0% found this document useful (0 votes)

66 views266 pages

Part Matchings

M [?] E is a matching if each node appears in at most one edge in M. A matching is perfect if it is of cardinality M = V / 2.

Uploaded by

hisuin

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

66 views266 pages

Part Matchings

M [?] E is a matching if each node appears in at most one edge in M. A matching is perfect if it is of cardinality M = V / 2.

Uploaded by

hisuin

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 266

Part V

Matchings
EADS
Ernst Mayr, Harald Rcke 540/604
Matching

Input: undirected graph G = (V, E).

M E is a matching if each node appears in at most one

edge in M.

Maximum Matching: nd a matching of maximum

cardinality
Bipartite Matching

Input: undirected, bipartite graph G = (L R, E).

M E is a matching if each node appears in at most one

edge in M.

Maximum Matching: nd a matching of maximum

cardinality
1
2
3
4
5
L R

5
EADS 16 Denition
Ernst Mayr, Harald Rcke 542/604
Bipartite Matching

Input: undirected, bipartite graph G = (L R, E).

M E is a matching if each node appears in at most one

edge in M.

Maximum Matching: nd a matching of maximum

cardinality
1
2
3
4
5
L R

5
EADS 16 Denition
Ernst Mayr, Harald Rcke 543/604
Bipartite Matching

A matching M is perfect if it is of cardinality M = V/2.

For a bipartite graph G = (L R, E) this means

M = L = R = n.
1
2
3
4
5
L R

5
EADS 16 Denition
Ernst Mayr, Harald Rcke 544/604
17 Bipartite Matching via Flows

Input: undirected, bipartite graph G = (L R s, t, E

Direct all edges from L to R.

Add source s and connect it to all nodes on the left.

Add t and connect all nodes on the right to t.

All edges have unit capacity.

s t
1
2
3
4
5
L R

5
Proof
Max cardinality matching in G value of maxow in G

Given a maximum matching M of cardinality k.

Consider ow f that sends one unit along each of k paths.

f is a ow and has cardinality k.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 546/604
Proof
Max cardinality matching in G value of maxow in G

Given a maximum matching M of cardinality k.

Consider ow f that sends one unit along each of k paths.

f is a ow and has cardinality k.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 546/604
Proof
Max cardinality matching in G value of maxow in G

Given a maximum matching M of cardinality k.

Consider ow f that sends one unit along each of k paths.

f is a ow and has cardinality k.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 546/604
Proof
Max cardinality matching in G value of maxow in G

Given a maximum matching M of cardinality k.

Consider ow f that sends one unit along each of k paths.

f is a ow and has cardinality k.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 546/604
Proof
Max cardinality matching in G value of maxow in G

Let f be a maxow in G

of value k

Integrality theorem k integral; we can assume f is 0/1.

Consider M= set of edges from L to R with f(e) = 1.

Each node in L and R participates in at most one edge in M.

M = k, as the ow must use at least k middle edges.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 547/604
Proof
Max cardinality matching in G value of maxow in G

Let f be a maxow in G

of value k

Integrality theorem k integral; we can assume f is 0/1.

Consider M= set of edges from L to R with f(e) = 1.

Each node in L and R participates in at most one edge in M.

M = k, as the ow must use at least k middle edges.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 547/604
Proof
Max cardinality matching in G value of maxow in G

Let f be a maxow in G

of value k

Integrality theorem k integral; we can assume f is 0/1.

Consider M= set of edges from L to R with f(e) = 1.

Each node in L and R participates in at most one edge in M.

M = k, as the ow must use at least k middle edges.

s t
1
2
3
4
5
L R
G

5
1
2
3
4
5
L R
G

5
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 547/604
17 Bipartite Matching via Flows
Which ow algorithm to use?

Generic augmenting path: O(mval(f

)) = O(mn).

Capacity scaling: O(m

2
logC) = O(m
2
).
EADS 17 Bipartite Matching via Flows
Ernst Mayr, Harald Rcke 548/604
18 Augmenting Paths for Matchings
Denitions.

Given a matching M in a graph G, a vertex that is not

incident to any edge of M is called a free vertex w. r. .t. M.

For a matching M a path P in G is called an alternating path

if edges in M alternate with edges not in M.

An alternating path is called an augmenting path for

Given a matching M in a graph G, a vertex that is not

incident to any edge of M is called a free vertex w. r. .t. M.

For a matching M a path P in G is called an alternating path

if edges in M alternate with edges not in M.

An alternating path is called an augmenting path for

Given a matching M in a graph G, a vertex that is not

incident to any edge of M is called a free vertex w. r. .t. M.

For a matching M a path P in G is called an alternating path

if edges in M alternate with edges not in M.

An alternating path is called an augmenting path for

Given a matching M in a graph G, a vertex that is not

incident to any edge of M is called a free vertex w. r. .t. M.

For a matching M a path P in G is called an alternating path

if edges in M alternate with edges not in M.

An alternating path is called an augmenting path for

matching M if it ends at distinct free vertices.
Theorem 1
A matching M is a maximum matching if and only if there is no
augmenting path w. r. t. M.
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 549/604
Augmenting Paths in Action
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 550/604
Augmenting Paths in Action
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 550/604
Augmenting Paths in Action
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 550/604
Augmenting Paths in Action
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 550/604
Augmenting Paths in Action
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 550/604
Augmenting Paths in Action
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 550/604
18 Augmenting Paths for Matchings
Proof.
If M is maximum there is no augmenting path P, because
we could switch matching and non-matching edges along P.
This gives matching M

= M P with larger cardinality.

Suppose there is a matching M

with larger cardinality.

Consider the graph H with edge-set M

M (i.e., only edges

that are in either M or M

but not in both).

Each vertex can be incident to at most two edges (one from
M and one from M

). Hence, the connected components are

alternating cycles or alternating path.
As M

> M there is one connected component that is a

path P for which both endpoints are incident to edges from
M

. P is an alternating path.
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 551/604
18 Augmenting Paths for Matchings
Proof.
If M is maximum there is no augmenting path P, because
we could switch matching and non-matching edges along P.
This gives matching M

= M P with larger cardinality.

Suppose there is a matching M

with larger cardinality.

