Game Theoretic Modelling of the International Relations System

The thesis models interactions in the system of states. Fundamental research question asked what consequences for success of strategies and prospects of cooperative behavior have particular settings and properties of the system. Thesis includes two features peculiar to international relations that did not appear anywhere else before: (i) determination of interaction occurrence with help of distance and power; and (ii) emergence of (dis)trust out of the previous interactions. The model is based on three elements: agents, environment, and rules. Players interacted in the Hobbesian Prisoner's Dilemma environment as described by realists, but thanks to payoff shift representing emergence of (dis)trust I also formalized constructivist argument of different cultures of anarchy and of mutually constitutive agent-structure relationship. Multi-agent computer simulations set within the abductive reasoning framework were chosen because lack of heterogeneous enough data and impossibility of experiments made this data generating method a necessity. The source code is written in C#. I translated 62 Axelrod's behavioral rules and then added several others that seemed promising. Three new strategies mirroring usual behavior of states were proposed too. To secure robustness of the results, application was run..