- From: Danny Ayers <danny.ayers@gmail.com>
- Date: Tue, 8 Mar 2005 22:56:11 +0100
- To: Jim Hendler <hendler@cs.umd.edu>
- Cc: "IC Dept.- MIT-Maqbool Al Maimani" <maqbool@oas.com.om>, Abir Qasem <abir.qasem@gmail.com>, semantic-web@w3.org, public-owl-dev@w3.org
Many thanks Jim, sounds exactly what I was looking for. On Tue, 8 Mar 2005 10:30:38 -0500, Jim Hendler <hendler@cs.umd.edu> wrote: > > At 11:48 +0100 3/8/05, Danny Ayers wrote: > >A little side question: > > > >On Mon, 7 Mar 2005 08:12:53 +0400, IC Dept.- MIT-Maqbool Al Maimani > > > >> In our work Lehigh we use HTN planning to compose Semantic Web services. > To > >> avoid redundancy in service composition we use Local Closed World > reasoning > >> (LCW). > > > >I've not looked very closely at HTN, but I was wondering how far its > >semantics can be expressed/applied in an OWL DL system. Presumably the > >descriptive constructs used can be expressed reasonably cleanly in > >OWL, but are there any algorithms that can be rearranged to work with > >DL reasoning? > > > >My main reason for asking is that I'm hoping to put together a > >planner-like system for goal-oriented project management. Ok, that's > >rather a grandiose way of putting it - all I'm after is a simple > >system that will accept to-do lists, help you decompose them into > >discrete tasks & subgoals and then tell you which task you should > >probably do next. I'm still at the stage of putting together the basic > >infrastructure, but had anticipated using an OWL DL reasoner (or even > >just RDFS) for fact-expansion alongside an auxiliary rule engine > >(maybe Prolog-based, maybe Rete-based, as a last resort hard-coded) to > >do prioritisation. But the more I can get under the DL umbrella the > >happier I'll be. Suggestions appreciated. > > > >Cheers, > >Danny. > > > > > I'm just running out the door and saw this - let me suggest you check out > the work of Evren Sirin, my student, who is doing his PhD on the combination > of HTN planners and Sem Web services. He, Bijan Parsia, and I are giving a > tutorial on this topic at WWW - if you go to > http://www.mindswap.org/papers and look for the papers by Sirin and also the > ones with Kuter and with Wu (both of whom are Dana Nau's students) you'll > find a number of papers on different aspects of the combination of HTN > planners and OWL > The paper http://www.mindswap.org/papers/SHOP-JWS.pdf is probably the best > oe to start with - abstract is: > > > Automated composition ofWeb Services can be achieved by using AI planning > techniques. > Hierarchical Task Network (HTN) planning is especially well-suited for this > task. In this > paper, we describe how HTN planning system SHOP2 can be used with OWL-S Web > Service > descriptions. We provide a sound and complete algorithm to translate OWL-S > service > descriptions to a SHOP2 domain.We prove the correctness of the algorithm by > showing the > correspondence to the situation calculus semantics of OWL-S. We implemented > a system > that plans over sets of OWL-S descriptions using SHOP2 and then executes > the resulting > plans over the Web. The system is also capable of executing > information-providing Web > Services during the planning process. We discuss the challenges and > difficulties of using > planning in the information-rich and human-oriented context of Web > Services. -- > Professor James Hendler Director, Semantic Web and Agent > Technologies > Maryland Information and Network Dynamics Lab. 301-405-2696 > 8400 Baltimore Ave, Suite 200 301-314-9734 (Fax) > College Park, MD 20742 > http://www.cs.umd.edu/users/hendler -- http://dannyayers.com
Received on Tuesday, 8 March 2005 21:56:12 UTC