A Novel Latency-Aware Resource Allocation and Offloading Strategy With Improved Prioritization and DDQN for Edge-Enabled UDNs

Abstract

Driven by the vision of 6G, the need for diverse computation-intensive and delay-sensitive tasks continues to rise. The integration of mobile edge computing with the ultra-dense network is not only capable of handling traffic from a large number of smart devices but also delivers substantial processing capabilities to the users. This combined network is expected as an effective solution for meeting the latency-critical requirement and will enhance the quality of user experience. Nevertheless, when a massive number of devices offload tasks to edge servers, the problem of channel interference, network load and energy shortage of user devices (UDs) would increase. Therefore, we investigate the joint uplink and downlink resource allocation and task offloading optimization problem in terms of minimizing the overall task delay while sustaining the UD battery life. Thus, to achieve long-term gains while making quick decisions, we propose an improved double deep Q-network scheme named Prioritized double deep Q-network. In this, the prioritized experience replay has been improved by considering the experience freshness factor along with temporal difference error to achieve fast and efficient learning. Extensive numerical results prove the efficacy of the proposed scheme by analyzing delay and energy consumption. Especially, our scheme can considerably decrease the delay by 11.86%, 26.22%, 48.56%, and 61.04% compared to the OELO scheme, DQN scheme, LOS, and EOS, respectively, when the number of UDs varied from 30 to 180.