07 Network Flow I

7.
N ETWORK F LOW I
‣ max-flow and min-cut problems

‣ Ford–Fulkerson algorithm
‣ max-flow min-cut theorem
‣ capacity-scaling algorithm
‣ shortest augmenting paths
‣ Dinitz’ algorithm
‣ simple unit-capacity networks
Lecture slides by Kevin Wayne 
Copyright © 2005 Pearson-Addison Wesley 
http://www.cs.princeton.edu/~wayne/kleinberg-tardos
Last updated on 1/14/20 2:18 PM

7. N ETWORK F LOW I

SECTION 7.1
Flow network
A flow network is a tuple G = (V, E, s, t, c).

・Digraph (V, E) with source s ∈ V and sink t ∈ V.
・Capacity c(e) ≥ 0 for each e ∈ E. assume all nodes are reachable from s
Intuition. Material flowing through a transportation network;

material originates at source and is sent to sink.
capacity
9
4 15 15 10
10
s 5 8 10 t
15
4 6 15 10
16
3
Minimum-cut problem
Def. An st-cut (cut) is a partition (A, B) of the nodes with s ∈ A and t ∈ B.
Def. Its capacity is the sum of the capacities of the edges from A to B.
cap(A, B) = c(e)
e A
10
s 5 t
15
capacity = 10 + 5 + 15 = 30
4
Minimum-cut problem
cap(A, B) = c(e)
e A
10
s 8 t
don’t include edges

from B to A
16
capacity = 10 + 8 + 16 = 34
5
Minimum-cut problem
cap(A, B) = c(e)
e A
Min-cut problem. Find a cut of minimum capacity.
10
s 8 t
10
capacity = 10 + 8 + 10 = 28
6
Network flow: quiz 1
Which is the capacity of the given st-cut?
A. 11 (20 + 25 − 8 − 11 − 9 − 6)
B. 34 (8 + 11 + 9 + 6)
C. 45 (20 + 25)
D. 79 (20 + 25 + 8 + 11 + 9 + 6)
capacity
s 20 8 10
6 12 8 11 9 8
6
1 16 25 t
7
Maximum-flow problem
Def. An st-flow (flow) f is a function that satisfies:

・For each e ∈ E : 0 f (e) c(e) [capacity]
・For each v ∈ V – {s, t} : e v
f (e) =
e v
f (e) [flow conservation]
flow capacity
inflow at v = 5 + 5 + 0 = 10
5/9 outflow at v = 10 + 0 = 10
5 5
10 0/4 /1 0 / 15 /
10
/ 5
10
s 5/5 5/8 v 10 / 10 t
10
/ 0 0 / 15 10
15 0/4 /6 /
10
10 / 16
8

f (e) =
e v
Def. The value of a flow f is: val(f ) = f (e) f (e)

e s e s
5/9
5 5
10 0/4 /1 0 / 15 /
10
/ 5
10
s 5/5 5/8 10 / 10 t
10
/ 0 0 / 15 10
15 0/4 /6 /
10
value = 5 + 10 + 10 = 25
10 / 16
9

f (e) =
e v
Def. The value of a flow f is: val(f ) = f (e) f (e)

e s e s
Max-flow problem. Find a flow of maximum value.
8/9
2 8
10 0/4 /1 0 / 15 /
10
/ 5
10
s 5/5 8/8 10 / 10 t
13
/ 3 0 / 15 10
15 0/4 /6 /
10
value = 10 + 5 + 13 = 28
13 / 16
10
7. N ETWORK F LOW I

SECTION 7.1
Toward a max-flow algorithm
Greedy algorithm.
・Start with f (e) = 0 for each edge e ∈ E.
・Find an s↝t path P where each edge has f (e) < c(e).
・Augment flow along path P.
・Repeat until you get stuck.
flow capacity
flow network G and flow f
0/4
0
0 /
10 0/2 /8 0/6 10
/
0 value of flow
s 0 / 10 0/9 0 / 10 t 0
12
Greedy algorithm.
0/4
0
0 /
10 0/2 /8 0/6 10
/
0
s 0 / 10 0/9 0 / 10 t 0
13
Greedy algorithm.
0/4
8
— 0
0 /
8 10 0/2 /8 0/6 10
/
0
—
8
s 0 / 10 0/9 —
0 / 10 t 0 +8=8
14
Greedy algorithm.
0/4
0
8 /
10 10 2 —
0/2 /8 0/6 10
/
8
—
2 10
s 0 / 10 —
0/9 —8 / 10 t 8 + 2 = 10
15
Greedy algorithm.
0/4
6
—
0
10 2/2 8
/8 6 —
0/6 /
/ 10
10
6 8
s —
0 / 10 —
2/9 10 / 10 t 10 + 6 = 16
16
Greedy algorithm.
ending flow value = 16
0/4
6
10 2/2 8
/8 6/6 /
/ 10
10
s 6 / 10 8/9 10 / 10 t 16
17
Greedy algorithm.
but max-flow value = 19
3/4
9
10 0/2 7
/8 6/6 /
/ 10
10
s 9 / 10 9/9 10 / 10 t 19
18
Why the greedy algorithm fails
Q. Why does the greedy algorithm fail?

A. Once greedy algorithm increases flow on an edge, it never decreases it.
Ex. Consider flow network G .

・The unique max flow f * has f *(v, w) = 0.
・Greedy algorithm could choose s→v→w→t as first path.
flow network G
v 2 t
2 1 2
s 2 w
Bottom line. Need some mechanism to “undo” a bad decision. 19

Residual network
Original edge. e = (u, v) ∈ E. original flow network G

・Flow f (e). 6 / 17
・Capacity c(e).
u v
flow capacity
Reverse edge. ereverse = (v, u).

・“Undo” flow sent.
residual network Gf residual
Residual capacity. capacity
u 11 v
c(e) f (e) e E
cf (e) = 6
f (e) e E
reverse edge
edges with positive

residual capacity
Residual network. G f = (V, Ef , s, t, cf ). where flow on a reverse edge
・Ef = {e : f (e) < c(e)} ∪ {ereverse : f (e) > 0}.

negates flow on
corresponding forward edge
・Key property: f ʹ is a flow in G f iff f + f ʹ is a flow in G.

