Incidents cause non-recurring congestion along freeways and contribute to increased delay for motorists. To manage severe incidents, traffic diversion does play an important role in moderating the delay and minimizing the impact area. The ultimate goal of the traffic diversion plan under a non-recurring congestion situation is not only to divert the traffic to alleviate the congestion, but also to control diverted traffic volumes effectively to prevent traffic breakdowns on alternative routes. Normally, the incident management procedure is based on the past experience and also on the use of historical data collected and stored for the purpose. Although historical data on incidents and their impacts may help in deriving a traffic diversion plan, it seems that better and quicker plans can be derived if the data is thoroughly analyzed using state-of-the- art tools such as data mining. The opportunities that data mining could offer in analyzing incident situations and in deriving at better traffic diversion plans are not fully utilized yet. An effort to demonstrate how data mining can be used in deriving a traffic diversion plan is presented in this paper.
Abstract