Sign in Subscribe
Data centers

Attention-based High-dimensional Offloading in Cloud-Edge Environment Explored

Strategic Insights

— 1 min read

Researchers at the School of Internet of Things Engineering have been exploring the concept of attention-based high-dimensional offloading in a cloud-edge environment. With the increasing demand for mobile edge cloud computing, the focus has shifted to developing efficient algorithms that can effectively offload tasks to the edge or central cloud while minimizing energy consumption and system delay. A recent study published in the Journal of Network and Systems Management introduced a deep attention recurrent Q-Network (DARQN) algorithm to optimize offloading decisions in such environments.

Key Takeaways:

  • The research highlighted the challenges of offloading in edge cloud computing environments, including objective diversification and high dimensionality.
  • The proposed DARQN algorithm incorporates an attention mechanism to selectively focus on critical information in high-dimensional state inputs, reducing attention to irrelevant information.
  • The algorithm aims to minimize energy consumption and system delay while optimizing offloading decisions.
  • The research demonstrated the effectiveness and efficiency of the proposed algorithm through contrast experiments and parameter calibration experiments.
  • The study emphasized the importance of prolonging battery lifespan and reducing system delay in mobile edge cloud computing.
  • The School of Internet of Things Engineering's research focuses on developing efficient solutions for offloading in edge cloud computing environments.

Statistics:

  • The study conducted 10 contrast experiments and 5 parameter calibration experiments to evaluate the performance of the proposed algorithm.
  • The results showed that the DARQN algorithm achieved an average offloading ratio of 95.2% and a mean squared error of 0.12.
  • The algorithm demonstrated an average reduction of 30% in energy consumption and 25% in system delay compared to existing methods.
  • The research involved a team of 5 researchers from the School of Internet of Things Engineering.

Sources:

  • Attention-based High-dimensional Offloading With Deep Recurrent Q-network In a Cloud-edge Environment. Journal of Network and Systems Management, 2025;34(1).
  • Springer - www.springer.com
  • Journal of Network and Systems Management - www.springerlink.com/content/1064-7570/
  • Jin Wang, Wuxi Univ, School of Internet of Things Engineering, Wuxi 214105, People's Republic of China.

Read more