Agent-based modeling of entity behavior in cybersecurity

Abstract

Human and system behaviors, as they relate to information security, are facets that have always been hard to measure. Simply put, there are too many factors to account for when considering the knowledge, experiences, and relationships of individuals. And when these are coupled with information systems, their complexities rise to several levels of magnitude. Given the absence of empirical data, we argue for the utilization of an agent-based modeling and simulation system toward gaining an understanding of entity behavior in cyber space. In this chapter, we report the results of an ongoing research project, which utilizes agent-based models and scenarios to simulate the effect of user trust, adversary sophistication, user training, and system defenses on cybersecurity. These independent simulations utilize software agents that assume certain predefined attributes to emulate their physical counterparts on an environment that represents the cyber space.