[DL] DL's semantics and model cheking questions ...

[Enrico Franconi]

On Wednesday, Sep 3, 2003, at 06:28 Europe/Rome, Antonio Aviña wrote:
> 1.- Why simple model cheking is not enough when querying a DL KB?, I 
> found a short explanation in a paper on Query Answering (by Tessaris 
> and Horrocks Ian) but to be honest this matter remains unclear to me 
> :-) .

Because an ABox does not devise a unique model. In general a TBox+ABox 
KB admits a set of models, and queries should be answer with what is 
true in all of them. This is deduction,  not model checking.

> 2.-  By assumming either a closed world assumption or epistemic-like 
> queries, simple model checking  would be enough for querying a DL KB?

This depends. There is a good description on when this happens with 
epistemic queries in the DL handbook in the chapter 6 on Extensions to 
Description Logics by Franz Baader, Ralph Kuesters, Frank Wolter. Or 
the original paper: Francesco M. Donini, Maurizio Lenzerini, Daniele 
Nardi, Werner Nutt, Andrea Schaerf: An Epistemic Operator for 
Description Logics. Artificial Intelligence 100(1-2): 225-274 (1998).

> 3.- Where I can find more info about Tarski declarative semantics in 
> order to get a better understanding on DL Semantics?

In the DL handbook you will find all the answers to your questions :-)

cheers
--e.