re: project 3

[Subbarao Kambhampati <rao@asu.edu>]

The conenction between project 1 and project 3 is just A* search
code. In Proj 1, you enhanced the A* search code (so it will print
search stats, path information etc), and added a child-generation
function for 8-puzzle problem.

In this project, you take the enhanced A* search code, and add a
child-generation function for doing regression search. Specifically
the "state" here will be a list of literals. You will generate the
children states by first getting the ground operators for the problem
(using the accessor function I supplied), and then regressing the
state over each of the ground operators (some of the results of the
regression may be nil, signifying that there is no child corresponding
to that move).

Hope this helps

Rao
[Nov 02, 2001]


>---------- Forwarded message ----------
>Date: Thu, 1 Nov 2001 20:23:30 -0700 (MST)
>From: snehalp@imap4.asu.edu
>To: Subbarao Kambhampati <rao@asu.edu>
>Subject: project3
>
>
>I am a little confused on the context of project3. In class you discussed that
>the project would be an extension project 3, but after reading the project3
>handout it seems as though project 3 is dealing with the blocks
>world........(or in that case any world specified by the problem and
>domain)  Is this correct? Can you please explain to me the correlation between
>project1 and project3. I can see that the way the child generation occurs is
>similar but apart from that i am not seeing any distinct similarities.  Maybe
>i have been staring at the project3 handout for too long........can you please
>clarify?  
>
>
>Also, when doing regression and checking for inconsistent states do we check
>for mutual exclusion between more than two literals?? 
>
>
>
>i would appreciate if i could have these questions clarified.
>
>
>thank you
>
>
>-snehal patel