ABSTRACT

In previous work we have developed the DASH agent platform aimed at simulating human behavior in order to understand and mitigate the negative consequences of human security decision-making for security protocols. The DASH architecture, shown below, includes a dualprocess approach with modules for rational and instinctive behavior, helping account for differing human responses to situations depending on factors such as level of attention, fatigue, stress or affect. It includes a reactive planning module that allows the agent to exhibit goaldirected behavior that is responsive to changes in the environment and support for different mental models that humans may use to reason about a situation.