Scribd
Upload
236 views1 page

Flow Network Synopsis

This is a synopsis on Flow Networks. An overview of the crucial theoretic concepts is given on one sheet using succint mathematical notations and fittingly short descriptions of the matters at hand. In this way major ideas and concepts like residual networks, network cuts, the minimum cost-flow problem, or the Ford-Fulkerson algorithm to compute a network's maximum flow are comprehensively summarized and easily conveyed to the reader.

Uploaded by

Sören Wellhöfer
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
236 views1 page

Flow Network Synopsis

This is a synopsis on Flow Networks. An overview of the crucial theoretic concepts is given on one sheet using succint mathematical notations and fittingly short descriptions of the matters at hand. In this way major ideas and concepts like residual networks, network cuts, the minimum cost-flow problem, or the Ford-Fulkerson algorithm to compute a network's maximum flow are comprehensively summarized and easily conveyed to the reader.

Uploaded by

Sören Wellhöfer
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 1

Sören Wellhöfer

Flow Networks
Abstract
A flow or transportation network is a finite directed graph where each edge has a maximum capacity
limiting the flow. The flow received by each node equals the out-going amount of flow of that node;
exceptions are flow-producing sources and flow-consuming sinks. Flow networks are used in modeling
pipe networks, commodity flows, traffic networks, market flows, ...
Definitions
Directed graph: G V , E 
– u , v∈ E weighted edges, u , v ∈V vertices
– c u , v  real-valued capacity (weight) of edge
– f u , v net flow from edge u to v
– s , t ∈V sources and sinks

1) Capacity-constraints: f u ,v ≤c u , v
2) Skew symmetry: f u , v=− f v , u
3) Flow conservation: ∑ f  u , w =0, w ∈V , u≠s∧u≠t
Sample flow network N
c  s , v 1, ... , v n , t =min c s ,v 1  , ... , c v n , t 
→ capacity of any path from source to sink

Residual Network: Gr  V , E r 
– c r u , v=c u , v− f u , v → residual capacity of edge
– Path p a = v 1, v 2,. .. , v n  with v 1=s , v n=t
and c r v i , v i 1 0 → augmenting path

Network cuts:
– Partition of G into two sets S and T s∈S∧t ∈T
– Removing all u , v∈ E , u∈S ∧v∈T yields f  s , t =0
– Capacity of cut c  S , T = ∑ c u , v  u∈S ∧v∈T
Residual network of N
Problems / Solutions (s, c, b, t) → augmenting path of
Minimum-cost-flow problem: capacity c r  s , c , b , t =1
– Each edge u , v∈ E costs k  u , v⋅f  u , v
– Minimize total cost ∑ f  u , v⋅k  u ,v

Ford-Fulkerson algorithm:
– 1956, L. Ford, D. Fulkerson: compute maximum flow in O ∣E∣⋅ f  ; termination not guaranteed
– Flow f  s , t  at max when no augmenting paths in G r
– Idea: As long as there are augmenting paths, increase flow along these paths.
I) ∀ u , v∈ E f u , v   0
II) While ∃ p a ∈G r (find augmenting path in residual network using Breadth/Depth-First-Search)
1. Calculate c r  p a 
2. ∀ u , v∈ p a f u , v  f  u , vc r  p a  and f  v ,u  f  v , u−c r  p a  (send flow)

Maximum-flow minimum-cut theorem:


– Minimum cut: cut with smallest capacity in network
– Maximum flow = Minimum (capacity) cut: f max  s , t =c min  S ,T 
2
– Finding min-cuts is hard; fastest deterministic algorithm O ∣V ∣ log ∣V ∣
References
“Graphs, Flows, and the Ford-Fulkerson Algorithm” - Vince Vatter, 2004
“Network flows” - Slides by Kevin Weyn, 2005
“A simple Min-cut algorithm” - Mechthild Stoer, Frank Wagner, Freie Universität Berlin, 1997
“Graph Theory” - Bondy, Murty, Graduate Texts in Mathematics, Springer, 2008
“Flow network”, “Ford-Fulkerson algorithm”, “”Maximum-flow minimum-cut theorem” - http://en.wikipedia.org/wiki, 2009

You might also like