An architecture for modeling emotions and personality to simulate human interactions

Emotions and personality are vital factors affecting human interactions. This thesis aims to build an integrated model of emotions and personality to simulate human interactions. Such simulations can be used in a wide variety of applications. This thesis addresses several problems and limitations of typical emotional systems and describes an architecture that addresses those problems and, thereby, brings more flexibility and power of expression to such systems.; This research focuses on designing a simple, computationally efficient emotional system that uses a minimal set of variables and a set of monotonic functions to generate a wide range of emotions. The architecture addresses the design issues involved in how conversational events affect emotions and how emotions in turn affect the conversational behavior of the system. The architecture is implemented for a MATES application, wherein two agents are simulated with a couple's personality parameters and, given a situation, perform a conversation.