From: Simon S. <sim...@ar...> - 2012-01-24 10:26:30
|
Hi Jakob, I updated the wiki to show the new option name. Thanks for the hint. (By the way, how often is the web doc updated by the wiki content?) Originally, we wanted the output of ActivityGen to be a trip file. However at this time, a trip could not contain intermediate stops, only the start and the end point. So we decided to write a broken route file. This is not nice though. I guess Daniel mentioned at this time that intermediate stops could come some when in the future. Do you know more about it? Will such an extended trip definition be possible? I would change the output of ActivityGen then. Best regards, Simon Am Dienstag, den 24.01.2012, 08:39 +0100 schrieb Jakob Erdmann: > Hello, > I'm afraid you were hit by a known bug in the --remove-loops code > [https://sourceforge.net/apps/trac/sumo/ticket/335]. > The goal of removing loops is to avoid unneeded detours. Unfortunately > the pruning is a bit overambitious and may remove needed edges, thus > invalidating routes. Nothing bad will happen if you leave that option > switched off, though. > > --continue-on-unbuild (which is deprecated and now called > '--ignore-errors') keeps running even if some routes could not be > built (i.e. if you graph is unconnected). > > regards, > Jakob > > 2012/1/23 Conrad (alec) Gosse <al...@th...>: > > Hello, > > > > Thank you for looking into this. After examining your commands below, I was also using two more options for duarouter: --remove-loops --continue-on-unbuild > > When I removed these, things work as expected. I believe I copied those options from one of the examples on the website since they sounded like a good idea. It appears that when some routes have loops removed they aren't valid any more, which must be the source of fictitious departure positions. > > > > What happens when I don't remove loops in routes? > > > > Best, > > Conrad (Alec) Gosse, MUEP > > Doctoral Candidate: Civil & > > Environmental Engineering > > University of Virginia > > Charlottesville, VA > > > > On Jan 23, 2012, at 9:14 AM, Simon Siemens wrote: > > > >> Dear Conrad, > >> > >> I took your files and executed the following commands on them. But no > >> errors appeared, also no warning related to your problem. Do I do > >> something wrong or differently from you? (I used the latest version from > >> SVN.) > >> > >> activitygen --net-file EPMS.net.xml --stat-file EPMS.stat.xml > >> --output-file EPMS.tmp.rou.xml > >> > >> duarouter --net-file EPMS.net.xml --route-files EPMS.tmp.rou.xml > >> --output-file EPMS.rou.xml --repair > repair.log 2>&1 > >> > >> grep -v "Repaired route of vehicle" repair.log (To see whether > >> some errors except for repairing appear. No errors appeared.) > >> > >> sumo --net-file EPMS.net.xml --route-files EPMS.rou.xml --end > >> 20000 > sumo.log 2>&1 > >> > >> grep -i warning sumo.log | grep -vi teleporting (To see whether > >> some warings except of teleporting vehicles appear. None > >> appeared.) > >> > >> > >> Do you have any idea, why I cannot reproduce your errors with the above > >> procedure? > >> > >> Best regards, Simon > >> > >> > >> > >> > >> Am Donnerstag, den 19.01.2012, 11:03 -0500 schrieb Conrad (alec) Gosse: > >>> Hello, > >>> > >>> I've attached the files I was having trouble with. Please let me know if something else seems amiss. I've just started with SUMO and certainly don't have everything perfect on my end. I think this might explain some of the bizarre looking travel patterns I've seen watching the GUI view of my network under simulation though. > >>> > >>> Are you effectively maintaining activitygen? > >>> > >>> Thanks, > >>> Conrad (Alec) Gosse, MUEP > >>> Doctoral Candidate: Civil & > >>> Environmental Engineering > >>> University of Virginia > >>> Charlottesville, VA > >>> > >>> On Jan 19, 2012, at 2:26 AM, Simon Siemens wrote: > >>> > >>>> Dear Conrad, dear Joerg, > >>>> > >>>> I try to follow the postings about ActivityGen and to fix bugs. Sorry if > >>>> I did not see a related help request. > >>>> > >>>> Conrad, could you please send me your population file and your network, > >>>> so I can test it and fix the bug? > >>>> > >>>> Thanks, Simon > >>>> > >>>> > >>>> Am Mittwoch, den 18.01.2012, 23:22 +0100 schrieb Joerg Schweizer: > >>>>> Hi Conrad, > >>>>> I also ran into trouble with activitygen, for different reasons