This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 6.889 Lecture 11: MultipleSource Shortest Paths Christian Sommer csom@mit.edu (figures by Philip Klein) October 19, 2011 SingleSource Shortest Path (SSSP) Problem : given a graph G = ( V,E ) and a source vertex s V , compute shortestpath distance d G ( s,v ) for each v V (and encode shortestpath tree) MultipleSource Shortest Path (MSSP) Problem : given a graph G = ( V,E ) and a source set S V , compute shortestpath distance d G ( s,v ) for some ( s,v ) S V (and encode shortestpath trees rooted at each s S ) Assumption (all of Lecture 11) planar G (extends to bdd. genus ), nonnegative edge lengths : E R + Straightforward SSSP for each source s S , time and encoding size O (  S  n ) This Lecture if all s S on single face f , time and encoding size O ( n log n ) ( independent of  S  / face size!) Why? one important application: allpairs shortest paths between boundary nodes of a piece in r division. requires only time O ( r log r ) (instead of O ( r 3 / 2 ) ) How? Main Idea compute one explicit shortestpath tree rooted at a root r i f , then modify tree to obtain shortestpath tree rooted at neighbor r i +1 f . tree changes: some edges not in tree anymore, some new edges join. modify using dynamic trees , each modification can be done in time O (log n ) how many modifications? which edges to modify? 1 How many modifications? Claim. Going around the face f , for each edge e E : e joins the tree at most once and e leaves the tree at most once....
View
Full
Document
 Fall '11
 ErikDemaine
 Algorithms

Click to edit the document details