hide
Free keywords:
Computer Science, Distributed, Parallel, and Cluster Computing, cs.DC
Abstract:
It follows from the Marcus-Spielman-Srivastava proof of the Kadison-Singer
conjecture that if $G=(V,E)$ is a $\Delta$-regular dense expander then there is
an edge-induced subgraph $H=(V,E_H)$ of $G$ of constant maximum degree which is
also an expander. As with other consequences of the MSS theorem, it is not
clear how one would explicitly construct such a subgraph.
We show that such a subgraph (although with quantitatively weaker expansion
and near-regularity properties than those predicted by MSS) can be constructed
with high probability in linear time, via a simple algorithm. Our algorithm
allows a distributed implementation that runs in $\mathcal O(\log n)$ rounds
and does $\bigO(n)$ total work with high probability.
The analysis of the algorithm is complicated by the complex dependencies that
arise between edges and between choices made in different rounds. We sidestep
these difficulties by following the combinatorial approach of counting the
number of possible random choices of the algorithm which lead to failure. We do
so by a compression argument showing that such random choices can be encoded
with a non-trivial compression.
Our algorithm bears some similarity to the way agents construct a
communication graph in a peer-to-peer network, and, in the bipartite case, to
the way agents select servers in blockchain protocols.