I meant something different. Assume that 1. there is not enough space on the eastbound lane. 2. due to the no-block heuristic, the vehicle that wants to move from south to east stops in front of the intersection then as a consequence, all the other traffic from the south is still blocked by the vehicle that wants to move east.
However, if you would want to reroute the vehicle instead of waiting to turn east, then it would indeed be an improvement to stop before the intersection. (I've taken note of this issue here: https://github.com/eclipse/sumo/issues/6263). This leads to the question whether all the other vehicles in the eastbound jam should have been rerouted before turning east as well. To achieve this, you could activate automated rerouting ( https://sumo.dlr.de/docs/Demand/Automatic_Routing.html) regards, Jakob Am Sa., 16. Nov. 2019 um 06:52 Uhr schrieb Hao Zhou <[email protected]>: > Thank you Jakob. When you say ' If the eastbound vehicle were to stop > before entering the intersection', do you mean the '*no-block-heuristic' > cannot apply to stopped vehicles? Is there any workaround? I assume if we > check the available space of vehicle's target lane every time step, and > then decide whether or not to reroute them, it should also work for stopped > vehicles. Is it correct? * > > On November 15, 2019 at 6:24 AM, Jakob Erdmann <[email protected]> > wrote: > > Since the eastbound vehicle does not fit fully on the eastbound lane its > rear portion is treated as if it were still on the previous lane (incoming > from the south). Due to this it blocks the following traffic. > > The type of check that you mention is already implemented: > https://sumo.dlr.de/docs/Simulation/Intersections.html#junction_blocking > However, it is only active when the simulation is running with internal > links (to prevent cross traffic from being blocked). > The check also would not help in your scenario: If the eastbound vehicle > were to stop before entering the intersection, all the traffic following > behind it would still be blocked. > > regards, > Jakob > > Am Fr., 15. Nov. 2019 um 15:09 Uhr schrieb Hao Zhou <[email protected] > >: > >> Thank Jacob. Actually I already deleted the internal links. So it looks >> weird to me there exists such a vehicle in the end of the eastbound lane. >> >> To prevent this, can we work around by first checking the free space on >> target lane before letting vehicles enter the intersection? Do you think it >> is feasible? >> >> Best, >> Hao >> >> On Nov 14, 2019, at 04:08, Jakob Erdmann <[email protected]> wrote: >> >> There is no way to completely prevent this time of blockage where a >> vehicle cannot exit the junction due to downstream jamming. >> If your simulation is using internal links then you can set the option >> --ignore-junction-blocker <TIME> to work around it. >> However, it doesn't look like a gridlock to me. The westbound vehicles >> can still cross the intersection and this might allow the eastbound jam to >> clear. >> >> regards, >> Jakob >> >> Am Di., 12. Nov. 2019 um 18:42 Uhr schrieb Hao Zhou < >> [email protected]>: >> >>> Hey sumo-users, >>> Does anyone know how to prevent the following gridlock? Note that I >>> already omitted the internal links. But the misplaced vehicle on the >>> eastbound lane seems to block the northbound lane. The northbound vehicle >>> cannot move forward. >>> >>> How should I fix this? >>> >>> >>> >>> Thanks, >>> Hao_______________________________________________ >>> sumo-user mailing list >>> [email protected] >>> To change your delivery options, retrieve your password, or unsubscribe >>> from this list, visit >>> https://www.eclipse.org/mailman/listinfo/sumo-user >>> >> _______________________________________________ >> sumo-user mailing list >> [email protected] >> To change your delivery options, retrieve your password, or unsubscribe >> from this list, visit >> https://www.eclipse.org/mailman/listinfo/sumo-user >> >> _______________________________________________ >> sumo-user mailing list >> [email protected] >> To change your delivery options, retrieve your password, or unsubscribe >> from this list, visit >> https://www.eclipse.org/mailman/listinfo/sumo-user >> > _______________________________________________ > sumo-user mailing list > [email protected] > To change your delivery options, retrieve your password, or unsubscribe > from this list, visit > https://www.eclipse.org/mailman/listinfo/sumo-user > > _______________________________________________ > sumo-user mailing list > [email protected] > To change your delivery options, retrieve your password, or unsubscribe > from this list, visit > https://www.eclipse.org/mailman/listinfo/sumo-user >
_______________________________________________ sumo-user mailing list [email protected] To change your delivery options, retrieve your password, or unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user
