Crowd Panic Behavior Simulation Using Multi-Agent Modeling

Abstract

This research introduces a novel approach to crisis management by implementing a multi-agent algorithm within a strategic decision system. The proposed system harnesses multiple agents’ collective intelligence and adaptive capabilities to enhance decision-making processes during critical situations. The study first investigates the theoretical foundations of crisis management and multi-agent systems, emphasizing the need for an integrated approach that combines strategic decision-making with autonomous agents. Subsequently, the research presents the design and implementation of the multi-agent algorithm, outlining its ability to gather, process, and analyze diverse data sources in real time. The multi-agent algorithm is specifically tailored to adapt to dynamic crisis scenarios, ensuring a resilient decision-making framework. Experimental simulations present the implementation of a panic simulator and prediction of evacuation and intervention routes using multi-agent artificial intelligence algorithms. The results demonstrate the multi-agent algorithm-driven decision system’s superiority in response time, resource allocation, and overall crisis mitigation. Furthermore, the research explores the system’s scalability and adaptability to different crisis types, illustrating its potential applicability across diverse domains.