Clearanced-based Hierarchical Pathfinding (HPA*)

Aigamedev.com seems to always have something interesting to read. This week it is using Annotated A* to find a path of at least minimum clearance through a graph. They also go into how you can reduce the graph to an annotated hierarchical pathfinding graph for near-optimal results at much greater speeds. The article is well explained and comes out with a very practical result.

If your not convinced maybe their talk on zombie brain AI from last month will convince you.

Goal-Oriented Action Planning

I found a very good paper on AI planning by the guy that did the F.E.A.R. AI architecture.

Jeff Orkin's page on GOAP contains lots more links to even more details.

Learn something!