The Study On Emergency Supplies Prepositioning And Allocation Problem Based On Type-2 Fuzzy Theory

Sudden diasaters, such as natural disasters, man-made disasters, public health events and social security events, usually cause great destruction to the society and economy. To reduce both human life losses and material damages, it plays an important role in pre-crisis preparedness phases and post-crisis response phases to deliver emergency supplies to the a?ected areas quickly and e?ciently. However,in real-world problems, a paramount challenge in creating an e?cient and e?ective supplies plan is uncertainty about whether or not the disasters will occur, and if they do, where and with what magnitude. In this thesis, we employ fuzzy theory to study the prepositioning and allocation problem of emergency supplies under uncertain information.This thesis proposes the critical value reduction methods and discusses the properties of two-stage fuzzy programming in order to represent the uncertain information and control the risk in the prepositioning and allocation problem for emergency supplies. Four classes of fuzzy prepositioning and allocation models are built. They are one-stage single objective expected value model, value-at-risk model,minimum risk model and two-stage multi-objective expected value model. In the first three classes of models, the acquisiton cost, transportation cost, supply, demand and capacity are unknown prior to the extraordinary even, which are represented as fuzzy parameters with variable possibility distributions. The variable possibility distributions are obtained through our possibility critical value or credibility critical value reduction methods for type-2 fuzzy variables. In the fourth model, the supply,demand and availability of relief-allocation path are uncertain, and the decisions are divided into two stages according to the realizations of the uncertain parameters.Furthermore, on the basis of the emergency logistics’ characteristics about fairness,timeliness and economy, the proposed model focuses on minimizing the proportion of unmet demands, the expected travel time and the total cost incurred in the whole process. For each model, we address the cases that the objective function and the credibility constraints can be turned into their equivalent parametric forms. Taking the advantage of structural characteristics of the equivalent optimization model,we design the parameter-based domain decomposition method or goal programming method to divide the original optimization problem into its corresponding mixedinteger parametric models, which can be solved by standard optimization software.Finally, to explore the viability of the developed models and the solution approaches,we perform some computational experiments on realistic scale case problems.The major novel work of this thesis includes the following five aspects:(i)The possibility critical value and credibility critical value reduction methods are proposed in order to deal with the uncertain information;(ii) The interconnections about two types of risk aversion two-stage fuzzy optimization models are discussed,and two important indices for two-stage fuzzy programming are established;(iii)Four kinds of fuzzy prepositioning and allocation models for emergency supplies are presented, which are one-stage single objective expected value model, value-at-risk model, minimum risk model and two-stage multi-objective expected value model.(iv) The basic properties of proposed models are analysed, which facilitate us to turn the original models into their equivalent parametric or mixed integer programming forms. The parameter-based domain decomposition method is developed to solve our models;(v) Numerical experiments are provided to demonstrate the advantages of parametric optimization method and two-stage programming methodology as well as the e?ectiveness of the designed solution methods.