[sumo-user] Rerouting traffic with traci

[Iris B. <Iri...@br...>]

Dear SUMO list,

I am still busy with setting up some experiments with routing traffic. 
Originally I intended to do all the route-calculations myself, and then 
assign the new routes to vehicles using traci. However, SUMO also 
supports its own rerouting. Time to investigate!

As usual, I find myself getting stuck with details. So let's start: 
changeTarget(self, vehID, edgeID) reroutes the car to the chosen edge. I 
belief the routing algorithm used is the one selected with the 
--routing-algorithm option when initializing SUMO? Is it true that the 
Dijkstra's algorithm being used is not randomized? And does this reroute 
use distance or expected travel times?

Second, I wonder about rerouteTravelTime() and rerouteEffort(). The 
latter reroutes based on the amount of energy is required to finish the 
route I presume? Do I understand correctly that rerouteTravelTime() on a 
car traveling from A to B should give the same result as calling 
changeTarget() on a car standing at A heading to B?

Also, is there a way of having a car automatically using the routing 
algorithm when it gets added? Right now, I intend to use 
traci.vehicle.addFull(vehID, "ID of edge car is currently on", ...) and 
then changeTarget() to get this behaviour.

I expect that having SUMO do the routing will be quicker than calling 
all the edge's adapted travel times through traci, finding a route, and 
then sending that back through traci. However, I will try and run some 
benchmarks in the near future. Also, I am interested in how SUMO 
calculates the adapted travel time of an edge, though I expect I can 
find this on the wiki if I look for it more thoroughly.

As always, thank you very much for all your help,

Iris Bekker