Research On Vehicle Routing Problems For Spreading Of Deicing Salt

In the north of China, the snowfall is usually large and has a longer duration. Even in some south areas, it also occurred and brought a lot of bad influence. The disaster caused by snowstorm in southern China in 2008, for example, brought direct economic loss reaching up to millions of RMB. Snow and ice on the road makes bad impacts on the transportation of the transport system, and the snow removal has become a vital task of winter road maintenance in our country. The deicing salt can melt snow effectively with low energy, which is widely used but has bad environmental effects. Thus, optimization of vehicle routing for snow spreading plays an important role of improving the operational efficiency and reducing the operational cost.This paper analyses the characteristics of vehicle routing problems for systematically related to spreading operations by considering their main components, such as: transportation network, road segments, sectors, vehicles and vehicle routing. Combined the working characteristics of the spread truck, vehicle routing problem for snow salting operation is abstracted as Capacitated Arc Routing Problem. Considering the network structure, capacity, load balancing, service continuity and other factors, the general CARP model is improved and expanded in order to according with the characteristics of vehicle routing problems for snow salting operation. Besides, the paper considered the problem of refill point that the vehicle in our study is divided into operational vehicles and transport vehicles. That transport vehicles select a suitable point as a temporary additional point of deicing according to the location of the operating vehicles. Finally, a model was established to minimize the total travel distance of operational vehicles and transport vehicles.CARP has been proved as an NP-hard problem and it is very difficult to adopt precise algorithm for the practical solution of large-scale road network, so an improved genetic algorithm for solving the model is proposed. At last, we obtained the best route of the two models by experiment test, one of which considers the interim supplement states. Comparing the results of the two models, we reach the conclusion that the total travel distance calculated by the model with interim supplement states is less than that the total travel distance made by the model which does not consider the interim supplement states, whats'more, the deadline distance in the first modle is 80.7% percent less than that made by the other one.