Under CopyrightJürjens, JanSchmitz, AndreasMichel, MarcelMarcelMichel2022-03-0730.1.20142013https://publica.fraunhofer.de/handle/publica/27987710.24406/publica-fhg-279877Nowadays risk analysis plays an important role in identifying and evaluating different types of vulnerabilities. The results of a risk analysis are used to develop adequate prevention strategies. However there is a serious weakness in classic risk analysis methodology. The classic version uses constant and unchangeable impact factors, which could result in aberrant model behaviour against intelligent opponents. On the other hand is game theory, which takes into account the concept of intelligent, strategically interacting agents. Unfortunately, many concepts in this area require a common knowledge about the game and therefore about the other players. In most cases, this concept of the common knowledge is not applicable due to a lack of information about the other actors. To address this shortcoming, the Adversarial Risk Analysis (ARA) framework was designed to evaluate model with uncertainties that characterize the approximate knowledge of several actors. At the moment these models are solved manually. There are no general algorithms to compute an optimal strategy for the models. All in all, the construction of an algorithm that can successfully compute this strategy is desirable.de004master thesis