Optimizing Transmission Power and Computational Resource Allocation for Multi-User Task Offloading in MEC to Enhance QoE

With the advent of the mobile edge computing paradigm, there is a shift in computational model to improve the Quality of Experience (QoE) by surrogating resource-hungry workloads from under-resourced user equipment or mobile devices to high computational servers such as mobile edge computing servers (MECS). The inclusion of MEC in a centralized computing model augments the cloud architecture by optimizing communication latency, delay, energy consumption, security, and privacy, etc, and thereby increasing computational efficiency. However, there still exist challenges in latency, energy efficiency when multiple users with multiple tasks are processed in a static MEC environment. Considering the above challenges, we propose a Distributed Deep Learning based Multiple Users Multiple Tasks Offloading (DDL-MUMTO) framework that efficiently allocates computational and transmission resources by optimizing task offloading decisions using Deep-Q-Network(DQN). Our study focuses on minimizing task execution time and energy consumption by addressing edge server resource allocation and algorithm convergence, and optimizing these factors to improve QoE further. A comparative analysis states that a considerable amount of delay and energy consumption is reduced using our proposed algorithm.