Distributed Graph Algorithms with Predictions.

Joan Boyar, Faith Ellen, and Kim S. Larsen.
Information and Computation. Accepted for publication.

We initiate the study of deterministic distributed graph algorithms with predictions in synchronous message passing systems. The process at each node in the graph is given a prediction, some extra information about the problem instance, which may be incorrect and which it may use to help obtain a solution. The goal is to develop algorithms that are faster when the information is mostly good and are never much slower than algorithms without predictions, even when the information is mostly bad. Concepts from the more general area of algorithms with predictions, such as error measures, consistency, robustness, and smoothness, are adapted to distributed graph algorithms with predictions.

We present a framework for evaluating distributed graph algorithms with predictions and methods for transforming existing algorithms without predictions to effectively use this extra information. Our approach is illustrated by developing algorithms with predictions for the Maximal Independent Set problem.

---

publication

Link to the publication at the publisher's site - subscription may be required.

open access (522 KB)

The same as the publisher's version, when the publisher permits. Otherwise, the author's last version before the publisher's copyright; this is often exactly the same, but sometimes fonts, page numbers, figure numbers, etc. are different. It may also be a full version. However, it is safe to read this version, and at the same time cite the official version, as long as references to concrete locations, numbered theorems, etc. inside the article are avoided.

other publications

Other publications by the author.