Optimization Of Relief Supply Allocation Decision Under Unconventional Natural Disasters

Various natural disasters have been occurring more and more frequently in recent years and have brought about huge damage both to the society and the population living in it. As a discipline to cope with natural disasters and their influence, emergency management as well as its applications in practical relief operations has become a hot topic. The management of emergency resources is a key component in emergency management. As a part of emergency resource management, the allocation of relief supplies supports the rescue work to a large extent. This dissertation carried out research into the allocation of rescue supplies to the affected areas in the first time in the constraint of supply shortage.This thesis focused on two aspects of the allocation optimization problem. For the one thing, the uncertainty of demand due to the complexity and unpredictability of natural disasters was discussed. For the another thing, the behavioral factors of the decision-makers in the process of allocating the limited supplies were taken into considerationOn basis of the analysis of the disaster context, we carried out research into the optimization of relief supply allocation through analyzing the demand from affected areas, problem formulation and making rapid decisions in the viewpoint of decision makers. The main contents of the study can be summarized as the following three aspects.1. The current status and the trend of the natural disasters were studied, based on which we pointed out the characteristics of the demand and supply for relief supplies and figured out the main factors influencing the matching of demand and supplies in the allocation process. We categorized the allocation into the pushing and pulling modes and the latter mode is the common one applied to the first time allocation of resources.2. We took the scenario of secondary and derivative disasters into consideration and analyzed the change in demand brought by these disasters. The uncertain demand was converted into the potential demand through applying the probability programming method. A nonlinear mixed-integer model was formulated with multiple objectives to solve the problem of the assignment planning in the context of multi-demand-point, multi-storage-depot with limited amount of rescue supplies.3. To simulate the practical allocation process in the immediate aftermath of natural disasters and measure the bounded rational behaviors of the decision-makers both in the distribution centers and in the affected areas, we furthered the proposed model into a bi-level programming one with the objectives of maximizing the equity and the degree of satisfaction respectively in the upper and lower level.