How to understand the major control parameters within Paramics

UPDATED: November 22, 2017

Control parameters can be divided into five main sections:
1. Global - affecting the entire network, demand and vehicle population
2. Category - affecting specific categories of link
3. Vehicle - affecting specific vehicle types
4. Link - affecting specific links, and
5. Node - affecting specific nodes or intersections.

Usually the major adjustments to the model would be to change the geometry of the network to match as accurately as possible the actual road layout (moving kerb and stop-line control points) and coding forced lane changes to override default Paramics lane changing.

Network changes may include:

* Routing/category factors
* Signposting and signposting range distances
* Restrictions
* Speed Controls
* Gradients/gradient model
* Junction visibility
* Category headway factors
* Link reaction factors
* Link headway factors
* Considering link headway factor
* Link end speeds/link stop times
* Nextlanes
* Force across/merges
* Stay in lane
* Lane and turn restrictions
* Lane choices

Additionally there are other parameters that can be set to control the overall behaviour of the model e.g. increasing or decreasing the "mean headway", the "mean reaction time" or the DVU aggression and awareness distributions. The basic vehicle behaviour of aggression and awareness is set for each vehicle when it is released onto the network. These aggression and awareness levels fall within a normal distribution that was calibrated against loop detection information taken from sites in the UK.

These levels can be adjusted but we would recommend that this is only done if the user has some background data which suggests that vehicle behaviour are their sites is not consistent with UK findings. It is normally possible to locally influence driver behaviour satisfactorily without adjusting the global aggression and awareness factors, which may influence the overall network performance in different ways.

The calibration and validation is done by comparison of model results to observed data (link flows, queues etc as mentioned below) but also using visualization of the vehicles moving through the junction (comparison to videos, users knowledge of how the junction/network operates). This gives a mixture of quantitative and qualitative analysis. The direct comparison of results and model output should conform to standards specific to the country in which the project is based, such as those published in Traffic Appraisal in Urban Areas (for example).


  • No Downloads