Core Concepts
The authors develop a deterministic round elimination framework to prove tight lower bounds for fundamental distributed graph problems in the Supported LOCAL model, including maximal matching, maximal independent set, ruling sets, arbdefective colorings, and generalizations thereof.
Abstract
The paper studies the complexity of fundamental distributed graph problems in the Supported LOCAL model, where nodes have complete information about the underlying communication network before the start of the computation.
The authors develop a new technique for proving lower bounds in the Supported LOCAL model, which reduces the task of proving a lower bound for a problem Π to the graph-theoretic task of proving non-existence of a solution to another problem Π' that can be derived from Π in a mechanical manner. This allows them to avoid the use of randomness, which was previously required in the round elimination framework for obtaining deterministic lower bounds.
Using this new technique, the authors establish substantial and asymptotically tight Supported LOCAL lower bounds for a variety of fundamental graph problems, including:
Maximal matching and its variants: The authors show that the same tight bounds known for the LOCAL model also hold in the Supported LOCAL model.
Arbdefective coloring: The authors prove that the same lower bounds known for the LOCAL model also hold in the Supported LOCAL model.
Arbdefective colored ruling sets: The authors prove tight lower bounds for this general problem family, which subsumes many important problems as special cases, such as maximal independent set and sinkless orientation.
The authors also provide a lifting theorem that allows them to turn their deterministic lower bounds into randomized lower bounds, extending the celebrated lifting theorem from the LOCAL model to the Supported LOCAL model.