20
Augmenting path
Def. An augmenting path is a simple s↝t path in the residual network G f .
Def. The bottleneck capacity of an augmenting path P is the minimum

residual capacity of any edge in P.
Key property. Let f be a flow and let P be an augmenting path in G f .

Then, after calling f ʹ ← AUGMENT( f, c, P), the resulting f ʹ is a flow and
val( f ʹ) = val( f ) + bottleneck(G f, P).
AUGMENT( f, c, P)
________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
δ ← bottleneck capacity of augmenting path P.

FOREACH edge e ∈ P :
IF (e ∈ E) f (e) ← f (e) + δ.
ELSE f (ereverse) ← f (ereverse) – δ.
RETURN f.
________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
21
Which is the augmenting path of highest bottleneck capacity?
A. A→F→G→H
B. A→B→C→D →H
C. A→F→B→G→H
D. A→F→B→G→ C→D→H
residual capacity
5
A 9 B 8 C 6 D
source
5 7 8 7 4 5 6 8 5
E 5 F 2 G 3 H
5 target 22
Ford–Fulkerson algorithm
Ford–Fulkerson augmenting path algorithm.

・Find an s↝t path P in the residual network G f .
FORD–FULKERSON(G)
________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
_
FOREACH edge e ∈ E : f (e) ← 0.

G f ← residual network of G with respect to flow f.
WHILE (there exists an s↝t path P in G f )
f ← AUGMENT( f, c, P).
Update G f. augmenting path
RETURN f.
23
7. N ETWORK F LOW I

‣ Dinitz' algorithm
SECTION 7.2
Relationship between flows and cuts
Flow value lemma. Let f be any flow and let (A, B) be any cut. Then,
the value of the flow f equals the net flow across the cut (A, B).
val(f ) = f (e) f (e)

e A e A
net flow across cut = 5 + 10 + 10 = 25
5/9
5 5
10 0/4 /1 0 / 15 /
10
/ 5
10
s 5/5 5/8 10 / 10 t value of flow = 25
10
/ 0 0 / 15 10
15 0/4 /6 /
10
10 / 16
25
val(f ) = f (e) f (e)

e A e A
net flow across cut = 10 + 5 + 10 = 25
5/9
5 5
10 0/4 /1 0 / 15 /
10
/ 5
10
s 5/5 5/8 10 / 10 t value of flow = 25
10
/ 0 0 / 15 10
15 0/4 /6 /
10
10 / 16
26
val(f ) = f (e) f (e)

e A e A
net flow across cut = (10 + 10 + 5 + 10 + 0 + 0) – (5 + 5 + 0 + 0) = 25
5/9
edges from B to A
5 5
10 0/4 /1 0 / 15 /
10
/ 5
10
s 5/5 5/8 10 / 10 t value of flow = 25
10
/ 0 0 / 15 10
15 0/4 /6 /
10
10 / 16
27
Which is the net flow across the given cut?
A. 11 (20 + 25 − 8 − 11 − 9 − 6)
B. 26 (20 + 22 − 8 − 4 − 4)
C. 42 (20 + 22)
D. 45 (20 + 25)
flow capacity
s 20 / 20 8/8 4 / 10
5 4
1/6 / 8/8 / 4/9 6 4/8
12 11 /
0
1/1 14 / 16 22 / 25 t
28
val(f ) = f (e) f (e)

e A e A
Pf. val(f ) = f (e) f (e)

e s e s
by flow conservation, all terms

except for v = s are 0
= f (e) f (e)
v A e v e v
val(f ) = f (e) f (e)

▪
e A e A
29
Weak duality. Let f be any flow and (A, B) be any cut. Then, val( f ) ≤ cap(A, B).
Pf.
val(f ) = f (e) f (e)
e A e A
f (e)
flow value
lemma e A
c(e)
e A
= cap(A, B) ▪
8/9
2 8
/1 /
10 0/4 5 0 / 15 10
/ 10
10
s 5/5 7/8 9 / 10 t s 5 t
12
2
/
15 0/4 /6 0 / 15 10 15
/
10
12 / 16
value of flow = 27 ≤ capacity of cut = 30 30

Certificate of optimality
Corollary. Let f be a flow and let (A, B) be any cut.

If val( f ) = cap(A, B), then f is a max flow and (A, B) is a min cut.
weak duality
Pf.
・For any flow f ʹ: val( f ʹ) ≤ cap(A, B) = val( f ).
・For any cut (Aʹ, Bʹ): cap(Aʹ, Bʹ) ≥ val( f ) = cap(A, B). ▪
weak duality
8/9
2 8
/1 /
10 0/4 5 0 / 15 10
/ 10
10
s 5/5 8/8 10 / 10 t s 8 t
13
3
/
15 0/4 /6 0 / 15 10
/ 10
10
13 / 16
value of flow = 28 = capacity of cut = 28 31

Max-flow min-cut theorem
Max-flow min-cut theorem. Value of a max flow = capacity of a min cut.
strong duality
1956 IRE TRANXACTIONX ON INFORiMATION THEORY 117
A Note on the Maximum Flow Through a Network*

P. ELIASt, A. FEINSTEINI, AND C. E. SHANNON!
Summary--This note discusses the problem of maximizing the from one terminal to the other in the original network
rate of flow from one terminal to another, through a network which
consists of a number of branches, each of which has a !imited capa- passes through at least one branch in the cut-set. In the
city. The main result is a theorem: The maximum possible flow from network above, some examples of cut-sets are (d, e, f),
left to right through a network is equal to the minimum value among and (b, c, e, g, h), (d, g, h, i) . By a simple cut-set we will
all simple cut-sets. This theorem is applied to solve a more general
problem, in which a number of input nodes and a number of output mean a cut-set such that if any branch is omitted it is no
nodes are used. longer a cut-set. Thus (d, e, f) and (b, c, e, g, h) are simple
cut-sets while (d, g, h, ;) is not. When a simple cut-set is
ONSIDER a two-terminal network such as that deleted from a connected two-terminal network, the net-
of Fig. 1. The branches of the network might work falls into exactly two parts, a left part containing the
c represent communication channels, or, more left terminal and a right part containing the right terminal. 32
generally, any conveying system of limited capacity as, We assign a value to a simple cut-set by taking the sum of
for example, a railroad system, a power feeding system,

