Home Page Seminar Planning Systems SS2005

	Institute of Computer Science — Knowledge-Based Systems Research Group
	Seminar "Planning Systems" SS 2005

General

Course No.:	6.780
Extent:	2 Hrs./week
Teacher:	Prof.Dr. Joachim Hertzberg
Date:	Tuesday, 8:30-10:00
Room:	31/322
Seminar Language:	English
Precondition for participation:	Basic knowledge in Informatics (at least Informatik A or comparable); basic knowledge (introduction level) in AI
First Meeting:	Tue, April 5. Contents: Presentation of the seminar concept; presentation and assignment of topics

Locally within the University of Osnabrück, use the Stud.IP system to get additional information and material!

Abstract

Action planning is among the classical topics of Artificial Intelligence. In recent years, the state of the art has seen a boost in performance of planning systems, owing to some new algorithms as well as to successful efforts in defining a family of planning domain description languages that have made algorithms and systems directly comparable and hence have considerably promoted the exchange of research results. In consequence, planning systems, in particular those that combine planning with scheduling, have recently become quite successful in applications, which range from logistics to space.

The seminar will cover the topic by means of ten selected parts (sets of chapters) from the book

M. Ghallab, D. Nau, P. Traverso: Automated Planning -- Theory and Practice. Morgan Kaufmann, 2004,

so there are ten presentation slots. The topics provide a bit of historical background, but emphasis is on state-of-the-art planning systems and algorithmics as well as on systems that are at the current frontier of the field.

All participants are much encouraged to make use of the PLANET repository (see Web link below) whenever possible, to download and experiment with the respective system they have to present.

Note to Prospective Participants

In addition to covering these ten topics, there will be much emphasis on learning how to present such material, both orally and written. There will be an introductory part by myself about how to give an oral presentation and about how to write a seminar paper (which is in many respects not much different from writing a bachelor or master thesis). Each and every seminar paper written by a participant will be reviewed by two other participants who have to provide a peer review on the presentation (clarity, style etc.). That means, all participants have to present their respective topics in a talk, write a seminar paper, and write two reviews about fellow students' seminar papers. The language to be used in all cases is English.

Each talk is centered around a particular book part. It is expected that additional material (e.g., papers from the references, recent developments concerning the paper topic, code from the PLANET repository) is considered for deepening and broadening the paper presentation. If in doubt, ask your instructor.

Each talk should take ca. 45 minutes plus discussion/questions. Your seminar paper should typically be around ten pages long, applying common-sense formatting (DIN A4, 10-12pt font, regular text margins). A particularly concise presentation or usage of particularly many drawings and figures may cause exceptions to that default.

Seminar Topics/Papers

Selected chapters from the book

M. Ghallab, D. Nau, P. Traverso: Automated Planning – Theory and Practice. Morgan Kaufmann, 2004

Background Reading

S.J. Russell, P. Norvig: "Artificial Intelligence. A Modern Approach", 2nd Edition, Prentice Hall, 2003, Kapitel 11, 12, 16, 17 (Planung)
I. Parberry: A Guide for New Referees in Theoretical Computer Science. Information and Computation, Vol. 112, No. 1, pp. 96-116, 1994 [Download PS]
J. Zobel: Writing in Computer Science. Springer, 2004 (2nd Ed.)

WWW-Resources

PLANET planning reserach and systems repository
Web Site of the Automated Planning textbook
Russell/Norvig's Web Site (AIMA)
Justin Zobel's Web page "Guides for technical writing", comprising links to texts (including a book by himself) about scientific writing in Computer Science and nearby field, giving technical presentations, scientific ethics, and more.