The Location And Scheduling Strategy Of Emergency Resources Based On The Robust Optimization

China is one of the countries in the world with many natural disasters happening each year. Large demand on the emergency resources follows each emergency. The appropriate location for the emergency material reserve centers, the reasonable scheduling strategy and the timely distribution of the emergency resources, are crucial on the whole rescue action and its social benefits.After the incident occurs, things like the road condition, goods demand and other uncontrollable external factors are unpredictable. It is really difficult to develop a timely and effective rescue program. In the thesis, we build the mathematical models by the robust optimization, and analyze the main uncertain factors in the emergency. Under our modeling, all the constraint on the uncertainties can be satisfied, with all the possible values taken. It is guaranteed, even in the worst situation, to optimize the objective function. Our thesis contains the following: 1. The locations of emergency material reserve centers:The ability to support the emergency resources is the key factor for a successful rescue. The locations of emergency material reserve centers are decisive to maximize, not only the ability of supplying goods, but also the capability for rapid reaction in the rescue operation. For the location of the emergency material reserve centers, we should consider many factors like the demand, facilities cost, transportation cost and others. Therefore, we will disperse the locations of the emergency material reserve centers to cover the needs of emergency resources, and to provide adequate services. Starting from of the strategic, economic aspect and service level of the facility locations, under the consideration of the uncertainty of the emergency, we study the following things in our thesis:(1) The multi-stage robust optimization model for locations with capacity limitation, only considering the transportation cost.By considering the changing demand at different stages, the restriction of the maximal inventory capacity of the emergency material reserve centers, and the cost of the transportation, we study the location problem of the emergency material reserve centers. We start at a multi-stage location model, under the capacity limitation on the inventory and only the cost of the transportation, with a definite demand of emergency resources. Based on this, we then build a robust optimization model with uncertain demand on supplies, and finally we transform it to a robust corresponding model. And in the last model, we take the change of the demand, transportation cost at each stage into account, and try to maximize the social benefit of the whole rescue system.(2) The multi-stage robust optimization model for locations with capacity limitation, only considering the cost of inventoryInstead of the cost of transportation considered in (1), we only consider the cost of inventory here. We start at a multi-stage location model, under the capacity limitation on the inventory and only the cost of inventory, with a definite demand of emergency resources. Based on this, we then build a robust optimization model with uncertain demand on resources, and finally we transform it to a robust corresponding model. And in the last model, we take the change of the demand, inventory cost at each stage into account, and try to maximize the social benefit of the whole rescue system.(3) The multi-stage robust optimization model with capacity limitation, considering the cost of both the transportation and the inventory.Gathering the consideration both in (1) and (2), we start at a multi-stage location model with the capacity limitation and a definite demand of emergency resources, considering the cost of both transportation and inventory, considering the facilities failure,we making a multi-stage location deterministic model with the capacity limitation and a definite demand of emergency resources, considering transportation, inventory and failure cost to improve the reliability of emergency system. Based on these, we then build robust optimization models with uncertain demand on resources, and finally we transform them to robust corresponding models.In this thesis, we apply the branch-bound algorithm for solving the above model. We conclude that the robust optimization model has more advantages than the corresponding location model with definite demand, like less emergency material reserve centers, bigger covered area, more inventory and higher satisfaction on demand. It provides decision makers with better solutions, and improves social benefits of the emergence rescue.2. The emergency schedulingA quick and efficient rescue operation depends heavily on the emergency scheduling. Things, like the quick arrival of the rescue workers, the availability of the rescue and relief for lives and properties, and the timely distribution of emergency supplies, are crucial to social benefits of the rescue operation. The emergency dispatch faces many complicated situations and unknown factors. It is very essential to build a quick, dependable and effective emergency scheduling strategy, facing the uncertainty of the demand and complexity of road conditions.Under the emergency, the road condition is not clear, and the communication facilities, are blocked up. So the resources demand and the time for transportation are uncertain. In the thesis, we start at a single-garage definite model with the maximal utilization rate of vehicles and the minimal time delay; then we build a multi-garage definite model with the maximal utilization rate of vehicles and the minimal time delay. And next, based on the above, we have a robust optimization model under uncertain resources demand and transportation time. At last we transform it to a robust corresponding model, and solve it by the branch-and-bound method. By comparative analysis on examples, we conclude that the robust scheduling model has better performance under the uncertainty of the time. Under the consideration of the vehicle utilization, with the increase of demand uncertainty, we get lower vehicle utilization, more time delay and the decrease of the objective function values. Then we will have a different optimal route. So in order to enhance the social benefits, we have to improve the accuracy of our prediction on the demand.