Project Overview

This project introduces a reinforcement learning–based proactive entanglement swapping (PES) framework for quantum networks.
Instead of performing entanglement swapping only after routing requests arrive, the proposed approach anticipates future demand and extends entanglement along frequently used path segments in advance.

By reducing the effective path length and swapping operations during routing, PES significantly improves network throughput and scalability.

Key Ideas

• Proactive swapping: Performs entanglement swapping ahead of time on high-demand segments
• RL-based decision making: Learns which nodes should perform swapping based on network state
• Reduced routing overhead: Shorter paths and fewer swaps during request servicing

Proactive vs Reactive Swapping

Proactive swapping creates longer entangled links before requests arrive, enabling faster and more reliable routing.

Learning Framework

The RL agent observes network topology, current entanglement states, and request statistics, and selects nodes where swapping should be performed proactively to maximize long-term throughput.

Performance Highlights

• Up to 100%+ improvement in request success rate over random and shortest-path baselines
• Lower routing latency due to reduced swapping during path construction
• Consistent performance gains under high traffic loads

Publication

Tasdiqul Islam, Md Arifuzzaman, Engin Arslan
Reinforcement Learning–Based Proactive Entanglement Swapping for Quantum Networks
IEEE QCNC 2024

BibTeX:

@inproceedings{islam2024reinforcement,
  title={Reinforcement Learning Based Proactive Entanglement Swapping for Quantum Networks},
  author={Islam, Tasdiqul and Arifuzzaman, Md and Arslan, Engin},
  booktitle={2024 International Conference on Quantum Communications, Networking, and Computing (QCNC)},
  pages={135--142},
  year={2024},
  organization={IEEE}
}