Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.31388 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913173849767936 |
|---|---|
| author | Park, Giseung Nam, Hyunyoung Byeon, Woohyeon Leshem, Amir Sung, Youngchul |
| author_facet | Park, Giseung Nam, Hyunyoung Byeon, Woohyeon Leshem, Amir Sung, Youngchul |
| contents | Multi-Objective Reinforcement Learning (MORL) extends standard RL by optimizing policies with respect to multiple, often conflicting, objectives. While max-min MORL has emerged as an effective approach for promoting fairness, its applicability remains limited, particularly when constraints must be incorporated. In this paper, we propose a MORL framework that integrates the max-min criterion with explicit constraint satisfaction. We establish a theoretical foundation for the proposed framework and validate the resulting algorithm through convergence analysis and experiments in tabular settings. We further demonstrate the practical relevance of our approach in simulated building thermal control, multi-objective locomotion control, and greenhouse-gas-emission-aware traffic management. Across these domains, our method effectively balances fairness and constraint satisfaction in multi-objective decision-making. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_31388 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Constrained Multi-Objective Reinforcement Learning with Max-Min Criterion Park, Giseung Nam, Hyunyoung Byeon, Woohyeon Leshem, Amir Sung, Youngchul Machine Learning Multi-Objective Reinforcement Learning (MORL) extends standard RL by optimizing policies with respect to multiple, often conflicting, objectives. While max-min MORL has emerged as an effective approach for promoting fairness, its applicability remains limited, particularly when constraints must be incorporated. In this paper, we propose a MORL framework that integrates the max-min criterion with explicit constraint satisfaction. We establish a theoretical foundation for the proposed framework and validate the resulting algorithm through convergence analysis and experiments in tabular settings. We further demonstrate the practical relevance of our approach in simulated building thermal control, multi-objective locomotion control, and greenhouse-gas-emission-aware traffic management. Across these domains, our method effectively balances fairness and constraint satisfaction in multi-objective decision-making. |
| title | Constrained Multi-Objective Reinforcement Learning with Max-Min Criterion |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2605.31388 |