The subject of mathematical modeling has expanded considerably in the
past twenty years. This is in part due to the appearance of the text by
Kemeny and Snell, "Mathematical Models in the Social Sciences," as well
as the one by Maki and Thompson, "Mathematical Models and Applica-
tions. " Courses in the subject became a widespread if not standard part
of the undergraduate mathematics curriculum. These courses included var-
ious mathematical topics such as Markov chains, differential equations,
linear programming, optimization, and probability. However, if our own
experience is any guide, they failed to teach mathematical modeling;
that is, few students who completed the course were able to carry out
the mod- eling paradigm in all but the simplest cases. They could be
taught to solve differential equations or find the equilibrium
distribution of a regular Markov chain, but could not, in general, make
the transition from "real world" statements to their mathematical
formulation. The reason is that this process is very difficult, much
more difficult than doing the mathemat- ical analysis. After all, that
is exactly what engineers spend a great deal of time learning to do. But
they concentrate on very specific problems and rely on previous
formulations of similar problems. It is unreasonable to expect students
to learn to convert a large variety of real-world problems to
mathematical statements, but this is what these courses require.