[Developers] Fwd: method set_initial_values(...) not working for param_init_number_vector

William Stockhausen - NOAA Federal william.stockhausen at noaa.gov
Mon Jun 16 11:13:39 PDT 2014


Here's Dave's solution to my problem (for some reason he wasn't able to
post to the developers list). His full solution is attached.

---------- Forwarded message ----------
From: dave fournier <davef at otter-rsch.com>
Date: Mon, Jun 16, 2014 at 10:00 AM
Subject: Re: [Developers] method set_initial_values(...) not working for
param_init_number_vector
To: William Stockhausen - NOAA Federal <william.stockhausen at noaa.gov>,
developers <developers at admb-project.org>



It appears that this function was some duped code which was never finished
properly.
I think it should look like this at the top of the file model48.cpp.

There may also be other set_initial_value functions which were not finished.

 void param_init_bounded_number_vector::set_initial_value(const
double_index_type& _it)
 {
   int mmin=indexmin();
   int mmax=indexmax();
   for (int i=mmin;i<=mmax;i++)
   {
     (*this)(i)=ad_double(_it(i));
   }
 }




On 06/16/2014 07:08 AM, William Stockhausen - NOAA Federal wrote:

Hi developers,

 I'm running ADMB 11.1 on MacOSX and Windows 7. I'm trying to set the
initial values for a param_init_number_vector (p, say) from a dvector (vls,
say) in a function using the following:

 p.set_initial_value(vls);

 The actual function is:

 void setInitVals(BoundedNumberVectorInfo* pI,
param_init_bounded_number_vector& p){
    int np = pI->getSize();
     if (np){
        dvector vls = pI->getInitVals();
        p.set_initial_value(vls);
        cout<<"vls = "<<vls<<endl;
         cout<<"p   = "<<p<<endl;
     }
}

 where pI is a pointer to an object whose method getInitVals() returns a
dvector.

 Everything compiles fine, but when I run the code,  the initial values for
p are not equal to vls (they're equal to the default initial values) after
running setInits.

 Fortunately this caused my code to produce some really strange results so
I realized there was a problem. Using the ADMB website, I tracked down the
following:

The 'set_initial_value' method for param_init_bounded_number_vector is
defined in model45.cpp as

 00008 00009 <http://www.admb-project.org/documentation/api/classparam__init__bounded__vector__vector.html#ad696a4660a01ad50266fcbafa7d3beb7>
void param_init_bounded_vector_vector::set_initial_value
<http://www.admb-project.org/documentation/api/classparam__init__bounded__vector__vector.html#ad696a4660a01ad50266fcbafa7d3beb7>(00010
  const double_index_type
<http://www.admb-project.org/documentation/api/classdouble__index__type.html>&
_it)00011  {00012     it
<http://www.admb-project.org/documentation/api/classparam__init__bounded__vector__vector.html#a106e1b4f5049a08386645afbb658384b>=new
double_index_type
<http://www.admb-project.org/documentation/api/classdouble__index__type.html>(_it);00013
 }00014


 and the conversion of a dvector to double_index_type is defined in
doublind.cpp as

00057 00062 <http://www.admb-project.org/documentation/api/classdouble__index__type.html#a70db7015028be71be42cbc4c37b59319>
double_index_type::double_index_type
<http://www.admb-project.org/documentation/api/classdouble__index__type.html#a2265b2e72fbd497b8cd4d4688398957a>(const
dvector <http://www.admb-project.org/documentation/api/classdvector.html>&
x)00063   {00064     p
<http://www.admb-project.org/documentation/api/classdouble__index__type.html#ad1bebc88ffff9232f7502f2c764ee3f2>
= new dvector_index
<http://www.admb-project.org/documentation/api/classdvector__index.html>((const
dvector <http://www.admb-project.org/documentation/api/classdvector.html>&)(x));00065
  }00066


 but the constructor for the dvector_index uses the copy constructor for
dvector

 00184 00189 <http://www.admb-project.org/documentation/api/classdvector__index.html#aba2c20e6f7432a5b4441e828427ce6e7>
dvector_index::dvector_index
<http://www.admb-project.org/documentation/api/classdvector__index.html#aba2c20e6f7432a5b4441e828427ce6e7>(const
dvector <http://www.admb-project.org/documentation/api/classdvector.html>&
v) : dvector <http://www.admb-project.org/documentation/api/classdvector.html>(v)00190
{00191 }00192


 which makes a shallow copy of v. Not sure why this doesn't work, but it
doesn't.

 My immediate issue is that I want to put setInitVals(...) in a code
library--i.e., not in the tpl code and the usual methods for setting the
initial values (i.e., simple assignment by element) are private and only
work when the function is defined inside the tpl.

 I've also tried the set_initial_value(...) method in the tpl code but it
doesn't work there either. Setting the initial values element-by-element in
the tpl code does work, so the problem is not on my end (e.g., the dvector
vls is all 0's).

 Hope this is helpful.

 Buck Stockhausen
  ***************************************************
* Dr. William T. Stockhausen                      *
***************************************************
* Resource Ecology and Fisheries Management       *
* Alaska Fisheries Science Center                 *
* National Marine Fisheries Service               *
* National Oceanic and Atmospheric Administration *
* 7600 Sand Point Way N.E.                        *
* Seattle, Washington 98115-6349                  *
***************************************************
* email: William.Stockhausen at noaa.gov             *
* voice: 206-526-4241 fax: 206-526-6723           *
* web  : http://www.afsc.noaa.gov                 *
***************************************************
All models are wrong, some are useful.--G.E.P. Box
Beware of geeks bearing equations.    --W. Buffett
***************************************************
Disclaimer: The opinions expressed above are personal
and do not necessarily reflect official NOAA policy.




_______________________________________________
Developers mailing
listDevelopers at admb-project.orghttp://lists.admb-project.org/mailman/listinfo/developers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/developers/attachments/20140616/27cc6868/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: model48.cpp
Type: text/x-c++src
Size: 6828 bytes
Desc: not available
URL: <http://lists.admb-project.org/pipermail/developers/attachments/20140616/27cc6868/attachment-0001.cpp>


More information about the Developers mailing list