Latency and energy-aware adaptive service migration in mobile edge computing

Area of Science:

• Computer Science
• Network Engineering
• Artificial Intelligence

Background:

• Mobile Edge Computing (MEC) enhances user experience by deploying edge servers.
• Service migration is crucial for maintaining low latency with user mobility.
• High migration frequency and resource competition increase energy consumption and costs.

Purpose of the Study:

• To develop a method for adaptive service migration in MEC that optimizes both latency and energy consumption.
• To address the challenges of dynamic environments, resource competition, and real-time decision-making in MEC service migration.

Main Methods:

• Modeled service migration as a bi-objective optimization problem (latency and energy consumption).
• Formulated the problem as a mixed-integer nonlinear programming problem.
• Designed a Deep Q-Network (DQN) based model, NPER-D3QN, integrating D3QN, Noisy Net, and prioritized experience replay for adaptive migration strategies.

Main Results:

• The NPER-D3QN model generates adaptive migration strategies for low latency and energy consumption.
• Simulation experiments show superior performance compared to existing methods in latency and energy efficiency.
• Demonstrated advantages in dynamic adaptation and multi-objective collaborative optimization.

Conclusions:

• The proposed delay- and energy-aware adaptive service migration method effectively addresses MEC challenges.
• NPER-D3QN offers a robust solution for optimizing service migration in dynamic and resource-competitive MEC environments.
• The method provides significant improvements in both service latency and energy efficiency for MEC providers.