Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 43337 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Habibi, Jafar; Safari, Mohammad Ali
- Abstract:
- Intelligent transportation systems (ITS) have been widely studied in recent years. One of the challenging problems in this field is routing on urban networks, which are largescale networks that are almost planar and sparse. In the routing problem, a person wants to journey from a known source to a known destination. He may have multiple choices as to the path to traverse, and wants to discover the best path, designated in terms of its travel time. The routing problem can itself be a subproblem of another larger problem, and so the running time of the routing algorithm is important. In addition, the existence of traffic flow on urban roads has great impact on travel times. The network is dynamic, and road travel times depend on the departure time. A person may prefer a longer path that has a lower traffic flow, to minimize his travel time. Hence, a suitable routing service should be aware of the traffic flow on the streets. A time-dependent routing serive needs to be integrated with a traffic forecast and estimation system to have an acceptable behavior. During the routing process, when the travel time of a road that will be reached in a near future is needed, the system asks the traffic forecast system to predict its travel time. Without a suitable traffic forecast and estimation system, the routing network can not be expected to work properly. In this work, the problem of routing on urban networks while taking into account traffic data is studied. Although Dijkstra’s algorithm produces the exact answer for this problem, this algorithm is time consuming on a large-scale urban network. Therefore the A∗ algorithm with landmarks is used to accelerate the routing. This algorithm is tested using 1152 test cases in which we observe that the running time is a forth to a fifth of that of Dijkstra’s algorithm, keeping in mind around one percent of error in the time distance between the source and the destination. The test cases are grouped into 24 groups, each consisting of 48 test cases, based on the Euclidean distance between the source and the destination. In each group, the test cases are given to the algorithm in half hour intervals simulating a full 24 hour period
- Keywords:
- Routing ; Traffic ; Time-Dependent Routing ; Dijkstra Algorithm ; Goal-Directed Search
Digital Object List-
محتواي پايان نامه - view
Bookmark