In traffic management a clear understanding of traffic situations now and in the future is useful for operational and planning purposes. Traffic systems in many countries are characterised by a highquality network of sensors and data systems, which together supply an assessment of the traffic situation at a particular time. Based on this assessment and the expected effects of specific measures, predictions can be made on how traffic will behave in the near or far future. This enables choices to be made the on the best measures to implement or to put in operation, based on predefined criteria for traffic flow, safety and/or environmental effects. In order to choose the most applicable tool to make these traffic predictions and to make a choice of measures, it is essential to have a good understanding of the available tools and the range of applications. Traffic prediction models are often applied to determine the effectiveness of infrastructural, design, mobility and traffic management measures. However traffic prediction modelling is a vast domain in which a wide range of models exist. There are many models which each have a specific purpose and approach to modelling traffic. Often the applied approach is deliberately developed with that purpose in mind and therefore only performs well for that purpose. Various trends have seen models developed with a more general purpose, but sometimes at the cost of specific accuracy and flexibility for certain measures and uses. Other trends have focussed on developing high quality and accurate models for specific applications, however with a higher level integration as part of modelling suites. The goal of this report is to give the reader a feeling of the different types of traffic models and their applications. To understand which options are available for modelling traffic and which applications are relevant, it is necessary to make a distinction between the various different applications. To be able to do this it is also necessary to give a global overview of the type of models that exist. This process is performed in chapter 2. An overview is given of what prediction entails, which is followed by a review of applications for prediction models in traffic management along with an overview of the main types of traffic prediction models. The current state of practice in model forecasting is given in chapter 3 and gives a description of the current models that are in use in The Netherlands and internationally. In many cases there will be a great deal of overlap, as the traffic model market is generally also universal. Nevertheless there are some local preferences. The specific application areas of different forecasting models is important, as not every model is able to capture and reproduce all traffic characteristics. The benefits and application on different policy levels are therefore described in chapter 4. Furthermore it is apparent from the current state of art that a shift in modelling approaches is ongoing in which modelling packages and suites are considered important to avoid inconsistency between different levels of forecasting. However other developments are also evolving in the core of models which reflect developments in traffic flow, but also in science. These future perspectives are described in chapter 5. (Author/publisher)
Samenvatting