Building a better RTS AI. A second map
-
To get things working you need two identical maps side by side. One map is the usual RTS map, it’s the map for real things, the place where the conclusions of the improved AI are put into practice. The second map is the place where the imaginary things get tested. The imaginary things get tested on fast forward, units don’t move and fight in real time. The type of operations on the second map are range operations, path operations. You can figure out the outcome of battles using autocalc ( the notion is borrowed from Heroes of Might and Magic) To give an example 4 marines equals one tank, a battle between 10 marines and four tanks will result in all marines killed and two tanks lost. If you think about it an improved AI may turn out relatively cheap computation wise.
-
To get things working you need two identical maps side by side. One map is the usual RTS map, it’s the map for real things, the place where the conclusions of the improved AI are put into practice. The second map is the place where the imaginary things get tested. The imaginary things get tested on fast forward, units don’t move and fight in real time. The type of operations on the second map are range operations, path operations. You can figure out the outcome of battles using autocalc ( the notion is borrowed from Heroes of Might and Magic) To give an example 4 marines equals one tank, a battle between 10 marines and four tanks will result in all marines killed and two tanks lost. If you think about it an improved AI may turn out relatively cheap computation wise.
thank you. I was wondering about that.
>64 Some days the dragon wins. Suck it up.