Congestion Charging in Urban Networks : Modelling Issues and Simulated Effects

Abstract: One of the major challenges cities face today, in their development towards sustainable urban areas, is the need for an efficient and environmentally friendly transport system. This transport system should manage to tie together the city without strong adverse impact on urban environment, air-quality and climate change. The specialized labour (and leisure) market, typical of a large urban area, exaggerates the need for efficient travel, as it is increasingly difficult to live and work within short distances.    The use of demand management tools has become more frequent in transport planning with this development towards more sustainable cities. Whereas investing in new capacity was previously the main response to increased demand for travel, there is a much broader range of policies in use today. One of these demand management tools is congestion charging. Singapore was first to implement congestion charging and during the last decade it was followed by London and Stockholm, with increasing support from the citizens as a consequence. Many other cities have performed feasibility studies for introduction of congestion charging.  The development of transport models for prediction of demand management tools, such as congestion charging, has however not been able to keep up with this change in kind of policy. Transport models that were developed for prediction and evaluation of infrastructure investments, such as new motorways, are often used to forecast effects of policies aimed at managing demand, which too often results in poor prediction. This thesis focuses on the needs for modelling of congestion charging. The state-of-practice models used before implementation in Singapore, London and Stockholm are reviewed, as well as more advanced dynamic models developed for prediction of congestion charging and other demand management tools. A number of gaps in the modelling of congestion charging are described and a new model called SILVESTER is developed, which closes some of these gaps. In particular, SILVESTER involves dynamic mesoscopic modelling of traffic flows, flexible departure times and users with heterogeneous preferences. The thesis describes the implementation of SILVESTER and considers and compares different methods of demand aggregation in order to reduce run-time of the large-scale dynamic model (Paper I). It also describes how preferred departure times of road users can be determined in calibration such that consistency exists between the departure time choice model and dynamic traffic flows which are input to assignment (Paper II). The unique implementation of congestion charging in Stockholm gives the possibility to validate SILVESTER on real-world measurement of reductions in traffic flow and behavioural adjustments to the charges (Paper III). SILVESTER is then used to analyse several modified versions of the Stockholm congestion charging scheme and to compare welfare and equity effects of the different schemes. It is shown that the welfare of the current scheme could be improved if charges were allowed to differ by location and driving direction (Paper IV). It is shown that the benefits of congestion charges calculated using SILVESTER are greater than the benefits calculated with a static model. Finally, the reasons for the greater benefits are investigated (Paper V).

  This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.