Distribution procedure classes#

AequilibraE’s distribution module comprises three different classes: GravityApplication, GravityCalibration, and Ipf.

GravityApplication#

This class, as its own name explains, applies a synthetic gravity model, using one of the available deterrence funcions: EXPO, POWER, or GAMMA. It requires some parameters, such as:

  • Synthetic gravity model (which is an instance of SyntheticGravityModel)

  • Impedance matrix (AequilibraeMatrix);

  • Vector (Pandas.DataFrame) with data for row and column totals;

  • Row and column fields, which are the names of the fields that contain the data for row and column totals.

The synthetic gravity model instance can be either created or loaded, if you have already calibared a model.

Plase check other arguments and parameters that are passed to GravityApplication in its documentation.

GravityCalibration#

Calibrate the model consists in checking if all the parameters set are appropriate. This class, as its own name explains, calibrates a traditional gravity model, using one of the available deterrence funcions: EXPO, POWER, or GAMMA. It requires some arguments such as:

  • Matrix containing the base trips (AequilibraeMatrix);

  • Impedance matrix (AequilibraeMatrix);

  • Deterrence function name.

Plase check other arguments and parameters that are passed to GravityCalibration in its documentation.

See also

Ipf#

IPF is an acronym for Iterative Proportial Fitting, also known as Fratar or Furness. The IPF procedure is used to “distribute” future trips based on a growth factor. The procedure can be run with or without an AequilibraE model, with the latter using one of AequilibraE matrices or NumPy arrays as data input.

In the following section, we present the validation of the results produced with AequilibraE’s IPF.