Re: [sumo-user] Rerouting traffic with traci

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

Hi Jakob,

Thank you for your tips. I have looked both at the routing using devices 
and duarouter. Is my assumption that the devices use a randomized 
dijkstra correct? Or do they spread out over the different routes in a 
different manner. I also have some questions regarding DUAROUTER, but 
should I start a new threat for that?

My goal is to compare routing with a shortest-path (or quickest-path) 
algorithm, where a car just picks the route best for itself, with 
routing using some real time centralized router which should aim at 
decreasing the global travel time. At least, that was the idea I had 
when I started... Making a smart router looks more difficult by the second.

Kind regards,
Iris


On 10/10/16 13:44, Jakob Erdmann wrote:
> Hello,
> - changeTarget uses the router configured with --routing-algorithm, 
> that one is not randomized. it uses travelTime. The travel times 
> obtained as described here: 
> http://sumo.dlr.de/wiki/Simulation/Routing#Travel-time_values_for_routing
> - rerouteTravelTime() works the same way but uses the current 
> travelTimes (without any smoothing)
> - rerouteEffort() is currently broken 
> (http://sumo.dlr.de/trac.wsgi/ticket/2580)
>
> For automatic routing I highly recommend this: 
> http://sumo.dlr.de/wiki/Demand/Automatic_Routing.
>
> regards,
> Jakob
>
>
> 2016-10-07 17:51 GMT+02:00 Iris Bekker <Iri...@br... 
> <mailto: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
>
>
>     ------------------------------------------------------------------------------
>     Check out the vibrant tech community on one of the world's most
>     engaging tech sites, SlashDot.org! http://sdm.link/slashdot
>     _______________________________________________
>     sumo-user mailing list
>     sum...@li...
>     <mailto:sum...@li...>
>     https://lists.sourceforge.net/lists/listinfo/sumo-user
>     <https://lists.sourceforge.net/lists/listinfo/sumo-user>
>
>