EXCALIBUR Adaptive Constraint-Based Agents in Artificial Environments |
[REFINEMENT SEARCH] | [Total-Order Planning] [Partial-Order Planning] [Hierarchical Planning] [Maximal Graphs] |
[ Please note: The project has been discontinued as of May 31, 2005 and is superseded by the projects of the ii Labs. There won't be further updates to these pages. ] |
(Related publication: [PUBLink])
Partial-order planning focuses on relaxing the temporal order of actions. In a refinement step, the position of a new action must not be totally ordered with respect to the plan's other actions. However, the commitment may include a decision on additional ordering relations that are necessary to ensure the consistency of the refinement. All unnecessary choice options for potential orderings are ruled out by the propagation process (sometimes called least commitment).
Partial-order planning is less committed than total-order planning, a total-order planning refinement subset always being a subset of (or equal to) a corresponding partial-order planning refinement subset. However, this does not necessarily result in the superiority of partial-order planning (see [PUBLink] for a detailed discussion).
Examples of partial-order planners are NOAH [PUBLink], TWEAK [PUBLink] and UCPOP [PUBLink].
[REFINEMENT SEARCH] | [Total-Order Planning] [Partial-Order Planning] [Hierarchical Planning] [Maximal Graphs] |
For questions, comments or suggestions, please contact us.
Last update:
May 19, 2001 by Alexander Nareyek