The second part of the course will concern "automated planning" (a.k.a. "AI planning").

Organisational Stuff

Assignement of students to topics (or topics to students):

TopicCovered by
1. Representation Debdeep, Jason, Yogesh
2. Search and Heuristics Debdeep, Neil, Priscilla
3. Planning with Time and Scheduling Amanda, Yogesh
4. Planning in the Real WorldPatrik
5. The Planning Competition Amanda, Jason
6. Learning in Planning Neil, Priscilla

Schedule of presentations: To take place during regular lecture hours, last two weeks in october (15th - 19th; 22nd - 26th). Order of the presentations does not necessarily have to be consistent with the numbering of the topics.


All IJCAI papers can be found at
Papers in the Journal of AI Research can be found at
Many papers are also available from the authors webpages.

Topic 1: Representation
Planning problem description languages; expressivity and complexity; compilation and conversion between languages.

Preprocessing, analysis and other kinds of inference.

Topic 2: Search and Heuristics
Heuristic state-space search.

Partial-order planning, encodings in SAT/CSP, and the like.

From the cutting edge... At this years ICAPS conference (taking place 22nd - 26th of september), will be held a workshop called Heuristics for Domain-Independent Planning: Progress, Ideas, Limitations, Challenges. The workshop program contains many interesting papers, but I particularly recommend the following: Note: Papers should be available on the workshop webpage soon. If not, ask me for a copy.

Topic 3: Planning with Time and Scheduling
Planners that deal with (some aspects of) temporal planning.

Temporal planning models: PDDL2.1, and it's relation to classical planning.


More to come...

Topic 4: Planning in the Real World

Topic 5: The Planning Competition
Summary papers from the first four competitions.
Various kinds of analysis of benchmark domains/problems.
Comments and criticisms related to the competition.

Topic 6: Learning in Planning
Note: I'm not an expert on machine learning, or its use in planning. I've selected the following papers because I believe they illustrate diverse possibilities for integrating learning into planning (i.e., they're about learning different things). There may be other, and better, works on these topics that I'm unaware of.