Revolutionizing load harmony in edge computing networks with probabilistic cellular automata and Markov decision...

Area of Science:

• Computer Science
• Distributed Systems
• Artificial Intelligence

Background:

• Edge computing networks face load balancing challenges due to distributed nature, dynamic workloads, and limited resources.
• Conventional methods like Round-Robin and Threshold-based load balancing lack scalability and flexibility in edge environments.

Purpose of the Study:

• To propose a new framework for achieving steady-state load balance and dynamic adaptation in edge networks.
• To address the limitations of existing load balancing approaches in highly variable edge environments.

Main Methods:

• A novel framework combining Principal Component Analysis (PCA) for distributed, self-organizing load balancing and Markov Decision Process (MDP) for node decision-making.
• PCA classification leverages stochasticity to model interactions between neighboring nodes based on local load.
• MDP framework optimizes load offloading policies using rewards for balance and penalties for exceeding capacity.

Main Results:

• The proposed PCA-MDP system achieves dynamic load balancing with low resource usage variability.
• Experimental results demonstrate higher load distribution efficiency, reward function stability, and faster convergence compared to existing approaches.
• Key performance metrics including load variance, convergence time, and scalability validate the model's robustness.

Conclusions:

• The PCA-MDP model offers a robust and scalable solution for dynamic load balancing in edge computing.
• Optimized resource exploitation and load harmony minimize latency, benefiting real-time applications like autonomous vehicles and IoT.
• This framework provides a foundation for next-generation, scalable edge-computing load-balancing solutions.