Online Optimization with Predictions and Switching Costs: Fast Algorithms and the Fundamental Limit

TL;DR

This paper introduces two gradient-based online algorithms for optimization with predictions and switching costs, demonstrating near-optimal regret bounds and exponential decay with prediction window length.

Contribution

Proposes receding horizon gradient algorithms with proven regret bounds and establishes near-optimality through fundamental lower bounds.

Findings

Regret bounds decay exponentially with prediction window length

RHAG algorithm achieves near-optimal regret performance

Numerical experiments validate theoretical results

Abstract

This paper studies an online optimization problem with a finite prediction window of cost functions and additional switching costs on decisions. We propose two gradient-based online algorithms: Receding Horizon Gradient Descent (RHGD), and Receding Horizon Accelerated Gradient (RHAG). Both algorithms only require a finite number of projected gradient evaluations at each stage. We provide upper bounds on the dynamic regrets of the proposed algorithms and show that the regret upper bounds decay exponentially with the length of the prediction window. Moreover, we study the fundamental lower bound on the dynamic regret for a broad class of deterministic online algorithms. The lower bound is close to RHAG's regret upper bound, indicating that our gradient-based RHAG is a near-optimal online algorithm. Finally, we conduct numerical experiments to complement our theoretical analysis.