Robust planning models for uncertain emergency evacuation problem

In this paper, we formulate the evacuation network planning problem as robust models in which the resulting evacuation planning solutions are immune to the demand uncertainty without known probability distributions in the complex emergency evacuation context. Specifically, this paper considers three types of robust models in which the uncertainty sets are characterised as the box uncertainty set, the ellipsoidal uncertainty set and the polyhedral uncertainty set, respectively. Performance evaluations of the robust models are conducted on a small-sized evacuation network to illustrate the applicability. The results show that the robust solutions are preferable, since it is convenient for planners to obtain the guaranteed feasible solutions within the prescribed uncertainty set. Even if the actual demand realisation exceeds the given set, planners could still find out a desirable evacuation solution with a smaller violation probability. In general, our study provides some useful insights for the emergency evacuation management.