Research On Urban Emergency Shelter Planning Models With Consideration Of Varying Shelter Demand

Once an earthquake takes place in a large city,without proper disaster operations management,the earthquake can quickly turn into a disaster for a city.This dissertation investigates the planning of emergency shelters with consideration of varying shelter demand.It aims to reduce the number of earthquake victims and improve their resiliency.1.Resident evacuation behavior after an earthquake in urban areas is analyzed,and an estimation method is developed to calculate the varying shelter demand.First,the effects of physical(geologic and geographic factors,earthquake magnitude),engineering-related(e.g.building structure)and human factors(tolerance to water supply cut-off and choice of emergency shelter)on shelter demand are analyzed.Second,based on these factors,an estimation method is developed to calculate the varying shelter demand.Finally,the method is demonstrated using the central area of Shanghai as example.The results show that there is a significant difference between results from the proposed model and those from models based only on building structure.2.A two-phase model for planning shelters in urban areas with consideration of varying shelter demand is formulated,and a novel,cross-entropy-local search algorithm is developed.First,a two-phase model for planning emergency shelter is formulated.At the first phase,it formulates an emergency shelter location model that minimizes the total shelter setup cost.At the second phase,it formulates a model to allocate the evacuees to shelters to minimize their total evacuation distance.Then,an efficient cross entropy with local search algorithm is developed to solve large scale integer programming problems.The computation results show that the optimal gap is less than 5%.Finally,an emergency shelter planning application based on a case study of Shanghai,China is used to make policy suggestions.3.A bi-objective model is proposed to minimize the total construction cost of emergency shelters and to maximize number of surviving residents in urban areas,and a particle swarm optimization algorithm based on non-dominated sorting is developed.First,a bi-objective model is proposed to jointly address construction cost minimization and survival rate maximization in emergency shelter planning.Then,a non-dominated sorting particle swarm optimization algorithm based on the Pareto-optimal strategy and the feasibility-based rule is developed.Finally,using the emergency shelter planning in Xuhui District of Shanghai as basis for a case study,a set of Pareto-optimal solutions is proposed.4.This section formulates a two-phase and hierarchical model,as well as an efficient hybrid cross-entropy method to solve the shelters location model and an allocation scheme with a swap procedure to allocate victims.First,the hierarchical characteristic of varying shelter demand is examined and an estimation method is developed to calculate the maximal number of victims.Second,emergency shelter locations with a nested hierarchy and the allocation of victims to the shelters are modeled.Furthermore,we employ an efficient hybrid cross-entropy method to solve the location model and develop an allocation scheme with a swap procedure to allocate victims.Finally,empirical results from the application on the Xuhui District in Shanghai,China show that emergency shelter planning based on varying demand can reduce the construction cost of shelters by 28.9%and the average traveled evacuation distance by 11.9%,compared to the current policy based on fixed demand.