Difference between revisions of "Open Problems:11"

From Open Problems in Sublinear Algorithms
Jump to: navigation, search
m (updating header)
Line 4: Line 4:
 
}}
 
}}
 
Given a stream in which edges are inserted and deleted to/from an unweighted, undirected graph, how well can we count triangles and other sub-graphs? Most of the previous work has focused on the case of insertions {{cite|BarYossefKS-02|JowhariG-05|BuriolFLMS-06}} although it appears that one of the algorithms in {{cite|JowhariG-05}} may work when edges can be deleted. Is it possible to match the insert-only bounds when edges are inserted and deleted?
 
Given a stream in which edges are inserted and deleted to/from an unweighted, undirected graph, how well can we count triangles and other sub-graphs? Most of the previous work has focused on the case of insertions {{cite|BarYossefKS-02|JowhariG-05|BuriolFLMS-06}} although it appears that one of the algorithms in {{cite|JowhariG-05}} may work when edges can be deleted. Is it possible to match the insert-only bounds when edges are inserted and deleted?
 +
 +
== Update ==
 +
 +
An algorithm for dynamic streams, i.e., when both insertion and deletion of edges are permitted, was proposed in {{cite|AhnGM-12b}} that matches the best known bounds for insertion-only streaming algorithms {{cite|BuriolFLMS-06}}.

Revision as of 14:37, 20 April 2017

Suggested by Stefano Leonardi
Source Kanpur 2006
Short link https://sublinear.info/11

Given a stream in which edges are inserted and deleted to/from an unweighted, undirected graph, how well can we count triangles and other sub-graphs? Most of the previous work has focused on the case of insertions [BarYossefKS-02,JowhariG-05,BuriolFLMS-06] although it appears that one of the algorithms in [JowhariG-05] may work when edges can be deleted. Is it possible to match the insert-only bounds when edges are inserted and deleted?

Update

An algorithm for dynamic streams, i.e., when both insertion and deletion of edges are permitted, was proposed in [AhnGM-12b] that matches the best known bounds for insertion-only streaming algorithms [BuriolFLMS-06].