This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Max Flow In optimal paths, every arc has a length, cost, or value. Now, every arc has a width. vi vj bi,j is the amount of flow that can go from vi to vj . 1 / 13 Max Flow 1 1 S 3 1 3 1 1 S 3 1 2 1 1 2 2 T 2 1 2 2 T 2 / 13 Max Matching S T 3 / 13 Ford Fulkerson i max v = ? +v xi,j  xj,k = 0 k v 0 xi,k bi,k j=s j = s, t j=t 4 / 13 Ford Fulkerson Max flow algorithm X X s t Capacity of a cut: C X, X =
iX,jX bi,j 5 / 13 Ford Fulkerson Labeling Process Step 0. vs X Step 1. If vi X and xi,j < bi,j then vj X Step 2. If vi X and xj,i > 0 then vj X 6 / 13 Max Flow a S 3 2 c 2 1 1 1 b 2 2 d 3 T S T 7 / 13 MultiTerminal Max Flow Undirected edges only. i.e. bi,j = bj,i Given edge capacities j 3 i Result of Max Flows j 7 i 8 7 k 5 4 k 8 / 13 Max Flow For any three nodes, Fi,j , Fj,k , Fi,k satisfy Fi,k min(Fi,j , Fj,k ) X X i k Fi,k = C[ X, X ] 9 / 13 Max Flow Fi,j min(Fi,k , Fk,j ) Fj,k min(Fj,i , Fi,k ) Among any three values, two values must be the same, and the third one is bigger or also the same value. 10 / 13 Max Flow Representation of j 7 i 8 7 k Given n = 100  node network
n 2 Max Flows = (n  1) Max Flows 11 / 13 Max Flow Edge Capacities e 2 j 3 k 1 1 2 s i Result of Max Flows e 3 j 3 k 2 s i 3 12 / 13 Max Flows For any undirected network, there is a treeshape network which has the same max flows and the same n  1 cut value. GomoryHu cut tree G  H tree Not only flow  equivalent, but also cut equivalent. 13 / 13 ...
View
Full
Document
This note was uploaded on 02/19/2008 for the course CSE 202 taught by Professor Hu during the Fall '06 term at UCSD.
 Fall '06
 Hu
 Algorithms

Click to edit the document details