynamics modeling. Aspects covered in the lectures include systematic computations, validation, mathematical formalization, visualization and the integration of psychological mechanisms.
A first lecture will provide an overview over agent-based modeling (ABM) and opinion dynamics in particular. It will address questions related to conceptual and empirical validation and illustrate concepts such as pattern-oriented modeling using a specific model in connection with real data. In a second lecture I would like to introduce Markov chains as a mathematical framework for the formalization of ABMs and other computational models. A particular focus will be on the transition from a microscopic to a macroscopic dynamical description. I will also introduce a series of information-theoretic measures to quantify the information loss associated to aggregation and relate it to ideas from computational emergence. A third lecture will be a more practical one. Step by step I would like to set-up a simple agent-based model (probably opinion dynamics) including systematic model analysis and endogenous network formation. I would like to share some of my working steps including a pipeline for visualizing and exploring complex interaction networks. Participants are invited to take an active part (the following tools will be used: MatLab, Pajek, Gephi). In a fourth lecture I will discuss the recent trend in social simulation research to take into account psychological traits in the agent architecture and to incorporate psychologically-founded mechanisms into interaction behavior. The order of the lectures is provisional and will be decided upon in the first lecture.
Understand what social simulation is about. Where are its strengths and weaknesses.
My hope is that participants will be able to develop and analyze their own model after attending the course.
Banisch, S., Markov Chain Aggregation for Agent-Based Models, Understanding Complex Systems, Springer, 2016. (PDF can be provided on request, preliminary version available at http://pub.uni-bielefeld.de/publication/2690117)
Castellano, C.; Fortunato, S. & Loreto, V., Statistical physics of social dynamics,
Reviews of Modern Physics, APS, 2009, 81, 591-646
Sven Banisch studied media system science at the Bauhaus University of Weimar, the University of Girona and the Lisbon School of Economics and Management. In 2014 he obtained his PhD in physics from the Bielefeld University (Germany) in an interdisciplinary project at the interface of linguistics and physics. Since 2014 he has been postdoctoral fellow at the Max Planck Institute for Mathematics in the Sciences (Leipzig) continuing his work on aggregation and the micro-macro link in agent-based models. His current interests include the socio-theoretical and mathematical formalization of computational models, application of opinion dynamics modeling to political data, computational operationalization of interpersonal behavior.Website