| Title: |
Divide and Conquer: Provably Unveiling the Pareto Front with Multi-Objective Reinforcement Learning |
| Authors: |
Röpke, Willem; Reymond, Mathieu; Mannion, Patrick; Roijers, Diederik M.; Nowé, Ann; Rădulescu, Roxana; Sub Intelligent Systems; Vorobeychik, Yevgeniy; Das, Sanmay; Nowe, Ann |
| Publication Year: |
2025 |
| Subject Terms: |
Multi-objective; Pareto front; Reinforcement learning; Artificial Intelligence; Software; Control and Systems Engineering |
| Description: |
An important challenge in multi-objective reinforcement learning is obtaining a Pareto front of policies to attain optimal performance under different preferences. We introduce Iterated Pareto Referent Optimisation (IPRO), which decomposes finding the Pareto front into a sequence of constrained single-objective problems. This enables us to guarantee convergence while providing an upper bound on the distance to undiscovered Pareto optimal solutions at each step. We evaluate IPRO using utility-based metrics and its hypervolume and find that it matches or outperforms methods that require additional assumptions. By leveraging problem-specific single-objective solvers, our approach also holds promise for applications beyond multi-objective reinforcement learning, such as planning and path finding. |
| Document Type: |
book part |
| File Description: |
application/pdf |
| Language: |
English |
| ISSN: |
1548-8403 |
| Relation: |
https://dspace.library.uu.nl/handle/1874/482980 |
| Availability: |
https://dspace.library.uu.nl/handle/1874/482980 |
| Rights: |
info:eu-repo/semantics/OpenAccess |
| Accession Number: |
edsbas.64385EAF |
| Database: |
BASE |