Augmenting path theorem. A flow f is a max flow iff no augmenting paths.
Pf. The following three conditions are equivalent for any flow f :
i. There exists a cut (A, B) such that cap(A, B) = val( f ).
ii. f is a max flow.
if Ford–Fulkerson terminates,
iii. There is no augmenting path with respect to f. then f is max flow
[ i ⇒ ii ]
・This is the weak duality corollary. ▪
33

Augmenting path theorem. A flow f is a max flow iff no augmenting paths.
Pf. The following three conditions are equivalent for any flow f :
i. There exists a cut (A, B) such that cap(A, B) = val( f ).
ii. f is a max flow.
iii. There is no augmenting path with respect to f.
[ ii ⇒ iii ] We prove contrapositive: ¬ iii ⇒ ¬ ii.

・Suppose that there is an augmenting path with respect to f.
・Can improve flow f by sending flow along this path.
・Thus, f is not a max flow. ▪
34
[ iii ⇒ i ]
・Let f be a flow with no augmenting paths.
・Let A = set of nodes reachable from s in residual network G f.
・By definition of A: s ∈ A.
・By definition of flow f: t ∉ A. edge e = (v, w) with v ∈ B, w ∈ A
must have f(e) = 0
original flow network G
val(f ) = f (e) f (e)
A B
e A e A
flow value
lemma = c(e) 0 t
<latexit sha1_base64="ZTUKVGmMjB/WcCE6Tg9AU4sscGI=">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</latexit>
e A
= cap(A, B) s
edge e = (v, w) with v ∈ A, w ∈ B

must have f(e) = c(e)
35
Computing a minimum cut from a maximum flow
Theorem. Given any max flow f , can compute a min cut (A, B) in O(m) time.
Pf. Let A = set of nodes reachable from s in residual network G f . ▪
argument from previous slide implies that

capacity of (A, B) = value of flow f
8
13
10 15 2
4
2
A s 5 8 10 t
15
2 3 1 6
10
13
16
36
7. N ETWORK F LOW I

‣ Dinitz' algorithm
SECTION 7.3
Analysis of Ford–Fulkerson algorithm (when capacities are integral)
Assumption. Every edge capacity c(e) is an integer between 1 and C.
Integrality invariant. Throughout Ford–Fulkerson, every edge flow f (e)

and residual capacity cf (e) is an integer.
Pf. By induction on the number of augmenting paths. ▪ consider cut A = { s }
(assumes no parallel edges)
Theorem. Ford–Fulkerson terminates after at most val( f *) ≤ n C

augmenting paths, where f * is a max flow.
Pf. Each augmentation increases the value of the flow by at least 1. ▪
Corollary. The running time of Ford–Fulkerson is O(m n C).

Pf. Can use either BFS or DFS to find an augmenting path in O(m) time. ▪
f(e) is an integer for every e
Integrality theorem. There exists an integral max flow f *.

Pf. Since Ford–Fulkerson terminates, theorem follows from integrality
invariant (and augmenting path theorem). ▪
38
Ford–Fulkerson: exponential example
Q. Is generic Ford–Fulkerson algorithm poly-time in input size?

m, n, and log C
A. No. If max capacity is C, then algorithm can take ≥ C iterations.

・s→v→w→t
・s→w→v→t each augmenting path
sends only 1 unit of flow
・s→v→w→t (# augmenting paths = 2C)
s
・s→w→v→t
・…
・s→v→w→t
C
C
・s→w→v→t
v 1 w
C
t
39
The Ford–Fulkerson algorithm is guaranteed to terminate if the edge

capacities are …
A. Rational numbers. Let D denote the product (or lcm) of the denominators.
Then, every edge flow f (e) and every residual capacity cf (e)
B. Real numbers. is a multiple of 1 / D.
C. Both A and B.
D. Neither A nor B.
40
Choosing good augmenting paths
Use care when selecting augmenting paths.

・Some choices lead to exponential algorithms.
・Clever choices lead to polynomial algorithms.
Pathology. When edge capacities can be irrational, no guarantee

that Ford–Fulkerson terminates (or converges to a maximum flow)!
Goal. Choose augmenting paths so that:

・Can find augmenting paths efficiently.
・Few iterations.
41
Choosing good augmenting paths
Choose augmenting paths with:

・Max bottleneck capacity (“fattest”). how to find?
・Sufficiently large bottleneck capacity. next
・Fewest edges. ahead
Theoretical Improvements in Algorithmic Efficiency

for Network Flow Problems
JACK EDMONDS
University of Waterloo, Waterloo, Ontario, Canada

AND
RICHARD M. K A R P
University of California, Berkeley, California
ABSTRACT. This paper presents new algorithms for t h e m a x i m u m flow problem, the Hitchcock
t r a n s p o r t a t i o n problem, and t h e general m i n i m u m - c o s t flow problem. U p p e r bounds on the
numbers of steps in these algorithms are derived, and are shown to compale favorably with
upper bounds on t h e numbers of steps required by earlier algorithms.
First, the paper states the m a x i m u m flow problem, gives the F o r d - F u l k e r s o n labeling method
for its solution, and points out t h a t an improper choice of flow a u g m e n t i n g p a t h s can lead to
Edmonds-Karp 1972 (USA)

