Update: I did some tests to identify if there is a pattern in writing vehicles that are after the save state timestep. As a result it seems that writes all the loaded vehicles (until the save state timestep) minus the arrived instead of the (departed - arrived) + waiting.
Στις Κυρ 3 Δεκ 2023 στις 2:18 π.μ., ο/η manos kampitakis < [email protected]> έγραψε: > Update: > For case 1 if the same route file without route distributions is given > everything works properly (still are saved vehicles that have "depart" > after the save state timestep & the issue I mentioned for the traffic > lights). > Clarification: The route file with route distributions (from marouter) has > only one route within each routeDistribution of a vehicle. It seems there > is something wrong with marouter because regardless of the number of max > alternatives I give, it always returns route distributions with only one > route. > > Στις Παρ 1 Δεκ 2023 στις 11:25 μ.μ., ο/η manos kampitakis < > [email protected]> έγραψε: > >> Hi all, >> >> Bug 1: >> in marouter when the option --ignore-taz is activated, this is written >> (example) for a vehicle: >> <vehicle depart="28814.71" departLane="best" departSpeed="avg" >> fromTaz="551932" id="3745" toTaz="552814" type="car"> >> There is no "from" or "to". And the fromTaz and toTaz has the source and >> origin edges of the vehicle (interestingly no error is reported) when the >> route file is used. >> Note: The trip file that was given to marouter had both from, to and >> fromTaz, toTaz correctly written. >> >> Bug 2: >> It seems that there is a bug in sumo 1.18 in saveState (Traci-Libsumo) >> >> Note for all cases: >> Load state is with a sumo configuration where rerouting is activated and >> routes from marouter are used. >> >> Case 1: >> Save state -> route file (marouter - route distributions) & activated >> rerouting >> When I load the saved state the 2 following errors are printed: >> - Loading state from '...\warm_up_state_200.xml' failed, check whether >> SUMO versions match. Unknown route '!warm_up_10421#0' in distribution >> '!warm_up_10421'. >> - Storage::writeUnsignedByte(): Invalid value, not in [0, 255] >> >> Case 2: >> Save state -> route file (marouter - route distributions) & deactivated >> rerouting >> Everything works fine. >> >> Case 3: >> Save state -> trip file & deactivated rerouting >> Load state works properly but after some timesteps (different each time) >> sometimes traci throws a FatalTraciError and with libsumo the kernel dies. >> Note: if rerouting is deactivated in the load state phase it seems that >> it works (tested for 20 runs). >> >> Case 4: >> Save state -> route file (marouter - route distributions) & activated >> rerouting >> Load state works properly but after some timesteps (different each time) >> sometimes traci throws a FatalTraciError and with libsumo the kernel dies. >> Note: if rerouting is deactivated in the load state phase it seems that >> it works (tested for 20 runs). >> >> It seems to me that there must be a conflict between the rerouting >> devices of the saved state and the ones that are added afterwards (load >> state) (Case 2, 3, 4). >> >> Other issues: >> - In all cases info is saved about vehicles that enter the simulation >> (depart) after the save state timestep. For example, I save the state at >> timestep 28720 but this is written in my state file (see depart): >> <vehicle id="peak_hour_1036" type="car" depart="28806.90" ... >> >> - In the beginning of the simulation I activate for some traffic lights >> the programID "no_control" and for some others I set the Program to "off" >> (deactivate). In the state file if a traffic light has two programs both >> are written and not only the one that was set. As for the traffic lights >> that were deactivated ("off") this information is not kept, rather the >> default programID is written (i don't know if this is an >> intended behavior). For example: >> <tlLogic id="clusterJ830_J841" programID="control" phase="0" >> duration="10000"/> >> <tlLogic id="clusterJ830_J841" programID="no_control" phase="7" >> duration="2000"/> >> >> Best regards, >> Manos >> >
_______________________________________________ sumo-user mailing list [email protected] To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user
