I am looking to implement a threaded version of GLPK in which each thread is just another instance of GLPK solving LP, but perhaps using different pivoting strategies or a different initial pivot. I do not want each thread to incur the cost of parsing or presolving the problem.
To do this, I wanted to essentially duplicate the contents at memory glp_prob *lp that is passed to glp_simplex. However, a simple cloning would not work as it seems like variables in glp_prob (i.e. void *cps, void *tree, DMP *pool) point to particular places in memory. I was essentially wondering if anyone could think of a painless way to do this. Thank you in advance. -- Parsiad Azimzadeh http://cs.uwaterloo.ca/~pazimzad _______________________________________________ Help-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-glpk
