Quantum-Train-Based Distributed Multi-Agent Reinforcement Learning

This paper proposes Dist-QTRL, a quantum-train-based framework for Distributed Multi-Agent Reinforcement Learning (RL) that addresses scalability and training complexity. Key contributions include:

• Parameterized quantum circuits reduce trainable parameters from ( N ) to ( \text{polylog}(N) ), boosting efficiency.
• Parallel training across multiple agents modeled as QPUs enhances convergence speed and scalability.
• A hybrid quantum–HPC workflow, enabling quantum training with classical inference using CPUs/GPUs for real-world deployment.

Empirical results validate the convergence and performance gains over centralized RL models, highlighting the promise of quantum-enhanced RL in distributed, high-dimensional environments.

Recommended citation: Chen, Kuan-Cheng; Chen, Samuel Yen-Chi; Liu, Chen-Yu; & Leung, Kin K. (2025). "Quantum-Train-Based Distributed Multi-Agent Reinforcement Learning." Proceedings of the 2025 IEEE Symposium for Multidisciplinary Computational Intelligence Incubators (MCII Companion), 1–5.
Download Paper