CHANGE OF LOCATION: Master thesis defense
10 views
Skip to first unread message
Thomas Bolander
Oct 30, 2012, 9:00:23 AM10/30/12
to <algolog@googlegroups.com>, Allan Johnsen, Mads Jensen, Mikkel Birkegaard Andersen, Daniel Svendsen, Martin Holm Jensen, Torben Brauner, Be-te Elsebeth Strøm
Dear all,
The location of Mikko's master thesis defense on Friday has changed to: room S10, building 101.
Thomas
On 29/10/2012, at 08.50, Thomas Bolander wrote:
> Master thesis defense in Algolog. All are welcome.
>
> Time: Friday 2 November, 13.45.
> Place: build. 322, room 030.
> Speaker: Mikko Berggren Ettienne
> Title of talk: Pattern Database Heuristics for Automated Planning
>
> Abstract: A* with admissible heuristics is the leading approach to optimal planning.
> Pattern database (PDB) heuristics are admissible heuristics based on abstractions
> of the search space and have recently had a breakthrough as general heuristics
> for automated planning. The selection of appropriate abstractions is of
> paramount importance to the informedness of a PDB heuristic.
> Based on a combination of novel and well-known techniques, we show how to
> efficiently constrain abstractions which leads to increased informedness
> of PDB heuristics. State-of-the-art in PDB heuristics iteratively selects promising abstractions
> from a the search space of all possible abstractions using modified local search techniques.
> We introduce an approach called variable pruned mutex constrained extended pattern database generation
> that has several theoretical advantages over the state-of-the-art approach.
> Some of which lead to more informed PDB heuristics,
> while others lead to reduced computation time without affecting informedness.
> Experimental evaluations show that our approach
> also improves state-of-the-art for PDB heuristics in practice.
>
The location of Mikko's master thesis defense on Friday has changed to: room S10, building 101.
Thomas
On 29/10/2012, at 08.50, Thomas Bolander wrote:
> Master thesis defense in Algolog. All are welcome.
>
> Time: Friday 2 November, 13.45.
> Place: build. 322, room 030.
> Speaker: Mikko Berggren Ettienne
> Title of talk: Pattern Database Heuristics for Automated Planning
>
> Abstract: A* with admissible heuristics is the leading approach to optimal planning.
> Pattern database (PDB) heuristics are admissible heuristics based on abstractions
> of the search space and have recently had a breakthrough as general heuristics
> for automated planning. The selection of appropriate abstractions is of
> paramount importance to the informedness of a PDB heuristic.
> Based on a combination of novel and well-known techniques, we show how to
> efficiently constrain abstractions which leads to increased informedness
> of PDB heuristics. State-of-the-art in PDB heuristics iteratively selects promising abstractions
> from a the search space of all possible abstractions using modified local search techniques.
> We introduce an approach called variable pruned mutex constrained extended pattern database generation
> that has several theoretical advantages over the state-of-the-art approach.
> Some of which lead to more informed PDB heuristics,
> while others lead to reduced computation time without affecting informedness.
> Experimental evaluations show that our approach
> also improves state-of-the-art for PDB heuristics in practice.
>
Reply all
Reply to author
Forward
0 new messages