A practical system for optimized assignment of shelters to evacuees

Abstract

This paper considers the problem of assigning safe shelters for temporary evacuation from an earthquake or a big fire. By studying the state-of-the-art theoretical research and the real situation in Japan, a system based on the latest service computing and optimization is proposed. It integrates a former approach for static assignment to residents (Zhao et al., Optimal Assignment of Wide-Area Evacuation Centers for Kyoto City (Abstract), March 2017) and an ongoing project for dynamic assignment to evacuees including visitors. In both situations, a service (Google's Distance Matrix is used but others are possible as well) is employed to reduce the heavy cost for distance calculation and maintenance. For the dynamic assignment, a novel algorithm is proposed to optimize the number of requests to the service using preprocessing. Comparison against existing approaches shows that the proposed system is more practical for a rapid and effective response in case of a disaster.