The aim of this study is to develop a technique of searching the optimal evacuation route in real-time, at a huge earthquake, employing Ant Colony Optimization (ACO) which is one of the metaheuristics techniques. After a severe earthquake in a city area, quick refuge to temporary or extensive evacuation area should be required occasionally. However, the occurrence of building collapse and/or fire gives the serious problem with being impossible to utilize the preexisting route to evacuation area. In addition, it happens that the re-movement to the second refuge is compelled by the fire occurrence of the first refuge. In such a case, another evacuation route should be reconsidered. But due to numerous routes to be reexamined for the optimal route, it is very difficult to find it on time. In order to overcome the difficulty due to numerous routes, this study examined the optimal route search by coupling technique through ACO and backtracking method. As a result of the numerical calculation, computation time by ACO shows excellent reduction effects with around 1/100, as compared to calculating all the combination of routes. These results show that the proposed method based on the ACO technique is a useful one in searching the optimal evacuation route in real-time.
View full abstract