I would have approached this problem this way: set a lower bound say 3. Then we wish to prove there doesn't exist any n,m>=3 that satisfies n^m=m^n. This is much more easily proved using just double induction. Then you can just search for a case where n=2.