Newsgroups: comp.constraints
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!agate!hpg30a.csc.cuhk.hk!eng_ser1!wltam
From: wltam@cs.cuhk.hk (Vincent Tam)
Subject: Questions about EHP...
Message-ID: <D2sqKA.BIt@eng_ser1.erg.cuhk.hk>
Lines:  41
Sender: news@eng_ser1.erg.cuhk.hk
Organization: Engineering Faculty, The Chinese U. of Hong Kong
X-Newsreader: TIN [version 1.2 PL2]
Date: Sun, 22 Jan 1995 07:38:33 GMT

	I'm currently writing my MPhil thesis on constraint solving
approach. And I've applied my constraint solving technique to the EHP 
which Patrick has mentioned and included in this discussion group. 
Now, concerning the EHP, I've 2 questions to ask :

>From: pat@cs.strath.ac.uk (Patrick Prosser)
>Date: 2 Sep 1994 12:48:01 +0100
>Newsgroups: comp.constraints
>Subject: exceptionally hard problems
>
>There has been considerable interest in exceptionally hard problems 
>(ehp's), that is problems that occur within the easy region,
>but they turn out to be many orders of magnitude harder than 
>other problems in that region. I have encountered these,
>but I have found that the phenomena is peculiar to a
>number of algorithms. That is, so far, I believe that 
>ehp are in fact existence proofs for eba's (exceptionally
>bad algorithms). I would like to be proved wrong.

1. Where's the source of the EHP ?

>Here is an exceptionally hard problem for forward checking with 
>dynamic variable ordering (choose variable with smallest 
>current domain). I pulled the plug on fc-dvo after 711 million 
>consistency checks. fc-cbj-ff found the following solution
>with 9588 consistency checks.
>
> 1 1 2 1 2 1 1 1 1 2 1 5 2 1 1 1 3 2 3 1 1 1 1 3 1 
>  1 1 3 8 2 1 1 2 8 1 6 1 3 5 6 1 2 7 1 4 3 1 3 2 4
>
>Note that the constraint graph IS connected. 

For the "fc-dvo" & "fc-cbj-ff" you've mentioned in the paragraph,
actually what system(s) (for example CHIP ?) you're using to find the
solution or do benchmarking ??

Thank you for your attention.

Cheers,
Vincent.
