DQN-based Multi-user Multi-task Offloading in Mobile Edge Computing
Aman Saurav, Biswadip Bandyopadhyay, Pratyay Kuila, Mahesh Chandra Govil · 2024
The proliferation of complex applications has resulted in a significant increase in energy-intensive apps, exerting substantial pressure on device battery life. Mobile Edge Computing (MEC), an emerging technology, addresses this challenge. This paper aims to enhance battery life by minimizing energy consumption through efficient offloading decisions in a multi-user edge computing environment. Given the NP-hard nature of this problem, we propose a Deep Q Network (DQN) model with a prioritized experience replay buffer to optimize the solution. The effectiveness of the proposed algorithm is validated through extensive simulations and comparative analysis across various realistic environments. Experimental results demonstrate an average reduction in energy consumption ranging from 9% to 22%.