The Power of Online Learning in Stochastic Network Optimization

TL;DR

This paper introduces online learning-based control algorithms for stochastic network optimization that achieve near-optimal utility-delay tradeoffs and fast convergence, demonstrating both theoretical guarantees and practical effectiveness.

Contribution

It presents the first algorithms that integrate online learning into stochastic network optimization with explicit near-optimal delay and sub-linear convergence guarantees.

Findings

Achieve near-optimal utility-delay tradeoff of O(ε), O((log(1/ε))^2)

OLAC2 converges in O(ε^{-2/3}) time

Algorithms outperform existing methods in simulations

Abstract

In this paper, we investigate the power of online learning in stochastic network optimization with unknown system statistics {\it a priori}. We are interested in understanding how information and learning can be efficiently incorporated into system control techniques, and what are the fundamental benefits of doing so. We propose two \emph{Online Learning-Aided Control} techniques, $\mathtt{OLAC}$ and $\mathtt{OLAC2}$, that explicitly utilize the past system information in current system control via a learning procedure called \emph{dual learning}. We prove strong performance guarantees of the proposed algorithms: $\mathtt{OLAC}$ and $\mathtt{OLAC2}$ achieve the near-optimal $[O(\epsilon), O([\log(1/\epsilon)]^2)]$ utility-delay tradeoff and $\mathtt{OLAC2}$ possesses an $O(\epsilon^{-2/3})$ convergence time. $\mathtt{OLAC}$ and $\mathtt{OLAC2}$ are probably the first algorithms that simultaneously possess explicit near-optimal delay guarantee and sub-linear convergence time. Simulation results also confirm the superior performance of the proposed algorithms in practice. To the best of our knowledge, our attempt is the first to explicitly incorporate online learning into stochastic network optimization and to demonstrate its power in both theory and practice.