Identifying optimal earthquake evacuation routes using genetic algorithm

Abstract

The disaster response team of Metropolitan Manila prepares for the 'Big One,' focusing on information dissemination as to what must be done when the earthquake happens. In order to make a quick and safe evacuation, it is necessary to formulate a clear-cut evacuation plan, specifically, to create evacuation routes. In this study, earthquake evacuation was simulated from Philippine General Hospital (PGH) to Rizal Park to identify optimal evacuation routes quantitatively using genetic algorithm (GA) and geospatial data. The problem was treated as a multi-objective optimization problem wherein the evacuation distance was minimized and the arrival probability maximized. Road networks were mapped using Geographical Information System (GIS) and information on road lengths and road blockage probability were imported to python. GA was used to search for optimal evacuation routes. The algorithm yielded a front of Pareto-optimal solutions. Subsequently, analytic hierarchical process (AHP) was applied to select the best optimal evacuation route according to preference. The best route identified has a distance of 1089.32 m and an arrival probability of 0.504. The model contributes to the preparation and planning of evacuation in the event of the 'Big One' ensuring the safest and most efficient evacuation route.