Consider the graph H with edge-set M

M (i.e., only edges

that are in either M or M

but not in both).

Each vertex can be incident to at most two edges (one from
M and one from M

). Hence, the connected components are

alternating cycles or alternating path.
As M

> M there is one connected component that is a

path P for which both endpoints are incident to edges from
M

= M P with larger cardinality.

Suppose there is a matching M

with larger cardinality.

Consider the graph H with edge-set M

M (i.e., only edges

that are in either M or M

but not in both).

Each vertex can be incident to at most two edges (one from
M and one from M

). Hence, the connected components are

alternating cycles or alternating path.
As M

> M there is one connected component that is a

path P for which both endpoints are incident to edges from
M

= M P with larger cardinality.

Suppose there is a matching M

with larger cardinality.

Consider the graph H with edge-set M

M (i.e., only edges

that are in either M or M

but not in both).

Each vertex can be incident to at most two edges (one from
M and one from M

). Hence, the connected components are

alternating cycles or alternating path.
As M

> M there is one connected component that is a

path P for which both endpoints are incident to edges from
M

. P is an alternating path.
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 551/604
18 Augmenting Paths for Matchings
Algorithmic idea:
As long as you nd an augmenting path augment your matching
using this path. When you arrive at a matching for which no
augmenting path exists you have a maximum matching.
Theorem 2
Let G be a graph, M a matching in G, and let u be a free vertex
w.r.t. M. Further let P denote an augmenting path w.r.t. M and
let M

= M P denote the matching resulting from augmenting

M with P. If there was no augmenting path starting at u in M
then there is no augmenting path starting at u in M

.
The above theorem allows for an easier implementation of an augment-
ing path algorithm. Once we checked for augmenting paths starting
from u we dont have to check for such paths in future rounds.
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 552/604
18 Augmenting Paths for Matchings
Algorithmic idea:
As long as you nd an augmenting path augment your matching
using this path. When you arrive at a matching for which no
augmenting path exists you have a maximum matching.
Theorem 2
Let G be a graph, M a matching in G, and let u be a free vertex
w.r.t. M. Further let P denote an augmenting path w.r.t. M and
let M

= M P denote the matching resulting from augmenting

M with P. If there was no augmenting path starting at u in M
then there is no augmenting path starting at u in M

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
P

EADS 18 Augmenting Paths for Matchings

Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
P

EADS 18 Augmenting Paths for Matchings

Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
P

P
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
u

e
P

P
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
u

e
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
u

e
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
u

e
P
1
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
u

e
P
1
P

1
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
18 Augmenting Paths for Matchings
Proof

Assume there is an augmenting

path P

w.r.t. M

starting at u.

If P

and P are node-disjoint, P

is
also augmenting path w.r.t. M ().

Let u

be the rst node on P

that
is in P, and let e be the matching
edge from M

incident to u

splits P into two parts one of

which does not contain e. Call this
part P
1
. Denote the sub-path of P

from u to u

with P

1
.

P
1
P

1
is augmenting path in M ().
u
u

e
P
1
P

1
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 553/604
How to nd an augmenting path?
Construct an alternating tree.
u
y
even nodes
odd nodes
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 554/604
How to nd an augmenting path?
Construct an alternating tree.
u
y
x
even nodes
odd nodes
Case 1:
y is free vertex not
contained in T
you found
alternating path
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 554/604
How to nd an augmenting path?
Construct an alternating tree.
u
x
y
even nodes
odd nodes
Case 2:
y is matched vertex
not in T; then
mate[y] T
grow the tree
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 555/604
How to nd an augmenting path?
Construct an alternating tree.
u
x
y
even nodes
odd nodes
Case 3:
y is already contained
in T as an odd vertex
ignore successor y
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 556/604
How to nd an augmenting path?
Construct an alternating tree.
u
x
y
even nodes
odd nodes
Case 4:
y is already contained
in T as an even vertex
cant ignore y
does not happen in
bipartite graphs
EADS 18 Augmenting Paths for Matchings
Ernst Mayr, Harald Rcke 557/604
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

, E)
S = 1, . . . , n
S

= 1

, . . . , n

Algorithm 50 BiMatch(G, match)

1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
start with an
empty matching
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
free: number of
unmatched nodes in
S
r: root of current tree
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
as long as there are
unmatched nodes and
we did not yet try to
grow from all nodes we
continue
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
r is the new node that
we grow from.
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
If r is free start tree
construction
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

have parent pointers.

Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
Q is a queue (BFS!!!).
aug is a Boolean that
stores whether we
already found an
augmenting path.
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
as long as we did not
augment and there are
still unexamined leaves
continue...
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
take next unexamined
leaf
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
if x has unmatched
neighbour we found an
augmenting path (note
that y r because we
are in a bipartite graph)
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
do an augmentation...
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
setting aug = true
ensures that the tree
construction will not
continue
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
reduce number of free
nodes
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
if y is not in the tree yet
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
...put it into the tree
Algorithm 50 BiMatch(G, match)
1: for x V do mate[x] 0;
2: r 0; free n;
3: while free 1 and r < n do
4: r r +1
5: if mate[r] = 0 then
6: for i = 1 to m do parent[i

] 0
7: Q ; Q. append(r); aug false;
8: while aug = false and Q do
9: x Q. dequeue();
10: for y A
x
do
11: if mate[y] = 0 then
12: augm(mate, parent, y);
13: aug true;
14: free free 1;
15: else
16: if parent[y] = 0 then
17: parent[y] x;
18: Q. enqueue(mate[y]);
add its buddy to the set
of unexamined leaves
19 Weighted Bipartite Matching
Weighted Bipartite Matching/Assignment

Input: undirected, bipartite graph G = L R, E.

an edge e = (, r) has weight w

e
0

nd a matching of maximum weight, where the weight of a

matching is the sum of the weights of its edges
Simplifying Assumptions (wlog [why?]):

assume that L = R = n

assume that there is an edge between every pair of nodes

(, r) V V
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 559/604
Weighted Bipartite Matching
Theorem 3 (Halls Theorem)
A bipartite graph G = (L R, E) has a perfect matching if and
only if for all sets S L, (S) S, where (S) denotes the set
of nodes in R that have a neighbour in S.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 560/604
19 Weighted Bipartite Matching
s t
1
2
3
4
5
L R

5
S
Halls Theorem
Proof:
Of course, the condition is necessary as otherwise not all
nodes in S could be matched to dierent neigbhours.
For the other direction we need to argue that the minimum
cut in the graph G

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

S
) L
S
gives that this is at least L.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 562/604
Halls Theorem
Proof:
Of course, the condition is necessary as otherwise not all
nodes in S could be matched to dierent neigbhours.
For the other direction we need to argue that the minimum
cut in the graph G

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

is at least L.

