Solution to part of the 1999 midterm exam:

4) FALSE, for uncertain domains the project axioms can either not be written,
   or will be wrong sometimes, or will be too weak - execution-time monitoring
   is crucial.

6) FALSE, rotation will cause non-linear distortion in the configuration
   space, and the description of free-space in the configuration space is
   no longer polygonal.


7) FALSE, hierarchical planning is at least NP-hard.

8) FALSE, because finding an optimal plan without a plan library, especially
   with real-valued variables, takes MUCH more than O(n*n) - in fact MAY
   be UNDECIDABLE !


9) Need to know the transition matrix A to know whether the system
   is observable. (Extra information was given during the actual exam).

10) TRUE, we did not even specify the dynamics of the system, so
    there is no way to tell if it is controllable.