A control-theoretic approach for dynamic adaptive video streaming over HTTP is proposed to improve user-perceived quality-of-experience (QoE) by optimizing bitrate adaptation. The paper addresses the limitations of existing commercial solutions and proposes a novel model predictive control (MPC) algorithm that combines throughput and buffer occupancy information to outperform traditional approaches. The key contributions include: (1) a principled control-theoretic model for bitrate adaptation, (2) a novel MPC algorithm that optimally combines throughput and buffer-based feedback signals, and (3) a practical implementation of the algorithm in a reference video player, dash.js, validated using realistic trace-driven emulations. The paper also evaluates the performance of different algorithms under various network conditions and shows that the proposed MPC approach outperforms existing solutions in terms of median QoE. The implementation is efficient and has low overhead, with a fast and low-memory FastMPC mechanism that approximates the performance of an exact MPC approach. The paper also discusses the challenges of implementing MPC in practice, including computational overhead and deployment issues, and proposes a solution using table enumeration and compression to reduce the complexity of the optimization problem. The results show that the proposed approach achieves significant improvements in QoE compared to existing solutions, particularly in high-throughput variability scenarios.A control-theoretic approach for dynamic adaptive video streaming over HTTP is proposed to improve user-perceived quality-of-experience (QoE) by optimizing bitrate adaptation. The paper addresses the limitations of existing commercial solutions and proposes a novel model predictive control (MPC) algorithm that combines throughput and buffer occupancy information to outperform traditional approaches. The key contributions include: (1) a principled control-theoretic model for bitrate adaptation, (2) a novel MPC algorithm that optimally combines throughput and buffer-based feedback signals, and (3) a practical implementation of the algorithm in a reference video player, dash.js, validated using realistic trace-driven emulations. The paper also evaluates the performance of different algorithms under various network conditions and shows that the proposed MPC approach outperforms existing solutions in terms of median QoE. The implementation is efficient and has low overhead, with a fast and low-memory FastMPC mechanism that approximates the performance of an exact MPC approach. The paper also discusses the challenges of implementing MPC in practice, including computational overhead and deployment issues, and proposes a solution using table enumeration and compression to reduce the complexity of the optimization problem. The results show that the proposed approach achieves significant improvements in QoE compared to existing solutions, particularly in high-throughput variability scenarios.