Recently, the Dutch government has published the "Nota Mobiliteit". This policy document defines the key elements of the Dutch transport policy for the forthcoming decennia. One of the goals for 2020 is not only to improve the journey times of travelers, but also to improve the reliability of these journey times. It states the ambition that during rush hours 95% of the trips are "on time". In this context "on time" means that the difference with the expected travel time is less than 20% or less than 10 minutes delay (whichever is the greatest). The Transport Research Centre of the Ministry of Transport recognized the need to develop a tool to assess transport policies on their contribution to reach the required reliability. The Transport Research Centre uses a National/Regional Model System to assess transport policies at the national/regional level. The objective of the present study was to develop a supplementary tool to assess the changes in reliability using these model systems. This is a novel approach, for which it was necessary to estimate a relation between reliability and parameters available in the model system. This relation has been estimated using traffic measurements. A considerable part of the Dutch road network is equipped with induction loops. On the selected network some 200 routes were identified and used for the estimation. From an entire year of data about 150 days of valid data have been selected. Although official public holidays and weekends are omitted, the data does cover both normal workdays as well as the summer holiday period. The study showed that the most important internal explanatory variables are the average speed, the maximum speed and the length of the trip. Although several different relations have been tested, a logistic function provides the best fitting relation between average speed and reliability. If the average speed over the year is slightly lower than the free flow speed, this has no impact on the reliability. However, below a certain speed, the reliability starts dropping and soon this becomes a linear decrease. A preliminary study showed that external factors like road works, weather and (large) accidents also have an impact on reliability. With the use of a database containing information on the occurrence of these externalities, we were able to estimate their effect on reliability. The large accidents have the biggest influence, but because large accidents do not occur very often accidents in general are the most important external factor. The final result of this project is a practical instrument that allows a user to compare the reliability of the network given a certain transport policy compared to a reference case. For the covering abstract please see ITRD E135207.
Abstract