Let S denote a minimum cut and let L

S
= L S and
R
S
= R S denote the portion of S inside L and R,
respectively.

Clearly, all neighbours of nodes in L

S
have to be in S, as
otherwise we would cut an edge of innite capacity.

This gives R
S
(L
S
).

The size of the cut is L L

S
+R
S
.

Using the fact that (L

S
) L
S
gives that this is at least L.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 562/604
Algorithm Outline
Idea:
We introduce a node weighting x. Let for a node v V, x
v
0
denote the weight of node v.

Suppose that the node weights dominate the edge-weights

in the following sense:
x
u
+x
v
w
e
for every edge e = (u, v).

Let H( x) denote the subgraph of G that only contains

edges that are tight w.r.t. the node weighting x, i.e. edges
e = (u, v) for which w
e
= x
u
+x
v
.

Try to compute a perfect matching in the subgraph H( x). If

you are successful you found an optimal matching.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 563/604
Algorithm Outline
Idea:
We introduce a node weighting x. Let for a node v V, x
v
0
denote the weight of node v.

Suppose that the node weights dominate the edge-weights

in the following sense:
x
u
+x
v
w
e
for every edge e = (u, v).

Let H( x) denote the subgraph of G that only contains

edges that are tight w.r.t. the node weighting x, i.e. edges
e = (u, v) for which w
e
= x
u
+x
v
.

Try to compute a perfect matching in the subgraph H( x). If

Suppose that the node weights dominate the edge-weights

in the following sense:
x
u
+x
v
w
e
for every edge e = (u, v).

Let H( x) denote the subgraph of G that only contains

edges that are tight w.r.t. the node weighting x, i.e. edges
e = (u, v) for which w
e
= x
u
+x
v
.

Try to compute a perfect matching in the subgraph H( x). If

Suppose that the node weights dominate the edge-weights

in the following sense:
x
u
+x
v
w
e
for every edge e = (u, v).

Let H( x) denote the subgraph of G that only contains

edges that are tight w.r.t. the node weighting x, i.e. edges
e = (u, v) for which w
e
= x
u
+x
v
.

Try to compute a perfect matching in the subgraph H( x). If

you are successful you found an optimal matching.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 563/604
Algorithm Outline
Reason:

The weight of your matching M

(u,v)M

w
(u,v)
=

(u,v)M

(x
u
+x
v
) =

v
x
v
.

Any other matching M has

(u,v)M
w
(u,v)

(u,v)M
(x
u
+x
v
)

v
x
v
.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 564/604
Algorithm Outline
What if you dont nd a perfect matching?
Then, Halls theorem guarantees you that there is a set S L,
with (S) < S, where denotes the neighbourhood w.r.t. the
subgraph H( x).
Idea: reweight such that:

the total weight assigned to nodes decreases

the weight function still dominates the edge-weights

If we can do this we have an algorithm that terminates with an
optimal solution (we analyze the running time later).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 565/604
Algorithm Outline
What if you dont nd a perfect matching?
Then, Halls theorem guarantees you that there is a set S L,
with (S) < S, where denotes the neighbourhood w.r.t. the
subgraph H( x).
Idea: reweight such that:

the total weight assigned to nodes decreases

the weight function still dominates the edge-weights

the total weight assigned to nodes decreases

the weight function still dominates the edge-weights

If we can do this we have an algorithm that terminates with an
optimal solution (we analyze the running time later).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 565/604
Changing Node Weights
Increase node-weights in (S) by +, and decrease the
node-weights in S by .

Total node-weight decreases.

Only edges from S to R (S)

decrease in their weight.

Since, none of these edges is

tight (otw. the edge would be
contained in H( x), and hence
would go between S and (S))
we can do this decrement for
small enough > 0 until a new
edge gets tight.
L R
S
(S)

+
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 566/604
Changing Node Weights
Increase node-weights in (S) by +, and decrease the
node-weights in S by .

Total node-weight decreases.

Only edges from S to R (S)

decrease in their weight.

Since, none of these edges is

tight (otw. the edge would be
contained in H( x), and hence
would go between S and (S))
we can do this decrement for
small enough > 0 until a new
edge gets tight.
L R
S
(S)

+
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 566/604
Changing Node Weights
Increase node-weights in (S) by +, and decrease the
node-weights in S by .

Total node-weight decreases.

Only edges from S to R (S)

decrease in their weight.

Since, none of these edges is

tight (otw. the edge would be
contained in H( x), and hence
would go between S and (S))
we can do this decrement for
small enough > 0 until a new
edge gets tight.
L R
S
(S)

+
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 566/604
Changing Node Weights
Increase node-weights in (S) by +, and decrease the
node-weights in S by .

Total node-weight decreases.

Only edges from S to R (S)

decrease in their weight.

Since, none of these edges is

tight (otw. the edge would be
contained in H( x), and hence
would go between S and (S))
we can do this decrement for
small enough > 0 until a new
edge gets tight.
L R
S
(S)

+
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 566/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
= 1
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
= 1
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Weighted Bipartite Matching
Edges not drawn have weight 0.
5
2
6
4
2
4
1
6
3
1
3
0
6
2
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
1
2
2
4
6
4
1
2
5
3
3
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 567/604
Analysis
How many iterations do we need?

One reweighting step increases the number of edges out of

S by at least one.

Assume that we have a maximum matching that saturates

the set (S), in the sense that every node in (S) is matched
to a node in S (we will show that we can always nd S and a
matching such that this holds).

This matching is still contained in the new graph, because

all its edges either go between (S) and S or between L S
and R (S).

Hence, reweighting does not decrease the size of a

maximum matching in the tight sub-graph.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 568/604
Analysis
How many iterations do we need?

One reweighting step increases the number of edges out of

S by at least one.

