Fuzzy Location-Routing Problem In Post-Earthquake Distribution Of Relief Materials

As one of the most frequently attacked countries by earthquakes, China suffers dramatic loses from earthquakes every year. In 2008, tragic death of thousands of people and collapse of houses were led by Wenchuan earthquake and so did huge impact to the society as a whole. Admittedly, the timely forecast of earthquake is essential. Yet, efficient approaches should be implemented since earthquakes inevitably result to disruptive consequences. Emergent needs of kinds of supplies from affected areas must be met after earthquakes. How to optimize the distribution system efficiently and scientifically is the key point in keeping the stability, reliability and time effect of the emergency aid system.In the existing researches on emergency logistics system, location problem and vehicle routing problem were studied separately. While because of the internal relations between the two kinds of problems, they should be studied jointly, which means Location-Routing Problem should be drawn attention to. In this thesis, relevant results of LRP in the literature were discussed with the background of emergency distribution problem after earthquake. Additionally, we focus on fuzzy LRP in the emergency distribution system by briefly summarizing and analyzing such system.Firstly, the thesis investigates a fuzzy multi-objective static LRP at the initial phase of emergency distribution under earthquake rescue. From the viewpoint of integrated emergency logistics system, the demand requested by different demand point is mostly fuzzy considering sudden occurrence of earthquake and traffic damage and jam come along in the meantime. We propose a fuzzy multi-objective static model for LRP in emergency distribution system, and then conduct an improved genetic algorithm. Through numerical results we verify the feasibility of our model and algorithm.Secondly, we study a fuzzy multi-objective LRP under emergency support after earthquake. Considering continued efforts in rescue, decision makers are able to make dynamic decisions given the real-time updated information on demand at each demand point and condition of traffic facility. We propose a fuzzy multi-objective dynamic model for LRP in emergency distribution system, and then translate it into traditional static model. At last, a two-stage heuristic algorithm for solving the problem is conducted, and we illustrate the model and algorithm by employing a numerical study.At last, a brief summary of the thesis is developed to point out the limitations and extending directions of the research. Existing studies on LRP of the post-earthquake emergency materials distribution are still rare at present, the author hopes this study can serve as an ice breaker in this area.