Utility-Based Reinforcement Learning: Unifying Single-objective and Multi-objective Reinforcement Learning
Machine Learning
2024-02-06 v1
Abstract
Research in multi-objective reinforcement learning (MORL) has introduced the utility-based paradigm, which makes use of both environmental rewards and a function that defines the utility derived by the user from those rewards. In this paper we extend this paradigm to the context of single-objective reinforcement learning (RL), and outline multiple potential benefits including the ability to perform multi-policy learning across tasks relating to uncertain objectives, risk-aware RL, discounting, and safe RL. We also examine the algorithmic implications of adopting a utility-based approach.
Cite
@article{arxiv.2402.02665,
title = {Utility-Based Reinforcement Learning: Unifying Single-objective and Multi-objective Reinforcement Learning},
author = {Peter Vamplew and Cameron Foale and Conor F. Hayes and Patrick Mannion and Enda Howley and Richard Dazeley and Scott Johnson and Johan Källström and Gabriel Ramos and Roxana Rădulescu and Willem Röpke and Diederik M. Roijers},
journal= {arXiv preprint arXiv:2402.02665},
year = {2024}
}
Comments
Accepted for the Blue Sky Track at AAMAS'24