The Max-Min Formulation of Multi-Objective Reinforcement Learning: From Theory to a Model-Free Algorithm

2024/06/12 by Giseung Park, Woohyeon Byeon, Park, Giseung +9 · 2 citations
Computer Science · Decision Sciences · #Advanced Multi-Objective Optimization Algorithms #Artificial Intelligence (cs.AI) #Auction Theory and Applications #FOS: Computer and information sciences #Machine Learning (cs.LG) #Reinforcement Learning in Robotics

paper · pdf · doi:10.48550/arxiv.2406.07826

openalex publication_date 2024/06/12 · openalex created_date 2024/06/14 · openalex updated_date 2026/07/28

Abstract

In this paper, we consider multi-objective reinforcement learning, which arises in many real-world problems with multiple optimization goals. We approach the problem with a max-min framework focusing on fairness among the multiple goals and develop a relevant theory and a practical model-free algorithm under the max-min framework. The developed theory provides a theoretical advance in multi-objective reinforcement learning, and the proposed algorithm demonstrates a notable performance improvement over existing baseline methods.

Cited by

• Multi-Objective Reinforcement Learning with Max-Min Criterion: A Game-Theoretic Approach
• FairDICE: Fairness-Driven Offline Multi-Objective Reinforcement Learning

Related

• Multi-Objective Reinforcement Learning with Max-Min Criterion: A Game-Theoretic Approach
• On Multi-objective Policy Optimization as a Tool for Reinforcement Learning: Case Studies in Offline RL and Finetuning
• Min Max Generalization for Two-stage Deterministic Batch Mode Reinforcement Learning: Relaxation Schemes
• Provable Multi-Objective Reinforcement Learning with Generative Models
• Meta-Learning for Multi-objective Reinforcement Learning