A two-stage hierarchical model for spatial location and evacuation allocation problem of urban earthquake shelters: a case study in Central urban area of Yangbi county, Yunnan province, China

The scientific planning of urban shelters can greatly reduce property losses and casualties, and can improve the efficiency and safety of post-disaster evacuation. Based on the hierarchical characteristics of shelters and the shortcomings of previous studies on evacuation demands and blocked roads, this study proposes a two-stage multi-hierarchy planning method for shelters. In the first stage, the construction cost of shelters is taken as the objective function to complete the spatial configuration of shelters. In the second stage, the evacuation allocation model for refugees is established, which minimizes the total evacuation distance. The planning method also establishes the calculation method for dynamic high-precision refuge demands, and considers the blockage of evacuation roads caused by damaged buildings. The traditional genetic algorithm is improved and hierarchical 0-1 coding is adopted to solve the spatial configuration model of hierarchical shelters. An "exchange" heuristic algorithm is proposed to solve the hierarchical evacuation allocation model. Finally, Yangbi County in Yunnan Province, China, is taken as an empirical case to verify the applicability and accuracy of the proposed method. When the algorithm iterates to 150 rounds, the fitness function value converges to 3040675, indicating that 61 new shelters should be constructed to satisfy the refuge demands, and the minimum construction cost of these shelters is 3,040,675 RMB. The minimum evacuation distances in the first, second, and third evacuation stages are determined to be 51645192, 17351731, and 16136090 m, respectively.