The maximum or minimum of a linear function in several va-
raibles are calculated. The constraints are given as inequali-
ties ("less then"). The routine does not handle degeneracy or
solutions constrained to be natural numbers.
Rem. If such constraints are given and a deci-
mal number is the answer, one cannot simply
round off to nearest natural number. This will
not always give the optimal solution.
The algorithm used is the simplex method and it finds only
the minimum of an object function. Any problem can be writ-
ten as a minimum problem. If a maximum value of f(x) is
going to be found, one may simply change the sign and find
the minimum of -f(x).
7. Linear programming