severe c o m p u t a t i o n a l difficulties. T h e n rules of choice t h a t avoid these difficulties are given.
We show t h a t , if each flow a u g m e n t a t i o n is made along an a u g m e n t i n g p a t h h a v i n g a minimum Dinitz 1970 (Soviet Union)
n u m b e r of arcs, t h e n a m a x i m u m flow in an n-node network will be o b t a i n e d a f t e r no more t h a n
~(n a - n) a u g m e n t a t i o n s ; and t h e n we show t h a t if each flow change is chosen to produce a
m a x i m u m increase in the flow value then, provided the capacities are integral, a m a x i m u m flow
will be d e t e r m i n e d within at most 1 + logM/(M--1) if(t, S) a u g m e n t a t i o n s , wheref*(t, s) is the
value of the maximum flow and M is the m a x i m u m n u m b e r of arcs across a cut.
Next a new algorithm is given for the m i n i m u m - c o s t flow problem, in which all s h o r t e s t - p a t h
c o m p u t a t i o n s are performed on networks with all weights nonnegative. In particular, this
a l g o r i t h m solves the n X n assigmnent problem in O(n3) steps. Following t h a t we explore a invented in response to a class
" s c a l i n g " technique for solving a minimum-cost flow problem by t r e a t i n g a sequence of derived
problems w i t h "scaled d o w n " capacities. It is shown t h a t , using this technique, the solution of exercises by Adel’son-Vel’skiĭ
a I i i t c h c o c k t r a n s p o r t a t i o n problem w i t h m sources and n sinks, m ~ n, and m a x i m u m flow B,
requires at most (n + 2) log2 (B/n) flow a u g m e n t a t i o n s . Similar results are also given for the
general minimum-cost flow problem. 42
An a b s t r a c t s t a t i n g the main results of the present paper was presented at the Calgary
Capacity-scaling algorithm
Overview. Choosing augmenting paths with “large” bottleneck capacity.

・Maintain scaling parameter Δ. though not necessarily largest
・Let G f (Δ) be the part of the residual network containing

only those edges with capacity ≥ Δ.
・Any augmenting path in G f (Δ) has bottleneck capacity ≥ Δ.
s s
0
0
10
10
11
11
2
2
0
0
12
12
17
17
2
2
t t
Gf Gf (Δ), Δ = 100 43
Capacity-scaling algorithm
CAPACITY-SCALING(G)
__________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
FOREACH edge e ∈ E : f (e) ← 0.

Δ ← largest power of 2 ≤ C.
WHILE (Δ ≥ 1)
G f (Δ) ← Δ-residual network of G with respect to flow f .
WHILE (there exists an s↝t path P in G f (Δ))
Update G f (Δ). Δ-scaling phase
Δ ← Δ / 2.
RETURN f.
__________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
44
Capacity-scaling algorithm: proof of correctness
Assumption. All edge capacities are integers between 1 and C.
Invariant. The scaling parameter Δ is a power of 2.

Pf. Initially a power of 2; each phase divides Δ by exactly 2. ▪
Integrality invariant. Throughout the algorithm, every edge flow f (e) and
residual capacity cf (e) is an integer.
Pf. Same as for generic Ford–Fulkerson. ▪
Theorem. If capacity-scaling algorithm terminates, then f is a max flow.

Pf.
・By integrality invariant, when Δ = 1 ⇒ G f (Δ) = G f .
・Upon termination of Δ = 1 phase, there are no augmenting paths.
・Result follows augmenting path theorem ▪
45
Capacity-scaling algorithm: analysis of running time
Lemma 1. There are 1 + ⎣log2 C⎦ scaling phases.

Pf. Initially C / 2 < Δ ≤ C; Δ decreases by a factor of 2 in each iteration. ▪
Lemma 2. Let f be the flow at the end of a Δ-scaling phase.

Then, the max-flow value ≤ val( f ) + m Δ.
Pf. Next slide.
Lemma 3. There are ≤ 2m augmentations per scaling phase.

or equivalently,
Pf. at the end
・Let f be the flow at the beginning of a Δ-scaling phase. of a 2Δ-scaling phase
・Lemma 2 ⇒ max-flow value ≤ val( f ) + m (2 Δ).

・Each augmentation in a Δ-phase increases val( f ) by at least Δ. ▪
Theorem. The capacity-scaling algorithm takes O(m2 log C) time.

Pf.
・Lemma 1 + Lemma 3 ⇒ O(m log C) augmentations.
・Finding an augmenting path takes O(m) time. ▪ 46
Capacity-scaling algorithm: analysis of running time
Lemma 2. Let f be the flow at the end of a Δ-scaling phase.

Then, the max-flow value ≤ val( f ) + m Δ.
Pf.
・We show there exists a cut (A, B) such that cap(A, B) ≤ val( f ) + m Δ.
・Choose A to be the set of nodes reachable from s in G f (Δ).
・By definition of A: s ∈ A.
・By definition of flow f: t ∉ A. edge e = (v, w) with v ∈ B, w ∈ A
must have f(e) < Δ
original flow network
val(f ) = f (e) f (e) A B

e A e A
flow value t
lemma (c(e) )
e A e A
c(e) s
e A e A e A
cap(A, B) m
edge e = (v, w) with v ∈ A, w ∈ B
must have f(e) > c(e) – Δ
47
7. N ETWORK F LOW I

SECTION 17.2
Shortest augmenting path
Q. How to choose next augmenting path in Ford–Fulkerson?

A. Pick one that uses the fewest edges.
can find via BFS
SHORTEST-AUGMENTING-PATH(G)
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
FOREACH e ∈ E : f (e) ← 0.
G f ← residual network of G with respect to flow f.
WHILE (there exists an s↝t path in G f )
P ← BREADTH-FIRST-SEARCH(G f ).
Update G f.
RETURN f.
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
49
Shortest augmenting path: overview of analysis
Lemma 1. The length of a shortest augmenting path never decreases.

Pf. Ahead.
number of edges
Lemma 2. After at most m shortest-path augmentations, the length of a

shortest augmenting path strictly increases.
Pf. Ahead.
Theorem. The shortest-augmenting-path algorithm takes O(m2 n) time.

