Patrick, in a linear program there can't be any product of variables... The
term "w[i] * y[i,j,k]" which figures in the constraint concordance[j,k] does
not respect this. Try rewriting this constraint with other variables, maybe
adding some other constraints to represent the same idea of "w[i] *
y[i,j,k]" without multiplying variables.
[]s,
Andre
On 4/26/07, Patrick Meyer <[EMAIL PROTECTED]> wrote:
Hi,
As you will see, I'm quite a newbie in optimisation and glpk. I have
searched the archives, but I cannot find a solution to my problem.
I have a constraint which looks as follows:
concordance{j in X, k in X}: sum{i in N} (w[i] * y[i,j,k]) = c[j,k];
where w,y and c are defined as follows:
param c{j in X, k in X};
var w{i in N} >=0, <=1;
var y{i in N, j in X, k in X} binary;
y is a variable which depends on three dimensions, for each element of N
and each pair of X^2
glpsol stops on this constraint and says:
multiplication of linear forms not allowed
Does anybody know how to overcome this problem?
Here's the constraint written in latex:
\sum_{i=1}^n w_i y_i(a,b) = c(a,b) \forall i\in N, (a,b) \in X\times X
Thank you for any help!
Patrick
_______________________________________________
Help-glpk mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-glpk
--
Andre M M de Lima
"Ut omnes unum sint."
_______________________________________________
Help-glpk mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-glpk
