An approach to generate MDPs using HTN representations

Felipe Meneguzzi; Katia Sycara; Yuqing Tang; Simon Parsons

An approach to generate MDPs using HTN representations

Felipe Meneguzzi, Katia Sycara, Yuqing Tang, Simon Parsons

Computing Science

Research output: Chapter in Book/Report/Conference proceeding › Published conference contribution

Abstract

Planning for deterministic and probabilistic domains differ significantly in representations they require, the algorithms that solve them and the way in which results are represented. Hierarchical Task Networks (HTN) and Markov Decision Process (MDPs) are representative formalisms of, respectively, deterministic and stochastic planning. Stochastic domain specifications can easily become opaque to a human designer, especially as the domain size increases. Our research aims to develop algorithms for lossless and automatic mapping of HTN models that are easily intelligible to humans into MDPs. In this paper we develop algorithms to convert deterministic planning domains with HTN domain knowledge and an action error model into MDPs that can then be solved, while maintaining a bound on the number of MDP states.

Original language	English
Title of host publication	Decision Making in Partially Observable, Uncertain Worlds: Exploring Insights from Multiple Communities
Number of pages	6
Publication status	Published - 2011

Access to Document

http://www.meneguzzi.eu/felipe/pubs/dmpouw-htn-mdp-2011.pdf

Cite this

@inproceedings{ff782b37b6de4fb68b9c0815f7db1f59,

title = "An approach to generate MDPs using HTN representations",

abstract = "Planning for deterministic and probabilistic domains differ significantly in representations they require, the algorithms that solve them and the way in which results are represented. Hierarchical Task Networks (HTN) and Markov Decision Process (MDPs) are representative formalisms of, respectively, deterministic and stochastic planning. Stochastic domain specifications can easily become opaque to a human designer, especially as the domain size increases. Our research aims to develop algorithms for lossless and automatic mapping of HTN models that are easily intelligible to humans into MDPs. In this paper we develop algorithms to convert deterministic planning domains with HTN domain knowledge and an action error model into MDPs that can then be solved, while maintaining a bound on the number of MDP states. ",

author = "Felipe Meneguzzi and Katia Sycara and Yuqing Tang and Simon Parsons",

year = "2011",

language = "English",

booktitle = "Decision Making in Partially Observable, Uncertain Worlds: Exploring Insights from Multiple Communities",

}

TY - GEN

T1 - An approach to generate MDPs using HTN representations

AU - Meneguzzi, Felipe

AU - Sycara, Katia

AU - Tang, Yuqing

AU - Parsons, Simon

PY - 2011

Y1 - 2011

N2 - Planning for deterministic and probabilistic domains differ significantly in representations they require, the algorithms that solve them and the way in which results are represented. Hierarchical Task Networks (HTN) and Markov Decision Process (MDPs) are representative formalisms of, respectively, deterministic and stochastic planning. Stochastic domain specifications can easily become opaque to a human designer, especially as the domain size increases. Our research aims to develop algorithms for lossless and automatic mapping of HTN models that are easily intelligible to humans into MDPs. In this paper we develop algorithms to convert deterministic planning domains with HTN domain knowledge and an action error model into MDPs that can then be solved, while maintaining a bound on the number of MDP states.

AB - Planning for deterministic and probabilistic domains differ significantly in representations they require, the algorithms that solve them and the way in which results are represented. Hierarchical Task Networks (HTN) and Markov Decision Process (MDPs) are representative formalisms of, respectively, deterministic and stochastic planning. Stochastic domain specifications can easily become opaque to a human designer, especially as the domain size increases. Our research aims to develop algorithms for lossless and automatic mapping of HTN models that are easily intelligible to humans into MDPs. In this paper we develop algorithms to convert deterministic planning domains with HTN domain knowledge and an action error model into MDPs that can then be solved, while maintaining a bound on the number of MDP states.

M3 - Published conference contribution

BT - Decision Making in Partially Observable, Uncertain Worlds: Exploring Insights from Multiple Communities

ER -

An approach to generate MDPs using HTN representations

Abstract

Access to Document

Fingerprint

Cite this