For simplification often programs that can be used for minimization and maximization problems use internally only one sense of the optimization, e.g., maximization. Within a framework this strategy is dangerous, because if we access internal results, e.g., the reduced costs, from an application, we might misinterpret them. Therefore, ABACUS also works internally with the true sense of optimization. The value of the best known feasible solution is denoted primal bound , the value of a linear programming relaxation is denoted dual bound if all variables price out correctly. The functions lowerBound() and upperBound() interpret the primal or dual bound, respectively, depending on the sense of the optimization. An equivalent method is also used for the local bounds of the subproblems.