When solving a linear-programming problem using the simplex method, it is computationally efficient to select a small number, say 5, possible candidate vectors from which one would be chosen to enter the basis. The candidate set consists of columns with large (most negative or most positive) reduced costs, and the vector in this set that yields the largest change in the objective function is selected. Succeeding iterations only consider candidate basis vectors from the vectors that remain in the set that have properly signed reduced costs. When all vectors in the set are chosen or none can serve to change the objective function in the proper direction, a new set is determined. Partial pricing; Simplex method.