primal heuristic that uses dualvalues for successive switching variable values
This heuristic tries to find solutions by taking the LP or NLP, rounding solution values, fixing the variables to the rounded values and then changing some of the values.To determine which variable is changed we give each variable a ranking dependent on its dualvalue. We work with a transformed problem that is always feasible and has objective = 0 iff the original problem is also feasible. Thus we cannot expect to find really good solutions.
Definition in file heur_dualval.h.
|SCIP_EXPORT SCIP_RETCODE||SCIPincludeHeurDualval (SCIP *scip)|
|SCIP_EXPORT SCIP_RETCODE||SCIPapplyHeurDualval (SCIP *scip, SCIP_HEUR *heur, SCIP_RESULT *result, SCIP_SOL *refpoint)|