Article ID: | iaor1997919 |
Country: | Netherlands |
Volume: | 71 |
Issue: | 2 |
Start Page Number: | 237 |
End Page Number: | 256 |
Publication Date: | Dec 1993 |
Journal: | European Journal of Operational Research |
Authors: | Kirschfink Heribert |
Keywords: | artificial intelligence |
A knowledge-based system for the collection, check and completion of traffic data is presented. The completion of traffic data, which, for economic reasons, are only detected for a few links of the road network, is carried out on the basis of different knowledge sources like topological relations within the network, statistic traffic counts, physical relations and the experience of traffic engineers. A blackboard architecture is used for the knowledge-based system. All knowledge is concentrated in the central blackboard, with access for the knowledge-based system for data completion as well as for traffic management systems that are based on it. The knowledge-based system for data completion has three tasks: first, the incoming, detected traffic data are checked for their plausibility; second, the traffic volumes are disseminated for each section on the basis of the detected values; third, completion of the traffic densities. The dissemination of the traffic volumes is carried out by constraint propagation while the densities are completed on the basis of heuristics. The system is currently being tested at a site in Hamburg.