Pf.
・O(m) time to find a shortest augmenting path via BFS.
・There are ≤ m n augmentations.
- at most m augmenting paths of length k Lemma 1 + Lemma 2
- at most n−1 different lengths ▪
augmenting paths are simple paths
50
Shortest augmenting path: analysis
Def. Given a digraph G = (V, E) with source s, its level graph is defined by:
・ℓ(v) = number of edges in shortest s↝v path.
・LG = (V, EG) is the subgraph of G that contains only those edges (v, w) ∈ E
with ℓ(w) = ℓ(v) + 1.
graph G
s t
level graph LG
s t
ℓ= 0 ℓ= 1 ℓ= 2 ℓ= 3 51
Which edges are in the level graph of the following digraph? 
A. D→F.
B. E→F.
C. Both A and B.
D. Neither A nor B.
1 3
B D
source A C E F sink
0 1 2 3
52
Def. Given a digraph G = (V, E) with source s, its level graph is defined by:
・ℓ(v) = number of edges in shortest s↝v path.
・LG = (V, EG) is the subgraph of G that contains only those edges (v, w) ∈ E
with ℓ(w) = ℓ(v) + 1.
Key property. P is a shortest s↝v path in G iff P is an s↝v path in LG.
level graph LG
s t
ℓ= 0 ℓ= 1 ℓ= 2 ℓ= 3 53
Lemma 1. The length of a shortest augmenting path never decreases.

・Let f and f ʹ be flow before and after a shortest-path augmentation.
・Let LG and LG ʹ be level graphs of G f and G f ʹ .
・Only back edges added to G f ′
(any s↝t path that uses a back edge is longer than previous length) ▪
level graph LG
s t
ℓ= 0 ℓ= 1 ℓ= 2 ℓ= 3
level graph LG′
s t
54

・At least one (bottleneck) edge is deleted from LG per augmentation.
・No new edge added to LG until shortest path length strictly increases. ▪
level graph LG
s t
ℓ= 0 ℓ= 1 ℓ= 2 ℓ= 3
level graph LG′
s t
55
Shortest augmenting path: review of analysis
Lemma 1. Throughout the algorithm, the length of a shortest augmenting

path never decreases.

Theorem. The shortest-augmenting-path algorithm takes O(m2 n) time.
56
Shortest augmenting path: improving the running time
Note. Θ(m n) augmentations necessary for some flow networks.

・Try to decrease time per augmentation instead.
・Simple idea ⇒ O(m n2 ) [Dinitz 1970] ahead
・Dynamic trees ⇒ O(m n log n) [Sleator–Tarjan 1983]

JOURNAL OF COMPUTER AND SYSTEM SCIENCES 26, 362-391 (1983)
A Data Structure for Dynamic Trees
DANIEL D. SLEATOR AND ROBERT ENDRE TARJAN
Bell Laboratories, Murray Hill, New Jersey 07974
Received May 8, 1982; revised October 18, 1982
A data structure is proposed to maintain a collection of vertex-disjoint trees under a

sequence of two kinds of operations: a link operation that combines two trees into one by
adding an edge, and a cut operation that divides one tree into two by deleting an edge. Each
operation requires O(log n) time. Using this data structure, new fast algorithms are obtained
for the following problems:
(1) Computing nearest common ancestors.
(2) Solving various network flow problems including finding maximum flows, blocking
flows, and acyclic flows.
(3) Computing certain kinds of constrained minimum spanning trees.
(4) Implementing the network simplex algorithm for minimum-cost flows.
The most significant application is (2); an O(mn log n)-time algorithm is obtained to find a
maximum flow in a network of n vertices and m edges, beating by a factor of log n the fastest
algorithm previously known for sparse graphs.
1. INTR~DIJCTI~N 57
7. N ETWORK F LOW I

SECTION 18.1
Dinitz’ algorithm
Two types of augmentations.

・Normal: length of shortest path does not change.
・Special: length of shortest path strictly increases.
within a phase, length of shortest
Phase of normal augmentations. augmenting path does not change
・Construct level graph LG.

・Start at s, advance along an edge in LG until reach t or get stuck.
・If reach t, augment flow; update LG; and restart from s.
・If get stuck, delete node from LG and retreat to previous node.
construct level graph
s t
level graph LG
59
Dinitz’ algorithm

Phase of normal augmentations.
advance
s t
level graph LG
60
Dinitz’ algorithm

augment
remove from level graph
edges with bottleneck capacity
s t
level graph LG
61
Dinitz’ algorithm

advance
s t
level graph LG
62
Dinitz’ algorithm

retreat
s t
level graph LG
63
Dinitz’ algorithm

advance
s t
level graph LG
64
Dinitz’ algorithm

augment
s t
level graph LG
65
Dinitz’ algorithm

advance
s t
level graph LG
66
Dinitz’ algorithm

retreat
s t
level graph LG
67
Dinitz’ algorithm

retreat
s t
level graph LG
68
Dinitz’ algorithm

end of phase
s t
level graph LG
69
Dinitz’ algorithm (as refined by Even and Itai)
INITIALIZE(G, f ) ADVANCE(v)
_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________ ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
LG ← level-graph of G f. IF (v = t)
P ← ∅. AUGMENT(P).
GOTO ADVANCE(s). Remove saturated edges from LG.
_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
P ← ∅.
GOTO ADVANCE(s).
RETREAT(v)
_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
IF (there exists edge (v, w) ∈ LG)

IF (v = s)
STOP. Add edge (v, w) to P.
ELSE GOTO ADVANCE(w).
Delete v (and all incident edges) from LG.

ELSE
Remove last edge (u, v) from P.
GOTO RETREAT(v).
GOTO ADVANCE(u). ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
70
How to compute the level graph LG efficiently?
A. Depth-first search.
B. Breadth-first search.
C. Both A and B.
D. Neither A nor B.
1 3
B D
source A C E F sink
0 1 2 3
71
Dinitz’ algorithm: analysis
Lemma. A phase can be implemented to run in O(m n) time.

