Purgatory/TopicVehicleRerouting
Contents
Vehicle Rerouting
Within the next subchapter, we want to try to give some advices on how investigations on vehicle rerouting should be prepared and evaluated. Though most of these are based on prior work, any further ideas or processes and other information is appreciated.
Basic Description
The research descibed in here deals with methods for computing vehicle routes, mainly in the case of unforeseenable events. Normally, we assume traffic is in an "equlibrium state" - a route change does not yield in a better performance of the road network. Nonetheless, the traffic may change due to:
- accidents and resulting jams
- additional event traffic
- street closures
- changing destinations (new shopping center)
In such cases, certain edges of the road network will be either completely unsuable or the time needed to pass them raises, either directly, or because of the additional demand of vehicles that want to pass them.
Things to keep in Mind
- The simulated traffic should be in equilibrium mode before the incident is introduced
- Explanation: Using fastest routes in an empty network, like obtained when using a simple Dijkstra (done when converting trips to routes via DUAROUTER) is not realistic. In this case, all vehicles would use the fastest edges in an empty network, which would yield in jam situation on these edges. I such cases, it is easy to find a new best route during simulation - but the initial situation would be not realistic. Having a proper traffic assignment is needed for realistic evaluations.