Hierarchical Lookahead for Automated Planning
| Bhaskara Marthi | Massachusetts Institute of Technology |
Notice: hosted by David Israel
Date: Tuesday April 29, 2008 at 11:00
Location: EJ228 (SRI E building) (Directions)
|
|
Hierarchical structure in actions has long been considered an important source of leverage in solving large sequential decision problems. Approaches such as HTNs and MAXQ are all based on identifying high level actions (HLAs), which are decomposed into lower level actions, and eventually into primitive actions that can actually be performed in the world. But previous approaches have not answered the question of what it means to have a transition model for an HLA, thus restricting the algorithms that can be used. In recent work, we have proposed an "angelic semantics" for high level actions. This opens the door to a number of a new offline and online planning algorithms that are able to safely commit to or reject plans at the high level. I will also briefly describe an application to robotic motion planning, where our semantics allows a principled combination of a symbolic high-level planner with roadmap-based methods at the low leve l. |
|
|
Please arrive at least 10 minutes early in order to sign in and be escorted to the conference room. SRI is located at 333 Ravenswood Avenue in Menlo Park. Visitors may park in the visitors lot in front of Building E, and should follow the instructions by the lobby phone to be escorted to the meeting room. Detailed directions to SRI, as well as maps, are available from the Visiting AIC web page. ![]()
©2013 SRI International 333 Ravenswood Avenue, Menlo Park, CA 94025-3493
SRI International is an independent, nonprofit corporation. Privacy policy |