+- Qbasicnews.com (http://qbasicnews.com/newforum)
+-- Forum: QBasic (http://qbasicnews.com/newforum/forum-4.html)
+--- Forum: QB Discussion & Programming Help (http://qbasicnews.com/newforum/forum-11.html)
+--- Thread: a concrete problem with my pathfinder, this time: (/thread-337.html)
a concrete problem with my pathfinder, this time: - Agamemnus - 03-04-2003
I figured out how to do it, theoretically.
1) First, I get the island borders.
2) Then, I get unique landmasses within the first island border level.
3) Then, I run an algorithm to figure out which landmass is connected to which.
4) Then, I run an algorithm that calculates the paths and distance between adjacent landmasses from each landmass by using even smaller landmasses (circle patches) within the landmass. And I also need to do (3) on this as well.
5) Then I make a search tree for the larger landmasses and find the path from the unit to the target landmass.
6) Then I do the same thing as (4), except I calculate the path from the border of the landmass to the unit or the target.
7) Then I combine those 3 paths to get the big path.
(edit: 1, 2, and 3 done; on to 4)
This is the only workable approach I know, plus there is the added advantage that I can reuse the paths gained in steps 2-6 every time I make a new path!! Thus the path calculation is fast for 10, 100, or 1000 units! :bounce:
http://www.geocities.com/pisforpi/weird3.gif
What do you people think? :|