Inicio  /  Algorithms  /  Vol: 15 Par: 7 (2022)  /  Artículo
ARTÍCULO
TITULO

Learning-Based Online QoE Optimization in Multi-Agent Video Streaming

Yimeng Wang    
Mridul Agarwal    
Tian Lan and Vaneet Aggarwal    

Resumen

Video streaming has become a major usage scenario for the Internet. The growing popularity of new applications, such as 4K and 360-degree videos, mandates that network resources must be carefully apportioned among different users in order to achieve the optimal Quality of Experience (QoE) and fairness objectives. This results in a challenging online optimization problem, as networks grow increasingly complex and the relevant QoE objectives are often nonlinear functions. Recently, data-driven approaches, deep Reinforcement Learning (RL) in particular, have been successfully applied to network optimization problems by modeling them as Markov decision processes. However, existing RL algorithms involving multiple agents fail to address nonlinear objective functions on different agents? rewards. To this end, we leverage MAPG-finite, a policy gradient algorithm designed for multi-agent learning problems with nonlinear objectives. It allows us to optimize bandwidth distributions among multiple agents and to maximize QoE and fairness objectives on video streaming rewards. Implementing the proposed algorithm, we compare the MAPG-finite strategy with a number of baselines, including static, adaptive, and single-agent learning policies. The numerical results show that MAPG-finite significantly outperforms the baseline strategies with respect to different objective functions and in various settings, including both constant and adaptive bitrate videos. Specifically, our MAPG-finite algorithm maximizes QoE by 15.27%" role="presentation">15.27%15.27% 15.27 % and maximizes fairness by 22.47%" role="presentation">22.47%22.47% 22.47 % compared to the standard SARSA algorithm for a 2000 KB/s link.

 Artículos similares

       
 
Wan Teng Tey, Tee Connie, Kan Yeep Choo and Michael Kah Ong Goh    
Traditional methods used to identify and monitor insect species are time-consuming, costly, and fully dependent on the observer?s ability. This paper presents a deep learning-based cicada species recognition system using acoustic signals to classify the ... ver más
Revista: Algorithms

 
Jianying Wang, Yuanpei Wu, Ming Liu, Ming Yang and Haizhao Liang    
Considering the high-efficient trajectory planning requirements for hypersonic vehicles, this paper proposes a real-time trajectory optimization method based on a deep neural network. First, the trajectory optimization model of the hypersonic vehicle ree... ver más
Revista: Aerospace

 
Valerie Bukas Marcus,Noor Azean Atan,Sanitah Mohd Yusof,Umi Mastura     Pág. pp. 100 - 115
Impact brought by COVID-19 changes the whole classroom culture from face to face to the extreme online learning and this includes a course that normally implements traditional face-to-face Service Learning. Educators are required to shift the learning in... ver más

 
Yufei Liu, Feng Zhou, Gang Qiao, Yunjiang Zhao, Guang Yang, Xinyu Liu and Yinheng Lu    
A deep learning-based cyclic shift keying spread spectrum (CSK-SS) underwater acoustic (UWA) communication system is proposed for improving the performance of the conventional system in low signal-to-noise ratio and multipath effects. The proposed deep l... ver más

 
Faiz Hasyim,Tjipto Prastowo,Budi Jatmiko     Pág. pp. 31 - 41
Covid-19 spurs teachers to carry out online learning. This study aimed to analyze the improvement of students' critical thinking skills through online learning based on Android-based PhET Simulation. This research was Quasi-Experimental using one group p... ver más