Assume that we have a maximum matching that saturates

the set (S), in the sense that every node in (S) is matched
to a node in S (we will show that we can always nd S and a
matching such that this holds).

This matching is still contained in the new graph, because

all its edges either go between (S) and S or between L S
and R (S).

Hence, reweighting does not decrease the size of a

maximum matching in the tight sub-graph.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 568/604
Analysis
How many iterations do we need?

One reweighting step increases the number of edges out of

S by at least one.

Assume that we have a maximum matching that saturates

the set (S), in the sense that every node in (S) is matched
to a node in S (we will show that we can always nd S and a
matching such that this holds).

This matching is still contained in the new graph, because

all its edges either go between (S) and S or between L S
and R (S).

Hence, reweighting does not decrease the size of a

maximum matching in the tight sub-graph.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 568/604
Analysis
How many iterations do we need?

One reweighting step increases the number of edges out of

S by at least one.

Assume that we have a maximum matching that saturates

the set (S), in the sense that every node in (S) is matched
to a node in S (we will show that we can always nd S and a
matching such that this holds).

This matching is still contained in the new graph, because

all its edges either go between (S) and S or between L S
and R (S).

Hence, reweighting does not decrease the size of a

maximum matching in the tight sub-graph.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 568/604
Analysis

We will show that after at most n reweighting steps the size

of the maximum matching can be increased by nding an
augmenting path.

This gives a polynomial running time.

EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 569/604
How to nd an augmenting path?
Construct an alternating tree.
u
y
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 570/604
How to nd an augmenting path?
Construct an alternating tree.
u
y
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 570/604
Analysis
How do we nd S?

Start on the left and compute an alternating tree, starting at

any free node u.

If this construction stops, there is no perfect matching in

the tight subgraph (because for a perfect matching we need
to nd an augmenting path starting at u).

The set of even vertices is on the left and the set of odd
vertices is on the right and contains all neighbours of even
nodes.

All odd vertices are matched to even vertices. Furthermore,

the even vertices additionally contain the free vertex u.
Hence, V
odd
= (V
even
) < V
even
, and all odd vertices are
saturated in the current matching.
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 571/604
Analysis
How do we nd S?

Start on the left and compute an alternating tree, starting at

any free node u.

If this construction stops, there is no perfect matching in

the tight subgraph (because for a perfect matching we need
to nd an augmenting path starting at u).

The set of even vertices is on the left and the set of odd
vertices is on the right and contains all neighbours of even
nodes.

All odd vertices are matched to even vertices. Furthermore,

Start on the left and compute an alternating tree, starting at

any free node u.

If this construction stops, there is no perfect matching in

the tight subgraph (because for a perfect matching we need
to nd an augmenting path starting at u).

The set of even vertices is on the left and the set of odd
vertices is on the right and contains all neighbours of even
nodes.

All odd vertices are matched to even vertices. Furthermore,

Start on the left and compute an alternating tree, starting at

any free node u.

If this construction stops, there is no perfect matching in

the tight subgraph (because for a perfect matching we need
to nd an augmenting path starting at u).

The set of even vertices is on the left and the set of odd
vertices is on the right and contains all neighbours of even
nodes.

All odd vertices are matched to even vertices. Furthermore,

The current matching does not have any edges from V

odd
to
outside of L \ V
even
(edges that may possibly be deleted by
changing weights).

After changing weights, there is at least one more edge

connecting V
even
to a node outside of V
odd
. After at most n
reweights we can do an augmentation.

A reweighting can be trivially performed in time O(n

2
)
(keeping track of the tight edges).

An augmentation takes at most O(n) time.

In total we otain a running time of O(n

4
).

A more careful implementation of the algorithm obtains a

running time of O(n
3
).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 572/604
Analysis

The current matching does not have any edges from V

odd
to
outside of L \ V
even
(edges that may possibly be deleted by
changing weights).

After changing weights, there is at least one more edge

connecting V
even
to a node outside of V
odd
. After at most n
reweights we can do an augmentation.

A reweighting can be trivially performed in time O(n

2
)
(keeping track of the tight edges).

An augmentation takes at most O(n) time.

In total we otain a running time of O(n

4
).

A more careful implementation of the algorithm obtains a

running time of O(n
3
).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 572/604
Analysis

The current matching does not have any edges from V

odd
to
outside of L \ V
even
(edges that may possibly be deleted by
changing weights).

After changing weights, there is at least one more edge

connecting V
even
to a node outside of V
odd
. After at most n
reweights we can do an augmentation.

A reweighting can be trivially performed in time O(n

2
)
(keeping track of the tight edges).

An augmentation takes at most O(n) time.

In total we otain a running time of O(n

4
).

A more careful implementation of the algorithm obtains a

running time of O(n
3
).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 572/604
Analysis

The current matching does not have any edges from V

odd
to
outside of L \ V
even
(edges that may possibly be deleted by
changing weights).

After changing weights, there is at least one more edge

connecting V
even
to a node outside of V
odd
. After at most n
reweights we can do an augmentation.

A reweighting can be trivially performed in time O(n

2
)
(keeping track of the tight edges).

An augmentation takes at most O(n) time.

In total we otain a running time of O(n

4
).

A more careful implementation of the algorithm obtains a

running time of O(n
3
).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 572/604
Analysis

The current matching does not have any edges from V

odd
to
outside of L \ V
even
(edges that may possibly be deleted by
changing weights).

After changing weights, there is at least one more edge

connecting V
even
to a node outside of V
odd
. After at most n
reweights we can do an augmentation.

A reweighting can be trivially performed in time O(n

2
)
(keeping track of the tight edges).

An augmentation takes at most O(n) time.

In total we otain a running time of O(n

4
).

A more careful implementation of the algorithm obtains a

running time of O(n
3
).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 572/604
Analysis

The current matching does not have any edges from V

odd
to
outside of L \ V
even
(edges that may possibly be deleted by
changing weights).

After changing weights, there is at least one more edge

connecting V
even
to a node outside of V
odd
. After at most n
reweights we can do an augmentation.

A reweighting can be trivially performed in time O(n

2
)
(keeping track of the tight edges).

An augmentation takes at most O(n) time.

In total we otain a running time of O(n

4
).

