Automated Generation of MDPs Using Logic Programming and LLMs for Robotic Applications

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Saccon, Enrico, De Martini, Davide, Saveriano, Matteo, Lamon, Edoardo, Palopoli, Luigi, Roveri, Marco
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915717763301376
author Saccon, Enrico
De Martini, Davide
Saveriano, Matteo
Lamon, Edoardo
Palopoli, Luigi
Roveri, Marco
author_facet Saccon, Enrico
De Martini, Davide
Saveriano, Matteo
Lamon, Edoardo
Palopoli, Luigi
Roveri, Marco
contents We present a novel framework that integrates Large Language Models (LLMs) with automated planning and formal verification to streamline the creation and use of Markov Decision Processes (MDP). Our system leverages LLMs to extract structured knowledge in the form of a Prolog knowledge base from natural language (NL) descriptions. It then automatically constructs an MDP through reachability analysis, and synthesises optimal policies using the Storm model checker. The resulting policy is exported as a state-action table for execution. We validate the framework in three human-robot interaction scenarios, demonstrating its ability to produce executable policies with minimal manual effort. This work highlights the potential of combining language models with formal methods to enable more accessible and scalable probabilistic planning in robotics.
format Preprint
id arxiv_https___arxiv_org_abs_2511_23143
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Automated Generation of MDPs Using Logic Programming and LLMs for Robotic Applications
Saccon, Enrico
De Martini, Davide
Saveriano, Matteo
Lamon, Edoardo
Palopoli, Luigi
Roveri, Marco
Robotics
Artificial Intelligence
I.2.8, I.2.9, I.6.3
We present a novel framework that integrates Large Language Models (LLMs) with automated planning and formal verification to streamline the creation and use of Markov Decision Processes (MDP). Our system leverages LLMs to extract structured knowledge in the form of a Prolog knowledge base from natural language (NL) descriptions. It then automatically constructs an MDP through reachability analysis, and synthesises optimal policies using the Storm model checker. The resulting policy is exported as a state-action table for execution. We validate the framework in three human-robot interaction scenarios, demonstrating its ability to produce executable policies with minimal manual effort. This work highlights the potential of combining language models with formal methods to enable more accessible and scalable probabilistic planning in robotics.
title Automated Generation of MDPs Using Logic Programming and LLMs for Robotic Applications
topic Robotics
Artificial Intelligence
I.2.8, I.2.9, I.6.3
url https://arxiv.org/abs/2511.23143