Multi-objective Optimization Of Relief Distribution In Emergency

In recent years, emergencies, such as natural disasters and man-made terrorists, have already been major fatal disasters that threaten humankind’s life. Rescue efforts can reduce the damage severity once emergencies occur. Relief distribution is the focus of rescue work. Different objectives, response time, total cost, demand fill rate etc. have to be considered by decision makers to find an optimal schedule for distributing relief according to the real situation.Based on the resuce time horizon, we classify the multi-objective relief distribution optimization problem into single-time period and multi-time period relief distribution optimizations. According to the types of the parameters and variables, the relief distribution problem is divided into static relief distribution and dynamic relief distribution. And with the contents of decision problems, we divide the relief distribution problems into vehicle routing problem and location and routing problem.Firstly, the dissertation studies the multi-objective static relief distribution vehicle routing problem in single time period. In detail, with characters of split delivery, fuzzy demand, flow balance, capaciticy constraints and damage routes, it establishes mixed integer model for multi-objective open vehicle routing problem with fairness. The fast non-dominated sorting differential evolution algorithm is applied with the defuzzy operation. The test instance of "Ya’an earthquake" is studied to illustrate the effectiveness and efficiency of the model and algorithm.Secondly, it studies the multi-objective static relief distribution location and routing problem in single time period. Specifically, taking minimization of relief distribution time, minimization of total cost and maximization of the reliability as objectives, it formulates multi-objective open location routing problem models with split delivery, which is the extension of tradition location and routing problem. With large demand of relief, the split delivery is required. With the uncertainty of network, the vehicles is not required to return the departing distribution centers, which is called open location routing problems. The non-dominated sorting genetic algorithm-II and the non-dominated sorting differential evolution algorithm are proposed to solve the model. And the case study of "Wenchuan Earthquke" is tested to express the practical of the model and algorithms.Thirdly, the dissertation studies multi-objective dynamic relief distribution vehicle routing problem in sigle time period. The high level of dynamic is the main character of emergencies. It establishes the multi-objective dynamic relief distribution vehicle routing problem model with fairness considered. The dynamic relief demand is estimated and conditions of networks are obtained according to the level of damage and the repair schedule. To adapt to the dynamic of relief demand and netoworks, it revises the non-dominated soring differential evolution algorithm. And the "Wenchuan earthquake" is tested to verify the performance of the algorithm.Finally, it studies the multi-objective dynamic relief distribution vehilcle routing problem in multi-time period. The running time for one relief distribution misson is defined as one time period. According to the relief distribution schedule, the rescue time horizon is divided into serveral time periods. it sets different objectives for relief distribution in different time periods. It models the fuzzy demand multi-objective dynamic relief distribution with split delivery in multi-time period as mixed integer program. With the defuzzy operation, the differential evolution algorithm is proposed to optimize the relief distribution in each time period. And "Ya’an earthquake" is studied to verify the practical and efficiency of the model and algoritym.With the above research, the emergency management methods could be enriched. It is expected to support the managers to make decisions scientifically and technically.