ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1904.08037
20
49

Improved Distributed Expander Decomposition and Nearly Optimal Triangle Enumeration

17 April 2019
Yi-Jun Chang
Thatchaphol Saranurak
ArXivPDFHTML
Abstract

An (ϵ,ϕ)(\epsilon,\phi)(ϵ,ϕ)-expander decomposition of a graph G=(V,E)G=(V,E)G=(V,E) is a clustering of the vertices V=V1∪⋯∪VxV=V_{1}\cup\cdots\cup V_{x}V=V1​∪⋯∪Vx​ such that (1) each cluster ViV_{i}Vi​ induces subgraph with conductance at least ϕ\phiϕ, and (2) the number of inter-cluster edges is at most ϵ∣E∣\epsilon|E|ϵ∣E∣. In this paper, we give an improved distributed expander decomposition. Specifically, we construct an (ϵ,ϕ)(\epsilon,\phi)(ϵ,ϕ)-expander decomposition with ϕ=(ϵ/log⁡n)2O(k)\phi=(\epsilon/\log n)^{2^{O(k)}}ϕ=(ϵ/logn)2O(k) in O(n2/k⋅poly(1/ϕ,log⁡n))O(n^{2/k}\cdot\text{poly}(1/\phi,\log n))O(n2/k⋅poly(1/ϕ,logn)) rounds for any ϵ∈(0,1)\epsilon\in(0,1)ϵ∈(0,1) and positive integer kkk. For example, a (0.01,1/polylog⁡n)(0.01,1/\text{poly}\log n)(0.01,1/polylogn)-expander decomposition can be computed in O(nγ)O(n^{\gamma})O(nγ) rounds, for any arbitrarily small constant γ>0\gamma>0γ>0. Previously, the algorithm by Chang, Pettie, and Zhang can construct a (1/6,1/polylog⁡n)(1/6,1/\text{poly}\log n)(1/6,1/polylogn)-expander decomposition using O~(n1−δ)\tilde{O}(n^{1-\delta})O~(n1−δ) rounds for any δ>0\delta>0δ>0, with a caveat that the algorithm is allowed to throw away a set of edges into an extra part which forms a subgraph with arboricity at most nδn^{\delta}nδ. Our algorithm does not have this caveat. By slightly modifying the distributed algorithm for routing on expanders by Ghaffari, Kuhn and Su [PODC'17], we obtain a triangle enumeration algorithm using O~(n1/3)\tilde{O}(n^{1/3})O~(n1/3) rounds. This matches the lower bound by Izumi and Le Gall [PODC'17] and Pandurangan, Robinson and Scquizzato [SPAA'18] of Ω~(n1/3)\tilde{\Omega}(n^{1/3})Ω~(n1/3) which holds even in the CONGESTED CLIQUE model. This provides the first non-trivial example for a distributed problem that has essentially the same complexity (up to a polylogarithmic factor) in both CONGEST and CONGESTED CLIQUE. The key technique in our proof is the first distributed approximation algorithm for finding a low conductance cut that is as balanced as possible. Previous distributed sparse cut algorithms do not have this nearly most balanced guarantee.

View on arXiv
Comments on this paper