The full text of this item is not available at this time because the student has placed this item under an embargo for a period of time. The Libraries are not authorized to provide a copy of this work during the embargo period, even for Texas A&M users with NetID.
Efficient Media Streaming Through Reinforcement Learning at a Wireless Edge
Abstract
This work proposes an approach to optimizing media streaming performance using reinforcement learning (RL) algorithms. The proposed approach leverages RL’s ability to learn from experience to adapt streaming policies in real-time based on the network conditions and user preferences. Specifically, we apply cross-layer optimization to the existing media streaming model and formulate the current model as a Markov Decision Process (MDP), where the goal is to compute and learn the optimal control policy and maximize the quality of experience. Unlike traditional approaches that rely on predetermined rules, RL algorithms learn to make decisions based on real-time feedback and can adapt to changes in the environment. This allows them to make more efficient use of network resources, leading to higher video quality and reduced buffering times. In this work, we also consider the impact of delayed and partial information sharing in the media streaming environment. In such scenarios, determining accurate state information becomes challenging and hence we propose the use of a prediction simulator referred as automaton to which the past action history data and last known ground truth states are fed as input. The automaton thus predicts the next state, improving the agent’s ability to make optimal decisions, even in situations where state information is partial. Overall, our proposed approach shows promise for improving the performance of media streaming systems in real-world applications, particularly in scenarios with delays or other forms of uncertainty in the environment. The proposed solution can be integrated into existing streaming systems to provide efficient and adaptive media streaming services.
Subject
Media StreamingReinforcement Learning
Markov Decision Process
Automaton
Optimal Policy
Proximal Policy Optimization
Partially Observable Markov Decision Process
Citation
Yegna Narayanan, Gayathri Narayan (2023). Efficient Media Streaming Through Reinforcement Learning at a Wireless Edge. Master's thesis, Texas A&M University. Available electronically from https : / /hdl .handle .net /1969 .1 /199136.