[Octave-bug-tracker] [bug #63858] fmincon fails in Octave 7.1.0, but not Octave 6.3

[anonymous]

URL:
  <https://savannah.gnu.org/bugs/?63858>

                 Summary: fmincon fails in Octave 7.1.0, but not Octave 6.3
                   Group: GNU Octave
               Submitter: None
               Submitted: Wed 01 Mar 2023 10:00:56 PM UTC
                Category: Octave Function
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: None
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
                 Release: 7.1.0
         Discussion Lock: Any
        Operating System: GNU/Linux
           Fixed Release: None
         Planned Release: None


    _______________________________________________________

Follow-up Comments:


-------------------------------------------------------
Date: Wed 01 Mar 2023 10:00:56 PM UTC By: Anonymous
Hi,

In the code below, I use fmincon for non-linear optimization. The code seems
to work in Octave 6.3, but hits the following error in 7.1.0: 

error: @<anonymous>: function called with too many inputs
error: called from
    @<anonymous>
    __dfdp__ at line 304 column 15
    fmincon>@<anonymous> at line 385 column 26
    __jacobian_constants__>@<anonymous> at line 135 column 11
    __octave_sqp_wrapper__>@<anonymous> at line 61 column 19
    sqp at line 364 column 7
    __octave_sqp_wrapper__ at line 76 column 60
    fmincon at line 451 column 24
    test_fmincon at line 35 column 22

Do you know what might be the problem? Does the functionality of fmincon
change with newer versions of Octave?


clear
pkg load optim

function alpha=alpha_poly(beta,mgn,msn,Lpn)

betapol=[beta;0];
K=length(beta);
mgn=mean(mgn); msn=mean(msn); Lpn=mean(Lpn,1);

  Hn=0;
for k=1:K
       j=K+1-k;
   Hn=Hn+beta(j)*((msn+log(Lpn)).^k-msn.^k);
endfor
Hn=log(0.01*sum(exp(Hn).*Lpn,2));

alpha=mgn-msn-polyval(betapol,msn)-Hn;
endfunction

function S=objfun_g(beta,mgn,msn,Lpn,param)
A=param(1); B=param(2);
alpha=A+B*beta;
S=alpha-alpha_poly(beta,mgn,msn,Lpn);
S=S'*S;
endfunction

mg = 8.2141
ms = 7.3109
Lps =[ 0.061017      0.065218 0.076858 0.088419 0.099907 0.11133   0.12269  
0.13399                  0.14525   0.15646   0.16762   0.17876   0.18987  
0.20095   0.21201   0.22306                  0.23409   0.24513   0.25616  
0.2672      0.27825   0.28931   0.30039   0.3115                  0.32264  
0.33381   0.34502   0.35628   0.36759   0.37895   0.39037   0.40187           
      0.41343   0.42508   0.43681   0.44863   0.46056   0.47258   0.48473  
0.49699                  0.50938   0.52191   0.53458   0.54741   0.5604     
0.57357   0.58692   0.60047                  0.61423   0.6282      0.64241  
0.65687   0.67159   0.68659   0.70189   0.71751                  0.73346  
0.74976   0.76645   0.78354   0.80107   0.81905   0.83753   0.85654           
      0.87612   0.8963      0.91715   0.93869   0.961        0.98414   1.0082 
    1.0332                  1.0593      1.0865      1.115        1.1449     
1.1764      1.2096      1.2447      1.2819                  1.3216      1.3639
     1.4094      1.4585      1.5117      1.5697      1.6334      1.7039       
          1.7826      1.8714      1.9729      2.0907      2.2299      2.3985  
   2.6092      2.8842                  3.2668      3.8574      4.9767     
14.11                          ]
params = [1.2235; -6.9294]
b0 = .2564
alpha_star=[]; beta_star=[];
options=optimset('Algorithm','octave_sqp');

[beta_star,f_star,e]=fmincon(@(beta)
objfun_g(beta,mg,ms,Lps,params),b0,[],[],[],[],-1,10,[],options);








    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?63858>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/