Abstract This article comments on the research paper "Management Decision and Integer Programming," which was used to illustrate the dangers of attempting to round a linear programming solution to obtain an integer answer. A number of examples have appeared in the literature to illustrate the dangers of rounding, the example in the paper is the only one that is not only small and easily understood but which illustrates a practical problem, contains a large number of rounding alternatives, and cannot be rounded to an optimal integer solution. By contrast, the previous examples which share property have been "textbook cases" involving few rounding possibilities, and thus may reasonably be questioned as to whether they truly illustrate that rounding can constitute a serious computational impediment in the real world. Nevertheless, it remains true that there are ways to round the LP solution to a feasible integer solution.
Fred Glover (Sat,) studied this question.