Finding Fuzzy Shortest Path by Integrating historical and Real-time Traffic Data

Because of the happenings such as accidents and road repairs, the traffic level of some links of the network can strongly be changed. Therefore, routing on the basis of merely historical traffic data is not always useful. On the other hand, routing based on real-time data is not adequate either. The reason is that the real-time data regarding near links can be trusted. However, this is not the case for the distant links; since their present traffic will certainly be changed before the vehicle gets there. The main goal of this research is to propose a new method of routing based on fuzzy integration of historical and real-time traffic data. In this method, the basis of routing for normal situations is the historical data. In case of an accident in the network, before getting to the vicinity of the accident, the shortest path will be recalculated. The weights allocated to the links are calculated based on the integration of historical and real-time traffic data, using a fuzzy inference system. These weights are variable and are a function of the distance between the present location of the vehicle and the accident location. The results of the research show that, using this method, in the case of accidents, the calculated path is clearly deviated from the location of the accident.