A more careful implementation of the algorithm obtains a

running time of O(n
3
).
EADS 19 Weighted Bipartite Matching
Ernst Mayr, Harald Rcke 572/604
A Fast Matching Algorithm
Algorithm 50 Bimatch-Hopcroft-Karp(G)
1: M
2: repeat
3: let 1 = P
1
, . . . , P
k
be maximal set of
4: vertex-disjoint, shortest augmenting path w.r.t. M.
5: M M (P
1
P
k
)
6: until 1 =
7: return M
We call one iteration of the repeat-loop a phase of the algorithm.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 573/604
Analysis
Lemma 4
Given a matching M and a maximal matching M

there exist
M

M vertex-disjoint augmenting path w.r.t. M.

Proof:

Similar to the proof that a matching is optimal i it does not

contain an augmenting paths.

Consider the graph G = (V, M M

), and mark edges in

this graph blue if they are in M and red if they are in M

The connected components of G are cycles and paths.

The graph contains k = M

M more red edges than

blue edges.

Hence, there are at least k components that form a path

starting and ending with a blue edge. These are
augmenting paths w.r.t. M.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 574/604
Analysis
Lemma 4
Given a matching M and a maximal matching M

there exist
M

M vertex-disjoint augmenting path w.r.t. M.

Proof:

Similar to the proof that a matching is optimal i it does not

contain an augmenting paths.

Consider the graph G = (V, M M

), and mark edges in

this graph blue if they are in M and red if they are in M

The connected components of G are cycles and paths.

The graph contains k = M

M more red edges than

blue edges.

Hence, there are at least k components that form a path

there exist
M

M vertex-disjoint augmenting path w.r.t. M.

Proof:

Similar to the proof that a matching is optimal i it does not

contain an augmenting paths.

Consider the graph G = (V, M M

), and mark edges in

this graph blue if they are in M and red if they are in M

The connected components of G are cycles and paths.

The graph contains k = M

M more red edges than

blue edges.

Hence, there are at least k components that form a path

there exist
M

M vertex-disjoint augmenting path w.r.t. M.

Proof:

Similar to the proof that a matching is optimal i it does not

contain an augmenting paths.

Consider the graph G = (V, M M

), and mark edges in

this graph blue if they are in M and red if they are in M

The connected components of G are cycles and paths.

The graph contains k = M

M more red edges than

blue edges.

Hence, there are at least k components that form a path

there exist
M

M vertex-disjoint augmenting path w.r.t. M.

Proof:

Similar to the proof that a matching is optimal i it does not

contain an augmenting paths.

Consider the graph G = (V, M M

), and mark edges in

this graph blue if they are in M and red if they are in M

The connected components of G are cycles and paths.

The graph contains k = M

M more red edges than

blue edges.

Hence, there are at least k components that form a path

there exist
M

M vertex-disjoint augmenting path w.r.t. M.

Proof:

Similar to the proof that a matching is optimal i it does not

contain an augmenting paths.

Consider the graph G = (V, M M

), and mark edges in

this graph blue if they are in M and red if they are in M

The connected components of G are cycles and paths.

The graph contains k = M

M more red edges than

blue edges.

Hence, there are at least k components that form a path

starting and ending with a blue edge. These are
augmenting paths w.r.t. M.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 574/604
Analysis

Let P
1
, . . . , P
k
be a maximal collection of vertex-disjoint,
shortest augmenting paths w.r.t. M (let = P
i
).

= M (P
1
P
k
) = M P
1
P
k
.

Let P be an augmenting path in M

.
Lemma 5
The set A = M (M

P) = (P
1
P
k
) P contains at least
(k +1) edges.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 575/604
Analysis

Let P
1
, . . . , P
k
be a maximal collection of vertex-disjoint,
shortest augmenting paths w.r.t. M (let = P
i
).

= M (P
1
P
k
) = M P
1
P
k
.

Let P be an augmenting path in M

.
Lemma 5
The set A = M (M

P) = (P
1
P
k
) P contains at least
(k +1) edges.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 575/604
Analysis

Let P
1
, . . . , P
k
be a maximal collection of vertex-disjoint,
shortest augmenting paths w.r.t. M (let = P
i
).

= M (P
1
P
k
) = M P
1
P
k
.

Let P be an augmenting path in M

.
Lemma 5
The set A = M (M

P) = (P
1
P
k
) P contains at least
(k +1) edges.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 575/604
Analysis

Let P
1
, . . . , P
k
be a maximal collection of vertex-disjoint,
shortest augmenting paths w.r.t. M (let = P
i
).

= M (P
1
P
k
) = M P
1
P
k
.

Let P be an augmenting path in M

.
Lemma 5
The set A = M (M

P) = (P
1
P
k
) P contains at least
(k +1) edges.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 575/604
Analysis
Proof.

The set describes exactly the symmetric dierence between

matchings M and M

Hence, the set contains at least k +1 vertex-disjoint

augmenting paths w.r.t. M as M

= M +k +1.

Each of these paths is of length at least .

EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 576/604
Analysis
Proof.

The set describes exactly the symmetric dierence between

matchings M and M

Hence, the set contains at least k +1 vertex-disjoint

augmenting paths w.r.t. M as M

= M +k +1.

Each of these paths is of length at least .

EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 576/604
Analysis
Proof.

The set describes exactly the symmetric dierence between

matchings M and M

Hence, the set contains at least k +1 vertex-disjoint

augmenting paths w.r.t. M as M

= M +k +1.

Each of these paths is of length at least .

EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 576/604
Analysis
Lemma 6
P is of length at least +1. This shows that the length of a
shortest augmenting path increases between two phases of the
Hopcroft-Karp algorithm.
Proof.

If P does not intersect any of the P

1
, . . . , P
k
, this follows
from the maximality of the set P
1
, . . . , P
k
.

Otherwise, at least one edge from P coincides with an edge

from paths P
1
, . . . , P
k
.

This edge is not contained in A.

Hence, A k +P 1.

The lower bound on A gives (k +1) A k +P 1,

and hence P +1.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 577/604
Analysis
Lemma 6
P is of length at least +1. This shows that the length of a
shortest augmenting path increases between two phases of the
Hopcroft-Karp algorithm.
Proof.

