Accelerated Backpressure Algorithm

TL;DR

The paper introduces an Accelerated Backpressure (ABP) algorithm that leverages Accelerated Dual Descent to improve convergence rates in network stability, especially under dynamic traffic conditions.

Contribution

It presents a novel ABP algorithm using ADD, providing a distributed, Newton-like approach for faster convergence in network queue management.

Findings

Significant improvement in convergence rate over traditional methods

Guarantees stable queues under varying traffic conditions

Effective in dynamic and stochastic network environments

Abstract

We develop an Accelerated Back Pressure (ABP) algorithm using Accelerated Dual Descent (ADD), a distributed approximate Newton-like algorithm that only uses local information. Our construction is based on writing the backpressure algorithm as the solution to a network feasibility problem solved via stochastic dual subgradient descent. We apply stochastic ADD in place of the stochastic gradient descent algorithm. We prove that the ABP algorithm guarantees stable queues. Our numerical experiments demonstrate a significant improvement in convergence rate, especially when the packet arrival statistics vary over time.