[ADMB Users] a conditional statement under PARAMETER_SECTION

[John Sibert]

If your code is generating compiler/linker errors you are probably 
referencing an undeclared variable (that is why you cannot put 
declarations inside a conditional block). So you if you really want to 
"ignore" a variable, you must completely banish if from your code. You 
can sometimes use conditional compilation for this purpose (ie #ifdef ). 
Otherwise, depending on what you really mean by "ignore", you could try 
something like this

DATA_SECTION

   // declare some variables for later use in the PARAMETER_SECTION
   number sel_par_min
   number sel_par_max
   init_int ignore_sel_par // read int from the .dat file to control 
estimation at run time;
                                           // 1 to ignore; 0 to estimate

  LOCAL_CALCS
   // set default values
   sel_par_min =  -5.0;
   sel_par_max =  5.0;
   sel_phz = 1; // or whatever phase you want
   // alternatively you could declare sel_phz to be an init_int and read 
it from the the .dat file
   if (ignore_sel_par)
   {
      // set sel_par_min and sel_par_max so their midpoint is the value 
you want for sel_par if you don't want 0.0
     sel_phz = -1; // forces the sel_par to NOT be active
   }
  END_CALCS

PARAMETER_SECTION
    init_bounded_number(sel_par_min, sel_par_max, sel_phz);

On 03/19/2011 09:14 AM, Saang-Yoon wrote:
> Hello, Steve.
> Thank you very much for your prompt response, which is helpful.  By
> the way, the conditional statement in your example is used below
> "init_bounded_matrix_vector sel_par".   Now if we want to ignore the
> parameter vector, "sel_par" by an option, then we may want to put the
> parameter declaration ("init_bounded_matrix_vector sel_par") under the
> optional statement, which would lead to a compile/link error.  I
> wonder about how you or others get around in "one" TPL file.  Again
> thank you.
> Saang-Yoon
>
>
> On Mar 19, 2:35 pm, Steve Martell<s.mart... at fisheries.ubc.ca>  wrote:
>> Sanng-Yoon,
>>
>> Statements such as this work fine for me in the PARAMETER_SECTION
>>
>>          //Selectivity parameters (A very complicated ragged array)
>>          init_bounded_matrix_vector sel_par(1,ngear,1,jsel_npar,1,isel_npar,-5.,5.,sel_phz);
>>          LOC_CALCS
>>                  //initial values for logistic selectivity parameters
>>                  //set phase to -1 for fixed selectivity.
>>                  for(int k=1;k<=ngear;k++)
>>                  {
>>                          if( isel_type(k)==1 || isel_type(k)==6 )
>>                          {
>>                                  sel_par(k,1,1) = log(ahat(k));
>>                                  sel_par(k,1,2) = log(ghat(k));
>>                          }
>>                  }
>>          END_CALCS
>>
>> On 2011-03-19, at 11:31 AM, Saang-Yoon wrote:
>>
>>> Hi, all.
>>> Whenever I put a conditional statement (e.g., if - else) under
>>> PARAMETER_SECTION, I run into an error.   I wonder whether ADMB is
>>> designed not to allow a conditional statement under PARAMETER_SECTION
>>> ever.  Sometimes I need to consider or ignore a parameter, depending
>>> on a model assumption.  If you were, how would you get around in
>>> ADMB?   Temporarily I make two TPL files (e.g., two EXE files) by the
>>> assumption.  But I suspect some of you may get around in "one" TPL
>>> file (i.e., "one" EXE file).
>>> Saang-Yoon
>>> _______________________________________________
>>> Users mailing list
>>> Us... at admb-project.org
>>> http://lists.admb-project.org/mailman/listinfo/users
>> _______________________________________________
>> Users mailing list
>> Us... at admb-project.orghttp://lists.admb-project.org/mailman/listinfo/users
> _______________________________________________
> Users mailing list
> Users at admb-project.org
> http://lists.admb-project.org/mailman/listinfo/users
>

-- 
John Sibert
Emeritus Researcher, SOEST
University of Hawaii at Manoa

Visit the ADMB project http://admb-project.org/