Pf.
・Initialization happens once per phase. O(m) using BFS
・At most m augmentations per phase. O(mn) per phase
(because an augmentation deletes at least one edge from LG)

・At most n retreats per phase. O(m + n) per phase
(because a retreat deletes one node from LG)

・At most m n advances per phase. O(mn) per phase
(because at most n advances before retreat or augmentation) ▪
Theorem. [Dinitz 1970] Dinitz’ algorithm runs in O(m n2) time.

Pf.
・By Lemma, O(m n) time per phase.
・At most n−1 phases (as in shortest-augmenting-path analysis). ▪
72
Augmenting-path algorithms: summary
year method # augmentations running time
1955 augmenting path nC O(m n C)
1972 fattest path m log (mC) O(m2 log n log (mC))
1972 capacity scaling m log C O(m2 log C) fat paths
1985 improved capacity scaling m log C O(m n log C)
1970 shortest augmenting path mn O(m2 n)
1970 level graph mn O(m n2 ) shortest paths
1983 dynamic trees mn O(m n log n )
augmenting-path algorithms with m edges, n nodes, and integer capacities between 1 and C
73
Maximum-flow algorithms: theory highlights
year method worst case discovered by
1951 simplex O(m n2 C) Dantzig
1955 augmenting paths O(m n C) Ford–Fulkerson
1970 shortest augmenting paths O(m n2) Edmonds–Karp, Dinitz
1974 blocking flows O(n3) Karzanov
1983 dynamic trees O(m n log n) Sleator–Tarjan
1985 improved capacity scaling O(m n log C) Gabow
1988 push–relabel O(m n log (n2 / m)) Goldberg–Tarjan
1998 binary blocking flows O(m3/2 log (n2 / m) log C) Goldberg–Rao
2013 compact networks O(m n) Orlin
2014 interior-point methods Õ(m n1/2 log C) Lee–Sidford
2016 electrical flows Õ(m10/7 C1/7) Mądry
20xx
max-flow algorithms with m edges, n nodes, and integer capacities between 1 and C 74
Maximum-flow algorithms: practice
Push–relabel algorithm (SECTION 7.4). [Goldberg–Tarjan 1988]

Increases flow one edge at a time instead of one augmenting path at a time.
A New Approach to the Maximum-Flow Problem
ANDREW V. GOLDBERG
Massachusetts Institute of Technology, Cambridge, Massachusetts
AND
ROBERT E. TARJAN
Princeton University, Princeton, New Jersey, and AT&T Bell Laboratories, Murray Hill, New Jersey
Abstract. All previously known efftcient maximum-flow algorithms work by finding augmenting paths,
either one path at a time (as in the original Ford and Fulkerson algorithm) or all shortest-length
augmenting paths at once (using the layered network approach of Dinic). An alternative method based
on the preflow concept of Karzanov is introduced. A preflow is like a flow, except that the total amount
flowing into a vertex is allowed to exceed the total amount flowing out. The method maintains a preflow
in the original network and pushes local flow excess toward the sink along what are estimated to be
shortest paths. The algorithm and its analysis are simple and intuitive, yet the algorithm runs as fast as
any other known method on dense.graphs, achieving an O(n)) time bound on an n-vertex graph. By
incorporating the dynamic tree data structure of Sleator and Tarjan, we obtain a version of the algorithm
running in O(nm log(n’/m)) time on an n-vertex, m-edge graph. This is as fast as any known method
for any graph density and faster on graphs of moderate density. The algorithm also admits efticient
distributed and parallel implementations. A parallel implementation running in O(n’log n) time using
n processors and O(m) space is obtained. This time bound matches that of the Shiloach-Vishkin
algorithm, which also uses n processors but requires O(n’) space.
Categories and Subject Descriptors: F.2.2 [Analysis of Algorithms and Problem Complexity]: Non-
numerical Algorithms and Problems; G.2.2 [Discrete Mathematics]: Graph Theory-graph algorithms; 75
Caveat. Worst-case running time is generally not useful for predicting or

comparing max-flow algorithm performance in practice.
Best in practice. Push–relabel method with gap relabeling: O(m3/2) in practice.
On I m p l e m e n t i n g P u s h - R e l a b e l M e t h o d
for the M a x i m u m Flow P r o b l e m EUROPEAN
JOURNAL
OF OPERATIONAL
Boris V. Cherkassky 1 and Andrew V. Goldberg 2 RESEARCH
ELSEVIER European Journal of Operational Research 97 (1997) 509-542
1 Central Institute for Economics and Mathematics,
Krasikova St. 32, 117418, Moscow, Russia
cher@eemi.msk.su
2 Computer Science Department, Stanford University Theory and Methodology
Stanford, CA 94305, USA
goldberg~cs. stanford, edu Computational investigations of maximum flow algorithms
Ravindra K . A h u j a a, M u r a l i K o d i a l a m b, A j a y K . M i s h r a c, J a m e s B . O r l i n d,.
A b s t r a c t . We study efficient implementations of the push-relabel method
Department t~'lndustrial and Management Engineering. Indian Institute of Technology. Kanpur, 208 016, India
for the maximum flow problem. The resulting codes are faster than the a
b AT& T Bell Laboratories, Holmdel, NJ 07733, USA

previous codes, and much faster on some problem families. The speedup c KA'F-ZGraduate School of Business, University of Pittsburgh, Pittsburgh, PA 15260, USA
is due to the combination of heuristics used in our implementations. We d Sloun School of Management, Massachusetts Institute of Technology. Cambridge. MA 02139. USA
also exhibit a family of problems for which the running time of all known
methods seem to have a roughly quadratic growth rate. Received 30 August 1995; accepted 27 June 1996
1 Introduction Abstract
The rnaximum flow problem is a classical combinatorial problem that comes up The maximum flow algorithm is distinguished by the long line of successive contributions researchers have made in
in a wide variety of applications. In this paper we study implementations of the obtaining algorithms with incrementally better worst-case complexity. Some, but not all, of these theoretical improvements
have produced improvements in practice. The purpose of this paper is to test some of the major algorithmic ideas developed
push-rdabel [13, 17] method for the problem. in the recent years and to assess their utility on the empirical front. However, our study differs from previous studies in
The basic methods for the maximum flow problem include the network sim- several ways. Whereas previous studies focus primarily on CPU time analysis, our analysis goes further and provides 76
plex method of Dantzig [6, 7], the augmenting path method of Ford and F~lker- detailed insight into algorithmic behavior. It not only observes how algorithms behave but also tries to explain why
Computer vision. Different algorithms work better for some dense

problems that arise in applications to computer vision.
In IEEE Transactions on PAMI, Vol. 26, No. 9, pp. 1124-1137, Sept. 2004 p.1
An Experimental Comparison of VERMA, BATRA: MAXFLOW REVISITED 1
Min-Cut/Max-Flow Algorithms for

