3 ms·
I love how with a little experimentation and no prior knowledge of these algorithms I was very quickly able to produce paths that produced sub optimal pathing w
by hartror 15y ago
I love how with a little experimentation and no prior knowledge of these algorithms I was very quickly able to produce paths that produced sub optimal pathing with each of the algorithms. It really highlights how bad relative to humans our algorithms are at this sort of thing.
This is a fantastic learning tool in many ways.
- orangecat 15y agoHmm, A-star is supposed to be optimal. Can you describe how you tricked it? Edit: never mind, I didn't see all the options. Using Manhattan distance when diagonal moves are allowed causes the heuristic to be invalid, so you can make it take an L path instead of a diagonal if the diagonal path would force it to move slightly backwards at the start.