Formal Multi-Objective Synthesis of Continuous-State MDPs

Sofie Haesaert; Petter Nilsson; Sadegh Soudjani

doi:10.1109/LCSYS.2020.3044263

Formal Multi-Objective Synthesis of Continuous-State MDPs

Sofie Haesaert (Corresponding author)
, Petter Nilsson
, Sadegh Soudjani

Research output: Contribution to journal › Article › Academic › peer-review

5 Citations (Scopus)

2 Downloads (Pure)

Abstract

This letter studies formal synthesis of control policies for continuous-state MDPs. In the quest to satisfy complex combinations of probabilistic temporal logic specifications, we derive a robust linear program for policy synthesis that is solved on a finite-state approximation of the system and is then refined back to a policy for the original system. This linear programming approach leverages occupation measures and enables the multi-objective optimizations needed to handle more complex probabilistic specifications.

Original language	English
Article number	9291398
Pages (from-to)	1765-1770
Number of pages	6
Journal	IEEE Control Systems Letters
Volume	5
Issue number	5
DOIs	https://doi.org/10.1109/LCSYS.2020.3044263
Publication status	Published - Nov 2021

Keywords

Approximate simulation relations
formal synthesis
robust linear programming
stochastic systems
temporal logic

Access to Document

10.1109/LCSYS.2020.3044263

Cite this

@article{2156be8303d847019345fc38b2626c7b,

title = "Formal Multi-Objective Synthesis of Continuous-State MDPs",

abstract = "This letter studies formal synthesis of control policies for continuous-state MDPs. In the quest to satisfy complex combinations of probabilistic temporal logic specifications, we derive a robust linear program for policy synthesis that is solved on a finite-state approximation of the system and is then refined back to a policy for the original system. This linear programming approach leverages occupation measures and enables the multi-objective optimizations needed to handle more complex probabilistic specifications.",

keywords = "Approximate simulation relations, formal synthesis, robust linear programming, stochastic systems, temporal logic",

author = "Sofie Haesaert and Petter Nilsson and Sadegh Soudjani",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.",

year = "2021",

month = nov,

doi = "10.1109/LCSYS.2020.3044263",

language = "English",

volume = "5",

pages = "1765--1770",

journal = "IEEE Control Systems Letters",

issn = "2475-1456",

publisher = "Institute of Electrical and Electronics Engineers",

number = "5",

}

TY - JOUR

T1 - Formal Multi-Objective Synthesis of Continuous-State MDPs

AU - Haesaert, Sofie

AU - Nilsson, Petter

AU - Soudjani, Sadegh

PY - 2021/11

Y1 - 2021/11

N2 - This letter studies formal synthesis of control policies for continuous-state MDPs. In the quest to satisfy complex combinations of probabilistic temporal logic specifications, we derive a robust linear program for policy synthesis that is solved on a finite-state approximation of the system and is then refined back to a policy for the original system. This linear programming approach leverages occupation measures and enables the multi-objective optimizations needed to handle more complex probabilistic specifications.

AB - This letter studies formal synthesis of control policies for continuous-state MDPs. In the quest to satisfy complex combinations of probabilistic temporal logic specifications, we derive a robust linear program for policy synthesis that is solved on a finite-state approximation of the system and is then refined back to a policy for the original system. This linear programming approach leverages occupation measures and enables the multi-objective optimizations needed to handle more complex probabilistic specifications.

KW - Approximate simulation relations

KW - formal synthesis

KW - robust linear programming

KW - stochastic systems

KW - temporal logic

UR - https://www.scopus.com/pages/publications/85097958976

U2 - 10.1109/LCSYS.2020.3044263

DO - 10.1109/LCSYS.2020.3044263

M3 - Article

AN - SCOPUS:85097958976

SN - 2475-1456

VL - 5

SP - 1765

EP - 1770

JO - IEEE Control Systems Letters

JF - IEEE Control Systems Letters

IS - 5

M1 - 9291398

ER -

Formal Multi-Objective Synthesis of Continuous-State MDPs

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this