Energy Minimization in Vision MaxFlow Revisited:
Yuri Boykov and Vladimir Kolmogorov∗ An Empirical Comparison of Maxflow
Algorithms for Dense Vision Problems
Abstract
After [15, 31, 19, 8, 25, 5] minimum cut/maximum flow algorithms on graphs emerged as Tanmay Verma IIIT-Delhi
an increasingly useful tool for exact or approximate energy minimization in low-level vision. tanmay08054@iiitd.ac.in Delhi, India
The combinatorial optimization literature provides many min-cut/max-flow algorithms with
Dhruv Batra TTI-Chicago
dbatra@ttic.edu Chicago, USA
different polynomial time complexity. Their practical efficiency, however, has to date been
studied mainly outside the scope of computer vision. The goal of this paper is to provide an
experimental comparison of the efficiency of min-cut/max flow algorithms for applications
in vision. We compare the running times of several standard algorithms, as well as a

Abstract
new algorithm that we have recently developed. The algorithms we study include both Algorithms for finding the maximum amount of flow possible in a network (or max-
Goldberg-Tarjan style “push-relabel” methods and algorithms based on Ford-Fulkerson flow) play a central role in computer vision problems. We present an empirical compari-
son of different max-flow algorithms on modern problems. Our problem instances arise
style “augmenting paths”. We benchmark these algorithms on a number of typical graphs
from energy minimization problems in Object Category Segmentation, Image Deconvo-
in the contexts of image restoration, stereo, and segmentation. In many cases our new lution, Super Resolution, Texture Restoration, Character Completion and 3D Segmen-
algorithm works several times faster than any of the other methods making near real-time tation. We compare 14 different implementations and find that the most popularly used
implementation of Kolmogorov [5] is no longer the fastest algorithm available, especially
performance possible. An implementation of our max-flow/min-cut algorithm is available
for dense graphs.
upon request for research purposes.
Index Terms — Energy minimization, graph algorithms, minimum cut, maximum
flow, image restoration, segmentation, stereo, multi-camera scene reconstruction. 1 Introduction

∗
Yuri Boykov is with the Computer Science Department at the University of Western Ontario, Canada,
yuri@csd.uwo.ca. Vladimir Kolmogorov is with Microsoft Research, Cambridge, England, vnk@microsoft.com.
This work was mainly done while the authors were with Siemens Corp. Research, Princeton, NJ.
Over the past two decades, algorithms for finding the maximum amount of flow possible in
a network (or max-flow) have become the workhorses of modern computer vision and ma- 77
chine learning – from optimal (or provably-approximate) inference in sophisticated discrete
Maximum-flow algorithms: Matlab
78
Maximum-flow algorithms: Google
79
7. N ETWORK F LOW I

Which max-flow algorithm to use for bipartite matching? 
A. Ford–Fulkerson: O(m n C).
B. Capacity scaling: O(m2 log C).
C. Shortest augmenting path: O(m2 n).
D. Dinitz’ algorithm: O(m n2).
we’ll show that Dinitz’ algorithm runs

in O(m n1/2) time for bipartite matching
SIAM J. CoMavx.
Vol. 4, No. 4, December 1975
NETWORK FLOW AND TESTING GRAPH CONNECTIVITY*

SHIMON EVEN" AND R. ENDRE TARJAN:I:
Abstract. An algorithm of Dinic for finding the maximum flow in a network is described. It is
then shown that if the vertex capacities are all equal to one, the algorithm requires at most O(IV[ 1/2 IEI)
time, and if the edge capacities are all equal to one, the algorithm requires at most O(I VI 2/3. IEI) time.
Also, these bounds are tight for Dinic’s algorithm.
These results are used to test the vertex connectivity of a graph in O(IVI 1/z. IEI 2) time and the
edge connectivity in O(I V[ 5/3. IEI) time.
Key words. Dinic’s algorithm, maximum flow, connectivity, vertex connectivity, edge connec-
tivity
81
1. Network flow. Let be a finite V is the set of
Simple unit-capacity networks
Def. A flow network is a simple unit-capacity network if:

・Every edge has capacity 1.
・Every node (other than s or t) has exactly one entering edge, node capacity = 1
or exactly one leaving edge, or both.
Property. Let G be a simple unit-capacity network and let f be a 0–1 flow.

Then, residual network G f is also a simple unit-capacity network.
Ex. Bipartite matching.

1
1 1
82
Shortest-augmenting-path algorithm.
・Normal augmentation: length of shortest path does not change.
・Special augmentation: length of shortest path strictly increases.
Theorem. [Even–Tarjan 1975] In simple unit-capacity networks,
Dinitz’ algorithm computes a maximum flow in O(m n1/2) time.
Pf.
・Lemma 1. Each phase of normal augmentations takes O(m) time.
・Lemma 2. After n1/2 phases, val( f ) ≥ val( f *) – n1/2.
・Lemma 3. After ≤ n1/2 additional augmentations, flow is optimal. ▪
Lemma 3. After ≤ n1/2 additional augmentations, flow is optimal.

