1)
- It is better to use a set for the resources (dollars, hours), rather
than have two constraints. This allows the addition of more resources in
the future. Also, the RHS should be a parameter of size 2.

- Only a fraction of each investment can be made. Therefore the decision
variable has an upper bound of 1. This has to appear in the graph as
well. It changes the RHS perturbation analysis as well as the plot of
z(b_1). Many people did not use the upper bound and so much of the work
done for this problem was wrong. In addition, the techniques used were
often wrong. Both parts 3 and 4 can be done using the geometry of the
feasible set from part 2.

2)
- Most people failed to notice that this is a minimization problem

3)
- Generally well done.
- The most common mistake is to have one variable for a pattern-flight pair
  without making sure that the flights covered by a given pattern are 
  either all used or all unused.

4)
- Many people forgot to quote the theorem they used in proving optimality
- Nobody got part 3