Path search finds only the best possible path in each interval

In Vissim it is assumed that not all drivers use only the best route from one parking lot to another, but that traffic is distributed across all known paths. For this, it would be useful to know the n best paths for each origin-destination relation. There are, however, no efficient methods to directly calculate the n best paths in context of a traffic assignment in a useful way.

The shortest path search finds the best path for each origin-destination pair.

Therefore, in each iteration of a simulation, the shortest path search of Vissim searches for the best path for each origin-destination relation. Due to the fact that over the course of an iteration the traffic situation and therefore the travel time on the edges changes until convergence is reached, different best paths can result in the iterations. As long as Alternative path search is not activated, the shortest path search carried out by Vissim never results in more than one best path for an OD pair.

Path file *.weg saves each best path.

All found paths, which qualify as the best paths in an iteration, are collected in Vissim and saved in the path file *.weg. These paths are then available for the following iterations.

Best path based on generalized costs

The criteria for the "best" path are the generalized costs. Due to the fact that the weighted coefficient for the generalized costs depends on the vehicle type, different best paths can be found for different vehicle types.

Path search at the beginning of each evaluation interval

The path search takes place at the beginning of the evaluation interval and uses the expected generalized costs which were determined for this evaluation interval in the previous iterations.

First simulation run uses path length

Because the first iteration does not yield any travel time information from the previous simulation, the length of the path [m] is used.

Default travel times for edges not yet used

For the following iterations, Vissim no longer uses path lengths, but enters a fictitious travel time of 0.1 seconds for edges not yet used by vehicles. This results in the use of paths with unused edges to appear attractive when searching according to route. It may be possible that only a few useful paths are found in the initial iterations. However, the collection of known paths (for which travel times were measured and generalized costs calculated) will grow more quickly in the path collection, if drivers are encouraged to try out unknown paths.

Weighting of travel distance helps avoid detours

This "Eagerness to experiment" of the driver may be influenced by a weighting of the distance traveled in the generalized cost functions. This results in long detours being avoided. Generally it is an advantage to find as many paths as possible. When unrealistic paths are found, these can be discarded in a later iteration. This can be defined in the options for path searches (Influencing path search and path choice).

Alternative path search

Optionally, you can carry out an additional search for Alternative path search with stochastic modifications of the edge evaluations or shortest path price increases (Performing an alternative path search).