Dear Mark,

A few brief comments:

In general I feel your text is rather hard to read for someone who does not know the Oz constraint programming model. So, I suggest you either simplify and completely leave out details like the notion of computation spaces, or you explain it in more detail.

Initially a problem (P) is modelled using integer values.

This is only a special (though most common) case. Do you want to mention that?

Constraint propagation (implemented by program entities known
as propagators) narrows down the possible values of the FD integers by
amplification of the constraint store.

You did not introduce the notion of a constraint store before. Also, the term "amplification" can lead to confusion (when I used it in a similar context in my thesis, a native English speaker marked it as "please replace" :).

Just my 2 cent..

Best
Torsten

--
Torsten Anders
Interdisciplinary Centre for Computer Music Research
University of Plymouth
Office: +44-1752-586219
Private: +44-1752-558917
http://strasheela.sourceforge.net
http://www.torsten-anders.de





On Feb 25, 2009, at 7:17 PM, mark richardson wrote:

Hi,
I have written the text below as part of my final year undergraduate
project.
I am trying to capture the essence of programming with constraints in Oz without the detail of the documentation. The intended audience is almost
certainly not aware of spaces, entailment, or even Oz for that matter.
I'm also not concerned yet about spelling, grammar or parts that can be
better expressed - just that what I have described is accurate.

My intention is for it to be a sort of basic 'walkthrough' describing an
'ideal' solution, where propagators work, things become stable,
solutions are produced,...etc

I wonder if anyone would be so kind as to have a quick glance at the
text below and let me know if I've said anything that is absolutely
ridiculous?
I would ask someone at my university to check it, but apparently I am
the only Oz 'expert' on campus. (and expert I am not!)


"There follows a simplified overview of the operation of a constraint
based program; the interested reader is referred to the official
Mozart-Oz documentation for a more detailed explanation.

Initially a problem (P) is modelled using integer values. Limits for the
possible values of these FD integers are applied. These are known as
basic constraints. A method known as 'propagate and distribute' is then employed. Constraint propagation (implemented by program entities known
as propagators) narrows down the possible values of the FD integers by
amplification of the constraint store. Propagators are connected to the
constraint store in a computational space (see fig2.1).
      When no further propagation is possible, the values of the FD
integers being non-determined, the program resorts to a distribution
method whereupon a constraint (C) is heuristically chosen. This results
in the creation of two new spaces. One space represents {P\/C}, the
other {P\/¬C}. Propagation then continues, if possible, in each new
space. This iteration of propagate and distribute continues until either
no further propagation is possible or all the FD integers can be
assigned values that satisfy all the constraints in the constraint
store. In the case of no further propagation being possible and there is
no assignment of values that will satisfy all constraints in the
constraint store, then that space is said to be failed and is removed
from the solution; this is analogous to the failure of a choice point in
languages such as Prolog. In the case of no further propagation being
possible but all the FD integers have all been assigned values that are
consistent with the constraint store, then the FD integers are said to
be determined and the space becomes both stable and a solution to the
problem P."

Thank you in advance
Regards
Mark

--
Mark Richardson
Final year undergraduate
University of Teesside

______________________________________________________________________ ___________ mozart-users mailing list mozart- [email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users

_________________________________________________________________________________
mozart-users mailing list                               
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users

Reply via email to