Approaches to Optimize Local Evacuation Maps for Helping Evacuation in Case of Tsunami

Abstract

Nowadays, most coastal regions face a potential risk of tsunami. Most of researches focus on people evacuation. However, there are always the part of evacuees (e.g. the tourist) who lack information of the city map, we then focus on the solution to guide people in evacuation.

With regards to the guiding system in evacuation, most of the studies focus on the placement of guidance signs. In fact, in panic situation the evacuees can get lost if the they do not receive guidance signs at a certain crossroad or corner. Then, if there are not enough signs at every crossroad or corner on a certain path to the safe places (called shelters), the guidance signs will become useless. In order to give evacuees a complete guidance of evacuation, we propose to place the local evacuation maps which highlight the shortest path from a place (where the panel of the map is placed) to the nearest shelter. At specific places in the city, once this kind of map is perceived, the evacuees should follow the shortest evacuation path to shelters. In this paper, we first present the method to optimize the local evacuation maps by using the genetic algorithm on candidate locations of the maps and the evaluating the percentage survivors as fitness. We then present two approaches to evaluate the fitness: one uses the agent-based simulation and other uses the linear programming formulation. By experimentation, our proposed approaches showed better results the approach of optimizing sign placement in the same scenario.