Parallel approximate undirected shortest paths via low hop emulators

Parallel approximate undirected shortest paths via low hop emulators
复制标题

通过低跳模拟器并行近似无向最短路径

DOI:
10.1145/3357713.3384321
复制
发表时间:
2020
期刊:
Symposium on Theory of Computing (STOC
影响因子:
--
通讯作者:
Zhong, Peilin
Zhong, Peilin
中科院分区:
--
文献类型:
--
作者:
Andoni, Alexandr;Stein, Clifford;Zhong, Peilin

文献摘要

参考文献

被引文献

相似文献

本文提出了一个计算无向图最短路的(1+ε)-近似并行算法,其中poly(logn)depth和mppoly(logn)算法适用于n-节点m-边图.虽然具有(接近)最优运行时间的顺序算法已经存在了几十年,但接近最优的并行算法已经成为一个更严峻的挑战。对于(1+ε)-近似,所有具有poly(logn)深度的现有算法对于某个常数tc>0至少执行Ω(mnc)工作。Cohen(STOC'94)提出的这个长期存在的上界的改进已经开放了25年。其中之一是超越跳集的新概念--低跳模拟器--一个多(logn)-近似模拟器图,其中每条最短路径至多有O(loglogn)跳(边)。低跳模拟器的直接应用是poly(logn)-近似单源最短路径(SSSP)、Bourgain嵌入、度量树嵌入和低直径分解的并行算法,所有这些算法都具有poly(logn)深度和mppoly(logn)工作。我们引入了可压缩预处理器并将其应用于谢尔曼的框架(SODA'17)中以解决无容量限制的最小成本流的更一般的问题(也称为,转运问题)。我们的算法使用mpoly(logn)功在poly(logn)深度上计算一个(1+ε)-近似无容量限制的最小费用流。因此,它还改进了最先进的顺序运行时间从mm·2 O(logn)到poly(logn)。
We present a (1+ε)-approximate parallel algorithm for computing shortest paths in undirected graphs, achievingpoly(logn) depth andmpoly(logn) work forn-nodesm-edges graphs. Although sequential algorithms with (nearly) optimal running time have been known for several decades, near-optimal parallel algorithms have turned out to be a much tougher challenge. For (1+ε)-approximation, all prior algorithms withpoly(logn) depth perform at least Ω(mnc) work for some constantc>0. Improving this long-standing upper bound obtained by Cohen (STOC’94) has been open for 25 years.We develop several new tools of independent interest. One of them is a new notion beyond hopsets — low hop emulator — apoly(logn)-approximate emulator graph in which every shortest path has at mostO(loglogn) hops (edges). Direct applications of the low hop emulators are parallel algorithms forpoly(logn)-approximate single source shortest path (SSSP), Bourgain’s embedding, metric tree embedding, and low diameter decomposition, all withpoly(logn) depth andmpoly(logn) work.To boost the approximation ratio to (1+ε), we introduce compressible preconditioners and apply it inside Sherman’s framework (SODA’17) to solve the more general problem of uncapacitated minimum cost flow (a.k.a., transshipment problem). Our algorithm computes a (1+ε)-approximate uncapacitated minimum cost flow inpoly(logn) depth usingmpoly(logn) work. As a consequence, it also improves the state-of-the-art sequential running time fromm· 2O(√logn)tompoly(logn).
DOI: 10.1145/800222.806745
发表时间: 1984-08
期刊: SIAM J. Comput.
影响因子: --
作者:
Faith Ellen;P. Ragde;A. Wigderson
通讯作者: Faith Ellen;P. Ragde;A. Wigderson
简短公告:Semi-MapReduce 遇到拥堵派系
DOI: --
发表时间: 2018
期刊: arXiv.org
影响因子: --
作者:
Soheil Behnezhad;Mahsa Derakhshan;M. Hajiaghayi
通讯作者: M. Hajiaghayi
简短公告:大规模并行近似距离草图
DOI: --
发表时间: 2019
期刊: International Symposium on Distributed Computing
影响因子: --
作者:
M. Dinitz;Yasamin Nazari
通讯作者: Yasamin Nazari
DOI: 10.1006/jagm.1997.0888
发表时间: 1997
期刊: J. Algorithms
影响因子: --
作者:
P. Klein;Sairam Subramanian
通讯作者: Sairam Subramanian
近似最短路径的更快并行算法
DOI: --
发表时间: 2019
期刊: Symposium on the Theory of Computing
影响因子: --
作者:
Jason Li
通讯作者: Jason Li