[Pellet-users] pellet + swrl == low performance??

[Michael Smith]

On Fri, 2008-06-20 at 13:17 +0200, Sergio Berná Niñerola wrote:

> 
> I have the same problem with adding a SWRL rule to an existing
> ontology with over 60 individuals. The expressivity of the ontology is
> SHIF, and the rule has 3 variables.
> 
> After some debugging I have narrowed the problem to the RuleStrategy
> implementation. The main problem is that for 60 individuals and 3
> variables (that neither of them satisfy the type restriction expressed
> in the rules), pellet generates all the  possible combinations of
> individuals and checks the rule with all of them resulting in an
> OutOfMemoryError.
> 
> Is this the default rule strategy or can it be caused by a bad rule or
> ontology specification?

RuleStrategy is the default rule strategy.  See faq entry at [1] for an
optimization that may address your problem.

Often, the details of the ontology or ruleset are relevant and a "bad
rule" can cause problems.  If the optimization does not address your
problem, please consider sending the ontology to help us debug.

-- 
Mike Smith

Clark & Parsia

[1] http://pellet.owldl.com/faq/tuning-rules/