Minimal Implicants from ROZDD or BDD in Prolog?

[burs...@gmail.com]

Dear All,

Kuniaki Mukai started showing boolean equations here:
https://groups.google.com/d/msg/swi-prolog/X2RrVc4YyTE/Q-jZljCPBQAJ

As a side problem, I started thinking about producing
a minimal DNF from a ROZDD or BDD in Prolog. Producing
just some DNF from a BDD is simple.

I am using this example from here:

Electrical Engineering - Logic Friday's minimized form
https://electronics.stackexchange.com/q/242721/177985

With just generating some DNF from the BDD, I get the following
result, which isn't bad, but which isn't exactly the minimial
from Logic Friday problem:

?- expr_tree(A*B*C + A*B*D + A*C*D + ~B* ~C*D +
A*C* ~D + ~A* ~B*D, T), expr_dnf(T, R).
R = A*C+A* ~C*D+ ~A* ~B*D

Anybody ever done computing minimal implicants in Prolog,
that would improve the above result? I think the minimal
solution uses implicants that are overlapping.

Best Regards

P.S.: Source code is here, that starts with BDD:
https://gist.github.com/jburse/503d14fc660d5bedeff4edd4c6f9f4ac#file-imp-p

There is an old article, but it doesn't use ROZDD or BDD I guess:

Minimal Boolean Sum-of-Product Forms in Prolog
A.K.Mitra in Computers & Mathematics with Applications
Volume 19, Issue 2, 1990, Pages 75-81
https://www.sciencedirect.com/science/article/pii/089812219090010H

[burs...@gmail.com]

How quickly the interest focus can shift. Currently
working on a new strand of problems, inspired
by Kuniaki Mukai, who brought up the 8x8 queens
problem as a CLP(B) problem.

See also here:
https://groups.google.com/d/msg/swi-prolog/ZQZV2i6TlN4/Sy55Jc-4AwAJ

The problem is hard for a CLP(B), since standard Prolog
backtracking gives already a good results. But some
insights have already been obtained by a CLP(FD) to
CLP(B) translator. Inner queens first seems better.

See also here:
https://plus.google.com/+JanBurse/posts/1gn6AvjdD1t

[burs...@gmail.com]

Kuniaki Mukai addressed me personally, but I am
afraid, I cannot answer. SWI-Prolog Google Groups
has already blocked my 3rd Account.

Now I don't have any accounts anymore. So you
need to make your homework by yourself. Tokyo 2020
is still some years ahead.

Ha Ha. What a bunch of morons.

From: Kuniaki Mukai <kun...@gmail.com>
Message-Id: <A5F21701-0CDB-43E6...@gmail.com>
Content-Type: multipart/alternative;
boundary="Apple-Mail=_D9C53125-5961-4F59-97E6-E74E54AE79D9"
Mime-Version: 1.0 (Mac OS X Mail 11.3 \(3445.6.18\))
Subject: CNF to DNF in ZDD (was Re: [SWIPL] the number of nodes of ROZDD)
Date: Thu, 24 May 2018 20:45:28 +0900
In-Reply-To: <da8808e9-e652-4a42...@googlegroups.com>
Cc: SWI-Prolog <swi-p...@googlegroups.com>
To: j4n bur53 <jan...@xlog.ch>

[burs...@gmail.com]

Hi Kuniaki,

Be careful about when you post "be quite" or "I will
not discuss any further", what bystanders then think.
They cannot deal with this Schizophrenia of yours.

If you want to address me, you can still post on
comp.lang.prolog, I will voluntarily answer there,
even on a daily basis. But I will definitely not answer

anymore on the SWI-Prolog Group.

Best Regards

[burs...@gmail.com]

Other google groups have daily or hourly limit,
for example sci.logic, sci.math, etc.. But they
don't delete posts of other people.

The show a warning and do not accept a post in
the Web GUI from Google, so one can save the content,
and maybe try again later.

Possibly the SWI-Prolog forum can be also configured
like that, and not that it simply swallow peoples
post and they disappear in a black hole.