[rules-users] Iterative diagnosis - how to determine what is the next best fact?

[Wolfgang Laun]

As an answer to your "big" question: no. There is no distinction between a
rule that isn't activated because one of its patterns doesn't match with any
of the existing facts or due to the absence of facts of a type. Hence, (A, B
and not C) with A, B and C that does not match is indistinguishable from A,
B and no C at all.

Are all your facts booleans or simple (integer, enum,...) values? "Answer to
question" seems to indicate this.

Note that there are rules (boolean expressions) determining a "solutiion"
and rules governing data acquisition. I'd try and model the second group,
possibly based on facts representing the individual "solution", what is
required to reach them, how much has been required, whether one is still
feasible, etc.

"Decision tree" seems to be the thing you need, but Drools doesn't have this
out of the box.

-W


On 21 August 2011 00:23, Kal <kalarhan at gmail.com> wrote:

> Hi,
>
> I am new to Drools. I read the documentation and I looked at the examples,
> but I still not sure how to solve this problem:
>
> 1) Kind: iterative diagnosis
>
> Rules:
> if (A, B and C) then print ("Solution 1")
> if (A, B and not C) then print ("Solution 2")
> if (A, D and E) then print ("Solution 3")
> and so on...
>
> The entry should be something like this:
> * no facts are know
> * user answers a question (lets say A = true), I update the Statefull
> session and fire the rules
> * system chooses a unknow fact based on condition for the best incomplete
> rule. In this case, it would be B (as it knows A, but it doesn't know B or
> C)
> * user answers a new question (B = false), I update the session + fireRules
> * system now knows that rule 1 (A, B and C) is not possible. It also knows
> that rule 2 (A, B and not C) is not possible. So, it chooses the second
> condition on rule 3 (A, D and E), which is: what is D?
> * user answers a new question (D = true)
> ....
> * user answers a new question (E= true)
> * system prints: Solution 3!
>
>
> My big question: can I extract from the working memory or rule manager or
> the agenda a list of the rules that are 'almost' good to fire, but that
> still requires some of its conditions to be know (by inserting facts)?
>
> Thanks!
>
>
> --
> View this message in context:
> http://drools.46999.n3.nabble.com/Iterative-diagnosis-how-to-determine-what-is-the-next-best-fact-tp3271709p3271709.html
> Sent from the Drools: User forum mailing list archive at Nabble.com.
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20110822/68e9144b/attachment.html