Assume we want to solve the so-called "Super-G" problem from Jean-Claude Constantin. The goal of the puzzle is to get the ring off the other components. Here is a picture of the problem:
The question here shall not be how to do that - that's trivial after you figured out. The question shall be, how we can write a simulation for this problem that seeks for the proper solution.
I can think of two completely different approaches, both of them non-trivial:
- May we find a way to represent all moves that we can do in an abstract way? The simulation would be trivial.
- Should we use a 3D simulation and use some random walk to move the ring? The set-up might be much simpler, but the simulation itself much harder, also because the line might be very hard to simulate.
We are not looking for a solution in the form of a code, but what approach should be taken.