Purely in principle, it shouldn’t be hard to solve a Rubik’s Cube with a computer, right?
Our program would have three parts:
1. A model of the Rubik’s Cube, that is, some data structure that represents a cube state.
2. Some functions which can simulate turns of each side.
3. A solving procedure which takes a scrambled cube, tries every possible turn sequence, and stops when solved.
It is more efficient to travel along the Hamilton circuit and visit each state of the cube. When you reach the solved state you stop traveling.
Bruce Norskog's cycle would work, but it will require executing somewhere around 43,252,003,274,489,855,999 moves (at worst) to arrive at the solution to a given scramble. :)
