Correcting demand matrices

Using Matrix correction, you can adjust the demand in the origin-destination matrix to the count data available in numeric link attributes, vehicle class-specific subattributes of the attribute Count data or user-defined attributes. The results of an assignment must be saved to a Path file.

Examples of Matrix correction use cases:

  • You have one OD-matrix and assignment results of dynamic assignment that include paths and their volumes as well as a set of count data in link attributes. Using Matrix correction, you automatically adjust the OD-matrix to the new path volumes, creating values that are closer to the count data. The count data cover an integer multiple of the evaluation interval. The OD-matrix does not change when count data is collected.
  • You receive an assignment result that includes multiple vehicle classes and the respective OD-matrices. The count data is listed separately by vehicle class. You use Matrix correction successively for the individual OD-matrices.

Alternatively, you can perform Matrix correction via the COM interface.

Method used

Vissim uses the least squares method. The total of squares of the differences between count data and volumes and the total of squares of the differences between the original and corrected matrix values is minimized. The number of iterations is set to 1000. OD relations with a volume of ZERO are not adjusted. The values in the other cells of the matrix can be edited.

Using matrices with realistic values

The matrix correction is not suitable for generating a matrix with realistic values from a "dummy" with unrealistic values. You need a matrix from a demand model, for example from Visum. Use this matrix in Vissim for simulation runs with dynamic assignment until the model converges. Then you will have current paths. Alternatively, adopt the paths from a Visum assignment. Then run the matrix correction with your count data. From a mathematical point of view, the matrix correction always provides a result based on the method of least squares. However, this result is not automatically meaningful and realistic. Make sure you save meaningful matrix values and count data.

Notes:  

  • Avoid correcting demand matrices in a Vissim network that also uses trip chain files to model transport demand.
  • Avoid correcting demand matrices in a Vissim network with vehicles that have been added to the Vissim network via the COM interface.
  • You cannot undo Matrix correction. Before using Matrix correction, save a backup of the OD-matrix you want to use as a basis for matrix correction.
  • The assignment result may become inconsistent when you use the following elements:
  • dynamic routing decisions
  • real parking lots
  • route guidance vehicles
  • Matrix correction can calculate values that are not integer.
  • The OD-matrix must contain integer values. These correspond to the values of the assignment results in the Path file.