If P does not intersect any of the P

1
, . . . , P
k
, this follows
from the maximality of the set P
1
, . . . , P
k
.

Otherwise, at least one edge from P coincides with an edge

from paths P
1
, . . . , P
k
.

This edge is not contained in A.

Hence, A k +P 1.

The lower bound on A gives (k +1) A k +P 1,

If P does not intersect any of the P

1
, . . . , P
k
, this follows
from the maximality of the set P
1
, . . . , P
k
.

Otherwise, at least one edge from P coincides with an edge

from paths P
1
, . . . , P
k
.

This edge is not contained in A.

Hence, A k +P 1.

The lower bound on A gives (k +1) A k +P 1,

If P does not intersect any of the P

1
, . . . , P
k
, this follows
from the maximality of the set P
1
, . . . , P
k
.

Otherwise, at least one edge from P coincides with an edge

from paths P
1
, . . . , P
k
.

This edge is not contained in A.

Hence, A k +P 1.

The lower bound on A gives (k +1) A k +P 1,

If P does not intersect any of the P

1
, . . . , P
k
, this follows
from the maximality of the set P
1
, . . . , P
k
.

Otherwise, at least one edge from P coincides with an edge

from paths P
1
, . . . , P
k
.

This edge is not contained in A.

Hence, A k +P 1.

The lower bound on A gives (k +1) A k +P 1,

If P does not intersect any of the P

1
, . . . , P
k
, this follows
from the maximality of the set P
1
, . . . , P
k
.

Otherwise, at least one edge from P coincides with an edge

from paths P
1
, . . . , P
k
.

This edge is not contained in A.

Hence, A k +P 1.

The lower bound on A gives (k +1) A k +P 1,

and hence P +1.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 577/604
Analysis
If the shortest augmenting path w.r.t. a matching M has edges
then the cardinality of the maximum matching is of size at most
M +
V
+1
.
Proof.
The symmetric dierence between M and M

contains
M

M vertex-disjoint augmenting paths. Each of these

paths contains at least +1 vertices. Hence, there can be at
most
V
+1
of them.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 578/604
Analysis
If the shortest augmenting path w.r.t. a matching M has edges
then the cardinality of the maximum matching is of size at most
M +
V
+1
.
Proof.
The symmetric dierence between M and M

contains
M

M vertex-disjoint augmenting paths. Each of these

paths contains at least +1 vertices. Hence, there can be at
most
V
+1
of them.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 578/604
Analysis
Lemma 7
The Hopcroft-Karp algorithm requires at most 2

V phases.
Proof.

After iteration |

V] the length of a shortest augmenting

path must be at least |

V] +1

Hence, there can be at most V/(

V +1)

V
additional augmentations.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 579/604
Analysis
Lemma 7
The Hopcroft-Karp algorithm requires at most 2

V phases.
Proof.

After iteration |

V] the length of a shortest augmenting

path must be at least |

V] +1

Hence, there can be at most V/(

V +1)

V
additional augmentations.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 579/604
Analysis
Lemma 8
One phase of the Hopcroft-Karp algorithm can be implemented
in time O(m).

Do a breadth rst search starting at all free vertices in the

left side L.
(alternatively add a super-startnode; connect it to all free vertices
in L and start breadth rst search from there)

The search stops when reaching a free vertex. However, the

current level of the BFS tree is still nished in order to nd a
set F of free vertices (on the right side) that can be reached
via shortest augmenting paths.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 580/604
Analysis

Then a maximal set of shortest path from the leftmost layer

of the tree construction to nodes in F needs to be
computed.

Any such path must visit the layers of the BFS-tree from left
to right.

To go from an odd layer to an even layer it must use a

matching edge.

To go from an even layer to an odd layer edge it can use

edges in the BFS-tree or edges that have been ignored
during BFS-tree construction.

We direct all edges btw. an even node in some layer to an

odd node in layer +1 from left to right.

A DFS search in the resulting graph gives us a maximal set

of vertex disjoint path from left to right in the resulting
graph.
EADS 20 The Hopcroft-Karp Algorithm
Ernst Mayr, Harald Rcke 581/604
s
s
s
s
How to nd an augmenting path?
Construct an alternating tree.
u
x
y
even nodes
odd nodes
Case 4:
y is already contained
in T as an even vertex
cant ignore y
The cycle w y x w
is called a blossom.
w is called the base of the
blossom (even node!!!).
The path u-w path is called
the stem of the blossom.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 583/604
How to nd an augmenting path?
Construct an alternating tree.
u
x
y
w
even nodes
odd nodes
Case 4:
y is already contained
in T as an even vertex
cant ignore y
The cycle w y x w
is called a blossom.
w is called the base of the
blossom (even node!!!).
The path u-w path is called
the stem of the blossom.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 583/604
How to nd an augmenting path?
Construct an alternating tree.
u
x
y
w
even nodes
odd nodes
Case 4:
y is already contained
in T as an even vertex
cant ignore y
The cycle w y x w
is called a blossom.
w is called the base of the
blossom (even node!!!).
The path u-w path is called
the stem of the blossom.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 583/604
Flowers and Blossoms
Denition 9
A ower in a graph G = (V, E) w.r.t. a matching M and a (free)
root node r, is a subgraph with two components:

A stem is an even length alternating path that starts at the

root node r and terminates at some node w. We permit the
possibility that r = w (empty stem).

A blossom is an odd length alternating cycle that starts and

terminates at the terminal node w of a stem and has no
other node in common with the stem. w is called the base
of the blossom.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 584/604
Flowers and Blossoms
Denition 9
A ower in a graph G = (V, E) w.r.t. a matching M and a (free)
root node r, is a subgraph with two components:

A stem is an even length alternating path that starts at the

root node r and terminates at some node w. We permit the
possibility that r = w (empty stem).

A blossom is an odd length alternating cycle that starts and

terminates at the terminal node w of a stem and has no
other node in common with the stem. w is called the base
of the blossom.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 584/604
Flowers and Blossoms
Denition 9
A ower in a graph G = (V, E) w.r.t. a matching M and a (free)
root node r, is a subgraph with two components:

A stem is an even length alternating path that starts at the

root node r and terminates at some node w. We permit the
possibility that r = w (empty stem).

