Defining simulated travel times

During a simulation, travel times of the vehicles are measured for each edge in the abstract network graph. All vehicles that have passed through an edge, report the time they have spent on it on leaving. All travel times delivered on an edge during an evaluation interval are averaged and thus yield the travel time for this edge in this evaluation interval.

When an edge is congested and vehicles have spent more than one evaluation interval on an edge, these vehicles report it at the end of the evaluation interval. The vehicles report this even if they have not managed to leave the edge until the end of the evaluation interval. Thus Vissim gets information also from heavily congested edges, even if due to the congestion no vehicle reaches the end to report about the congestion.

The travel times measured in the current iteration are not used directly for path selection in the same iteration, but are adopted only in the following iterations. This behavior is useful because, for example, for a path selection on a Tuesday between 9:00 to 10:00 the relevant travel time is not on the same day between 8:00 to 9:00, but rather the travel time between 9:00 to 10:00 on the Monday before.

For the simulation of the experience growing with time, not only the travel time of the previous iteration is taken into account, but in particular the travel times of all previous iterations.

You can set the more distant measurements to have less influence. For this the following methods can be used:

• exponential smoothing of the travel times (Selecting exponential smoothing of the travel times)
• MSA (Method of successive averages) (Selecting the MSA method for travel times)