Comments on: Pathfinder (AI) http://www.rubenswieringa.com/blog/pathfinder-ai Ruben writes flashy writings Thu, 28 Aug 2008 10:34:45 +0000 http://wordpress.org/?v=2.5.1 By: kim bo kastekniv http://www.rubenswieringa.com/blog/pathfinder-ai#comment-42206 kim bo kastekniv Sat, 19 Jul 2008 18:43:18 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-42206 i agree with bechte to some degree.. a* forgets about paths that didn't work in the past - but they could work later - so starting over again from scratch when reaching a deadend maybe isn't such a bad idea under some circumstances.. i agree with bechte to some degree.. a* forgets about paths that didn’t work in the past - but they could work later - so starting over again from scratch when reaching a deadend maybe isn’t such a bad idea under some circumstances..

]]> By: bechte http://www.rubenswieringa.com/blog/pathfinder-ai#comment-30659 bechte Sat, 24 May 2008 06:39:25 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-30659 Hi there, I think backtracking is potencial a possible first idea, but in the end you will end up in bad performance with high memory allocation for your stack, just if the map you are trying to find your way out becomes bigger as 5x5. Have you been looking for other path finder implementations, such as the A* Pathfinder algorithm. I think this is what you really need. :-) Cheerio bechte Hi there,
I think backtracking is potencial a possible first idea, but in the end you will end up in bad performance with high memory allocation for your stack, just if the map you are trying to find your way out becomes bigger as 5×5.
Have you been looking for other path finder implementations, such as the A* Pathfinder algorithm. I think this is what you really need. :-) Cheerio bechte

]]> By: Martin http://www.rubenswieringa.com/blog/pathfinder-ai#comment-26048 Martin Sat, 26 Apr 2008 16:29:16 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-26048 An A* algorithm would be much more practical. If you have no information (heuristic) about where the target might be, use dijkstra shortest path (essentially the same as A* with a constant heuristic). An A* algorithm would be much more practical. If you have no information (heuristic) about where the target might be, use dijkstra shortest path (essentially the same as A* with a constant heuristic).

]]> By: sean http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24469 sean Mon, 14 Apr 2008 21:53:14 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24469 This is Dijkstraaaaaaa!!!!! ( and not using backtracking only worsens the performance ) This is Dijkstraaaaaaa!!!!! ( and not using backtracking only worsens the performance )

]]> By: Ruben http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24294 Ruben Sun, 13 Apr 2008 09:05:42 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24294 Hey jack, thanks! I had a look, very interesting :) Hey jack, thanks! I had a look, very interesting :)

]]> By: jack http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24268 jack Sun, 13 Apr 2008 01:31:11 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24268 If you google "god's algorithm" you will find a path optimizer algorithm. I haven't looked for it in awhile but check it out. Cool stuff you are working on. Jack If you google “god’s algorithm” you will find a path optimizer algorithm.
I haven’t looked for it in awhile but check it out.

Cool stuff you are working on.

Jack

]]> By: Ruben http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24231 Ruben Sat, 12 Apr 2008 13:40:35 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24231 Haha, yes, I guess so, thanks. :) Haha, yes, I guess so, thanks. :)

]]> By: -- http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24229 -- Sat, 12 Apr 2008 13:27:13 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-24229 that's a lee algorithm, right? that’s a lee algorithm, right?

]]> By: Nate http://www.rubenswieringa.com/blog/pathfinder-ai#comment-23991 Nate Wed, 09 Apr 2008 23:11:51 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-23991 I maybe have found in issue in your pathfinder's logic. If we consider the directions as a compass and it chooses the path: East East South South West North it will fail rather than choose West to continue what could turn out to be a totally valid path. Is this something it was intended to do? I maybe have found in issue in your pathfinder’s logic. If we consider the directions as a compass and it chooses the path:

East
East
South
South
West
North

it will fail rather than choose West to continue what could turn out to be a totally valid path. Is this something it was intended to do?

]]> By: Andy http://www.rubenswieringa.com/blog/pathfinder-ai#comment-23746 Andy Mon, 07 Apr 2008 20:48:12 +0000 http://www.rubenswieringa.com/blog/pathfinder-ai#comment-23746 Here is an image of what I mean: <img src="http://www.rubenswieringa.com/blog/wp-content/uploads/2008/andy/path.jpg" alt="" /> Start is top left, target is bottom right. X represents a no go area. Algorithm simply increments naighbours until it hits the target, then follows it home via the least energy path. Here is an image of what I mean:

Start is top left, target is bottom right. X represents a no go area. Algorithm simply increments naighbours until it hits the target, then follows it home via the least energy path.

]]>