Combining evolutionary and cooperative game theory in an experimental simulation platform with applications to a fixed basic income guarantee
Teaser
This project investigates the question how the disciplines of evolutionary game theory and cooperative game theory can be combined. Classical evolutionary game theory focusses on pairwise contests whereas cooperative game theory is concerned with distributing the payoff of a grand coalition among its members. An evolutionary cooperative game theory should shift the focus from pairwise interaction to the idea of “playing the field”. Our goal was to build a simulation platform for experimentation in R building on the existing R packages CoopGame [1] for cooperative game theory and EvolutionaryGames [2] for evolutionary game theory. Evolutionary cooperative game is still in its infancy, i.e. there are only very few research papers, but it promises multidisciplinary applications like e.g. discussing the pros and cons of a fixed basic income guarantee.
Short description
The team started familiarizing with various point valued solution concepts from cooperative game as well as with a few dynamics from evolutionary game theory.
During the winter semester 2019/20 we scheduled weekly virtual meetings between the Kempten team and Prof. Dr. Arend Hintze at Michigan State University (MSU) using video-conferencing equipment and software. Given that Prof. Dr. Arend Hintze moved to Dalarna University in Sweden on January 1, 2020, we decided not to involve anybody else from Michigan State University in the project.
During the semester the team very successfully developed a powerful and flexible tool for combining point valued solution concepts from CoopGame [1] with different solution dynamics. In addition, the input mask allows users to specify predefined games, like e.g. apex games, an initial distribution of player types and the number of simulation steps. The output window visualizes the distribution of different player types as well as the ratio of coalitions which actually form.
The research on the combination of cooperative and evolutionary game theory is still in a very early stage and this project laid excellent foundations for future research in this exciting field.
The Kempten team very much enjoyed the scientific work together with Prof. Dr. Arend Hintze. We recommend this kind of experience to future Kempten students and exchange students spending a semester at Kempten University alike.
The project was offered as part of the master’s degree course in computer science. Students received 5 ECTS for the project (and not 15 ECTS as for bachelor’s projects with a larger scope).
Source reference
1 J. Staudacher, J. Anwander: Using the R package CoopGame for the analysis, solution and visualization of cooperative games with transferable utility. 2019. https://cran.r-project.org/package=CoopGame.
2 D. Gebele, J. Staudacher: EvolutionaryGames: Important Concepts of Evolutionary Game Theory, 2017. https://cran.r-project.org/package=EvolutionaryGames
Authors
Students taking part in the project: Hasan Gürses, Rosanna Incontro, Joachim Kienle, Andreas Mayrock, Ramona Neher
Supervising professors: Prof. Dr. Arend Hintze (MSU and Dalarna University) and Prof. Dr. Jochen Staudacher (HS Kempten)
Faculty: Faculty of Informatics
Project in cooperation with Michigan State University, USA, and Dalarna University, Sweden
date of realization: WS 2019/20