ULC403

Design and Analysis of Algorithms

The Maximum-Flow Problem and the

Ford-Fulkerson Algorithm

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 1 / 26
Transportation networks

■ Transportation networks: Networks whose edges carry some sort of

traffic and whose nodes act as “switches” passing traffic between dif-
ferent edges.
S.No Transport Networks Edges Nodes
1 Highway System Highways Interchanges
2 Computer Networks Links that can switches
carry packets
3 Fluid Networks Pipes that carry Junctures
liquid/fluid where pipes
are plugged
together.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 2 / 26
Graphs in modelling of Transportation Networks
■ Network models of this type have several ingredients:
• capacities on the edges, indicating how much they can carry;
• source nodes in the graph, which generate traffic;
• sink nodes in the graph, which can “absorb” traffic as it arrives
and
• the traffic/flow itself, which is transmitted across the edges.

Figure: A flow network, with source s and sink t. The numbers next to the edges
are the capacities.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 3 / 26
Flow Networks

■ A flow network is a directed graph G = (V, E) with the following fea-

tures.
• Associated with each edge < u, v >∈ E is a capacity, which is a
nonnegative number that we denote c(u, v) ≥ 0
• There is a single source node s ∈ V
• There is a single sink node t ∈ V
■ Assumptions about the flow networks
1) No edge enters the source s and no edge leaves the sink t,
2) There is at least one edge incident to each node,
3) All capacities are integers.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 4 / 26
Defining flow

■ What it means for our n/w to carry traffic/flow:

• an abstract entity (generated at source nodes → transmitted across
edges → absorbed at sink nodes).
• An s − t flow is a function f that maps each edge < u, v >∈ E
to a nonnegative real number, f : V × V → R+ .
• f (u, v) intuitively represents the amount of flow carried by
edge < u, v >∈ E.
■ A flow f must satisfy the following two properties.
(i) (Capacity conditions) For each < u, v >∈ E, we have
0 ≤ f (u, v) ≤ c(u, v).
(ii) (Conservation conditions) For each node u ∈ V other than s
and t, we have
∑ ∑ ∑ ∑
v∈V f (u, v) = v∈V f (v, u) or e into v f (e) = e out of ∈v f (e)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 5 / 26
Value of a flow

■ The value of a flow f , denoted v(f ), is defined to be the amount of

flow generated at the source:
∑ ∑
|f | = v(f ) = e out of s f (e) = e into t f (e)
∑ ∑
|f | = v(f ) = v∈V f (s, v) = v∈V f (v, t)

■ The flow into the sink node (t) is same as flow going out from the
source node (s) and thus, the flow is conserved.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 6 / 26
The Lucky Puck Company’s trucking problem.

Figure: A flow network G = (V, E) for LPCT.

+ The Jalandhar factory is the source s, and the Delhi warehouse is the
sink t.

+ Pucks are shipped through intermediate cities, but only c(u, v) crates
per day can go from city u to city v.

+ Each edge is labeled with its capacity.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 7 / 26
 . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 8 / 26
The Maximum-Flow Problem

■ Given a flow network G(V, E), with a capacity associated with each
edge in E and two distinguished nodes s and t known as source and
sink respectively,

■ Given a flow network G(V, E), find a flow of maximum possible

value.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 9 / 26
Designing the Algorithm

■ Always set the flow as high as the edge capacities would allow while
preserving the conservation conditions.
■ Is this the maximum possible for the graph in the figure?
• no other s − t path on which we can directly push flow without
exceeding some capacity
• yet we do not have a maximum flow. . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 10 / 26
More general way of pushing flow from s → t

■ Let’s perform the operation denoted by a dotted line.

■ Dark edges are carrying flow before the operation, and the dashed edges
form the new kind of augmentation

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 11 / 26
Systematically pushing flows: Residual graph

■ We can push forward on edges with leftover capacity, and

■ We can push backward on edges that are already carrying flow, to divert
it in a different direction.

■ The residual graph provides a systematic way to search for forward-

backward operations such as this.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 12 / 26
Residual Graph

■ Given a flow network G = (V, E) and a flow f , the residual network of

G induced by f is Gf = (V, Ef ), where

Ef = {(u, v) ∈ V × V : cf (u, v) >0} with residual capacity cf

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 13 / 26
Residual Graph Construction

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 14 / 26
Augmenting Paths in a Residual Graph

■ The way in which we push flow from s to t in Gf is as follows.

• An augmenting path p is a simple path from s to t on a residual
network Gf . Choose an augmenting path.
• Residual capacity: The maximum capacity by which we can
increase the flow on each edge in an p is known as the residual
capacity of p

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 15 / 26
Max-flow: Ford Fulkerson Algorithm

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 16 / 26
Complexity Analysis

■ A straightforward implementation of FORD-FULKERSON runs in time

O(E|f ∗ |), where f ∗ is the maximum flow found by the algorithm.
• Initialisation take time Θ(E) .
• The while loop is executed at most |f ∗ | times, since the flow value
increases by at least one unit in each iteration.
• The time to find a path in a residual network is O(E).
• The time to update capacity and flow values is O(1).

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 17 / 26
Edmonds-Karp algorithm

■ It’s similar to a Ford-Fulkerson method.

■ The algorithm is identical to the Ford–Fulkerson algorithm, except that

the search order when finding the augmenting path is defined.

■ It chooses the augmenting path as a shortest path from s to t in

the residual network.

■ Edmonds-Karp algorithm runs in O(V E 2 ) time.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 18 / 26
Max-flow and Min-cut
Are they related ?

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 19 / 26
The structure of the flow network & Max-flow

■ Note: the structure of the flow network places upper bounds on the
maximum value of an s-t flow.
■ A basic “obstacle” to the existence of large flows:
• Divide the nodes of the graph into two sets, S and T , so that
s ∈ S and t ∈ T .

• Any flow that goes from s to t must cross from S into T at some
point, and thereby use up some of the edge capacity from S to T .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 20 / 26
Cut-sets of Flow Networks
■ A cut in a network is a partition of V into S and T (T = V − S)
such that s (source) is in S and t (target) is in T .
■ Each such “cut” of the graph puts a bound on the maximum possible
flow value.
■ Capacity of a cut(S, T )

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 21 / 26
The upper bound on the flow

■ Let f be any s − t flow, and (S, T ) any s − t cut. Then v(f ) ≤ c(S, T ).

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 22 / 26
Min-cut

■ Min s − t cut (also called as a Min-Cut) is a cut of minimum capacity.

• The maximum-flow value equals the minimum capacity of
any such division, called the minimum cut.
• As a bonus, our algorithm will also compute the minimum cut.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 23 / 26
Max-flow min-cut theorem

■ If f is a flow in a flow network G = (V, E) with source s and sink t,

then the following conditions are equivalent:
1. f is a maximum flow in G.
2. The residual network Gf contains no augmenting paths.
3. |f | = c(S, T ) for some cut (S, T ) of G.
Note: If |f | = c(S, T ), then c(S, T ) is the required min-cut.

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 24 / 26
Applications

■ Application area of max-flow, min-cut is very vast.

■ Interested students may refer the document available at

MAX-FLOW-APPLICATIONS

. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 25 / 26
 . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
(DAA, Dr. Ashish Gupta) ULC403 : Unit-III 26 / 26