Deep Learning and Optimization Approaches in Task Scheduling and Computing Resource Allocation for VR Video Services in Advanced 6G Networks: A Review

Abstract

The rapid evolution of sixth-generation (6G) wireless networks is
enabling immersive applications such as Virtual Reality (VR) video
services, which demand ultra-low latency, high bandwidth, and efficient
resource utilization. Traditional cloud-based architectures often fail to
meet these requirements due to communication delays and scalability
limitations. To address these challenges, edge computing combined with
deep learning and optimization techniques has emerged as an effective
solution for task scheduling and computing resource allocation.
Deep learning models, particularly Deep Reinforcement Learning (DRL),
have shown strong capability in handling complex and dynamic
decision-making scenarios. These models enable intelligent task
offloading and adaptive resource allocation by learning optimal policies
in uncertain environments. Compared to conventional heuristic
approaches, DRL-based frameworks significantly reduce task
completion time and improve system efficiency. In addition,
optimization techniques such as Lyapunov and convex optimization
enhance performance by ensuring system stability and efficient
resource utilization while balancing multiple objectives, including
latency, energy consumption, and throughput.
Hybrid approaches that integrate deep learning with optimization
techniques further improve scalability and adaptability in edge-cloud
environments. However, VR video services introduce additional
challenges, including real-time rendering, high data transmission rates,
and strict Quality of Experience requirements. This review highlights
recent advancements, evaluates existing methods, and identifies key
challenges in developing efficient and scalable solutions for next
generation VR-enabled wireless networks.