Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: Motion 11 questions & comments...



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear Dan and all,

Dan Zuras Intervals wrote:
> 	Ah, thank you Nate.  I see I was confused about the
> 	nature of reverse operations altogether.
> 
> 	Then, Marco, I am also confused as to the purpose
> 	of these operations.
> 
> 	If they are formally equivalent to the existing
> 	forward operations, why not use those instead?
> 	Well, those together with the intersection
> 	operation in an expression?

As I already pointed out on this list (January 24th), Corollary 1 is
false. The counter-example I gave was:

A=[0,2]
B=[-1,1]
C=[1,1]

with \circ=\times

and compute \times_1^-(B,C,A) = hull({x\in[0,2]\mid\exists b\in[-1,1],
x\times b\in[1,1]})

You should get [1, 2]

Now, compute \times_1^-(B,C) = hull({x\in R\mid\exists b\in[-1,1],
x\times b\in[1,1]})

You should get hull([-\infty,-1]\cup[1,+\infty]), that is [-\infty,
+\infty]

If you intersect the last result with A=[0, 2], you get [0, 2], which is
different from [1, 2].


Reverse division is useful to compute Newton steps. All the other
reverse operations are useful to implement narrowing operators in
constraint programming algorithms. I shamelessly refer you to the first
two sections of the following paper:

Interval Extensions of Multivalued Inverse Functions. F. Goualard, 2007
http://cinerep.free.fr/refbase/publications/goualard-gaol2007.pdf

for a motivation of the introduction of reverse operations, and to the
following paper for an algorithm that relies exclusively on them:

Revising hull and box consistency. Benhamou, F., Goualard, F.,
Granvilliers, L., & Puget, J. - F. 1999.
http://goualard.free.fr/publications/files/BenhamouGoualardGranvilliersPuget-ICLP99.pdf

FG.
- --
Frédéric Goualard                                 LINA - UMR CNRS 6241
Tel.: +33 2 51 12 58 38    Univ. of Nantes - Ecole des Mines de Nantes
Fax.: +33 2 51 12 58 12            2, rue de la Houssinière - BP 92208
http://goualard.frederic.free.fr/               F-44322 NANTES CEDEX 3

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iD8DBQFLcQN9EJvxJgN8HkgRAvfNAJ9PdFd1xEI3ZNmFFdpuBhJMWD+B6gCgkdVq
2yNJ8im3MU7gwIhrOXhJMrY=
=PDN1
-----END PGP SIGNATURE-----