FLOWS AND CONNECTIVITY

SOME DEFINITIONS

 CAPACITATED NETWORK: A weighted directed graph with an integer

valued function c known as capacity function defined on its set of
arcs/edges is called a capacitated network.
 SOURCE VERTEX : A special vertex s with in-degree 0 is designated as
the source vertex
 DESTINATION VERTEX/SINK : A special vertex t with out-degree 0 is
designated as the destination vertex/sink
 All vertices other than the source vertex or the destination vertex are
INTERMEDIATE vertices
 Let c(e) be the cost of the edge e(u,v) is called the capacity of e. The
capacity of an edge is greater than 0
 A flow f in a network is an integer valued function defined on its set of
arcs such that 0 ≤ f(e) ≤ c(e) for each arc e.
 The sum of flows along all the arcs directed to the vertex i is the inflow
into i
 The sum of flows along all the arcs directed from the vertex i is the
outflow from i
 A flow is a feasible flow if it satisfies the feasible condition: For every
intermediate vertex (i.e. vertex other than source or sink) i, the inflow
into i is equal to the outflow from i.
 A feasible flow in a capacitated network such that the value of the flow
is as large as possible is called the maximum flow in the network.
 The problem of finding a feasible flow in a capacitated network such that
the value of the flow is maximum is known as maximum flow network.
 It is assumed that the digraph is asymmetric.
CUTS IN CAPACITATED NETWORK

Any partition of vertex set V in the weighted digraph G=(V,E) into two sets
S and T such that the source s is in S and sink t is in T and the set
(S,T)={(u,v)|u є S and v є T and (u,v) є E)} is called a cut or source-sink-cut in
the network since no flow can be sent from source to sink if all the arcs in
the cut are deleted.

A cut is called a mincut if its capacity does not exceed the capacity of any
other cut.

If f is a feasible flow in the network the sum of flows along the arcs in the
cut (S,T ) is the flow f(S,T) along the cut. Obviously, 0 ≤ f(S,T) ≤ c(S,T)

THEOREM-1 :If f is any feasible flow in the capacitated network G and if

(S,T) is any cut in the network, f(G)=f(S,T)-f(T,S)

PROOF

The vertex set V={1, 2, ………, n}.

S is any set of vertices that contains vertex 1 (source) and T is its

complement that contains the vertex n (the sink)

∑jf(i,j) - ∑jf(j,i) =f(G) when i=1

So, ∑i є S∑j f(i,j) - ∑i є S∑j f(j,i) =f(G). If i and j are both in S, the term
f(i,j) appears in first summation ∑i є S∑j f(i,j) as well as in the second
summation ∑i є S∑j f(j,i) .

So it is enough if we let the subscript j vary for all j in T

Hence, ∑i є S∑jєT f(i,j) - ∑i є S∑ jєT f(j,i) =f(G).

  COROLLARY-1 : The value of f(G) of any feasible flow f in a network is
also equal to the inflow to the sink
 COROLLARY-2 :If f is any feasible flow and if (S,T) is any cut, f(G) ≤
c(S,T)

If f is a feasible flow in a capacitated network the arc(i,j) is

 f-saturated if f(i,j) = c(i,j).

 f-zero or f-free if f(i,j) = 0
 f-positive if 0 < f(i,j) < c(i,j).

THEOREM-2:

a) If f is a feasible flow and if (S,T) is any cut, f(G) = c(S,T) if and

only if every arc in (S,T) is f-saturated and if every arc in the cut
(T,S) is f-zero.
b) If f is a feasible flow and if (S,T) is any cut such that f(G) =
c(S,T), f is the maximum flow and (S,T) is the minimum cut.

PROOF

a)f(G) = c(S,T) if and only if f(S,T) – f(T,S) =c(S,T). This is possible if and only
if every arc in cut (T,S) is f-zero and every arc in cut (S,T) is f-
saturated.

b)Let f’ be a maximum flow and (S’,T’) be a minimum cut in the network.

