Combined Locaiton-Transportation Problem Considering Stage Characteristics Of Post-earthquake Relief Supplies

In recent years, the global earthquake happened frequently, causing great damage and loss. As a key step of post-earthquake emergency logistics system, relief distribution must meet the logistics needs as soon as possible, under constraints of time, space and resources in order to minimize losses. How to locate the temporary transfer facilities and schedule multi-mode transportation are key problems in optimizing the emergency logistics system. The relationship between them is complex and controversial, so it is necessary to optimize and manage them together. By considering the relationships between location-allocation problem of emergency relief distribution centers and joint transportation problem of relief materials, Combined Location-Transportation Problem(CLTP) in post-earthquake relief distribution was investigated from the perspective of system optimization.Firstly, post-earthquake rescue activities are dynamic and real-time update process with obvious stage characteristics. Therefore, in order to improve the efficiency of relief distribution, this thesis analyzed the stage characteristics of post-earthquake relief distribution and divided it into three stages, i.e. initial stage, emergency rescue stage and emergency recovery stage.Then compared the three stages of emergency rescue with common logistics systems, got the features of the stages, held the main problem and the key factors.Secondly, the phenomenon and trait of the three stages of the emergency rescue were abstracted as mathematical symbols, and three models based on emergency supplies phases after earthquake were builded in proper parameters, variables and expressions, as follows:(1) In the initial stage after earthquake, the collection centers as the supply nodes supply different relief materials, the effective nodes as the demand nodes need different relief materials, different relief materials own different urgency, road network is interrupted, the transport capacity of each mode of transportation is different, communication equipment is demaged,’pull’distribution mode, relief materials shortage, etc. A bilevel programming model was proposed to describe the combined location-transportation problem in the initial stage of post-earthquake, which the upper objective is minimizing the delivery time, and the lower objective is maximizing the equity and fairness.(2) In the emergency rescue stage, the supply of the distribution centers are constantly changing, the need of each effective node is also changing, different relief materials own different urgency, road network is gradually repaired, the transport capacity of each mode of transportation is different, the equipment of communicaition basically returns to normal,’pull’ distribution mode, the relief materials are still short, time windows restrictions. When the decision before was not fit for the present missions, it is need to be re-optimized and open the distribution centers gradually, only in this way, more and more work can be managed efficiently. So a bilevel programming model was proposed to describe the combined location-transportation problem in the emergency rescue stage of post-earthquake, which the upper objective is maximizing the time-satisfaction, and the lower objective is maximizing the equity and fairness.(3) In the emergency recovery stage, expect the multi-modes of transportation, multi-kinds of relief materials and so on, the road net work and the communications equipment is back to normal completely, the demand can be meet, collection centers supply the materials on demand, the time is not pressing in this stage, the multi-period dynamic-decision making was considered because of the long time span of this stage. A multi-period dynamic-decision making model was proposed to describe the combined location-transportation problem in the emergency recovery stage of post-earthquake, which objective is minimizing the total cost.Thirdly, hybrid genetic algorithms combined with different heuristic rules were proposed based on the characteristics of each model. A serious of problems just as objectives and restrictions was resolved well, like two level object decision making problem with leader and follower, dynamical decision, multi-period dynamic-decision, the restriction of the transportation capacity, the throughput, time windows, the minimum satisfied percentage, the maximum quantity of distribution centers and so on.Finally, the validity of the model and algorithm were demonstrated by a numerical example based on the relief distribution process after Wenchuan earthquake. The results show that the proposed genetic algorithm has good performance and is effective for resolving the CLTP in relief distribution.The thesis set up powerful distribution system of emergency supplies and flexible contingency reaction mechanism based on the post-earthquake, gave auxiliary decision support to the guarantee of relief materials and provided the practical problem that urged to be settled after the emergency with scientific theory, method and tools.