Hello all, I have the following problem: I have a parameter A with (3 indices) and a decision variable X also with 3 indices (same matrix form). The idea is as follows: Parameter A is the old matrix, X is the new matrix. Now I need to be sure that at most 10 changes are made. So I would like to have at most 10 different X[a,b,c] to A[a,b,c] for all a,b,c.
Now I was used to programs that could work with absolute values (SUM[a,b,c]: abs(A[a,b,c]-X[a,b,c]) <= 10), however GLPK is not able to work with absolute values of decision variables. Does anyone have an idea how to solve this? I tried introducing a Dummy: D[a,b,c] = 1 for A[a,b,c]<=X[a,b,c], 0 else but that didn't work out.... Hope someone can help me out! Regards, Timo -------------------------------------------------------------------- mail2web.com - Microsoft® Exchange solutions from a leading provider - http://link.mail2web.com/Business/Exchange _______________________________________________ Help-glpk mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-glpk
