Skip to content

Tuning Modeling Parameters in the Mesoscopic Simulator

For calibration purposes, we can classify the vehicle modelling parameters according to their influence in the simulation outputs, into three groups:

  • Global parameters
  • Local Section and Turn Parameters
  • Particular Vehicle Type Parameters

Influence of Global Parameters

The number of parameters set by the experiment is greatly reduced in a mesoscopic simulation from the number set for a microsimulation experiment. There are options to use the slope model to affect speed and to change the variability of the look ahead distance variability in lane changing.

Influence of Section Parameters

Section parameters influence all vehicles, regardless of type, when driving in a particular section of the network. Examples of parameter influences include the following:

  • Jam Density: Sets the capacity of a section. When this value is exceeded, vehicles are unable to enter the section.

  • Reaction Time Factor: A local parameter for the global vehicle reaction time. It can be used to calibrate extreme situations; for example on links with gradients, sharp bends or other ad-hoc situations.

  • Lane Selection Model:

  • Penalize shared lanes.
  • Take into account fast/slow lanes.

Note that the lane selection mode can be disabled in mesoscopic simulation so all vehicles select the closest lane based on the connectivity of the network. This option should be used with caution.

Influence of Turn Parameters

The mesoscopic model offers a simpler version of the microsimulation turning yield parameters. These are documented in the Mesoscopic Modeling Section The parameters are found in the Road Types Editor and the reaction times can be modified for different vehicle types.

  • Look-ahead distance: Set by road type or for an individual section.
  • Look-ahead distance variability: Set in the experiment editor for all turns.
  • Reaction Time Factor: A local parameter for the vehicles using the turn.

Influence of Vehicle Parameters

Vehicle parameters influence all vehicles of a particular type when driving anywhere in the network. Examples of parameter influences include the following:

  • Mesoscopic yield reaction times.
  • Proportions assigned to paths or following OD routes or following SRC generated paths.