Incentivizing Safer Actions in Policy Optimization for Constrained Reinforcement Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Hazra, Somnath, Dasgupta, Pallab, Dey, Soumyajit
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918139120320512
author Hazra, Somnath
Dasgupta, Pallab
Dey, Soumyajit
author_facet Hazra, Somnath
Dasgupta, Pallab
Dey, Soumyajit
contents Constrained Reinforcement Learning (RL) aims to maximize the return while adhering to predefined constraint limits, which represent domain-specific safety requirements. In continuous control settings, where learning agents govern system actions, balancing the trade-off between reward maximization and constraint satisfaction remains a significant challenge. Policy optimization methods often exhibit instability near constraint boundaries, resulting in suboptimal training performance. To address this issue, we introduce a novel approach that integrates an adaptive incentive mechanism in addition to the reward structure to stay within the constraint bound before approaching the constraint boundary. Building on this insight, we propose Incrementally Penalized Proximal Policy Optimization (IP3O), a practical algorithm that enforces a progressively increasing penalty to stabilize training dynamics. Through empirical evaluation on benchmark environments, we demonstrate the efficacy of IP3O compared to the performance of state-of-the-art Safe RL algorithms. Furthermore, we provide theoretical guarantees by deriving a bound on the worst-case error of the optimality achieved by our algorithm.
format Preprint
id arxiv_https___arxiv_org_abs_2509_09208
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Incentivizing Safer Actions in Policy Optimization for Constrained Reinforcement Learning
Hazra, Somnath
Dasgupta, Pallab
Dey, Soumyajit
Machine Learning
Artificial Intelligence
Constrained Reinforcement Learning (RL) aims to maximize the return while adhering to predefined constraint limits, which represent domain-specific safety requirements. In continuous control settings, where learning agents govern system actions, balancing the trade-off between reward maximization and constraint satisfaction remains a significant challenge. Policy optimization methods often exhibit instability near constraint boundaries, resulting in suboptimal training performance. To address this issue, we introduce a novel approach that integrates an adaptive incentive mechanism in addition to the reward structure to stay within the constraint bound before approaching the constraint boundary. Building on this insight, we propose Incrementally Penalized Proximal Policy Optimization (IP3O), a practical algorithm that enforces a progressively increasing penalty to stabilize training dynamics. Through empirical evaluation on benchmark environments, we demonstrate the efficacy of IP3O compared to the performance of state-of-the-art Safe RL algorithms. Furthermore, we provide theoretical guarantees by deriving a bound on the worst-case error of the optimality achieved by our algorithm.
title Incentivizing Safer Actions in Policy Optimization for Constrained Reinforcement Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2509.09208