Compositional Learning for Modular Multi-Agent Self-Organizing Networks

TL;DR

This paper presents compositional learning methods for multi-agent self-organizing networks, improving scalability, safety, and efficiency by reducing failures and enhancing throughput through a modular framework.

Contribution

Introduces two compositional learning approaches and a modular framework for multi-agent networks, enhancing scalability, safety, and efficiency over traditional methods.

Findings

Reduced handover failures significantly

Improved network throughput and latency

Faster convergence and higher sample efficiency

Abstract

Self-organizing networks face challenges from complex parameter interdependencies and conflicting objectives. This study introduces two compositional learning approaches-Compositional Deep Reinforcement Learning (CDRL) and Compositional Predictive Decision-Making (CPDM)-and evaluates their performance under training time and safety constraints in multi-agent systems. We propose a modular, two-tier framework with cell-level and cell-pair-level agents to manage heterogeneous agent granularities while reducing model complexity. Numerical simulations reveal a significant reduction in handover failures, along with improved throughput and latency, outperforming conventional multi-agent deep reinforcement learning approaches. The approach also demonstrates superior scalability, faster convergence, higher sample efficiency, and safer training in large-scale self-organizing networks.