Multi-Agent Reinforcement Learning for Movable Antenna-aided Cell-Free Massive MIMO Systems
Abstract
The inherent non-convex minimum-separation constraints introduced by movable antennas present a formidable challenge to the joint optimization of antenna positions and transmission strategies, rendering conventional methods computationally infeasible, particularly in large-scale cell-free massive multiple-input multiple-output (MIMO). In this work, we propose the graph-based learning individual intrinsic reward heterogeneous-agent proximal policy optimization (GLIIR-HAPPO) algorithm, a novel heterogeneous multi-agent reinforcement learning (MARL) framework that fundamentally overcomes this impasse by systematically decomposing the original coupled optimization into coordinated subproblems. To ensure tractability, we embed the non-convex geometric constraints into a penalty-augmented reward structure and develop a specialized geometric solver that enables the positioning agents to efficiently navigate the high-dimensional action space. Specifically, we propose an architecture featuring a dynamic-interaction graph critic for adaptive cross-role coordination, together with role-conditioned federated distillation that synchronizes same-role policies through compact actor-output statistics. Beyond architectural design, we establish a rigorous theoretical analysis that derives monotonic performance improvement bounds and establishes convergence guarantees for the proposed bi-level optimization. Numerical simulations demonstrate that our framework yields significant sum-rate improvements over state-of-the-art MARL schemes. Moreover, the performance of our advanced architecture closely approaches its fully centralized counterpart, while drastically reducing communication overhead.
Related
- Mutual-Coupling-Aware Joint Precoding and Antenna Positioning for Movable-Antenna MU-MIMO Systems
- Joint Flexible Beamforming, Resource Allocation, and Task Buffering for Movable-Antenna Enabled MEC Systems
- Hierarchical Beam Training and Codebook Design for Movable Antenna-Assisted Near-Field Systems
- Low-Altitude Fluid Antenna Network with Multi-Agent Reinforcement Learning
- Movable Antenna Enabled Anti-Jamming: From Local Optimality Guaranteed Design to a Unified Antenna Position Optimization Framework
BibTeX
@misc{xu2026multiagent,
title = {{Multi-Agent Reinforcement Learning for Movable Antenna-aided Cell-Free Massive MIMO Systems}},
author = {Bokai Xu and Jiayi Zhang and Shuaifei Chen and Ziheng Liu and Huahua Xiao and Derrick Wing Kwan Ng and Bo Ai},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.08387},
eprint = {2610.08387},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.08387},
}
RIS
TY - UNPB
TI - Multi-Agent Reinforcement Learning for Movable Antenna-aided Cell-Free Massive MIMO Systems
AU - Bokai Xu
AU - Jiayi Zhang
AU - Shuaifei Chen
AU - Ziheng Liu
AU - Huahua Xiao
AU - Derrick Wing Kwan Ng
AU - Bo Ai
PY - 2026
DA - 2026/10/06
PB - arXiv
DO - 10.48550/arxiv.2610.08387
UR - https://arxiv.org/abs/2610.08387
AB - The inherent non-convex minimum-separation constraints introduced by movable antennas present a formidable challenge to the joint optimization of antenna positions and transmission strategies, rendering conventional methods computationally infeasible, particularly in large-scale cell-free massive multiple-input multiple-output (MIMO). In this work, we propose the graph-based learning individual intrinsic reward heterogeneous-agent proximal policy optimization (GLIIR-HAPPO) algorithm, a novel heterogeneous multi-agent reinforcement learning (MARL) framework that fundamentally overcomes this impasse by systematically decomposing the original coupled optimization into coordinated subproblems. To ensure tractability, we embed the non-convex geometric constraints into a penalty-augmented reward structure and develop a specialized geometric solver that enables the positioning agents to efficiently navigate the high-dimensional action space. Specifically, we propose an architecture featuring a dynamic-interaction graph critic for adaptive cross-role coordination, together with role-conditioned federated distillation that synchronizes same-role policies through compact actor-output statistics. Beyond architectural design, we establish a rigorous theoretical analysis that derives monotonic performance improvement bounds and establishes convergence guarantees for the proposed bi-level optimization. Numerical simulations demonstrate that our framework yields significant sum-rate improvements over state-of-the-art MARL schemes. Moreover, the performance of our advanced architecture closely approaches its fully centralized counterpart, while drastically reducing communication overhead.
ER -