non linear optimization returns a "Too many output arguments."

Hello all,

I am using the genetic algorithm to solve a MINL problem, I am calling the non-linear-constraints as:

nonlcon = @non_linear_constraint_all_ga;

I have around 60 constraints, so nonlcon supposes to be a vector of 60 elements..

the called function is as follows :

 function c1=non_linear_constraint_all_ga(x)

 [~ ,c1,~]=objective_finction_and_non_linear_contsrains_ga(x);

 end

i am taking only the second output from the BIG function "objective_finction_and_non_linear_contsrains_ga"

function [y ,c1,ceq1]=objective_finction_and_non_linear_contsrains_ga(x)

for (j=1 :Ns+2:(2+Ns)*Nc)

     for (i=0:Ns-1)
          k=i+1;  
          c1(1,m) = x(1,j+2+i).*sqrt(w((i+k),1)-x(1,j));
          m=m+1;
     end

    end
end

I keep getting the following error :

Too many output arguments.

Caused by:
  Failure in initial user-supplied nonlinear constraint function evaluation.

Any idea ?

best regards,

Answer 1

[Matt J]

The nonlinear constraint function must be equipped to return 2 arguments if requested, one for inequality and one for equality constraints.

 function [c1,ceq1]=non_linear_constraint_all_ga(x)

   [~ ,c1,ceq1]=objective_finction_and_non_linear_contsrains_ga(x);

 end

[Amani mahdy]

Thank you so much, IT WORKS =D

but at the end, nonlcon will equal to c1 since it is the first output , isnt it ?

[Matt J]

No, ga will always request both c1 and ceq1 as outputs. However, if it sees that ceq1 is empty, it will realize that you have no equality constraints.