A Structure-Sensitive Translation from Hybrid to Numeric Planning

Francesco Percassi; Enrico Scala; Mauro Vallati

doi:10.1007/978-3-031-47546-7_8

A Structure-Sensitive Translation from Hybrid to Numeric Planning

Francesco Percassi, Enrico Scala, Mauro Vallati

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

PDDL+ is an expressive planning formalism that enables the modelling of hybrid domains with both discrete and continuous dynamics. However, its expressiveness makes this language notoriously difficult to handle natively. To address this challenge, translations from time-discrete PDDL+ into numeric PDDL2.1 have been proposed as a way to reframe the rich expressiveness of PDDL+ into a simpler and more manageable formalism. In this work, we first analyse existing translations and provide a means to compare them in terms of induced state space and the size of the reformulated tasks. Secondly, we propose a novel translation leveraging the structure of the problem to generate a compact reformulation. Our experimental results indicate that the novel translation outperforms the existing ones on a range of benchmarks.

Original language	English
Title of host publication	AIxIA 2023 – Advances in Artificial Intelligence
Subtitle of host publication	XXIInd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023, Rome, Italy, November 6–9, 2023, Proceedings
Editors	Roberto Basili, Domenico Lembo, Carla Limongelli, Andrea Orlandini
Publisher	Springer, Cham
Pages	105-118
Number of pages	14
Volume	LNCS 14318
Edition	1st
ISBN (Electronic)	9783031475467
ISBN (Print)	9783031475450
DOIs	https://doi.org/10.1007/978-3-031-47546-7_8
Publication status	Published - 3 Nov 2023
Event	22nd International Conference of the Italian Association for Artificial Intelligence - Rome, Italy Duration: 6 Nov 2023 → 9 Nov 2023 Conference number: 22 http://www.aixia2023.cnr.it/

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer Cham
Volume	LNCS 14318
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	22nd International Conference of the Italian Association for Artificial Intelligence
Abbreviated title	AIxIA 2023
Country/Territory	Italy
City	Rome
Period	6/11/23 → 9/11/23
Internet address	http://www.aixia2023.cnr.it/

Access to Document

10.1007/978-3-031-47546-7_8Licence: Unspecified

Cite this

Percassi, F., Scala, E., & Vallati, M. (2023). A Structure-Sensitive Translation from Hybrid to Numeric Planning. In R. Basili, D. Lembo, C. Limongelli, & A. Orlandini (Eds.), AIxIA 2023 – Advances in Artificial Intelligence: XXIInd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023, Rome, Italy, November 6–9, 2023, Proceedings (1st ed., Vol. LNCS 14318, pp. 105-118). (Lecture Notes in Computer Science; Vol. LNCS 14318). Springer, Cham. https://doi.org/10.1007/978-3-031-47546-7_8

Percassi, Francesco ; Scala, Enrico ; Vallati, Mauro. / A Structure-Sensitive Translation from Hybrid to Numeric Planning. AIxIA 2023 – Advances in Artificial Intelligence: XXIInd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023, Rome, Italy, November 6–9, 2023, Proceedings. editor / Roberto Basili ; Domenico Lembo ; Carla Limongelli ; Andrea Orlandini. Vol. LNCS 14318 1st. ed. Springer, Cham, 2023. pp. 105-118 (Lecture Notes in Computer Science).

@inproceedings{4b973983885b4c088c9e758f31fbe1c2,

title = "A Structure-Sensitive Translation from Hybrid to Numeric Planning",

abstract = "PDDL+ is an expressive planning formalism that enables the modelling of hybrid domains with both discrete and continuous dynamics. However, its expressiveness makes this language notoriously difficult to handle natively. To address this challenge, translations from time-discrete PDDL+ into numeric PDDL2.1 have been proposed as a way to reframe the rich expressiveness of PDDL+ into a simpler and more manageable formalism. In this work, we first analyse existing translations and provide a means to compare them in terms of induced state space and the size of the reformulated tasks. Secondly, we propose a novel translation leveraging the structure of the problem to generate a compact reformulation. Our experimental results indicate that the novel translation outperforms the existing ones on a range of benchmarks.",

keywords = "AI Planning, Hybrid Planning, Model Translation",

author = "Francesco Percassi and Enrico Scala and Mauro Vallati",

note = "Funding Information: Francesco Percassi and Mauro Vallati were supported by a UKRI Future Leaders Fellowship [grant number MR/T041196/1]. Enrico Scala has been partially supported by AIPlan4EU, a project funded by EU Horizon 2020 research and innovation programme under GA n. 101016442. Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 22nd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023 ; Conference date: 06-11-2023 Through 09-11-2023",

