Towards Evacuation Planning of Groups with Genetic Algorithms

In crisis situations on board ships, it is of utmost importance to have the passengers safely evacuate to the lifeboats in an efficient manner. Existing methods such as marked escape routes and maps are not optimal as pre-planned escape routes may become heavily congested by passengers. Further, the closest lifeboat is not always feasible as lifeboat capacity can be exceeded. Also considering that some evacuees are strongly affiliated, such as families, and that they prefer to evacuate together as a group, it becomes a difficult problem to solve. This paper models the area to be evacuated as a time-expanded graph with hazard severities as probabilities of survivability for each node. The presented approach applies a multi-objective genetic algorithm with multiple fitness functions to maximize the over all survivability. Finally, the proposed method picks the best evacuation plan from a pool of potential solutions returned by the genetic algorithm. The solution generates better routing plans than comparable methods, specially in situations where grouping and congestions are considered. In essence this is an essential step towards automatic planning of evacuations which in turn contributes to smoother evacuations of crises situations and saving lives.