Deterministic Pareto-Optimal Policy Synthesis for Multi-Objective Reinforcement Learning

arXiv cs.LG Papers

Summary

This paper introduces a novel preference-conditioned Bellman operator based on Chebyshev scalarization to compute deterministic Pareto-optimal policies for Multi-Objective Markov Decision Processes, proving its convergence and effectiveness in capturing the entire Pareto frontier.

arXiv:2606.26397v1 Announce Type: new Abstract: Real-world decision-making often requires balancing multiple conflicting objectives, a challenge that standard Reinforcement Learning (RL) frequently addresses by aggregating rewards into a single scalar signal. While effective for simple tasks, this approach often fails to capture the full spectrum of optimal trade-offs, known as the Pareto frontier. In this paper, we introduce a novel preference-conditioned Bellman operator, motivated from the Chebyshev scalarization, designed to compute deterministic Pareto-optimal policies for Multi-Objective Markov Decision Processes (MOMDPs). We prove that this operator satisfies an enveloping property, where the estimated value functions upper-bound the true Pareto frontier, and demonstrate that it monotonically converges to a coverage set of this frontier. Furthermore, we also show how to extract deterministic policies from these converged Q-estimates. This ensures the agent can recover a policy for any given preference, capturing the entire Pareto-optimal frontier while guaranteeing each synthesized policy remains approximately Pareto-optimal. Experimental results validate that our algorithm successfully recovers complex trade-offs, providing a solution for deterministic Pareto-optimal policy synthesis.
Original Article
View Cached Full Text

Cached at: 06/26/26, 05:18 AM

# Deterministic Pareto-Optimal Policy Synthesis for Multi-Objective Reinforcement Learning
Source: [https://arxiv.org/abs/2606.26397](https://arxiv.org/abs/2606.26397)
[View PDF](https://arxiv.org/pdf/2606.26397)

> Abstract:Real\-world decision\-making often requires balancing multiple conflicting objectives, a challenge that standard Reinforcement Learning \(RL\) frequently addresses by aggregating rewards into a single scalar signal\. While effective for simple tasks, this approach often fails to capture the full spectrum of optimal trade\-offs, known as the Pareto frontier\. In this paper, we introduce a novel preference\-conditioned Bellman operator, motivated from the Chebyshev scalarization, designed to compute deterministic Pareto\-optimal policies for Multi\-Objective Markov Decision Processes \(MOMDPs\)\. We prove that this operator satisfies an enveloping property, where the estimated value functions upper\-bound the true Pareto frontier, and demonstrate that it monotonically converges to a coverage set of this frontier\. Furthermore, we also show how to extract deterministic policies from these converged Q\-estimates\. This ensures the agent can recover a policy for any given preference, capturing the entire Pareto\-optimal frontier while guaranteeing each synthesized policy remains approximately Pareto\-optimal\. Experimental results validate that our algorithm successfully recovers complex trade\-offs, providing a solution for deterministic Pareto\-optimal policy synthesis\.

## Submission history

From: Aniruddha Joshi \[[view email](https://arxiv.org/show-email/6ea3df43/2606.26397)\] **\[v1\]**Wed, 24 Jun 2026 21:28:49 UTC \(109 KB\)

Similar Articles

Regularized Offline Policy Optimization with Posterior Hybrid Bayesian Belief

arXiv cs.AI

This paper introduces Posterior Hybrid Bayesian Belief (PhyB), a framework that reformulates the expectation in Bayesian RL as a convex combination over dynamics models, enabling efficient regularized offline policy optimization with bounded objective discrepancy and state-of-the-art performance.