Compositional Reinforcement Learning for Discrete-Time Stochastic Control Systems

Autor: Abolfazl Lavaei, Mateo Perez, Milad Kazemi, Fabio Somenzi, Sadegh Soudjani, Ashutosh Trivedi, Majid Zamani

Publikováno v: IEEE Open Journal of Control Systems, Vol 2, Pp 425-438 (2023)

We propose a compositional approach to synthesize policies for networks of continuous-space stochastic control systems with unknown dynamics using model-free reinforcement learning (RL). The approach is based on implicitly abstracting each subsystem

Externí odkaz: https://doaj.org/article/45505ccfb81a4a648e3ed92d8009b464

Good-for-MDPs Automata for Probabilistic Analysis and Reinforcement Learning

Autor: Dominik Wojtczak, Ashutosh Trivedi, Fabio Somenzi, Ernst Moritz Hahn, Sven Schewe, Mateo Perez

Publikováno v: Tools and Algorithms for the Construction and Analysis of Systems
Tools and Algorithms for the Construction and Analysis of Systems ISBN: 9783030451899
TACAS (1)

We characterize the class of nondeterministic \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::14e920f53c41f3bd7aad4b2218f68718
http://europepmc.org/articles/PMC7439745

Formal Controller Synthesis for Continuous-Space MDPs via Model-Free Reinforcement Learning

Autor: Sadegh Soudjani, Ashutosh Trivedi, Abolfazl Lavaei, Majid Zamani, Fabio Somenzi

Publikováno v: ICCPS

A novel reinforcement learning scheme to synthesize policies for continuous-space Markov decision processes (MDPs) is proposed. This scheme enables one to apply model-free, off-the-shelf reinforcement learning algorithms for finite MDPs to compute op

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7a0c3e59e919047dec1e8c3a91b68494
https://doi.org/10.1109/iccps48487.2020.00017

Reinforcement Learning and Formal Requirements

Autor: Fabio Somenzi, Ashutosh Trivedi

Publikováno v: Numerical Software Verification ISBN: 9783030284220
NSV@CAV

Reinforcement learning is an approach to controller synthesis where agents rely on reward signals to choose actions in order to satisfy the requirements implicit in reward signals. Oftentimes non-experts have to come up with the requirements and thei

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::b4bd7da65aa57b6bf889af24e750b5c0
https://doi.org/10.1007/978-3-030-28423-7_2

Global Almost-Sure Reachability in Stochastic Constant-Rate Multi-Mode Systems

Autor: Fabio Somenzi, Behrouz Touri, Ashutosh Trivedi

Publikováno v: HSCC

A constant-rate multi-mode system is a hybrid system that can switch freely among a finite set of modes, and whose dynamics is specified by a finite number of real-valued variables with mode-dependent constant rates. We introduce and study a stochast

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::0a803f78f0a7859e78e2fd45e7749f7a
https://doi.org/10.1145/3178126.3178149