A blossom is an odd length alternating cycle that starts and

terminates at the terminal node w of a stem and has no
other node in common with the stem. w is called the base
of the blossom.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 584/604
Flowers and Blossoms
1
2
3
4
5
6
7
8
9
1 2 3 4 5
6
7
8
9
10
11
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 585/604
Flowers and Blossoms
Properties:
1. A stem spans 2 +1 nodes and contains matched edges
for some integer 0.
2. A blossom spans 2k +1 nodes and contains k matched
edges for some integer k 1. The matched edges match all
nodes of the blossom except the base.
3. The base of a blossom is an even node (if the stem is part of
an alternating tree starting at r).
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 586/604
Flowers and Blossoms
Properties:
1. A stem spans 2 +1 nodes and contains matched edges
for some integer 0.
2. A blossom spans 2k +1 nodes and contains k matched
edges for some integer k 1. The matched edges match all
nodes of the blossom except the base.
3. The base of a blossom is an even node (if the stem is part of
an alternating tree starting at r).
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 586/604
Flowers and Blossoms
Properties:
1. A stem spans 2 +1 nodes and contains matched edges
for some integer 0.
2. A blossom spans 2k +1 nodes and contains k matched
edges for some integer k 1. The matched edges match all
nodes of the blossom except the base.
3. The base of a blossom is an even node (if the stem is part of
an alternating tree starting at r).
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 586/604
Flowers and Blossoms
Properties:
4. Every node x in the blossom (except its base) is reachable
from the root (or from the base of the blossom) through two
distinct alternating paths; one with even and one with odd
length.
5. The even alternating path to x terminates with a matched
edge and the odd path with an unmatched edge.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 587/604
Flowers and Blossoms
Properties:
4. Every node x in the blossom (except its base) is reachable
from the root (or from the base of the blossom) through two
distinct alternating paths; one with even and one with odd
length.
5. The even alternating path to x terminates with a matched
edge and the odd path with an unmatched edge.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 587/604
Flowers and Blossoms
1 2 3 4 5
6
7
8
9
10
11
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 588/604
When during the alternating tree construction we discover a
blossom B we replace the graph G by G

= G/B, which is
obtained from G by contracting the blossom B.

Delete all vertices in B (and its incident edges) from G.

Add a new (pseudo-)vertex b. The new vertex b is

connected to all vertices in V \ B that had at least one edge
to a vertex from B.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 589/604
When during the alternating tree construction we discover a
blossom B we replace the graph G by G

= G/B, which is
obtained from G by contracting the blossom B.

Delete all vertices in B (and its incident edges) from G.

Add a new (pseudo-)vertex b. The new vertex b is

= G/B, which is
obtained from G by contracting the blossom B.

Delete all vertices in B (and its incident edges) from G.

Add a new (pseudo-)vertex b. The new vertex b is

connected to all vertices in V \ B that had at least one edge
to a vertex from B.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 589/604
Shrinking Blossoms

Edges of T that connect a node u

not in B to a node in B become
tree edges in T

connecting u to
b.

Matching edges (there is at most

one) that connect a node u not in
B to a node in B become
matching edges in M

Nodes that are connected in G to

at least one node in B become
connected to b in G

.
w
x y
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 590/604
Shrinking Blossoms

Edges of T that connect a node u

not in B to a node in B become
tree edges in T

connecting u to
b.

Matching edges (there is at most

one) that connect a node u not in
B to a node in B become
matching edges in M

Nodes that are connected in G to

at least one node in B become
connected to b in G

.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 590/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
5 6
14
2
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
5 6
14
2
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
5 6
14
2
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8
10 11
14
15
1
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Example: Blossom Algorithm
0 1 2 3
4 5 6
7
8 9
10 11 12 13
14
15 16 17
dierent
choices
dierent
choices
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 591/604
Assume that we have contracted a blossom B w.r.t. a matching
M whose base is w. We created graph G

= G/B with
pseudonode b. Let M

be the matching in the contracted graph.

Lemma 10
If G

contains an augmenting path p

starting at r (or the

pseudo-node containing r) w.r.t. to the matching M

then G
contains an augmenting path starting at r w.r.t. matching M.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 592/604
Assume that we have contracted a blossom B w.r.t. a matching
M whose base is w. We created graph G

= G/B with
pseudonode b. Let M

be the matching in the contracted graph.

Lemma 10
If G

contains an augmenting path p

starting at r (or the

pseudo-node containing r) w.r.t. to the matching M

then G
contains an augmenting path starting at r w.r.t. matching M.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 592/604
Proof.
If p

does not contain b it is also an augmenting path in G.

Case 1: non-empty stem

Next suppose that the stem is non-empty.

P
1
P
3
r
i b
q
P
1
P
3
r
i
w
k
q
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 593/604
Proof.
If p

does not contain b it is also an augmenting path in G.

Case 1: non-empty stem

Next suppose that the stem is non-empty.

P
1
P
3
r
i b
q
P
1
P
3
r
i
w
k
q
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 593/604
Proof.
If p

does not contain b it is also an augmenting path in G.

Case 1: non-empty stem

Next suppose that the stem is non-empty.

P
1
P
3
r
i b
q
P
1
P
3
r
i
w
k
q
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 593/604
Proof.
If p

does not contain b it is also an augmenting path in G.

Case 1: non-empty stem

Next suppose that the stem is non-empty.

P
1
P
3
r
i b
q
P
1
P
3
r
i
w
k
q
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 593/604

After the expansion must be incident to some node in the

blossom. Let this node be k.

If k w there is an alternating path P

2
from w to k that
ends in a matching edge.

P
1
(i, w) P
2
(k, ) P
3
is an alternating path.

