Two Applications of Adversarial Risk Analysis

Abstract

Adversarial risk analysis (ARA) attempts to apply statistical methodologyto game-theoretic problems and provides an alternative to the solution concepts in traditional game theory. Specifically, it uses a Bayesian model for the decision-making processes of one's opponents to develop a subjective distribution over their actions, enabling the application of traditional risk analysis to maximize the expected utility. This thesis applies ARA framework to network routing problems in an adversarial contexts and a range of simple Borel gambling games.