Originally posted by: SystemAdmin
The barrier method is a primal-dual method. That means, it simultaneously drives the primal and dual solution to feasibility. As a consequence, throughout the algorithm both vectors will usually be infeasible, and you reach primal and dual feasibility only at the very end.
I am not sure, but I think that this is true for barrier algorithms in general, not only the one implemented in CPLEX.
#CPLEXOptimizers#DecisionOptimization