Problem
Research questions and friction points this paper is trying to address.
Quantum Learning Methods
Multi-Agent Reinforcement Learning
Barren Plateaus
Multi-Agent Quantum Reinforcement Learning using Evolutionary Optimization
🤖 AI Summary
To address the “barren plateau” problem and parameter inefficiency in training variational quantum circuits (VQCs) for multi-agent reinforcement learning (MARL), this paper proposes a gradient-free evolutionary quantum reinforcement learning framework. It is the first to integrate three genetic algorithm variants into a multi-agent quantum learning paradigm, entirely eliminating quantum gradient computation and substantially mitigating training stagnation. Evaluated on the Coin Game benchmark, the method achieves competitive policy performance with only 2.12% of the trainable parameters required by large classical neural networks—i.e., a 97.88% reduction—while outperforming classical networks of comparable size. The core contribution is the establishment of the first evolutionary quantum optimization architecture tailored for MARL, offering high sample efficiency, strong robustness, and extreme parameter compression. This framework provides a viable quantum-enhanced solution for resource-constrained applications, such as vehicular edge intelligence.
📝 Abstract
Multi-Agent Reinforcement Learning is becoming increasingly more important in times of autonomous driving and other smart industrial applications. Simultaneously a promising new approach to Reinforcement Learning arises using the inherent properties of quantum mechanics, reducing the trainable parameters of a model significantly. However, gradient-based Multi-Agent Quantum Reinforcement Learning methods often have to struggle with barren plateaus, holding them back from matching the performance of classical approaches. While gradient free Quantum Reinforcement Learning methods may alleviate some of these challenges, they too are not immune to the difficulties posed by barren plateaus. We build upon an existing approach for gradient free Quantum Reinforcement Learning and propose three genetic variations with Variational Quantum Circuits for Multi-Agent Reinforcement Learning using evolutionary optimization. We evaluate our genetic variations in the Coin Game environment and also compare them to classical approaches. We showed that our Variational Quantum Circuit approaches perform significantly better compared to a neural network with a similar amount of trainable parameters. Compared to the larger neural network, our approaches archive similar results using $97.88%$ less parameters.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Quantum Learning
Variational Quantum Circuits
Evolutionary Optimization
🔎 Similar Papers
No similar papers found.
