As the demand for high-quality video streaming experiences continues to rise, the fusion of mathematical models with machine learning algorithms has emerged as a promising approach to enhance adaptive video streaming. This review paper provides a comprehensive exploration of the integration of mathematical models and machine learning in the context of adaptive video streaming decision-making. Beginning with an overview of traditional adaptive streaming and its limitations, we delve into the foundations of mathematical models and machine learning techniques. The paper proposes a conceptual framework for the seamless integration of these approaches, focusing on content prediction, user behavior modeling, and network condition forecasting. Through an examination of case studies and experiments, we showcase instances where this integration has demonstrated significant improvements in streaming performance. The review concludes by addressing current challenges, open issues, and outlining potential avenues for future research in this dynamic and evolving field. This synthesis aims to contribute to the broader understanding of how mathematical models and machine learning synergize to optimize adaptive video streaming and pave the way for a more immersive and personalized viewing experience.
