Path: utzoo!attcan!uunet!ncrlnk!ncrcae!hubcap!gatech!udel!burdvax!macbeth!lang From: lang@macbeth.PRC.Unisys.COM (Francois-Michel Lang) Newsgroups: comp.lang.prolog Subject: Re: The meaning of "declarative" Message-ID: <8326@burdvax.PRC.Unisys.COM> Date: 15 Nov 88 16:38:41 GMT References: <818@etive.ed.ac.uk> <590@dcl-csvax.comp.lancs.ac.uk> <859@etive.ed.ac.uk> <256@aipna.ed.ac.uk> <41315@linus.UUCP> <602@quintus.UUCP> <41581@linus.UUCP> <648@quintus.UUCP> Sender: news@PRC.Unisys.COM Organization: Unisys Corporation, Paoli Research Center; Paoli, PA Lines: 25 In article <648@quintus.UUCP> ok@quintus.UUCP (Richard A. O'Keefe) writes: ... Let's look at the monkey-and-bananas program which started this discussion. It's just a depth-first search, where states have the form state(MonkeyHoriz,MonkeyVert,BoxVert,HasBananas). ... There is an easy way of fixing this, and while it looks like a hack it is actually a principled strategy which competes well with breadth-first search. That is ITERATIVE DEEPENING. To keep this message short: ... I suggest that you consult a good AI text to find out when iterative deepening is appropriate. In this particular case, the iterative deepening version works well no matter how you shuffle the rules. ... Can somebody point to a reference for "iterative deepening"? I have looked in the indexes of a half-dozen or so standard AI textbooks and found nothing. I know this is a term and a search strategy that Richard discusses. Richard, did you come up with the name? If not, do you know who did, or where the strategy is referred to as such (other than your messages and tutorial notes)? ---------------------------------------------------------------------------- Francois-Michel Lang Paoli Research Center, Unisys Corporation lang@prc.unisys.com (215) 648-7256 Dept of Comp & Info Science, U of PA lang@cis.upenn.edu (215) 898-9511