A risk minimization based approach for planning evacuation trip distribution

Abstract

Most studies on evacuation strategies hitherto focus on routing and the resulting traffic pattern. While doing so, such studies generally consider evacuating people to the nearest shelter location (relief point). In this paper, we illustrate why a strategy of evacuating to the nearest relief point would not be the most risk-averse plan. We present a methodology for planning evacuation trip distribution to potential relief points in a risk-averse manner. The procedure explained would help policy makers to identify relief points where capacity must be improved while preparing for a disaster. We present a risk-minimizing formulation that prescribes evacuation trip distribution and corresponding capacity enhancement scheme (RETDC+) for the relief points. To assess the reduction in risk, we compare this formulation with two other formulations: one where risk-averse trip distribution is performed without enhancing the existing capacities (RETDC), and another where trips are distributed to the nearest relief points (PETDC). We envisage to employ failure probability values based on susceptibility maps that are often prepared for disaster response planning. The methodology is demonstrated for the city network of Gangtok, India which is prone to natural disasters such as landslides and flash floods. We illustrate that the evacuation trip matrix obtained by employing the proposed methodology is 20% more risk-averse for landslide and 27% for flash flood, respectively, than the evacuation trip matrix obtained by a naive distribution to the nearest relief points.