Research On The Multi-Objective Programming Theory For Emergency Logistics Under Mixed Uncertainty

In recent years,the frequent strikes of natural disasters have brought huge damages to human life and property,and have affected the social stability and economic development.Because of the suddenness,the destructiveness and unpredictability of natural disasters,it is necessary to adopt the scientific and reasonable rescue measures in order to reduce casualties and incidents.Emergency logistics planning is a key issue of emergency rescue,where emergency resources will be transported to the disaster areas at the fastest speed with the limited time,space and mixed uncertain factors,meeting the minimum demand for disaster areas.The research of emergency logistics planning theory in China starts relatively late and the foundation is weak.In order to ensure emergency rescue processes in an orderly and efficient manner,it is meaningful to scrutinize the location of distribution centers,vehicle route planning,emergency resources distribution and multi-modal transportation.This dissertation is based on the theory of robust stochastic optimization,probability and statistics,scenario decision and system optimization,which focuses on the pre-disaster preparation and post-disaster response phase.Using operational research,management,multi-objective optimization modeling method,and combining qualitative analysis and quantitative calculation,the problems of location-routing,location-allocation and multimodal transportation in emergency logistics under mixed uncertainty are studied.The paper is consisted by the following three parts:(1)The bi-level programming problem of emergency logistics location-routing based on transshipment strategy is studied under mixed uncertainty.The model is developed based on the principle of the nearest emergency transport strategy with the consideration of satisfying the minimum level of service for disaster areas,vehicle capacity constraints and the horizontal transshipment strategy between the distribution centers being taken.The pre-disaster preparation phase aims to determine the number and location of the distribution centers and the inventory strategy of emergency resources,meanwhile,the post-disaster response phase is about planning the disaster distribution plan and the vehicle routing.The upper model focuses on minimize the sum of the expected costs,and the lower level model aims to find the worst-case optimal solution of the response time among all scenarios.Finally,the proposed model is demonstrated as an efficient reference for the decision makers in the emergency rescue operations through the case study.(2)Under the fuzzy demand scenario,the robust stochastic programming approach based on fairness is proposed for the location-allocation problem in emergency logistics.According to the robust optimization theory of multiple scenarios,not only supplies and demands but also the road traffic condition is treated as the uncertain factors.Moreover,the fairness of emergency resource distribution is another factor to be taken care of.The multi-objective robust stochastic programming model attempts to minimize the weighted average response time and the total costs,concurrently,the variance of the two parts is taken into account.Finally,the robustness of the solution and model is analyzed through the example of Wenchuan earthquake,and the viability of the model is verified.(3)The probe based on the disaster information updating is deployed into the robust allocation strategy problem of multi-modal transportation in emergency logistics.To optimize the problem of emergency resources transportation in the post-disaster,the study knowing the location of the distribution centers is conducted on how to coordinate the transportation of emergency resources for a variety of disaster scenarios,relief resources and multiple rescue phase cycles by multi-modal transportation from the supply points to the distribution centers,the distribution centers to the disaster areas and two distribution centers.The minimum level of service and the priority for disaster areas are considered in the present work,combining with the characteristics of road,railway and air.The robust allocation strategy model of multi-modal transportation is developed based on the disaster information updating technology aiming to minimize:total cost,total response time and the maximum penalty cost of unmet demands.Finally,the validity of the model is verified by a numerical example,which is based on the dynamically updated information of the demand and the condition of the road network.With the above research,the related theories of the emergency logistics planning have further been enriched.The weakness of current emergency logistics management decision has been improved,and the applications of robust optimization theory in emergency logistics have been strengthened.Assistant decision support is provided for the effective scientific emergency rescue activities.