The paper proposes an efficient digital twin (DT) data processing scheme to reduce service latency for multicast short video streaming (MSVS). Key highlights:
A precise measurement model is developed to characterize the relationship among DT model size, user dynamics, and user clustering accuracy. The model size is analyzed based on the weights and biases of neural network layers, as well as the number of centroids in the clustering algorithm. User dynamics are represented by the data variation of user status.
A novel service latency model is proposed by incorporating the impact of user clustering accuracy. The user clustering accuracy affects users' average engagement time, which further impacts the transmission resource demands. The sequential process between video transcoding and multicast transmission is also considered.
A diffusion-based resource management algorithm is designed, which utilizes the denoising technique to improve the action-generation process of the deep reinforcement learning algorithm. This helps address the challenges of complex environments and high-dimensional action spaces.
Simulation results on real-world datasets demonstrate that the proposed DT data processing scheme can effectively reduce service latency compared to benchmark schemes under varying network capacities.
To Another Language
from source content
arxiv.org
Principais Insights Extraídos De
by Xinyu Huang,... às arxiv.org 04-23-2024
https://arxiv.org/pdf/2404.13749.pdfPerguntas Mais Profundas