Distributed Control and Quality-of-Service in Multihop Wireless Networks

TL;DR

This paper proposes a distributed control algorithm for multihop wireless networks that ensures end-to-end delay requirements and is proven to be throughput optimal, using a novel discrete review approach.

Contribution

It introduces a distributed control method using discrete review and dynamic weights to meet delay constraints, with analytical proof of throughput optimality.

Findings

The algorithm guarantees end-to-end delay requirements.

It is proven to be throughput optimal.

The control is implemented via an incremental gradient ascent method.

Abstract

Control of wireless multihop networks, while simultaneously meeting end-to-end mean delay requirements of different flows is a challenging problem. Additionally, distributed computation of control parameters adds to the complexity. Using the notion of discrete review used in fluid control of networks, a distributed algorithm is proposed for control of multihop wireless networks with interference constraints. The algorithm meets end-to-end mean delay requirements by solving an optimization problem at review instants. The optimization incorporates delay requirements as weights in the function being maximized. The weights are dynamic and vary depending on queue length information. The optimization is done in a distributed manner using an incremental gradient ascent algorithm. The stability of the network under the proposed policy is analytically studied and the policy is shown to be throughput optimal.