A* and nodes
Displaying 1-20 of 24 total.
12
next
RageCage
|
would anyone actually use a node pathfinding system that uses A* algorithms and want to use a verge node mapping thing?
I'm making a node mapper in verge for use in my game and was wondering if I should make it user-friendly enough for others...
anyone planning to make entities be able to get from point A-B with AI would probably need it.
Posted on 2004-05-06 16:37:27
|
zaril
|
i used to have a function made to return an entity string. it's funky so code it and make it availible?
Posted on 2004-05-06 17:20:28
|
mcgrue
|
R U SERIOS?
U CANOT B SERIOS.
...
Of course a user-friendly A* library would be desireable. A great boon to others, in fact. You'd be a leader, a prophet, but they're too blind to notice!
Posted on 2004-05-06 20:37:24
|
Rysen
|
yes! There was a point where I thought about trying it myself although I can't seem to remember what I needed it for....:P
Posted on 2004-05-06 20:40:03
|
mcgrue
|
I did some basic recursive pathfinding and movecode building in hero rancher, but nothing as awesome as A*.
Posted on 2004-05-06 21:06:32
|
resident
|
Good pathfinding is always nessecary.
Unless you're creating Zombies.
mmm, zombies.
Posted on 2004-05-06 22:21:47
|
rpgking
|
I'm guessing they use A* in games like Warcraft and Diablo...
Posted on 2004-05-07 00:57:15
|
grenideer
|
ya, and NeverWinter Nights. A* is pretty slow and you only wanna use it when it's necessary (there are complicated obstacles and possible places to get stuck in like dead ends or different rooms). I know NWN used a LOD pathfinding system where A* was really only used for enemies onscreen.
For verge applications, if you're gonna have some kind of pac-man maze or other coolness, A* may be necessary. But if you just wanna avoid trees or houses like an average rpg map, it'll probably be overkill.
Coolness tho. I say code it user-friendly and let vergers decide if they want to use it!
Posted on 2004-05-07 02:25:28
|
RageCage
|
well when I said A* I meant applying A* to a system of nodes and links to generate a table that would be used for getting form a to b.
for a more visual and better explained article:
http://gamedev.net/reference/articles/article1841.asp
anyway, I'll try to go ahead and make it a little more user friendly n stuff... and release the algorithms too ^_^
Posted on 2004-05-07 04:01:32 (last edited on 2004-05-07 04:06:32)
|
RageCage
|
ok well I've finished up making the A* algorithm. I decided to make it use every tile as a node and I'm about to change the sorting to binary heaps so that it will go significantly faster in finding paths. I'll be releasing a simple little demo soon.
But the reason I'm posting is because I'm to the point I need to find a better way to move entities. I want some way of making an entity walk to point x, y. It sounds easy enough but not with the current move script system. also the 4-directional movement as opposed to 8 directional makes things look pretty bad. I'm not really sure what to do about it. unless vecna makes a slightly more controllable movement system, I'll have to make my own.
... comments?
Posted on 2004-05-17 15:54:38
|
Mythril
|
Huh? Can't you just ignore diagonals when implementing the A* algorithm?
I hadn't even heard about this algorithm before you posted about it here, though.
Posted on 2004-05-17 16:11:40
|
RageCage
|
I guess you could just not go diagonal, but that's no fun when we have the great new v3 diagonal abilities ^_^
but diagonals isnt the real problem. it's the fact that using movement scripts arnt very good for this situation.
if a character has a speed of 100 and another has a speed of 150, the same "U1" movescript will move them different distances. I would like to avoid trying to control entity speed too much just to get accuracy.
Posted on 2004-05-17 16:34:58 (last edited on 2004-05-17 16:37:19)
|
Mythril
|
So movescripts doesn't move the entity tilewise in v3?
That sucks. Mehehe. :P
Posted on 2004-05-17 18:09:11
|
vecna
|
Quote:Originally posted by RageCage
if a character has a speed of 100 and another has a speed of 150, the same "U1" movescript will move them different distances.
What??? Are you sure? U1 is supposed to move you 16 pixels, always, regardless of speed. If not, that's a bug. But it works in every situation I've tested.
Posted on 2004-05-17 18:35:40
|
RageCage
|
Hmm, so it seems it doesnt work too well when you stack moveentity functions... well then... thanks
is there any chance diagonal move scripts will be implemented anytime soon? so I know if I should try to work around that or just wait.
Posted on 2004-05-18 01:46:19 (last edited on 2004-05-18 02:40:06)
|
Gayo
|
As long as we're making wishes, I wouldn't mind the ability to change an entity's speed in a movescript, either.
Posted on 2004-05-18 03:39:23
|
Mythril
|
Quote:Originally posted by RageCage
Hmm, so it seems it doesnt work too well when you stack moveentity functions... well then... thanks
Don't you just put all the moves into the same movescript?
I think if you use several calls to the movescript function, it could mess it up, yeah.
Posted on 2004-05-18 12:18:51
|
vecna
|
Gayo: 'S50' or something sets the speed. You fail again! :D That's worked since v1 too!
'Stacking' entitymove doesnt work. They're aren't queued, they overwrite each other.
I may be able to add diagonal movecodes, I'll look into it. My concern is that it would slow entity processing down to a crawl to perform the collision detection thats applied to the player to every single entity on the map (which becomes an exponential growth problem). But I think there's some shortcuts I can take.
Posted on 2004-05-18 13:29:24
|
RageCage
|
muchos gracias
(ahhA! I have dwarfed vecna's post count!! The world is MINE!)
Posted on 2004-05-18 13:39:29 (last edited on 2004-05-18 16:04:11)
|
RageCage
|
one more thaing and I'll be happy =p
I find it very important that we have a way of telling when a entity has finished moving to a movescript
Posted on 2004-05-18 23:01:27
|