year = "2023",

month = nov,

day = "3",

doi = "10.1007/978-3-031-47546-7_8",

language = "English",

isbn = "9783031475450",

volume = "LNCS 14318",

series = "Lecture Notes in Computer Science",

publisher = "Springer, Cham",

pages = "105--118",

editor = "Roberto Basili and Domenico Lembo and Carla Limongelli and Andrea Orlandini",

booktitle = "AIxIA 2023 – Advances in Artificial Intelligence",

address = "Switzerland",

edition = "1st",

url = "http://www.aixia2023.cnr.it/",

}

Percassi, F, Scala, E & Vallati, M 2023, A Structure-Sensitive Translation from Hybrid to Numeric Planning. in R Basili, D Lembo, C Limongelli & A Orlandini (eds), AIxIA 2023 – Advances in Artificial Intelligence: XXIInd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023, Rome, Italy, November 6–9, 2023, Proceedings. 1st edn, vol. LNCS 14318, Lecture Notes in Computer Science, vol. LNCS 14318, Springer, Cham, pp. 105-118, 22nd International Conference of the Italian Association for Artificial Intelligence, Rome, Italy, 6/11/23. https://doi.org/10.1007/978-3-031-47546-7_8

A Structure-Sensitive Translation from Hybrid to Numeric Planning. / Percassi, Francesco; Scala, Enrico; Vallati, Mauro.

AIxIA 2023 – Advances in Artificial Intelligence: XXIInd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023, Rome, Italy, November 6–9, 2023, Proceedings. ed. / Roberto Basili; Domenico Lembo; Carla Limongelli; Andrea Orlandini. Vol. LNCS 14318 1st. ed. Springer, Cham, 2023. p. 105-118 (Lecture Notes in Computer Science; Vol. LNCS 14318).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A Structure-Sensitive Translation from Hybrid to Numeric Planning

AU - Percassi, Francesco

AU - Scala, Enrico

AU - Vallati, Mauro

N1 - Conference code: 22

PY - 2023/11/3

Y1 - 2023/11/3

N2 - PDDL+ is an expressive planning formalism that enables the modelling of hybrid domains with both discrete and continuous dynamics. However, its expressiveness makes this language notoriously difficult to handle natively. To address this challenge, translations from time-discrete PDDL+ into numeric PDDL2.1 have been proposed as a way to reframe the rich expressiveness of PDDL+ into a simpler and more manageable formalism. In this work, we first analyse existing translations and provide a means to compare them in terms of induced state space and the size of the reformulated tasks. Secondly, we propose a novel translation leveraging the structure of the problem to generate a compact reformulation. Our experimental results indicate that the novel translation outperforms the existing ones on a range of benchmarks.

AB - PDDL+ is an expressive planning formalism that enables the modelling of hybrid domains with both discrete and continuous dynamics. However, its expressiveness makes this language notoriously difficult to handle natively. To address this challenge, translations from time-discrete PDDL+ into numeric PDDL2.1 have been proposed as a way to reframe the rich expressiveness of PDDL+ into a simpler and more manageable formalism. In this work, we first analyse existing translations and provide a means to compare them in terms of induced state space and the size of the reformulated tasks. Secondly, we propose a novel translation leveraging the structure of the problem to generate a compact reformulation. Our experimental results indicate that the novel translation outperforms the existing ones on a range of benchmarks.

KW - AI Planning

KW - Hybrid Planning

KW - Model Translation

UR - https://link.springer.com/book/10.1007/978-3-031-47546-7

UR - http://www.scopus.com/inward/record.url?scp=85177219105&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-47546-7_8

DO - 10.1007/978-3-031-47546-7_8

M3 - Conference contribution

SN - 9783031475450

VL - LNCS 14318

T3 - Lecture Notes in Computer Science

SP - 105

EP - 118

BT - AIxIA 2023 – Advances in Artificial Intelligence

A2 - Basili, Roberto

A2 - Lembo, Domenico

A2 - Limongelli, Carla

A2 - Orlandini, Andrea

PB - Springer, Cham

T2 - 22nd International Conference of the Italian Association for Artificial Intelligence

Y2 - 6 November 2023 through 9 November 2023

ER -

Percassi F, Scala E, Vallati M. A Structure-Sensitive Translation from Hybrid to Numeric Planning. In Basili R, Lembo D, Limongelli C, Orlandini A, editors, AIxIA 2023 – Advances in Artificial Intelligence: XXIInd International Conference of the Italian Association for Artificial Intelligence, AIxIA 2023, Rome, Italy, November 6–9, 2023, Proceedings. 1st ed. Vol. LNCS 14318. Springer, Cham. 2023. p. 105-118. (Lecture Notes in Computer Science). Epub 2023 Nov 2. doi: 10.1007/978-3-031-47546-7_8

A Structure-Sensitive Translation from Hybrid to Numeric Planning

Abstract

Publication series

Conference

Access to Document

Other files and links

Cite this