A multi-stage algorithm for earthquake shelter site selection using multi-criteria decision making methods

The management of earthquake damages plays a crucial role in sustainable development of a society and results in preparedness and mitigation of losses. The densely populated city of Tehran, Iran, coupled with the fragility of houses and infrastructure, highlight the necessity of a reliable earthquake damage assessment, preparedness against earthquake and planning for losses mitigation. This paper attempts to identify appropriate sites for constructing earthquake shelters based on land-use, demographic and street network data. Next, each building is assigned to the nearest shelter. Finally, all people in the study area are assigned to a shelter, even though the capacity of shelters is limited. In the proposed algorithm, first the distance matrix of parcels with a public usage is formed. Then, some solutions are presented as appropriate alternatives for shelter construction based on the distance matrix. Finally, the alternatives are ranked based on: a) the minimum variance of shelters’ populations, and b) the minimum total distances of shelters from each other, i.e., the number of shelters is minimum. After verification of all public parcels in the study area only eighteen of them had enough area to be considered as an option. The implementation of proposed algorithm generated four solutions for constructing shelters. Finally, these four alternatives were ranked based on the weighted average of two above criteria. Considering the efficiency of the proposed algorithm in the study area, it may be usable for other regions.