The Sichuan–Yunnan region of the Qinghai–Tibet Plateau in China is earthquake-prone. Ge et al. (2010) investigated the practice of earthquake disaster management in China through interdisciplinary efforts, introduced the disaster planning and management system established by the Chinese government, and proposed the principle of "people-oriented, cost-ignoring" disaster relief in China. A hierarchical relief supply warehouse construction system was built according to this principle. Based on the three key indices of earthquake emergency, response, and rescue (ERR), Lu and Xu (2014) comparatively analyzed the 2008 Ms 8.0 Wenchuan earthquake and the 2013 Ms 7.0 Lushan earthquake. They found that emergency rescue proceeded more efficiently, concertedly, and orderly in the Lushan earthquake than in the Wenchuan earthquake. This suggests that China's earthquake emergency rescue system has been continuously improved and perfected with the accumulation of experience and the emphasis by the government.
Relief supplies are necessities of life that are intended to meet the basic needs of victims. They are a form of humanitarian action taken by the government for disaster relief. Disaster relief can be divided into three stages, i.e., preparations, responses, and actions (Apte, 2009). According to this division, the site selection of relief supply warehouses is an essential task in the stage of preparations. Depending on timeliness, warehouses can be classified into permanent and temporary emergency facilities. Depending on the application field, they can be further classified into medical facilities, disaster relief warehouses, disaster shelters, etc (Trivedi and Singh, 2018; Cao and Chen, 2019). In this sense, the site selection of relief supply warehouses involves the disaster relief warehouses among permanent emergency facilities.
This study categorized existing literature in this field under three themes: earthquake-affected areas, maximal covering location, and multi-objective robustness analysis.
The prediction of earthquake-affected areas provides an essential basis for determining the method for the site selection of relief supply warehouses in earthquakes with Ms ≥ 7. In recent years, there has been a flood of literature prompted by deepening research on earthquakes (Gupta et al., 2022; Li et al., 2016; Liu et al., 2015). In the research field regarding the construction and site selection of relief supply warehouses, the literature about the prediction of earthquake-affected areas mainly involves an extensive prediction of earthquake-affected areas formed by seismic zones and seismic belts within a large region (Wang et al., 2020; Yang et al., 2016). Renkli et al. (2015) assumed the occurrence of an earthquake in Istanbul, predicted the damages to buildings according to their seismic capacity, and identified disaster-affected areas. Wang et al. (2020) introduced seismic resilience to judge the seismic capacity of disaster-affected areas and further determined their locations and scales. Many researchers have simulated the design of several earthquake-affected areas, and taken actual earthquakes as site selection models to validate the case studies (Geng et al., 2021; Guan et al., 2020; Jana et al., 2021).
Maximal covering location is an important content of on-site research selection. In earthquake emergency relief, maximal covering location is mainly used to assess the timeliness of transport of relief supplies from a relief supply warehouse, thus evaluating the scientificity of the site selection scheme of the relief supply warehouse. In terms of research on the site selection of relief supply warehouses in earthquake disasters, Han et al. (2011) studied the timeliness of transport of relief supplies from relief supply warehouses to disaster-affected areas by investigating the routing and scheduling of rescue vehicles. Based on exploring the impact of secondary disasters on the transport roads of relief supplies, Zhang et al. (2019) examined the site selection planning of relief supply warehouses. Geng et al. (2021) focused on the problem of pain perception by victims and proposed introducing victims' pain perception cost into modeling maximal covering location based on the concept of people-oriented sustainable development. Nedjati et al. (2016) designed a drone transport network and used it to quickly deliver post-earthquake necessities of life in designated, densely populated areas. Taking into account the effect of the post-earthquake transport roads of relief supplies, Sun et al. (2019) divided the mode of transport into two stages, i.e., road transport and helicopter transport, and built a path–site selection spatio-temporal model for transport of relief supplies.
Multi-objective robustness analysis is currently the primary method for research on the site selection of relief supply warehouses. On the strength of the two site selection methods of maximal covering location and optimal arrival time, multi-objective robustness adds multiple other constraint conditions to better align the site selection method with disaster relief practice and ultimately obtain a more scientific and reasonable site selection model using various algorithms. Yan et al. (2021) developed an optimal site selection scheme to provide a location for maximum coverage that determines the grading of relief supplies by maximizing rescue satisfaction and minimizing warehouse number and designed a heuristic multi-center clustering location algorithm for dual-objective models.Balick and Beamon (2008) proposed a variant model of the maximal covering location model, integrated facility location and inventory decisions, and considered budgetary constraints and capacity restrictions. He et al. (2017) defined a set of effectiveness-oriented criteria for the site selection of emergency warehouses, proposed a multi-criteria ranking method, and solved the problem of inaccurate weight information in stochastic pairwise dominant relations and the pruning procedure of the ELECTRE-II method. By interviewing management staff in several regions, Roh et al. (2015) summarized multiple criteria for the site selection of humanitarian relief warehouses, determined the weights of various criteria using the analytical hierarchy process, and ranked site locations based on fuzzy TOPSIS. To deal with earthquakes and other disasters, international organizations have also conducted some studies for the site selection of relief supply warehouses worldwide through setting constraint conditions using the maximal covering location method, such as CARE International (Duran et al., 2011), International Federation of Red Cross (Charles and Lauras, 2011), etc.
Regarding the three research fields involved, there appear to be numerous studies on the scientific prediction of disaster-affected areas that mainly concentrate on earthquakes. Still, previous studies rarely touch upon emergency rescue or the site selection of relief supply warehouses. Most studies that explore the site selection of relief supply warehouses in earthquake disasters usually assume disaster-affected areas, but scarcely demonstrate the scientificity of the assumption. Studies investigating the site selection of relief supply warehouses in earthquakes based on the maximal covering location method mainly focus on the emergency planning of the route and mode of transport in the assumed scenario of damaged roads (possibly due to secondary disasters), aiming to ensure the arrival of relief supplies within the time allowed. They rarely attempt to predict relevant factors such as the mode of transport of relief supplies or the limit of safe transport based on the occurrence laws of earthquake disasters in the region. Regarding the multi-objective robustness of the site selection of relief supply warehouses, most studies make assumptions about factors such as construction cost, transport cost, material cost, and psychological expectation and obtain the optimal site from scientific equations. However, no sufficient consideration has been given to either the principles of earthquake disaster relief or the declaration and acceptance standards of relief supply warehouses. As a result, these assumptions cannot match the practical considerations of site selection. In general, in existing studies on the site selection of relief supply warehouses in earthquake disasters, there is no scientific prediction of the mode of transport or earthquake-affected areas. While highly subjective assumptions are made for multiple constraint conditions of site selection, practical constraint conditions that match the site selection criteria of relief supply warehouses are often absent.
In this context, this study proposes a method of site selection for relief supply warehouses based on the prediction of earthquake disasters and the quantification of site selection criteria. By fully utilizing the achievements of scientific research on earthquakes for prediction in site selection, this study improved the scientificity of prediction. It also established constraint conditions through quantifying and classifying site selection criteria. Based on meeting the site selection criteria of relief supply warehouses, it avoids the defects of two subjective methods employed by existing studies, i.e., the assumption of constraint conditions and empirical assignment of weights.