We saw that planning graphs can handle only propositional actions. What if we want to use planning graphs for a problem with variables in the goal, such as At (P1, x) ^ At (P2, x), where x ranges over a finite domain of locations? How could you encode such a problem to work with planning graphs?