| Sumario: | The following survey paper reviews several methodologies across the fields of computational meteorology, dynamic programming, communications psychology, and narrative artificial intelligence that can provide effective messaging solutions to the ongoing problem of low evacuation turnout during emergencies like hurricanes. We focus on two main aspects of emergency messaging. First, we focus on dynamic programming as a technique to optimize message formats. This technique has been applied to emergency management in the past, as it provides researchers with state-based stochastic models that can lay a framework for developing communication optimization algorithms. Second, communications psychology suggests how the brain is hardwired for greater persuasion and information intake when facts are organized in the format of narratives, suggesting that story structure itself optimizes messaging. We review work in computational models of narrative generation for their potential to be applied to the generation of emergency messaging. The introduction and synthesis of relevant work from these disciplines and their application to emergency messaging could provide a novel approach for improving evacuation outcomes.
|
|---|