The impedance for a path leg with the ride sharing system takes the following factors into account:
- Journey time
The journey time is made up of the ideal journey time in PrT and an aggregated detour factor for the passenger resulting from pooling with other passengers.
- Wait time
The aggregated wait time is the average time the passenger waits on this OD pair for the vehicle assigned to him/her.
- Variability of the supply
The variability of the supply is reflected in the fact that not every trip request is served in the same way. The trips on an OD pair can take 5 minutes or twice the time. The same applies to the wait time. This uncertainty is perceived as additional impedance regarding the path.
This uncertainty is most evident in the uncertainty of not being transported. If the rejection rate of trip requests is high for an OD pair, this must have a particularly strong effect on the impedance. This additional impedance p is determined in such a way that the resulting fall in passenger numbers (with regard to the DRT transport system) is approximately identical to the number of passengers not carried.
The total impedance of a DRT path leg is therefore calculated using this formula:
impi,j= x•di,j • tPrT i,j + y•wi,j + a •σ(di,j) + b • σ(wi,j) + c • p
where
|
d |
Mean detour factor |
|
tPrT |
Journey time of the PrT path |
|
w |
Mean wait time |
|
p |
Impedance resulting from the rejection rate |
|
σ |
Standard deviation of d and w |
In addition to this impedance, a buffer for a DRT path leg can be considered, which becomes effective when a conventional public transport path leg is connected. This buffer reflects the planning behavior of the passenger, who has to combine the variable supply of the ride sharing service with the fixed supply of the timetable.
If the timetable provides many regular services, the passenger will plan with a smaller buffer. The selected buffer is included in the search impedance of the DRT path leg.
The impedance of a connection that has DRT paths can change in the course of several iterations. Therefore, if the dominance criteria are applied naively, it could happen that such a connection is dominated in an early search iteration, although this would not have happened with the final impedance.
In order to prevent such a faulty dominance, you can specify factors for the DRT impedance for the minimum and maximum search impedance. A connection is therefore only dominated if it would also be dominated after the application of the minimum factor. Accordingly, a connection only dominates if it would do so even after applying the maximum factor.
The factors for the search impedance are applied in the same way to the cut off rules and to the preselection.