however. > >>>>> It appears the developers of this tool (not the DLR) are no longer > >>>>> actively supporting it. > >>>>> > >>>>> So I am also planning to develop my own activity generator in Python, > >>>>> also because I would like to know *exactly* how the demand is generated > >>>>> (which models, assumptions, parameters). > >>>>> There are already a couple of software packages out there, but also not > >>>>> so well documented (unless you read the source). > >>>>> > >>>>> There are also efforts to combine the activity generator of MatSim with > >>>>> SUMO. I'd use the population and facilities specification xml files used > >>>>> in Matsim to generate the routes with a Python script, but I'm not there > >>>>> yet. I need first to extract facilities from Openstreet, just to see how > >>>>> good it works... > >>>>> > >>>>> Any thoughts? > >>>>> > >>>>> Joerg > >>>>> > >>>>> > >>>>> On 01/18/2012 10:53 PM, Conrad (alec) Gosse wrote: > >>>>>> Hello, > >>>>>> > >>>>>> I am using NETGEN to produce a spider network and running into trouble using ActivityGen to generate demand and ultimate produce routes. Specifically, many of the temporary routes produced by activitygen have vehicles departing from position values larger than the length of the edge in question. This situation is carried through the duarouter --repair step and results in numerous similar SUMO errors, e.g.: > >>>>>> Error: Vehicle 'h22958c1:1' will not be able to depart at the given position! > >>>>>> > >>>>>> The only thing that I'm generating with my own code in this process is the land use, which consists of many entries similar to this in the stats file: > >>>>>> <street edge="1/1to1/2" population="0.171651" workPosition="0.817595" /> > >>>>>> > >>>>>> I've just recompiled from SVN sources; why would activitygen originate trips from positions past the limits of an edge? > >>>>>> > >>>>>> Best, > >>>>>> Conrad (Alec) Gosse, MUEP > >>>>>> Doctoral Candidate: Civil& > >>>>>> Environmental Engineering > >>>>>> University of Virginia > >>>>>> Charlottesville, VA > >>>>>> > >>>>>> > >>>>>> ------------------------------------------------------------------------------ > >>>>>> Keep Your Developer Skills Current with LearnDevNow! > >>>>>> The most comprehensive online learning library for Microsoft developers > >>>>>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, > >>>>>> Metro Style Apps, more. Free future releases when you subscribe now! > >>>>>> http://p.sf.net/sfu/learndevnow-d2d > >>>>>> _______________________________________________ > >>>>>> sumo-user mailing list > >>>>>> sum...@li... > >>>>>> https://lists.sourceforge.net/lists/listinfo/sumo-user > >>>>> > >>>>> > >>>> > >>>> > >>>> > >>>> ------------------------------------------------------------------------------ > >>>> Keep Your Developer Skills Current with LearnDevNow! > >>>> The most comprehensive online learning library for Microsoft developers > >>>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, > >>>> Metro Style Apps, more. Free future releases when you subscribe now! > >>>> http://p.sf.net/sfu/learndevnow-d2d > >>>> _______________________________________________ > >>>> sumo-user mailing list > >>>> sum...@li... > >>>> https://lists.sourceforge.net/lists/listinfo/sumo-user > >>> > >> > >> > >> > >> ------------------------------------------------------------------------------ > >> Try before you buy = See our experts in action! > >> The most comprehensive online learning library for Microsoft developers > >> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, > >> Metro Style Apps, more. Free future releases when you subscribe now! > >> http://p.sf.net/sfu/learndevnow-dev2 > >> _______________________________________________ > >> sumo-user mailing list > >> sum...@li... > >> https://lists.sourceforge.net/lists/listinfo/sumo-user > > > > > > ------------------------------------------------------------------------------ > > Try before you buy = See our experts in action! > > The most comprehensive online learning library for Microsoft developers > > is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, > > Metro Style Apps, more. Free future releases when you subscribe now! > > http://p.sf.net/sfu/learndevnow-dev2 > > _______________________________________________ > > sumo-user mailing list > > sum...@li... > > https://lists.sourceforge.net/lists/listinfo/sumo-user |