You want to get the shortest path to the longest path -> indicates you can try BFS.
The path can't include the cycle -> you should storage the path have passed what vertices(when you do BFS).
The complexity: you can see the path don't include the cycle means the path can pass every node only one times. So the solution space is O(n!) and BFS may visit every path in the solution space in the worst case. So the complexity is O(n!).
You may be disappointed with that result. Okay, your problem is a famous problem that have been well studied. You can read this paper:
Finding the K Shortest Loopless Paths in a Network
Jin Y. Yen
Or the wikipedia pages for a intuitive view for the K-th shortest path problems.