Methodology of modeling the behavior processes of antagonistic agents in security systems

Abstract

The problem is formulated and the need for developing a methodology for modeling the behavior of antagonistic agents in security systems is shown. The presented concept is implemented at three levels, namely: the level of the security system as a whole, the level of individual agents and the level of the group of agents. Five stages of the concept implementation are presented. At the first stage, it is proposed to analyze protected business processes and threats to these processes. An ontological model is proposed as a basic model of this stage as a carrier of knowledge about the studied prelet region. An approach to the automation of ontology construction is presented, focused on the intellectual analysis of texts in natural languages, namely, texts of articles published in scientific journals. At the second and third stages of constructing the methodology, models of individual and group behavior of agents of cybersecurity systems are proposed. The presented models reflect the reflective properties of agents that affect the decision-making and learning processes. The developed models made it possible to form a model basis for the self-organization of the security system. A practical application of the described models is an algorithm for determining the implementation of the most probable threat, based on the cost indicators of threats and the probabilities of their implementation. This can ensure the efficient distribution of limited financial investment in cybersecurity.