Ring roads around major cities are intensively used, and congestion is often a problem. On these expressways congestion can back-propagate onto connecting freeways, and via on-ramps onto the urban road network . It is therefore beneficial to be able to suppress the growth of congestion or even prevent the onset of it as much as possible. Implementing route guidance is a means to steer the network state away from bad traffic conditions.
Finding the optimal route guidance settings is a complex task. A proactive approach implies the need for prediction, as the traffic conditions are constantly changing and control actions can only be taken at a limited number of locations, typically where traffic enters the network. It is important to regard potential unfavorable effects of route guidance, such as overreaction. Dissolving congestion on one road and inducing it on the other. That’s why coordination should be taken into account. Furthermore, the proposed route guidance algorithm should be able to operate in real-time, in order to timely react on incidents. Therefore, an efficient method is necessary.
The objective of this thesis is to develop a route guidance algorithm, that considers the features of prediction, coordination and efficiency. The intended algorithm is built upon the concept of back-pressure routing: a simple, distributed way to reach maximum throughput under stabile network conditions. Back-pressure routing has its origin in communication networks, and recently has also been studied for urban traffic signal controls. The question is, how this concept can be translated to perform well on the freeway routing problem. The concept of Network Fundamental Diagram is expected to be of use. A macroscopic simulation model will facilitate short-term traffic prediction. The applicability and effectiveness of the algorithm, and various configurations, are to be examined by testing it for several scenarios in a case study of Rotterdam.