Hi StarCrunch,
Cooperative pathfinding is something i’ve already considered, many times. But it doesn’t really suits to me as it end up a bit tricky to implement and to get working right with different search algorithms (in other works, I did not succeeded).
Actually, there is an alternative method, but that goes out of a simple pathfinding library purposes, yet it works really fine. For a group of units, one just runs the pathfinding algorithm (in a clever way so that it won’t drop the framerate), and then use “steering behaviours”. Something I might definitely try someday.
And about ThetaStar, I succeded in implementing it, as it wasn’t that hard, and the papers gave some very clear explanations about that. There are some things left to fix yet, such as finding an efficient line of sight algorithm as the original one described in the article fails miserably in some cases.
Thanks for your feedback and ideas, though. [import]uid: 142361 topic_id: 30317 reply_id: 145479[/import]