An Emergency Supplies Pre-positioning And Emergency Evacuation Research Based On Robust Optimization

In recent years,the frequent occurrence of emergencies has seriously affected the safety of people's life and property.In order to improve the ability to respond to emergencies and dimin-ish the possible impact of emergencies,there is an increasing interest in developing emergency management.The characteristics of emergencies are almost impossible to speculate the timing and the intensity,and it is difficult to exactly estimate the impact,damage and resource needs of emergencies.After the emergency,the related logistics,demand and other information may change unexpectedly during the response process.Such changes may have a great impact on the response plan.Starting from the preparation phase and response phase of emergency man-agement,this thesis mainly studies quick and effective emergency supplies pre-positioning and evacuation strategies to deal with emergencies.Firstly,the dynamic emergency supplies pre-positioning model is studied.The model is divided into the emergency supplies pre-positioning stage and the emergency supplies pre allo-cation stage.The goal of the model is to minimize the total cost of the two stages.The model considers the stochastic demand information,which is only partially known after the occurrence of emergency,and then the distributionally robust optimization method is applied to solve it.Fi-nally,a real case is presented to give some suggestions on the strategy of emergency supplies in Bohai area.Secondly,aiming at the problem of emergency evacuation after emergencies,a multiple-priority emergency evacuation problem based on cell transmission model is developed.A multiple-priority emergency evacuation network model is constructed to prioritize the trans-portation of the most severely injured persons,which can furthest reduce casualties caused by emergencies.Through the characterization of uncertain demands with the polyhedron uncer-tainty set,the decision maker can take a more flexible evacuation strategy according to the actual situation.Finally,the validity of the model is proved by a real case study.The major contribution of this thesis includes the following three aspects:(i)a chance con-straint programming model for dynamic emergency supplies pre-positioning problem is built.The model is transformed into a computationally tractable form by replacing the chance con-straint with its safe tractable approximations.The goal is to minimize the total cost of the pre-positioning network;(ii)a multiple-priority emergency evacuation model under the uncertain demand is built.According to the quantity and type of known information,the polyhedron un-certainty set is used to character the uncertain demand.Finally,the duality theorem is employed to convert the robust counterpart into a computationally tractable model;(iii)the validities of the proposed models are illustrated by the analyses of real cases.