Then f(G) ≤ f’(G) ≤ c(S’,T’) ≤ c(S,T)
 Now if f(G)=c(S,T), then it implies that f(G) = f’(G) = c(S’,T’) = c(S,T)
So, f is indeed a maximum flow and C(S,T) is a minimum cut

FLOW AUGMENTING PATH

 An alternating sequence P of vertices and arcs of the form v0, e1, v1, e2,
……….. , ek, vk, in which no vertex is repeated is called a semipath from v0
to vk.
 Arc ei in this path is a forward edge if it is directed to vi . Otherwise it is
a backward edge in P.
 A semipath P is said to be f-unsaturated if no forward path is f-
saturated and no backward edge is f-free
 An f-unsaturated path is path from source to sink is called an f-
augmenting path.
 For each arc ei in an f-augmenting path P, δ(ei) is defined as c(ei)-f(ei) if ei
is a forward arc and f(ei) if ei is a backward arc.
 The excess flow capacity of a semipath P is δ(P) = min {δ(ei) : ei є P}

If P is an f-augmenting path with excess flow capacity δ(P), a new feasible flow
f’ can be obtained by increasing the flow of each forward edge by δ(P), and
decreasing the flow along each backward edge by δ(P) such that f’(G)=f(G) +
δ(P). Once this is accomplished, the semipath is no longer an f-augmenting path.
In other words if there is an f-augmenting path in a network, flow f is not a
maximum flow as its flow value can be increased using this augmenting path. It
turns out the converse of this assertion is also true.

In the semipath given in FIGURE-4 the flow can be increased by increasing the
flow along forward arcs by 1 unit and decreasing the flow along backward arc by
1 unit as shown in FIGURE-5

THEOREM-3

A flow f in a capacitated network is a maximum flow if and only if there is

no f-augmenting path in the network.

PROOF
We need to show that if there are no f-augmenting path with respect to f, f is
indeed a maximum flow.
Let S be the set of all vertices i such that there is an f-unsaturated path from
the source vertex 1 in S to i.
Obviously 1 is a vertex in S and the sink (vertex n) is not in S.
Thus there is a cut (S,T) in the network.
Let <i,j> be any arc inn this cut.
Since there are no f-unsaturated path from source to j, arc <i,j> is necessarily
saturated.
Likewise any arc in (T,S) is f-zero
So the current flow value is equal to the capacity of cut (S,T). Hence the
current flow is maximum.

THEOREM-4(FORD-FULKERSON THEOREM)
In a capacitated network, the value of maximum flow is equal to the
capacity of the minimum cut. This theorem is also known as MAX-FLOW-
MIN-CUT THEOREM
PROOF
Let f be a maximum flow
Let S be the set of vertices i such that there is a f-unsaturated path from the
source to i.
Then the complement T of S is non-empty.
Thus there is a cut (S,T)
Each arc in this cut is f-saturated and each arc in the cut in (T,S) is f-zero.
So the flow value f(G) is equal to c(S,T) which is a minimum cut.

THE EDMONDS-KARP ALGORITHM to solve the maxflow problem

INPUT:

 A capacitated network with a capacity function c and an initial flow f .

 A vertex set V= {1,2,…………..,n] where 1 is the source vertex and n is the
sink.
 Capacity of arc > 0

Step1: Construct a digraph D(f) =(V,E’) as follows:

(a) If (i,j) is an f-saturated arc in E, (j,i) is an arc in E’

(b) If (i,j) is an f-zero arc in E, (i,j) is an arc in E’
(c) If (i,j) is an f-positive arc in E, both (i,j) and (j,i) are in E’

Step2: Starting from the source vertex apply a breadth first search (BFS)
algorithm in D(f) to obtain a directed path from source to sink. If there is no
such path, go to step 4

Step3: The directed path in the digraph D(f) from the source to sink n
obtained in step-2 is a semipath in G with positive excess flow capacity, so it is
an f-augmenting path. Increase the flow value in the network by using this path.
Go to step 1

Step4: The maximum flow value is the outflow from the source. Let S be
the set of vertices that can be reached from vertex 1 in D(f) and let T be its
complement. (S,T) is the minimum cut.