Pf. Each augmentation increases flow value by at least 1. ▪
Lemma 1 and Lemma 2. Ahead.
83
within a phase, length of shortest

Phase of normal augmentations. augmenting path does not change

・If get stuck, delete node from LG and go to previous node.
construct level graph
s t
level graph LG
84

advance
s t
level graph LG
85

remove from level graph
augment
all edges in augmenting path
s t
level graph LG
86

advance
s t
level graph LG
87

retreat
s t
level graph LG
88

advance
s t
level graph LG
89

augment
s t
level graph LG
90

end of phase (length of shortest augmenting path has increased)
s t
level graph LG
91
Simple unit-capacity networks: analysis

Lemma 1. A phase of normal augmentations takes O(m) time.
Pf.
・O(m) to create level graph LG.
・O(1) per edge (each edge involved in at most one advance, retreat, and
augmentation).
・O(1) per node (each node deleted at most once). ▪
92
Consider running advance–retreat algorithm in a unit-capacity network

(but not necessarily a simple one). What is running time? 
both indegree and outdegree
of a node can be larger than 1
useful for this
A. O(m). week’s homework!
B. O(m3/2).
C. O(m n).
D. May not terminate.
93
Lemma 2. After n1/2 phases, val( f ) ≥ val( f *) – n1/2.

・After n1/2 phases, length of shortest augmenting path is > n1/2.
・Thus, level graph has ≥ n1/2 levels (not including levels for s or t).
・Let 1 ≤ h ≤ n1/2 be a level with min number of nodes ⇒ ⎢Vh ⎢ ≤ n1/2.
level graph LG for flow f
s t
V1 Vh Vn1/2
94
Lemma 2. After n1/2 phases, val( f ) ≥ val( f *) – n1/2.

・After n1/2 phases, length of shortest augmenting path is > n1/2.
・Thus, level graph has ≥ n1/2 levels (not including levels for s or t).
・Let 1 ≤ h ≤ n1/2 be a level with min number of nodes ⇒ ⎢Vh ⎢ ≤ n1/2.
・Let A = {v : ℓ(v) < h} ∪ {v : ℓ(v) = h and v has ≤ 1 outgoing residual edge}.
・capf (A, B) ≤ ⎢Vh ⎢ ≤ n1/2 ⇒ val( f ) ≥ val( f *) – n1/2. ▪
unit-capacity
simple network
residual network Gf
residual edges
s t
V1 Vh Vn1/2
95
Simple unit-capacity networks: review
Theorem. [Even–Tarjan 1975] In simple unit-capacity networks,

Dinitz’ algorithm computes a maximum flow in O(m n1/2) time.
Pf.
・Lemma 1. Each phase takes O(m) time.
・Lemma 2. After n1/2 phases, val( f ) ≥ val( f *) – n1/2.
・Lemma 3. After ≤ n1/2 additional augmentations, flow is optimal. ▪
Corollary. Dinitz’ algorithm computes max-cardinality bipartite matching

in O(m n1/2) time.
96

07 Network Flow I

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

07 Network Flow I

Uploaded by

Copyright:

Available Formats

7.

‣ max-flow and min-cut problems

Last updated on 1/14/20 2:18 PM

‣ max-flow and min-cut problems

A flow network is a tuple G = (V, E, s, t, c).

Intuition. Material flowing through a transportation network;

Def. An st-cut (cut) is a partition (A, B) of the nodes with s ∈ A and t ∈ B.

Def. An st-cut (cut) is a partition (A, B) of the nodes with s ∈ A and t ∈ B.

don’t include edges

Def. An st-cut (cut) is a partition (A, B) of the nodes with s ∈ A and t ∈ B.

Min-cut problem. Find a cut of minimum capacity.

Which is the capacity of the given st-cut?

Def. An st-flow (flow) f is a function that satisfies:

Def. An st-flow (flow) f is a function that satisfies:

Def. The value of a flow f is: val(f ) = f (e) f (e)

Def. An st-flow (flow) f is a function that satisfies:

Def. The value of a flow f is: val(f ) = f (e) f (e)

Max-flow problem. Find a flow of maximum value.

‣ max-flow and min-cut problems

flow network G and flow f

flow network G and flow f

flow network G and flow f

flow network G and flow f

ending flow value = 16

flow network G and flow f

but max-flow value = 19

flow network G and flow f

Q. Why does the greedy algorithm fail?

Ex. Consider flow network G .

Bottom line. Need some mechanism to “undo” a bad decision. 19

Original edge. e = (u, v) ∈ E. original flow network G

Reverse edge. ereverse = (v, u).

edges with positive

Residual network. G f = (V, Ef , s, t, cf ). where flow on a reverse edge

・Ef = {e : f (e) < c(e)} ∪ {ereverse : f (e) > 0}.

・Key property: f ʹ is a flow in G f iff f + f ʹ is a flow in G.

Def. An augmenting path is a simple s↝t path in the residual network G f .

Def. The bottleneck capacity of an augmenting path P is the minimum

Key property. Let f be a flow and let P be an augmenting path in G f .

δ ← bottleneck capacity of augmenting path P.

Which is the augmenting path of highest bottleneck capacity?

Ford–Fulkerson augmenting path algorithm.

FOREACH edge e ∈ E : f (e) ← 0.

‣ max-flow and min-cut problems

val(f ) = f (e) f (e)

net flow across cut = 5 + 10 + 10 = 25

s 5/5 5/8 10 / 10 t value of flow = 25

val(f ) = f (e) f (e)

net flow across cut = 10 + 5 + 10 = 25

s 5/5 5/8 10 / 10 t value of flow = 25

val(f ) = f (e) f (e)

net flow across cut = (10 + 10 + 5 + 10 + 0 + 0) – (5 + 5 + 0 + 0) = 25

s 5/5 5/8 10 / 10 t value of flow = 25

Which is the net flow across the given cut?

val(f ) = f (e) f (e)

Pf. val(f ) = f (e) f (e)

by flow conservation, all terms

val(f ) = f (e) f (e)

value of flow = 27 ≤ capacity of cut = 30 30

Corollary. Let f be a flow and let (A, B) be any cut.

value of flow = 28 = capacity of cut = 28 31

Max-flow min-cut theorem. Value of a max flow = capacity of a min cut.

1956 IRE TRANXACTIONX ON INFORiMATION THEORY 117

Which edges are in the level graph of the following digraph? 