Crisis Management in the Era of the IoT, Edge Computing, and LLMs

The Internet of Things (IoT) combined with edge and cloud computing are employed extensively across various sectors, including manufacturing, transportation, and crisis management. In crisis management, specifically, the IoT plays a crucial role in data collection and analysis, aiding decision makers and first responders during emergency operations. As the available data increases, however, existing approaches for crisis management may encounter problems in ingesting large and diverse datasets and extracting relevant information efficiently. To address such problems, this paper introduces a novel decision support system capable of ingesting diverse data with low delay by utilizing Large Language Models (LLMs). This way, our system can support decision makers by answering relevant questions automatically. To show real-world applicability, we examine two use cases, namely, pandemic management and natural disaster management, and we discuss the process of using our system to make more informed decisions. To evaluate the proposed decision support system, we build a prototype and conduct experiments that show high accuracy and low cost of generating answers in the context of crisis management and the IoT.