Study On The Queueing Maximal Covering Location Model And Algorithm Of Emergency Facility

Recently, with the increasing frequencies of natural disaster and accident that occurred, the caused damage and impact on the society are also rising. On the condition that no accurate predictions about the time and location of the potential disaster and accident can be made, study on how to reduce the losses and impact that caused becomes the focus of emergency management. Previous studies had shown that pre-disaster materials reserves were more time saving and efficient economically, compared with the post-disaster rescue. Thus, it plays an important role in reducing the people and property losses effectively in the disasters.As a strategic decision, the siting of emergency service facilities has a lasting and far-reaching impact on the disaster relief. In numerous location models, the covering type model has a wide range of application in the emergency service facility location. This paper reviews the development process of the covering location model from the time when it was proposed to so far and makes some classifications. Specifically, the covering location models are mainly divided into location set covering and maximal covering model according to the objective function. Whereas deterministic model, probabilistic model and the probabilistic model applied with the queueing theory can be get according to the methods used. In order to deepen the understanding of the basic model and improve the practical ability to apply the models, further discussions, including the detail form of the models, the key assumptions and their drawbacks, are made on some important kinds.Subsequently, the more precise covering location models in the assumption and approach that applied the queueing theory are discussed in details. According to the differences of the queueing model applied, the M/M/s/s loss system and the M/G/s non-loss system are further studied, respectively. For the former, the queueing maximal covering location model with assured service reliability is built, in which demands that cannot be served immediately are get lost due to no waiting policy, whereas the possibility that every covered demands can get service are no less than the pre-determined value. Two covering models, the descriptive model MCLP-R1 and the integer linear model MCLP-R2, are established in the point view of covering qualities and an algorithm is designed when the demands can be partially met. The computational results show that a satisfactory solution can be obtained within a short period of time by the algorithm. For the latter, demands cannot be served immediately will wait in the line, but additional restriction on the waiting time is imposed. Considering the different assignment policies, the QMCLP-SA and QMCLP-NS models are established, based on the system assignment policy and customer nearest selection policy. An algorithm to solve the QMCLP-SA model has been devised and a satisfactory solution can be obtained within a short period of time, thus, providing a low bound for the problem. Meanwhile, the results obtained by the two different assignment policies are-compared with each other. It shows that the system assignment policy is more efficient, whereas the location result may be unfair from the point view of the customers and the opposite conclusion can be obtained by the customer nearest selection policy. The inadequacies and further research directions are discussed at last.