List of Tutorials

Tutorials Schedule

September 19th, morning (09:00-12:30), amphitheater 13	TUT3
September 19th, evening (14:00 - 17:30), amphitheater 13	TUT5
September 20th, morning (09:00-12:30), amphitheater 13	TUT1
September 20th, evening (14:00 - 17:30), amphitheater 13	TUT2
September 20th, evening (14:00 - 17:30), amphitheater 12	TUT4

TUT1: Heuristics for Classical Planning (With Costs)

Speakers: Emil Keyder and Blai Bonet

Duration: Half-day

Abstract: One of the most effective approaches to classical planning is heuristic search. Here we will review a number of heuristics proposed for this problem, with a heavy focus on suboptimal solutions to the delete relaxation. Heuristics based on other relaxations or decompositions of the problem will also be discussed. We will try to motivate each heuristic in terms of a clear solution to a model resulting from a well-defined relaxation of the problem.

The audience for this tutorial is researchers with a basic knowledge of planning who are interested in practical domain-independent heuristic techniques.

Slides (2.42MB)

TUT2: Petri Nets and Their Relation to Planning

Speakers: Sarah Hickmott, Blai Bonet, Patrik Haslum, Sylvie Thiebaux and Stefan Edelkamp

Duration: Half-day

Abstract: This tutorial presents Petri nets, a formalism for modelling discrete dynamical systems widely used in automated verification (model checking), along with some basic algorithmic tools for the analysis of Petri nets. The focus is on the relation between Petri nets and modelling formalisms used in planning, and the exchange of algorithmic techniques between the two fields.

Slides (2.16MB)

TUT3: Real-Time Planning in Dynamic and Partially-Known Domains

Speakers: Maxim Likhachev and Sven Koenig

Duration: Half-day

Abstract: The tutorial gives an overview of approaches to real-time planning in dynamic and partially-known domains, all of which gain drastic efficiency by planning with a series of A* variants. The tutorial explains the approaches, presents analytical results about their runtimes and plan qualities and demonstrates their application to various problems in AI and robotics, including symbolic planning and motion planning for high degree-of-freedom robot arms, outdoor ground robots and air robots.

Slides (6.41 MB)

TUT4: Representing, Eliciting, and Reasoning with Preferences

Speakers: Carmel Domshlak (tutorial created by Ronen I. Brafman and Carmel Domshlak)

Duration: Half-day

Abstract: When we design an agent that automatically shops on the web or controls a rover on Mars, we don’t want it to buy any item or conduct any experiment. We want it to buy the best available item and conduct the most useful experiment. In short, we want it to act optimally, or at least to strive doing so. But acting well on behalf of a user requires understanding of that user’s goals and preferences. How can an agent obtain this information efficiently when acting on behalf of a lay user? How can this be done with a minimal effort on the part of the user? How does one represent preference information compactly and reasons with it effectively? These questions drive the research conducted in the area of preference modeling, elicitation, representation, and reasoning techniques.  

The tutorial will survey some of the major developments in this area, discussing the problems of decision-making under certainty and uncertainty, and explaining some practical applications of each of these settings and their characteristics. Much emphasis will be placed on graphical models of preference and models of qualitative preferences that are especially suitable for lay users, as well as on algorithmic techniques for preference elicitation and reasoning. We will also try to connect between various knowledge-representation tools for preference handling, and their suitability to be used within action planning techniques.  

The audience for this tutorial is researchers who are interested in the topic in general, and its implication on over-subscription planning, planning with rich goal models, and applications of planning.

Slides (3.96MB)

TUT5: Practical Planning & Scheduling

Speakers: Ari Jonsson, Steve Chien and Mark Johnston

Duration: Half-day

Abstract: Automated planning & scheduling technology has shown considerable promise in a number of domains including space mission operations, production management, and vehicle fleet operations.  In this tutorial we will provide insights into a number of techniques that have been successfully deployed to real world applications, with a bias towards space applications.  These techniques include committed and local search for planning, constraint-based planning in various forms, constraint reasoning and mathematical programming.   The tutorial will focus on answering the following questions: What is automated planning & scheduling technology? How does it work in practice?  What requirements do applications place on planning & scheduling tools? What limitations are encountered and how are they overcome?  Who have used such technology and what were their experiences?

Audience:

• Planning & scheduling practitioners interested in learning more about applications.

• Planning & scheduling researchers interested in getting feedback from practitioners on useful techniques. 

• Junior researchers and students curious as to applied career paths.

• Funding agency program managers interested in technology transfer and deployment.

Short bios:

Dr. Ari Jonsson is Dean of the School of Computer Science, Reykjavik University, Iceland.  Previously he was a senior research scientist with RIACS/NASA Ames Research Center. He was one of the principal architects of the EUROPA planning system, which has been used in a number of applications at NASA.  Among those is the MAPGEN planning system, which supports operations of the Mars Exploration Rover(s) (MER). He holds a PhD. in Computer Science from Stanford University. 

Dr. Steve Chien is a Principal Scientist at the Jet Propulsion Laboratory, California Institute of Technology. He is a three time honoree in the NASA Software of the Year Competition, most recently as the team lead for the Autonomous Sciencecraft, co-winner in 2005.  He has been awarded NASA Medals in 1997, 2000, and 2006 for his efforts in developing and deploying advanced autonomous systems for NASA.  He has led the deployment of autonomy software to five spacecraft/rovers.  He holds a B.S. with Highest Honors in Computer Science, with minors in Mathematics and Economics, M.S., and Ph.D. degrees in Computer Science, all from the University of Illinois.

Dr. Mark Johnston is a Principal Scientist at the Jet Propulsion Laboratory, California Institute of Technology. Prior to joining JPL he has held several positions.  He was Chief Technical Officer for Optimal Planning Science and Interval Logic. Prior to that he was with the Space Telescope Sciences Institute where he played a key role in the development of the SPIKE system used in science planning for the Hubble Space Telescope.  He currently leads an effort to automate scheduling and resource allocation for the Deep Space Network.

Slides: Part 1 (8.41MB), Part 2 (17,3MB), Part 3 (6,32MB)

Tutorial Program Chairs

• Malte Helmert, University of Freiburg, Germany (helmert at informatik <dot> uni-freiburg <dot> de)

• Russell Knight, JPL, USA (knight at jpl <dot> nasa <dot> gov)