Multi-Agent Reinforcement Learning-Based Computation Offloading for Unmanned Aerial Vehicle Post-Disaster Rescue
2024

UAV Task Offloading for Disaster Rescue

publication 10 minutes Evidence: high

Author Information

Author(s): Wang Lixing, Jiao Huirong

Primary Institution: School of Computer Science and Engineering, Northeastern University, Shenyang, China

Hypothesis

Can multi-agent reinforcement learning improve task offloading decisions for UAVs in post-disaster scenarios?

Conclusion

The CER-MADDPG algorithm significantly reduces system overhead and improves stability in UAV task offloading compared to other algorithms.

Supporting Evidence

  • CER-MADDPG outperformed MADDPG and SGRA-PERs in terms of system overhead.
  • The algorithm effectively minimizes both time and energy consumption during UAV operations.
  • Simulations demonstrated improved stability and scalability of the proposed method.

Takeaway

This study shows how drones can work together better by sharing their experiences to make smarter decisions when helping in disasters.

Methodology

The study proposes a multi-agent deep deterministic policy gradient algorithm for UAV task offloading, verified through simulations.

Limitations

The study assumes honest behavior from UAVs and edge devices, which may not reflect real-world scenarios.

Digital Object Identifier (DOI)

10.3390/s24248014

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication