Re: trAInsported: Alpha testers needed
Posted: Tue Mar 05, 2013 3:01 pm
100 Comments! Btw this piece of work is just soooooooooooo amazing. I never fought, that with Love2D you can create this masterpiece!
Thanks! Yes, the Engine is great!TheScriptan wrote:100 Comments! Btw this piece of work is just soooooooooooo amazing. I never fought, that with Love2D you can create this masterpiece!
I don't know the person personally... just his friend. But the image upload service is great. If only it would allow multiple images!substitute541 wrote:Who made "somepic.someserver.de"? The name is just a bit hilarious.
You got mesubstitute541 wrote: Edit: Hehehehe....
Speaking of which, I noticed that some implementations take too long to compute and are aborted during their search by my game. This usually happens with larger maps.substitute541 wrote:Just noticed that AI phoenix's pathfinding algorithm is B* (Breath First Search or Best First Search). Ah, I'm having a hard time implementing A*...
It is actually slightly modified Dijkstra's algorithm .substitute541 wrote:Just noticed that AI phoenix's pathfinding algorithm is B* (Breath First Search or Best First Search). Ah, I'm having a hard time implementing A*...
What about function that returns percentage of time the AI used for computation. It could be used like this:Germanunkol wrote:Speaking of which, I noticed that some implementations take too long to compute and are aborted during their search by my game. This usually happens with larger maps.
Any experience or thoughts on this? I can't give the AIs much more time to compute, otherwise it'll slow down the game much more. Or should I rather limit the maximum amount of trains and give more time to each train's computation instead? Or force players to do more pre-computation in ai.init (where they already have 10 times as much time)
Does any one have thoughts/ideas on this?
Code: Select all
while not pathfindingFinished and getAITimeUsage() < 0.95 do
-- One step of algorithm
end
... yep, that's why you always find the shortest distance (without using u-turns).jonyzz wrote:It is actually slightly modified Dijkstra's algorithm .substitute541 wrote:Just noticed that AI phoenix's pathfinding algorithm is B* (Breath First Search or Best First Search). Ah, I'm having a hard time implementing A*...
Awesome idea. Will think about this.jonyzz wrote:Code: Select all
while not pathfindingFinished and getAITimeUsage() < 0.95 do -- One step of algorithm end