Evacuation routing algorithms commonly neglect the differences of evacuees in mobility, physical strength and resistance to hazard effects, and calculate routes using a single algorithm. Hence, congestion or block of paths may occur due to the diverse mobility, and unnecessary injuries or fatalities may be caused by the different resistance to hazard and fatigue. Owing to the fast-changing situation in the hazardous environment, in this paper, we present a distributed routing algorithm based on the Cognitive Packet Network (CPN) that performs parallel heuristic search to customise routes for evacuees with respect to their ongoing physical condition and the immediate environment. With the aid of portable devices, evacuees are divided into different groups in terms of a virtual ???health value??? and can switch groups during the evacuation process. Experimental results indicate that the presented routing algorithm and the dynamic grouping mechanism can reduce the mortality rate of evacuees.
展开▼
