Path Selection and Rate Allocation for Video Streaming in Multihop Wireless Networks

In this paper, we investigate the important problem of optimal path selection and rate allocation for concurrent video sessions in a multihop wireless network. We present a novel formulation of the problem, which optimizes the application level performance such as video distortion, while seamlessly incorporating the network layer parameters such as packet delay and loss. For the formulated problem, we propose a branch-and-bound framework, predicated on the powerful reformulation-linearization technique, that can produce (1-epsi) optimal solutions. We demonstrate through analytical and simulation results that the proposed mechanism can fully utilize the network capacity and improve the quality of reconstructed video. The proposed approach provides an important methodology for addressing non-convex, non-polynomial programming problems that arise in wireless networks, especially those involving cross-layer design and optimizations