13
7

Node and Edge Averaged Complexities of Local Graph Problems

Abstract

The node-averaged complexity of a distributed algorithm running on a graph G=(V,E)G=(V,E) is the average over the times at which the nodes VV of GG finish their computation and commit to their outputs. We study the node-averaged complexity for some distributed symmetry breaking problems and provide the following results (among others): - The randomized node-averaged complexity of computing a maximal independent set (MIS) in nn-node graphs of maximum degree Δ\Delta is at least Ω(min{logΔloglogΔ,lognloglogn})\Omega\big(\min\big\{\frac{\log\Delta}{\log\log\Delta},\sqrt{\frac{\log n}{\log\log n}}\big\}\big). This bound is obtained by a novel adaptation of the well-known KMW lower bound [JACM'16]. As a side result, we obtain the same lower bound for the worst-case randomized round complexity for computing an MIS in trees -- this essentially answers open problem 11.15 in the book of Barenboim and Elkin and resolves the complexity of MIS on trees up to an O(loglogn)O(\sqrt{\log\log n}) factor. We also show that, (2,2)(2,2)-ruling sets, which are a minimal relaxation of MIS, have O(1)O(1) randomized node-averaged complexity. - For maximal matching, we show that while the randomized node-averaged complexity is Ω(min{logΔloglogΔ,lognloglogn})\Omega\big(\min\big\{\frac{\log\Delta}{\log\log\Delta},\sqrt{\frac{\log n}{\log\log n}}\big\}\big), the randomized edge-averaged complexity is O(1)O(1). Further, we show that the deterministic edge-averaged complexity of maximal matching is O(log2Δ+logn)O(\log^2\Delta + \log^* n) and the deterministic node-averaged complexity of maximal matching is O(log3Δ+logn)O(\log^3\Delta + \log^* n). - Finally, we consider the problem of computing a sinkless orientation of a graph. The deterministic worst-case complexity of the problem is known to be Θ(logn)\Theta(\log n), even on bounded-degree graphs. We show that the problem can be solved deterministically with node-averaged complexity O(logn)O(\log^* n), while keeping the worst-case complexity in O(logn)O(\log n).

View on arXiv
Comments on this paper