When determining a new variable to enter the basis by the simplex method, it is somewhat computationally inefficient to price out all nonbasic columns, as is the way of the standard simplex algorithm or its multiple pricing refinement. The scheme of partial pricing starts by searching the nonbasic variables in index order until a set of candidate vectors has been found. These vectors are then used as possible vectors to enter the basis, as is done in multiple pricing. After the candidate set is depleted, another set is found by searching the nonbasic vectors from the point where the first set stopped its search. The process continues in this manner by searching and selecting candidate sets until the optimal solution is found. Although the total number of iterations to solve a problem usually increases, computational time is saved by this type of pricing strategy.