Evacuation-path-selection model of real-time fire diffusion in urban underground complexes.

Evacuation in the event of a fire in an urban underground complexes (UUC) is a big challenge. How to conduct limited evacuation in case of a fire in UUC is a problem worthy of attention. A hexagonal grid was used to balance the timing of each move to the adjacent grid to synchronize fire diffusion and crowd evacuation. This paper proposes an improved ant colony (IACO) algorithm to determine the shortest evacuation path in the case of a fire under an UUC considering real-time fire diffusion. The proposed model was verified under various sce-narios. The results show that compared with the traditional quadrilateral raster map, the hexagon raster map can ensure that the moving distance in any direction is equal. The proposed IACO contributes to the body of knowledge resulted in shorter evacuation paths and avoided detours into dangerous areas and "through the wall" errors. These findings could be used to improve fire safety management plans and select evacuation technology, while assisting in the development of intelligent emergency evacuation systems.