Crisis Control: Agent-Based Models with PyCX for Modeling Crowd Dynamics During Evacuations
摘要
On vessel emergencies and associated crises, effective evacuation strategies must be devised in order to ensure safety among passengers. In this regard, this research proposes an advanced ABM framework which would contribute to crisis management through realistic modeling of real dynamic interactions in vessel evacuations. Unlike the traditional static modeling approaches, the proposed framework will integrate individual passenger characteristics such as stress levels, injury status, age demographics, and navigation through obstacles into a realistic human behavior model in emergency situations. We simulated four different scenarios: (1) evacuation with obstacles and panicked agents, (2) evacuation with injured and uninjured agents, (3) evacuation using emergency doors, and (4) evacuation with consideration of age affecting agent speed. The results from all these scenarios reveal that personalization of agent behaviors significantly enhances the accuracy of evacuation simulations, providing more valuable information about the best crisis response strategies. The present paper systematically investigates evacuation scenarios and evidences that adaptive strategies are imperative in view of enhancing efficiency and the level of safety of evacuation. The ABM framework presented will therefore become a useful instrument for the development of better emergency preparedness and decision-making within a maritime context focused on saving human lives and minimizing risks during life-threatening situations.