Resarch On Strategic Decision Models Of Service Areas And Method For Optimize Addressing Of The Urban Shelters

The recent years have witnessed the occurrence of several heavy urban disasters causing widespread destruction and distress. These disasters have forced a majority of residents to be temporarily evacuated from their homes to centralized shelters. Generally, a shelter could be a construction specially built for fighting disasters or a multi--purpose construction for temporary use in emergence services, e.g. schools and gymnasiums. Government and local authorities are responsible to build or employ enough shelters and guarantees that their total capability meets the demand of accommodating all residents. However, it is not always appropriate to direct residents to the nearest shelter because each shelter has a finite capability. Therefore, a pre-designed evacuation map, in which each resident or a group of them (e.g. a city block) should be explicitly assigned to a shelter, is essential to protect residents from disasters. Such a map delineates service areas of shelters which should be mutually exclusive and must fully cover the entire residence area.Let us assume a scenario where all shelters have been sited and their total capability satisfies the demand of accommodating all residents. In order to determine their service areas, several parameters must be taken into account, namely, capability of each shelter, spatial distribution of residents, and travel cost of transporting residents to assigned shelters. A preliminary consideration is that residents should be assigned to a shelter as close to them as possible. If the closest shelter has no spare capability to accommodate these residents, they should be assigned to another shelter. However, a shelter’s service area might be spatially separated into several parts. This means that some residents have to travel through other shelters’ service areas before reaching the one assigned to them. An immediate consequence of this situation could be the increasing difficulty of efficiently transporting residents to shelters in a large-scale evacuation process:people are unwilling to move on when they have reached a shelter though it is not assigned to them, and there might be heavily conflicting traffic flow towards different shelters from their interlocked service areas. Therefore, along with reducing the total travel cost and being subject to capability constraint of each shelter, the spatial continuity of a service area should be also taken into account. Keeping spatial continuity means that each shelter’s service area should form a homogeneous region delineated with a single polygon, and these polygons fully covering the entire residence area are mutually exclusive in space.To date, a number of approaches have been proposed to deal with service areas, influence areas, dominant areas, buffers, hinterlands of variant facilities, etc. However, few of them have considered, to the best of our knowledge, all the aforementioned constraints and objectives in one integrated approach. Some assume that the capability of a facility is infinite and that residents are assigned to the nearest facility, which is very likely to cause overcrowding of shelters if they are not sited in an optimal manner. Others assume that all demand points are evenly distributed in space, and the higher the capability is, the larger the service area is. However, this assumption is inappropriate because residents are always unevenly distributed. Still others take the minimizing of the total travel cost as the main constraint, and thus the spatial continuity of service areas might be compromised.According to the residence evacuation plan for the city emergency, an algorithm for distributing the shelter for victim or spatial disaster cell is proposed based on shift insertion method in this paper. Considering the uneven distribution of the residents in space and the capability limitation of each shelter, the proposed algorithm minimizes the total travel cost and keeps the spatial service continuity of each shelter with capability constraint. As a result, it can provide the scientific basis and the decision support for emergency evacuation. Comparative experiments under different constraints in the paper have evaluated the algorithm’s availability and feasibility. Moreover, applications in engineering projects also prove the excellent practical utilities.