Spatial configurations and learning rules impact on the evolution of cooperative behaviour in the n-person iterative prisoner’s dilemma

Spatial configurations and learning rules impact on the evolution of cooperative behaviour in the n-person iterative prisoner’s dilemma

Gismondi, Roberta (TU Delft Electrical Engineering, Mathematics and Computer Science; TU Delft Intelligent Systems)

Yorke-Smith, N. (mentor)
Venkatesha Prasad, R.R. (graduation committee)

Delft University of Technology

Computer Science and Engineering

CSE3000 Research Project

2021-07-01

Decision-making dynamics and their impact of human behaviour have raised a large number of questions throughout the years. Traits like competition and collaboration amongst agents are often studied, in the context of Game Theory, by the medium of games such as the Iterative Prisoners’ Dilemma.Furthermore, many realistic scenarios and possible real world applications of the Iterative Prisoners’ Dilemma (e.g. socio-geographic and economic ones) can only be modelled ed by a more general instance of the game that allows for multiple numbers of players such as the n-person IPD. Work has been done to analyse the effect of spatial configuration on the outcome of the game.The goal of this research is to conduct further analysis on the possible confounding factors of these experimental settings. In particular, we investigate on the effect than different machine learning approaches for learning-rule inference (such as genetic algorithms and particle swarm optimisation)have on the correlation between the dependent variables of previous controlled experiments (the number of players, the scale of interaction and the initial percentage of cooperators and defectors) and the evolution of cooperative behaviour. The data that are relevant to answer the research question are gathered by means of repeated controlled experiments that aim to give insight on the effect that the factors under analysis have on the convergence of the environment to a state of almost full cooperation.

Prisoner’s Dilemma
Machine Learning
Game Theory
Particle Swarm Optimization
Genetic Algorithm
Evolutionary Algorithm - EA

http://resolver.tudelft.nl/uuid:ebfa9960-5442-4351-a2d3-af21dd75df54

Student theses

bachelor thesis

© 2021 Roberta Gismondi