As I said, we will discuss any outstanding questions on heuristics
(including probably a quick discussion on use of reachability
heuristics for
Plan-space planning), and then discuss bounded-length planning as a
combinatorial problem. What we did with Graphplan today will become
a simple subcase--with other cases being pushing planning as SAT/CSP/IP etc.
The additional reading for tomorrow's class is:
Chapters 6 and 7 in Nau et. al.'s text book
or if you want a shorter version, Chapter 11 (sections 4 and 5) in
Russell and Norvig (if you don't have the text, here is that chapter
from AIMA site:
http://aima.cs.berkeley.edu/newchap11.pdf )
-----------------
Also, if any of you felt a little lost today and wanted a more
structured lecture on the heuristics, check out
http://rakaposhi.eas.asu.edu/ICAPS06-tutorial/ICAPS06-1.mp3
which is the first part of the tutorial based closely on the paper you
read and slides you saw.
Rao
No comments:
Post a Comment