Finding the nearest facility for travel and waiting time in a transport network

One of user's queries from navigation service is to find the nearest facility in terms of time. The facility that is being questioned by the user as a destination may have a queuing service system (e.g. bank), which means that the cost function of the shortest path includes the waiting time at the destination as well as the travel time. This research conducts in the zone 1 of Mashhad with Bank at destination. In this research, we first calibrate the volume-travel time function to predict travel time by using history volume data of SCATS. The results of the analysis show the Moving-Average model with a period of 4 weeks is more precise to predict volumes and consequently travel time. Then we use Simulation-based method to predict waiting times in Bank. A* algorithm with different scenarios is applied to solve the shortest path problem. This algorithm is compared with the Dijkstra’s algorithm in different networks. Results show by increasing the nodes of network, the required time to solve the A* algorithm is significantly lower than the Dijkstra’s algorithm. In general, this study indicates the A* algorithm and the suggested heuristic function reduce run time for solving the shortest path problems.

Keywords:

Travel Time Prediction , Volume-travel time function , Waiting time , Simulation , shortest path