QoS-Guaranteed Adaptive Bandwidth Allocation for Mobile Multiuser Scalable Video Streaming

Bandwidth allocation plays an important role to guarantee the quality of service (QoS) in streaming real-time scalable video to multiuser over wireless time-varying channels. In this letter, an adaptive bandwidth allocation (ABA) scheme is presented to maximize the bandwidth utilization while provisioning each individual user's QoS. Cooperated with traffic rate adaptation, it jointly addresses the spatial diversity among multiuser and the temporal fluctuation of the individual channel condition. A continuous-time Markov decision processes model is introduced to formulate the QoS-guaranteed ABA problem as constrained stochastic optimization. Based on performance sensitivity analysis, a policy iteration algorithm is developed to find the optimal ABA policy online without any prior knowledge of channel statistics. The adaptability to unknown environments and high computation efficiency make it feasible to implement in running-time. Simulations results demonstrate the effectiveness of the proposed method.