RELATIONSHIP BETWEEN CONGESTION AND TRAFFIC ACCIDENTS ON EXPRESSWAYS AN INVESTIGATION WITH BAYESIAN BELIEF NETWORKS

Accidents and congestion are two frustrating events, which can be observed very frequently on roads. Accidents, especially on expressways, can trigger heavy traffic congestions imposing huge external costs and reducing the level of service. Therefore it is obvious that accidents clearly have an impact on congestion. But the opposite, i.e. the effect of congestion on occurrence of accidents, is less studied and still questionable . One can argue that congestion can reduce the high speeds on expressways and as a result of that the accident rate is reduced. But in a congested road section vehicles are closely packed and as a result of that rear-end collisions, back-up collisions as well as side collisions can occur. Therefore it is important to analyze the impact on the accidents by congestion so that the policy makers can implement relevant measures to reduce the external costs of both accidents and congestion. This paper investigates the effects of traffic congestion on the occurrence of accidents on 8 radial routes (inbound direction) of Metropolitan Expressway (MEX). Data were obtained from the International Traffic Database (ITDb) . Two softwares, namely WinMine Toolkit 2) and MSBNx , which use the concept of Bayesian Belief Networks (BBN), were used to model the interrelationships among occurrence of accidents and other variables such as congestion index (CI), traffic density and volume. 2. Relationship between congestion and accidents