Federated Edge-Semantic Learning for Decentralized Resilient Evacuation (FESL-DRE): Simulation Dataset

This dataset supports the research presented in the paper: “Federated Edge-Semantic Learning for Decentralized and Resilient Indoor Evacuation under Dynamic Hazards” The dataset was generated using a simulation framework designed to evaluate decentralized evacuation systems operating under dynamic hazards and partial infrastructure failures. The simulation models an indoor environment as a graph-based structure, where nodes represent spatial locations and edges represent navigable paths. Hazard conditions, such as fire intensity, are dynamically simulated using distance-based decay functions. Crowd density evolves during the simulation based on agent movement and interactions. The dataset is designed to enable reproducibility of experiments involving: decentralized coordination using gossip protocols semantic reasoning for hazard interpretation adaptive evacuation routing resilience analysis under node failures human-centric behavioral modeling