A Novena to St. Jude, or Four Edifying Case Studies in Mathematical Programming

In this author's opinion, the two most basic reasons for failure in the field of mathematical programming have nothing to do with modeling ability or technical expertise. The two reasons are: (1) Failure because the modeler does not realize that the method (mathematical programming) is a means to an end (the problem) rather than an end in itself, and (2) Failure because a wrong answer is extrapolated from what appears to be the right method. The case studies here have been selected because they seem to exemplify these two reasons rather well. Two were problems in which this author was personally involved; therefore the models have been altered somewhat to preserve the innocent (and the guilty) from the sneers of their fellow workers (who would, of course, never do anything like what is presented here). The other two case studies have been repeated to this author in various bars at various O.R. conventions for some years. The only thing the author has contributed in these cases is the considerable time it took to find out what really happened.