quote:
Originally posted by DirectorTsaarx:
[...]
Carryover of production ability might be nice, but I have no idea how difficult it would be to program. Or what the effect would be on the AI's production ability.
[...]
[WILD_SPECULATION]
Weeell... If we assume that the AI is filling its build queue in a knapsack-esque fashion (w/ a scheme like "weight" = cost, "capacity" = time, and "value" = importance to the AI given the situation), than it
seems like it ought to be trivial to convert the build queue to a
fractional knapsack algorithm, which would solve the AI portion of this problem quite nicely. 'Course, that first assumption might be a bit of a tall order...
[/WILD_SPECULATION]
(Anyone care to guess who ought to be studying for an Analysis of Algorithms final instead of reading the forums?)
E. Albright
[EDIT: Corrected parameters ("weight", etc.) in my knapsack analogy / speculation / whatever. Hmm... If my initial post reflected my understanding of knapsack problems, I
REALLY ought to be studying...]
[ 06 May 2002: Message edited by: ealbright ]