A combined Ratio-Emotional Architecture to Model Agent Behavior in Emergency Situation

Abstract

Throughout our lives, we are dealing with panic situations and often we are constrained to make crucial decisions in a limited delay. Especially in case of crowd evacuation; where human lives risk to be damaged or even ended following a disaster such as fire, explosions, and floods. These recent years, crowd modelling have been increasingly proposed. Unluckily, most of these models miss the consideration of victims' emotions. This paper propose a combined rational and emotional archi-tecture to simulate human behaviors in emergencies. Specifically, we focus on the dominant emotions and personality based on the OCEAN and the OCC models. Our aim is to show the influence of these models of the human behavior in crises. Our architecture contains also a metric module that computes the distance between an agent and the emergency exits, the outputs of this module affect the intensity of the agent's emotions and accordingly the agent's behavior. To make our simulation more realistic, we use Fuzzy Logic to model agent's behavior changes in emergency.