We are thinking of using the CarLayer as the holder of all RoadGraph nodes and connect it with all the other layer based on geometrical criteria to simplify the definition of the demand.
Proposition:
In this context, if we delete the RoadGraph and allow all the Layers to hold the geometry. It should :
simplify our graph structure (Only one type of graph)
simplify further optimization
decrease the complexity for a mnms user to define its own mobility graph
But:
Mid refactoring of the graph code base
Implement new method to define zones
No more shared links between layers => No microscopic simulation
Example
If we consider a simple mobility graph with 4 nodes of CarLayer and 3 nodes of BusLayer located on the same road we can represent it as follow:
The two layers are connected following rules on distances between nodes and potential exclusion between layer (ie I cant take a Bus and then my personal Car for example)
Context
We are thinking of using the CarLayer as the holder of all RoadGraph nodes and connect it with all the other layer based on geometrical criteria to simplify the definition of the demand.
Proposition:
In this context, if we delete the RoadGraph and allow all the Layers to hold the geometry. It should :
But:
Example
If we consider a simple mobility graph with 4 nodes of CarLayer and 3 nodes of BusLayer located on the same road we can represent it as follow:
The two layers are connected following rules on distances between nodes and potential exclusion between layer (ie I cant take a Bus and then my personal Car for example)