Heuristic Search Algorithms
About this Seminar
Organization, Content Overview, Exercises
J¨
org Hoffmann
Michael Katz
Peter Kissmann
Agenda
1
About Us
2
About the Organization
3
About the Content
4
About Your Seminar Work
Agenda
1
About Us
2
About the Organization
3
About the Content
4
About Your Seminar Work
Foundations of Artificial Intelligence (FAI) Group
Where?
E1 1, 3rd floor.
What?
Basic research in AI (much less applied than DFKI). Mainly
concerned with the development and analysis of general problem
solving mechanisms, in particular automatic planning and general
game playing.
Since when?
April 2012.
Who?
See next slides.
Prof. J¨
org Hoffmann
Saarland University
Foundations of Artificial Intelligence Group Campus E1 1, Room 3.18
66123 Saarbr¨ucken, Germany E-Mail:[email protected]
Subject Areas: Automatic planning. Also: model checking, SAT, web service composition, business process management, markov decision processes, network security.
Research Interests: Everything that’s to do with combinatorial search problems: modeling, algorithm design and analysis, effective
implementation, experiments, applications. As such, my research spans the whole breadth from maths to engineering. I’m especially familiar with the heuristic search approach to problem solving, but have worked on other forms of reasoning as well.
Dr. Michael Katz
Saarland University
Foundations of Artificial Intelligence Group Campus E1 1, Room 3.17
66123 Saarbr¨ucken, Germany E-Mail:[email protected]
Subject Areas: Automatic planning. Also: constraint satisfaction and optimization, combinatorics and graph algorithms.
Research Interests: Everything that’s to do with combinatorial search problems: modeling, algorithm design and analysis, effective
implementation, experiments, applications. I’m especially interested in the heuristic search approach to problem solving - automated heuristic derivation, pruning methods, etc.
Dr. Peter Kissmann
Saarland University
Foundations of Artificial Intelligence Group Campus E1 1, Room 3.17
66123 Saarbr¨ucken, Germany E-Mail:[email protected]
Subject Areas: General game playing and automatic planning. Also: symbolic search and binary decision diagrams.
Research Interests: Mainly symbolic search (using BDDs) applied to the above areas. Especially, finding optimal plans (for classical planning problems) and calculating solutions for general single- and
non-simultaneous two-player games. Apart from symbolic search for solving games I have also implemented a general game player. Another interesting topic is the theoretical analysis of BDD sizes for some typical planning and GGP benchmarks.
Ellen Wintringer
Saarland University Campus E1 7, Room 4.22 66123 Saarbr¨ucken, Germany
E-Mail:[email protected]
Secretary shared with Prof. Weickert.
The only member of the group who knows how Saarland University
works!
Agenda
1
About Us
2
About the Organization
3
About the Content
4
About Your Seminar Work
What, Where, When
That one’s easy:
Seminar talks and discussions, here, now.
(Building E1 1 room 3.06, Mondays 14:15–15:45.)
Seminar Web Page:
http:
//fai.cs.uni-saarland.de/teaching/winter12-13/heuristic-search.html
Seminar Mailing List:
Your Questions
Private questions about organization/content/general:
J¨
org Hoffmann.
I might decide to have a “Sprechstunde” (consultation hours) at
some point, but don’t for now.
Come to the front directly after the lecture (limited to 10 minutes
overall).
Or write me an email.
Questions about your paper and all aspects of your seminar work:
Your paper’s supervisor.
Questions to discuss with your fellow students:
Agenda
1
About Us
2
About the Organization
3
About the Content
4
About Your Seminar Work
Content
Date Algorithm Name/Description Student Supervisor
Mon, 15.10. GSAT Cholpon Degenbaeva Joerg Hoffmann
Mon, 22.10. WalkSAT Stefan Bier Joerg Hoffmann
Mon, 29.10. Local Search Topology Thomas Lamma Joerg Hoffmann
Mon, 5.11. A* Charles Hariman Peter Kissmann
Mon, 12.11. Weighted A* Sandy Heidrich Peter Kissmann Mon, 19.11. Optimistic Search Daniel Spanier Peter Kissmann Mon, 26.11. Explicit Estimation Search Nguyen Quynh Peter Kissmann Mon, 3.12. A* Performance Upper Bounds Pramod Mudrakarta Michael Katz Mon, 10.12. A* Performance Lower Bounds Robin Burghartz Michael Katz
Mon, 17.12. IDA* Martin Bromberger Peter Kissmann
Mon, 7.1. Breadth First Heuristic Search Eric Gliemmo Michael Katz Mon, 14.1. Frontier Search Ashkan Taslimi Michael Katz
Mon, 21.1. LRTA* Helge Dombrowski Joerg Hoffmann
Mon, 28.1. LifeLong Planning A* Aliaksandr Talaika Joerg Hoffmann
Do feel free to coordinate/discuss with fellow students working on
related subjects! (It’s up to you whether or not you do this.)
Agenda
1
About Us
2
About the Organization
3
About the Content
4
About Your Seminar Work
Simple Tips for Your Seminar Work
Your brief paper summary:
Summarize the three most important points of the paper.
Summarize the contribution of the paper.
Indicate problems / flaws of the paper.
Indicate further research directions.
Your talk:
30 minutes, PDF slides.
You should have between 15–30 slides.
Put the paper into the context of the other papers presented during
this seminar!
After your talk, there’ll be an open discussion.
Be prepared to answer questions!
Simple Tips for Your Seminar Work, ctd.
Other people’s talks:
Attending these talks is obligatory! (If you have a good reason to
not be able to come to one of the talks, contact your supervisor.)
Read the summary sent prior to the seminar slot; perhaps have a
quick look at the paper (available on the web page).
Pay attention during the talk, take notes reg possible questions.
Participate actively in the discussion.
Your seminar paper:
Longer summary of the paper, in your own words.
Put the paper into the context of the other papers presented during
this seminar!
Agenda
1
About Us
2
About the Organization
3
About the Content
4
About Your Seminar Work
Grading Rules
Your final grade will be based on:
Your brief paper summary.
How accurate, concise, insightful is the summary?
Your technical talk + following discussion.
How clear are the slides and talk?
How accurately is the paper content presented?
How precisely and carefully is the paper placed into the context of the other seminar talks?
How accurately are the discussion questions answered?
Participation in discussions of other papers.
How actively do you participate?
How insightful are your comments and questions?
Your seminar paper.
How accurate and insightful is the seminar paper?
How precisely and carefully is the paper placed into the context of the other seminar talks?