You are on page 1of 1

Sean Li CS 4820 Notes Spring 2013 Introduction to Algorithms Lecture 15 2/22/13 Network Flow. Let G = (V, E) be a directed graph.

h. Dene two special notes nodes s V as the source and t V as the sink. Finally, suppose we have a capacity function c : E N. Conservation of Flow. For each node other than the source or sink, the ow of the incoming edges equals the ow of the outgoing edges. A ow in (G, s, t, c) is a map f : E N, satisfying: 1. (Capacity constraint.) f (e) c(e). 2. (Conservation of internal ow.) For all v V, v {s, t}, f in (v) = f out (v), where / in out f (v) = (u,v)E f (u, v) and f (v) = (v,u)E f (v, u). However, s can create ow and t can absorb ow. Given any ow f , for any s, t-cut A, B (a cut is dened as a pair (A, B) such that A B = , A B = V, s A, t B), f (A, B) =
uA vB

f (u, v)
uB vA

f (u, v).

Given a path p, the bottleneck capacity of p is the min edge capacity on p. Theorem. The capacity of the max-ow is the capacity of the min-cut. (Will be proved later.) Ford-Fulkerson Algorithm. Find a path, nd bottleneck, push ow, take residual graph, repeat.

You might also like