Flow control for end-to-end delay and power constrained wireless multihop networks

In this paper, we solve the resource allocation problem of minimizing the total average power consumption for end-to-end delay constrained traffic in a multihop wireless network comprised of links with time-varying (Markov-modeled) wireless channels. We are given a set of source-destination pairs. Each source may use multiple routes to transport traffic to the destination at specified data rates. The traffic transported on each route between each src-dst pair is subject to an end-to-end delay guarantee. We present a 2-tier hierarchical solution to solve the above problem. At the bottom tier, each link transmits packets to minimize its long-term average power subject to long-term average delay constraints (BE Collins and RL Cruz, 1999). Given this packet transmission policy at every link and the associated energy-delay and energy-rate trade-off relations, we perform a network-wide optimization of traffic flows at the top tier. This problem is framed as a non-differentiable convex optimization problem is and solved using an incremental sub-gradient optimization technique. We implement our algorithm over sample network topologies and compare its performance with alternate algorithms, noting significant insight of the flow allocation policy as well as substantial gains in energy efficiency and throughput.