PTV VISSIM EXAMPLE DESCRIPTION
EVALUATION
GEH STATISTIC (UDA)
Short description
GEH statistic computed for node movements and data collections by means of user
defined attributes (UDA).
Objective
GEH statistic is a standard approach to compare two sets of traffic volumes (count data
vs. modeled data). It is defined as follows:
2 (𝑀𝑀−𝐶𝐶)2
𝐺𝐺𝐺𝐺𝐺𝐺 = � where M = modeled traffic volume, C = counted traffic volume
𝑀𝑀+𝐶𝐶
As Vissim offers a predefined function GEH (x, y) for use with formula UDAs, it is not
necessary to manually code the GEH function.
In this example, GEH statistic is calculated for node movements and for data collection
measurements. It also checks if the GEH value drops below a threshold value (here set to
5). In addition, a flow check is performed which checks if the difference between M and C
is below a threshold value.
UDAs & Formulas
There are several UDAs defined for various network object types:
For Links:
CountedData: ‘storage’ attribute for your own counted data for each link/connector.
The data can be edited in the link list and is used by the corresponding movements.
GEH: receives the resulting GEH value from each movement
For Movements:
CountsFromLinks: fetches the counted data of all the associated connectors
GEH: computes GEH results for each movement
PTV AG 2017 Page 1/2
GEHcheck: returns 1 if the GEH value is below the threshold value and 0 otherwise.
The type of this UDA is defined as Bool so that in a list the result is shown as a check
box rather than an integer value.
FlowCheck: returns 1 if the absolute or relative difference between counted and
modeled traffic volume is below the threshold values, and 0 otherwise. Defined as
Bool like GEHcheck.
For Nodes:
GEHlabel: Computes the percentage of movements which are below the thresholds
for GEH value and FLOW differences.
For Data collection measurements:
CountedData: ‘storage’ attribute for your own counted data for each data collection
measurement. The data can be edited in the data collection measurement list.
GEH: computes GEH results for each data collection measurement
GEHcheck: returns “YES” if the GEH value is below the threshold value, and “NO”
otherwise. This is another way of showing “true/false” information rather than bool as
used in GEHcheck for movements.
FlowCheck: Returns “YES” if the absolute or relative difference between counted and
modeled traffic volume is below the threshold values, and “NO” otherwise
Expressions used in the UDA formulas:
GEH(x;y): Computes the GEH statistic for x and y
IF(b;w;f): Conditional branch: Outputs w if condition b is true, and f otherwise
TableLookup(): Allows accessing attributes of other network object types
Visualization
As the results are visible only after a complete simulation run, we recommend
activating the quick mode while simulating.
The GEH results for movements are visualized through their corresponding connectors by
means of a link color scheme. The results for the data collection measurements are
visualized as a label color scheme. The following colors are used:
Green: GEH value is below 5 (condition fulfilled: no problem)
Yellow: GEH value is between 5 and 10 (attention, might need further investigation)
Red: GEH value above 10 (did not meet the GEH requirements, signalize problem)
The charts show the calculated GEH values for the links (= connectors representing the
movements in the node) and for the data collection measurements. In addition, the chart
layout GEH_Movements visualizes the data of the movements directly.
Limitations of this Example
All movements in the intersection are modelled by connectors which are placed
completely inside the node. Instead, if the model would have long links stretching over
several nodes, the traffic volume on such links could vary between the connectors. In
such a situation the evaluation approach point out in this example would not be
appropriate; neither could the movements be colored through a link/connector color
scheme.
PTV AG 2017 Page 2/2
