Back to the HomePage.

Literature on Fast Downward and related topics

This page links to some papers that may be of interest for people working on the Fast Downward codebase or using Fast Downward in their research. For papers that are not about the algorithms implemented within Fast Downward but rather about applications and uses of Fast Downward, check the Users page.

So far, this page does not intend to be comprehensive or even cover the most important papers. Rather, we're adding things on an as-needed basis. So being listed on this page doesn't necessarily mean that a paper is more important to Fast Downward than papers not listed here.

For many of the papers shown on this page, presentation slides are available on Malte's publications page. These may provide a gentler introduction to the important concepts than the more technical papers.

Fast Downward architecture

The following two papers describe the original version of Fast Downward (~2003-2004) as a whole. They are still very relevant to the current implementation, but note that many parts have been added to the planner since then (e.g. optimal planning, support for action costs, many more heuristics). The JAIR 2006 paper gives a complete account of the planner except for the translator component. The AIJ 2009 paper describes the translator component.

• Malte Helmert.
  The Fast Downward Planning System.
  Journal of Artificial Intelligence Research 26, pp. 191-246. 2006.

• Malte Helmert.
  Concise finite-domain representations for PDDL planning tasks.
  Artificial Intelligence 173, pp. 503-535. 2009.

Landmark heuristics

The original landmark heuristic was initially developed by Matthias Westphal as a semester project. It was later developed further by Silvia Richter, leading to the work described in the AAAI paper by Silvia Richter, Malte Helmert and Matthias Westphal. After that, Silvia Richter and Matthias Westphal developed the heuristics even further for their LAMA planner, which is described in detail in their JAIR paper. The LAMA implementation served as the basis for the work on admissible heuristics by Erez Karpas and Carmel Domshlak, which in turn was extended to a more general class of landmarks in the paper by Emil Keyder, Silvia Richter and Malte Helmert.

References:

• Matthias Westphal.
  Zielordnungen und Landmarken für SAS+-Planer (in German).
  Semester thesis, Albert-Ludwigs-Universität Freiburg. July 2007.

• Silvia Richter, Malte Helmert and Matthias Westphal.
  Landmarks Revisited.
  In Proceedings of the 23rd AAAI Conference on Artificial Intelligence (AAAI 2008), pp. 975-982. AAAI Press 2008.

• Silvia Richter and Matthias Westphal.
  The LAMA Planner: Guiding Cost-Based Anytime Planning with Landmarks.
  Journal of Artificial Intelligence Research 39, pp. 127-177. 2010.

• Erez Karpas and Carmel Domshlak.
  Cost-Optimal Planning with Landmarks.
  In Proceedings of the Twenty-first International Joint Conference on Artificial Intelligence (IJCAI-09), pp. 1728-1733. 2009.

• Emil Keyder, Silvia Richter and Malte Helmert.
  Sound and Complete Landmarks for And/Or Graphs.
  In Proceedings of the 19th European Conference on Artificial Intelligence (ECAI 2010), pp. 335-340. IOS Press 2010.