Details
| Original language | German |
|---|---|
| Title of host publication | Proceedings of the 1st ICAPS Workshop on Hierarchical Planning |
| Editors | Pascal Bercher, Daniel Höller, Susanne Biundo, Ron Alford |
| Pages | 31-39 |
| Number of pages | 9 |
| Publication status | Published - Jun 2018 |
| Externally published | Yes |
| Event | 1st ICAPS Workshop on Hierarchical Planning 2018 - Delft, Netherlands Duration: 24 Jun 2018 → 29 Jun 2018 |
Abstract
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Proceedings of the 1st ICAPS Workshop on Hierarchical Planning. ed. / Pascal Bercher; Daniel Höller; Susanne Biundo; Ron Alford. 2018. p. 31-39.
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Programmatic Task Network Planning
AU - Mohr, Felix
A2 - Lettmann, Theodor
A2 - Hüllermeier, Eyke
A2 - Wever, Marcel Dominik
A2 - Bercher, Pascal
A2 - Höller, Daniel
A2 - Biundo, Susanne
A2 - Alford, Ron
PY - 2018/6
Y1 - 2018/6
N2 - Many planning problems benefit from extensions of classical planning formalisms and modeling techniques, or even require such extensions. Alternatives such as functional STRIPS or planning modulo theories have therefore been proposed in the past. Somewhat surprisingly, corresponding extensions are not available for hierarchical planning, despite their potential usefulness in applications like automated service composition. In this paper, we present programmatic task networks (PTN), a formalism that extends classical HTN planning in three ways. First, we allow both operations and methods to have outputs instead of only inputs. Second, formulas may contain interpreted terms, in particular interpreted predicates, which are evaluated by a theory realized in an external library. Third, PTN planning allows for a second type of tasks, called oracle tasks, which are not resolved by the planner itself but by external libraries. For the purpose of illustration and evaluation, the approach is applied to a real-world use case in the field of automated service composition.
AB - Many planning problems benefit from extensions of classical planning formalisms and modeling techniques, or even require such extensions. Alternatives such as functional STRIPS or planning modulo theories have therefore been proposed in the past. Somewhat surprisingly, corresponding extensions are not available for hierarchical planning, despite their potential usefulness in applications like automated service composition. In this paper, we present programmatic task networks (PTN), a formalism that extends classical HTN planning in three ways. First, we allow both operations and methods to have outputs instead of only inputs. Second, formulas may contain interpreted terms, in particular interpreted predicates, which are evaluated by a theory realized in an external library. Third, PTN planning allows for a second type of tasks, called oracle tasks, which are not resolved by the planner itself but by external libraries. For the purpose of illustration and evaluation, the approach is applied to a real-world use case in the field of automated service composition.
M3 - Aufsatz in Konferenzband
SP - 31
EP - 39
BT - Proceedings of the 1st ICAPS Workshop on Hierarchical Planning
T2 - 1st ICAPS Workshop on Hierarchical Planning 2018
Y2 - 24 June 2018 through 29 June 2018
ER -