Social media platforms have a developing role in how people respond to hazards, providing a network to seek help and respond to help requests. Understanding the dynamics of behavior on social media networks can improve community-level hazard response efforts. People who experience damages may use social media to seek immediate help for debris removal, supplies delivery or emergency rescue, and peers connected on social media may respond by reposting the help request or providing help in person. This research develops an agent-based model (ABM) to simulate a community of individuals that use social media to seek help and respond to requests for help during a hurricane. Agents represent individuals that are in a community affected by a hurricane and share a social media network. Behavioral rules for seeking help and providing help are developed using the Theory of Planned Behavior and parametrized through analysis of a survey of social media use conducted in communities that were affected by 2018 Hurricanes Florence and Michael. The ABM simulates agents that post help requests, repost help requests, provide help in person, and receive help. A Design of Experiments approach is applied to assess how ABM parameters, including community size, connectivity of a network, damage rate, and propensity for using social media, influence the number of requests for help that are met through the social media network. Results demonstrate that high connectivity leads to rapid reposting and results in cascading responses to meet requests for help.

社交媒体平台在人们如何应对危害，提供寻求帮助和响应帮助请求的网络方面发挥着不断发展的作用。了解社交媒体网络上行为的动态可以改善社区一级的危害响应工作。遭受损失的人可以使用社交媒体寻求即时帮助以清除碎片，运送物品或进行紧急救援，而社交媒体上连接的同伴可以通过重新发布帮助请求或亲自提供帮助来进行响应。这项研究开发了一种基于主体的模型（ABM），以模拟一个社区，这些社区在飓风期间使用社交媒体寻求帮助并响应求助请求。代理代表在遭受飓风影响的社区中并共享社交媒体网络的个人。寻求帮助和提供帮助的行为规则是使用计划行为理论制定的，并通过对受2018年飓风佛罗伦萨和迈克尔影响的社区进行的社交媒体使用情况的调查分析来参数化。ABM模拟发布帮助请求，重新发布帮助请求，亲自提供帮助并接收帮助的代理。实验设计方法用于评估ABM参数（包括社区规模，网络的连通性，损坏率和使用社交媒体的倾向）如何影响通过社交媒体网络满足的求助请求数量。结果表明，高连接性导致快速重新发布，并导致级联响应以满足帮助请求。