>>330
Here's a classic valve paper that combines A* with source engine's navmesh system and a method to smooth the jagged path: https://steamcdn-a.akamaihd.net/apps/valve/2009/ai_systems_of_l4d_mike_booth.pdf
If you look at the navmesh, there's actually not very many nodes in comparison to the grid in that youtube video. :>

I remember this site was a pretty useful resource when I was implementing A* for the first time:

https://www.redblobgames.com/pathfinding/a-star/introduction.html

>>351
Well, more presentation slides than paper

>>353
Djirska a shit

While A* is one of the fastest algorithms to find a path from A to B, it doesn't guarantee that it is the shortest path. Dijkstra on the other hand guarantees the shortest path, but takes much more time to find it.

>>384
This isn't true, the path found by A* is guaranteed to be optimal so long as the heuristic used is "admissible", meaning it never in any case over-estimates the distance to the goal. In practice, in-admissible heuristics are often used as they produce good-enough paths faster than using the equivalent admissible heuristic.