Online Algorithms with Unreliable Guidance

TL;DR

This paper presents a unified framework for online algorithms with unreliable guidance, enabling the transformation of standard algorithms into learning-augmented ones with strong guarantees.

Contribution

It introduces the OAG model and the DTB compiler, providing a general approach to enhance online algorithms with learning-augmented capabilities.

Findings

Achieves new trade-offs for bipartite matching with adversarial arrivals.

Obtains optimal solutions for caching and metrical task systems.

Provides a generic compiler applicable to various online algorithms.

Abstract

This paper introduces online algorithms with unreliable guidance (OAG), a model for ML-augmented online decision-making that cleanly separates the predictive and algorithmic components, thus offering a single, well-defined analysis framework that depends only on the problem at hand. Formulated through the lens of request-answer games, the OAG model brings multiple concepts (predictions from the answer space, guide, anytime competitiveness) which enable learning-augmented algorithms to be analyzed independently of predictor-specific choices - such as prediction semantics, error functions, or probing strategies - that would otherwise restrict the algorithm's generality and applicability. The clean framework of the OAG model allows to build the first generic compiler, the drop-or-trust-blindly (DTB) compiler, that turns almost any standard, prediction-free online algorithm into a learning-augmented one. Although simple, we show that the DTB compiler produces new learning-augmented algorithms with strong consistency-robustness guarantees for three classic online problems: we achieve new trade-offs for bipartite matching with adversarial arrival order, and obtain optimal solutions for caching and uniform metrical task systems.