RE: UrbanSim 3 vs. 4

Hello,

I would first like to say that we have a functional UrbanSim 4/OPUS model working (including the recent ELCM changes). We have prepared a database for one county in our region that we are using as a pilot project. We have successfully run several different scenarios for a 30 year regional development forecast. We are currently preparing for an additional one-county pilot project before moving on to our full, seven county, regional forecast. We are also working on integrating our travel model with UrbanSim, with the help of CUSPA and our travel model developers.

For us, the results of using UrbanSim 4 have been positive. I think it is important to remember that UrbanSim 4 is still in Beta version. So there are going to be a lot of changes to the structure of the code to deal with. It can be very frustrating trying to figure out why a model does not work. For me, the reason is usually because our local "package" of code does not change as the UrbanSim code changes. So whenever I update to a new build and try to run, I get runtime errors. I have found that the best way to solve this is to look at how the code in the PSRC package (which is part of the UrbanSim code release) is structured and make the appropriate changes to our local package. The documentation can also be helpful, especially the configuration stuff.

So, while we do not have problems related to the lack of a sample database since we have our own, we do experience problems running the code when there is a major change or we do something wrong, etc. I really think that given the differences between versions 3 and 4, everyone is a "new user." Since we have UrbanSim 4 running, I am willing to help out if I can. The more people and groups that are able to run the software, the better the situation is for everyone. And hopefully sooner rather than later we can have some good discussion about improving the results of our output, rather than just getting it up and running!

Peter

Peter J. McNally
Planning Analyst
Southeast Michigan Council of Governments
535 Griswold, Suite 300
Detroit, MI 48226
313.324.3442
mcnally_at_semcog.org

>>> "Joseph Nunez" <joseph.nunez_at_nieir.com.au> 10/29/2006 8:12 pm >>>
Hi All,

I have been working on Urbansim for quite a few weeks now. I have briefly
seen Urbansim 3 before concentrating on Urbansim 4 and it seems to me that
Urbansim 4 is a lot more flexible and there is more functionality than
urbansim 3.

It seems to me that all the new users, including me are having problems with
running all the models.

I have managed to make "RUN" work but not yet "estimate":

pre_scheduled_events
events_coordinator
land_price_model
household_transition_model
household_relocation_model
household_location_choice_model

A lot of the times, the problems associated with running the models just
comes down to having the proper database format and running the models in
sequence as, One output of the model is an input to another. (i.e. the
household_reloction_model outputs the hrm_index which inturn is used by the
household location choice model)

For the meantime before the people at Urbansim will finish creating a
revised database, I would be happy to send anyone my revised
eugene_baseyear......(its mainly the converted US3 and then some.) ...that
will work on the models listed above.

If anyone else has working models i.e. the new employment_location_choice
model or developer models and wouldn't mind sharing their code or database
or part their of...that would be great.

I think everyone who just started using urbansim4 are in the same boat,a
little cooperation between the community will go a long way.

Best Regards,

Joe

-----Original Message-----
From: Aaran Stent <aaran.stent_at_nieir.com.au>
To: "'users_at_urbansim.org'" <users_at_urbansim.org>
Date: Mon, 30 Oct 2006 12:30:00 +1100
Subject: RE: UrbanSim 3 vs. 4

> Hi Jen,
>
> I think you have summarised the problem very well. Opus and UrbanSim 4
> provide a great set of new features and will offer ongoing support.
> However the biggest issue with UrbanSim 4 is that it simply will not
> work out of the box like UrbanSim 3.
>
> I have had two developers working on UrbanSim 4 for several weeks now
> and they are yet to get the test database and all of the models in
> working order. When I first looked at UrbanSim in Feb 2006 I was able
> to have a running simulation that I could then experiment with in a
> single day.
>
> My developers have stated there are a lot of hard coded references to
> the University server causing a lot of things to break. There are also
> numerous differences in what the documentation states versus the source
> code. The team is stepping through correcting a lot of code instead of
> working on models and scenarios. We are planning to share our changes
> via our SVN server once we have a working model running.
>
> The Java version is your best option in my honest opinion. It provides
> a working system in just a few hours. UrbanSim 4 will not.
>
> PS. I don't mean to come down hard on UrbanSim 4 as it is such a huge
> step forward from version 3, it simply isn't ready for widespread use
> like UrbanSim 3 is.
>
> Aaran Stent
> National Economics
> aaran.stent_at_nieir.com.au <mailto:aaran.stent_at_nieir.com.au>
> 416 Queens Parade
> Clifton Hill VIC 3068
> Ph: 0419 227 262
>
>
>
> -----Original Message-----
> From: users-bounces_at_urbansim.org [mailto:users-bounces_at_urbansim.org]On
> Behalf Of Jennifer C. Duthie
> Sent: Saturday, 28 October 2006 1:44 AM
> To: users_at_urbansim.org
> Subject: UrbanSim 3 vs. 4
>
>
> Hi,
>
> I'm working on a research project for TxDOT looking at what type of
> land
> use model is best for each type of region (defined by size, resources,
> need, ...) depending on a variety of factors. I had been trying to get
> UrbanSim 4 running for awhile and was having some of the same problems
> as Zachary, so I just this week switched to UrbanSim 3.
> Although there are many advantages to UrbanSim 4 (as listed
> http://www.urbansim.org/news/index.shtml#opus), we're worried that
> there
> are still some bugs in the code and the documentation has not yet been
> ironed out. Also, it is very helpful to have the Eugene test database
> running for the earlier version. As we're looking for general
> conclusions about types of models, we feel this should suffice. My
> only
> concern is that the Java version is no longer being maintained, so if
> we
> do have a problem it may be more difficult to get it resolved.
>
> I just wanted to see if anyone on this list had any convincing
> arguments
> as to why we shouldn't use the Java version for our research...
>
> Thanks!
> Jen Duthie
> PhD Candidate
> Transportation Engineering
> University of Texas, Austin
> _______________________________________________
> Users mailing list
> Users_at_urbansim.org
> http://www.urbansim.org/mailman/listinfo/users
>
>
>
> _______________________________________________
> Users mailing list
> Users_at_urbansim.org
> http://www.urbansim.org/mailman/listinfo/users

_______________________________________________
Users mailing list
Users_at_urbansim.org
http://www.urbansim.org/mailman/listinfo/users
Received on Mon Oct 30 2006 - 07:40:40 PST