> I'm writing you because, Using Glpk I found it useful to add a method : > > void glp_load_matrix_from_matrix(glp_prob *lp,double **matrix) > > because the the other input fuctions where so painful I didn't even > succeed in using them without seg faults, on a complex problem I was > computing. I understand thoses functions are very useful when the matrix > has many zeros, however I think it would be useful to actually embeed > both of these fonctions. > > I can of course give you the code I've writen, even though , I guess > it's no problem for you to write it very quickly, if you decide it > should be added to the library. >
Thank you for your suggestion. The problem is that there exist many different formats (data structures) to represent matrices, so implementing all such formats would be impractical. Do you mean some dense row-wise or column-wise format, where a matrix is coded as a plain one-dimensional array, or what? Andrew Makhorin _______________________________________________ Help-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-glpk
