Dynamic Resource Allocation for Streaming Services using Deep Reinforcement Learning: User-Centric Approach to Mobile Usage

In this work, we propose a novel method to dynamically adjust bandwidth, power, and quality while considering the user's residual data availability in a video streaming environment utilizing cellular data. The method is designed to ensure the best video quality without exceeding the user's data usage limit through deep reinforcement learning. The main goal is to maximize the user's quality of experience (QoE) through efficient resource allocation. To validate the proposed method, we conducted comparative experiments with existing algorithms that do not consider data usage and a fixed resource allocation method. The experimental results show that the proposed method increases QoE by 55.01% compared to the methods that do not consider data usage and by 444.21% compared to the fixed resource allocation method, providing users with a superior streaming experience.