Comments on: Dynamic pathfinding(Cutting edge research!) http://www.oddco.ca/zeroth/zblog/2008/07/29/dynamic-pathfindingcutting-edge-research/ Game design, development, technology, programming, and python Fri, 10 Sep 2010 03:24:13 +0000 hourly 1 http://wordpress.org/?v=3.1 By: Recent Faves Tagged With "pathfinding" : MyNetFaves http://www.oddco.ca/zeroth/zblog/2008/07/29/dynamic-pathfindingcutting-edge-research/comment-page-1/#comment-540 Recent Faves Tagged With "pathfinding" : MyNetFaves Sat, 13 Sep 2008 00:35:15 +0000 http://www.oddco.ca/zeroth/zblog/?p=105#comment-540 [...] public links >> pathfinding Dynamic pathfinding(Cutting edge research!) First saved by haleyhazardx | 1 days ago Fixing Pathfinding First saved by miik | 19 days ago [...] [...] public links >> pathfinding Dynamic pathfinding(Cutting edge research!) First saved by haleyhazardx | 1 days ago Fixing Pathfinding First saved by miik | 19 days ago [...]

]]> By: Zeroth http://www.oddco.ca/zeroth/zblog/2008/07/29/dynamic-pathfindingcutting-edge-research/comment-page-1/#comment-131 Zeroth Thu, 31 Jul 2008 01:11:21 +0000 http://www.oddco.ca/zeroth/zblog/?p=105#comment-131 Actually, I should note that if you use Foxit PDF reader, the pdf slides work. Adobe doesn't work. Well, for larger, more open spaces, then one could use, I believe, navigation meshes. Or, just simple lookahead. However, most of the pathfinding problems are in fact, due to small maze-like maps. Take a look at the actual paper, its a lot more refined. http://ircl.cs.ualberta.ca/files/webfm/lrts/pubs/bulitko08a.pdf Actually, I should note that if you use Foxit PDF reader, the pdf slides work. Adobe doesn’t work.

Well, for larger, more open spaces, then one could use, I believe, navigation meshes. Or, just simple lookahead. However, most of the pathfinding problems are in fact, due to small maze-like maps. Take a look at the actual paper, its a lot more refined. http://ircl.cs.ualberta.ca/files/webfm/lrts/pubs/bulitko08a.pdf

]]> By: zylum http://www.oddco.ca/zeroth/zblog/2008/07/29/dynamic-pathfindingcutting-edge-research/comment-page-1/#comment-129 zylum Wed, 30 Jul 2008 15:48:06 +0000 http://www.oddco.ca/zeroth/zblog/?p=105#comment-129 That's some interesting stuff. Definitely something I'd love to research. It's a shame the presentation slides are missing most of the text.. I was thinking of other ways of pathfinding and was wondering why most maps use tiles. This system is sort of crude and creates a lot of problems when there are large spaces with no obstacles (ie adds lots of unnecessary computation). I was wondering if there are any good algorithms based on computational geometry. Such as defining obstacles as polygons or whatever and find paths that are lines wrapping around the obstacles (sorta hard to explain what I'm thinking). Computational geometry is fairly mature so I'm guessing there are many techniques that could be used to make this process fast. Plus there would be few objects to run the algorithm on rather than thousands of tiles.. Of course this method would have some disadvantages. I don't see it working very well in a maze-like environment. That’s some interesting stuff. Definitely something I’d love to research. It’s a shame the presentation slides are missing most of the text.. I was thinking of other ways of pathfinding and was wondering why most maps use tiles. This system is sort of crude and creates a lot of problems when there are large spaces with no obstacles (ie adds lots of unnecessary computation). I was wondering if there are any good algorithms based on computational geometry. Such as defining obstacles as polygons or whatever and find paths that are lines wrapping around the obstacles (sorta hard to explain what I’m thinking). Computational geometry is fairly mature so I’m guessing there are many techniques that could be used to make this process fast. Plus there would be few objects to run the algorithm on rather than thousands of tiles.. Of course this method would have some disadvantages. I don’t see it working very well in a maze-like environment.

]]>