ABSTRACT
Since the 19th century, the industry of large luxury cruise ships carrying passengers has developed rapidly. Large luxury cruise ships have thousands of built-in cabins, up to hundreds of meters high, and can carry up to several thousand passengers simultaneously. Therefore, it is very important to evacuate people in case of fire and other hazards. Traditional cruise ship emergency evacuation strategies usually simulate the behavior in advance of a fire scenario or when a fire first occurs, and choose the best evacuation route as the evacuation method. However, this method does not consider dynamically changing environmental and personnel factors’ influence on the optimal evacuation route. Therefore, this paper proposes a dynamic evacuation method based on a neural network, which can update the evacuation plan synchronously and give the global optimal evacuation route according to the real-time changes of the disaster during the evacuation process. In order to improve the efficiency of the algorithm, the network flow model is optimized based on actual cruise ship drawings. Finally, Pathfinder evacuation simulation software is used for simulation experiments. The experimental results show that the evacuation method can achieve the demand of dynamic evacuation with high efficiency based on ensuring the shortest total evacuation time and can be a reference for the actual emergency evacuation work system strategy.
