Effectiveness of Prejudiced Strategies in Evolutionary Games

Avni, Nagpal (2022) Effectiveness of Prejudiced Strategies in Evolutionary Games. Masters thesis, Indian Institute of Science Education and Research Kolkata.

Text (MS dissertation of Nagpal Avni (Roll No.17MS026)) 17MS026_Thesis_file.pdf - Submitted Version Restricted to Repository staff only Download (473kB)

Abstract

People are often prejudiced and judge others on the basis of seemingly arbitrary attributes such as wealth and skin colour. Prejudices form a type of heuristic that individuals use to aid in decision-making when interacting with others. And while heuristics can be helpful, prejudices have their inefficiencies. The Chadefaux-Helbing model conceives prejudice as a strategy that players can adopt in evolutionary games. We start by reproducing the model’s results pertaining to the performance of these prejudiced strategies. We find that the performance of the prejudiced individuals depends on the expected number of interactions with other players. P(1) individuals are successful when the expected number of interactions are small, P(5) individuals when the expected number of interactions are large, and P(3) is an intermediate case between P(1) and P(5). We then attempt to broaden the scope of the Chadefaux-Helbing model by allowing for player strategies to evolve with time. We observe that, over a long time, the TFT strategy is adopted by a majority of the population and while the less prejudiced individuals can coexist with TFT individuals, the more prejudiced individuals eventually die out. In the near term, we hope to carry out more rigorous analysis to better our understanding of this result. v

Actions (login required)

View Item