If k = w then P
1
(i, w) (w, ) P
3
is an alternating path.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 594/604
Proof.
Case 2: empty stem

If the stem is empty then after expanding the blossom,

w = r.
P
3
b
q
P
3
w
k
q

The path r P
2
(k, ) P
3
is an alternating path.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 595/604
Proof.
Case 2: empty stem

If the stem is empty then after expanding the blossom,

w = r.
P
3
b
q
P
3
w
k
q

The path r P
2
(k, ) P
3
is an alternating path.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 595/604
Proof.
Case 2: empty stem

If the stem is empty then after expanding the blossom,

w = r.
P
3
b
q
P
3
w
k
q

The path r P
2
(k, ) P
3
is an alternating path.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 595/604
Proof.
Case 2: empty stem

If the stem is empty then after expanding the blossom,

w = r.
P
3
b
q
P
3
w
k
q

The path r P
2
(k, ) P
3
is an alternating path.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 595/604
Lemma 11
If G contains an augmenting path P from r to q w.r.t. matching
M then G

contains an augmenting path from r (or the

pseudo-node containing r) to q w.r.t. M

.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 596/604
Proof.

If P does not contain a node from B there is nothing to

prove.

We can assume that r and q are the only free nodes in G.

If P does not contain a node from B there is nothing to

prove.

We can assume that r and q are the only free nodes in G.

If P does not contain a node from B there is nothing to

prove.

We can assume that r and q are the only free nodes in G.

If P does not contain a node from B there is nothing to

prove.

We can assume that r and q are the only free nodes in G.

If P does not contain a node from B there is nothing to

prove.

We can assume that r and q are the only free nodes in G.

If P does not contain a node from B there is nothing to

prove.

We can assume that r and q are the only free nodes in G.

Case 1: empty stem
Let i be the last node on the path P that is part of the blossom.
P is of the form P
1
(i, j) P
2
, for some node j and (i, j) is
unmatched.
(b, j) P
2
is an augmenting path in the contracted network.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 597/604
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
Search for an augmenting path
starting at r.
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
A(i) contains neighbours of node i.
We create a copy

A(i) so that we later
can shrink blossoms.
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
found is just a Boolean that allows
to abort the search process...
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
In the beginning no node is in the tree.
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
Put the root in the tree.
list could also be a set or a stack.
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
As long as there are nodes with
unexamined neighbours...
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
...examine the next one
Algorithm 50 search(r, found)
1: set

A(i) A(i) for all nodes i
2: found false
3: unlabel all nodes;
4: give an even label to r and initialize list r
5: while list do
6: delete a node i from list
7: examine(i, found)
8: if found = true then return
If you found augmenting path
abort and start from next root.
Algorithm 50 examine(i, found)
1: for all j

A(i) do
2: if j is even then contract(i, j) and return
3: if j is unmatched then
4: q j;
5: pred(q) i;
6: found true;
7: return
8: if j is matched and unlabeled then
9: pred(j) i;
10: pred(mate(j)) j;
11: add mate(j) to list
Examine the neighbours of a node i
Algorithm 50 examine(i, found)
1: for all j

A(i) do
2: if j is even then contract(i, j) and return
3: if j is unmatched then
4: q j;
5: pred(q) i;
6: found true;
7: return
8: if j is matched and unlabeled then
9: pred(j) i;
10: pred(mate(j)) j;
11: add mate(j) to list
For all neighbours j do...
Algorithm 50 examine(i, found)
1: for all j

A(i) do
2: if j is even then contract(i, j) and return
3: if j is unmatched then
4: q j;
5: pred(q) i;
6: found true;
7: return
8: if j is matched and unlabeled then
9: pred(j) i;
10: pred(mate(j)) j;
11: add mate(j) to list
You have found a blossom...
Algorithm 50 examine(i, found)
1: for all j

A(i) do
2: if j is even then contract(i, j) and return
3: if j is unmatched then
4: q j;
5: pred(q) i;
6: found true;
7: return
8: if j is matched and unlabeled then
9: pred(j) i;
10: pred(mate(j)) j;
11: add mate(j) to list
You have found a free node which
gives you an augmenting path.
Algorithm 50 examine(i, found)
1: for all j

A(i) do
2: if j is even then contract(i, j) and return
3: if j is unmatched then
4: q j;
5: pred(q) i;
6: found true;
7: return
8: if j is matched and unlabeled then
9: pred(j) i;
10: pred(mate(j)) j;
11: add mate(j) to list
If you nd a matched node that is not
in the tree you grow...
Algorithm 50 examine(i, found)
1: for all j

A(i) do
2: if j is even then contract(i, j) and return
3: if j is unmatched then
4: q j;
5: pred(q) i;
6: found true;
7: return
8: if j is matched and unlabeled then
9: pred(j) i;
10: pred(mate(j)) j;
11: add mate(j) to list
mate(j) is a new node from
which you can grow further.
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
Contract blossom identied by
nodes i and j
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
Get all nodes of the blossom.
Time: O(m)
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
Identify all neighbours of b.
Time: O(m) (how?)
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
b will be an even node, and it has
unexamined neighbours.
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
Every node that was adjacent to a node
in B is now adjacent to b
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
Only for making a blossom
expansion easier.
Algorithm 50 contract(i, j)
1: trace pred-indices of i and j to identify a blossom B
2: create new node b and set

A(b)
xB

A(x)
3: label b even and add to list
4: update

A(j)

A(j) b for each j

A(b)
5: form a circular double linked list of nodes in B
6: delete nodes in B from the graph
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 600/604
Only delete links from nodes not in B to B.
When expanding the blossom again we can
recreate these links in time O(m).
Analysis

A contraction operation can be performed in time O(m).

Note, that any graph created will have at most m edges.

The time between two contraction-operation is basically a

BFS/DFS on a graph. Hence takes time O(m).

There are at most n contractions as each contraction

reduces the number of vertices.

The expansion can trivially be done in the same time as

needed for all contractions.

An augmentation requires time O(n). There are at most n

of them.

In total the running time is at most

n (O(mn) +O(n)) = O(mn
2
) .
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 601/604
Analysis

A contraction operation can be performed in time O(m).

Note, that any graph created will have at most m edges.

The time between two contraction-operation is basically a

BFS/DFS on a graph. Hence takes time O(m).

There are at most n contractions as each contraction

reduces the number of vertices.

The expansion can trivially be done in the same time as

needed for all contractions.

An augmentation requires time O(n). There are at most n

of them.

In total the running time is at most

n (O(mn) +O(n)) = O(mn
2
) .
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 601/604
Analysis

A contraction operation can be performed in time O(m).

Note, that any graph created will have at most m edges.

The time between two contraction-operation is basically a

BFS/DFS on a graph. Hence takes time O(m).

There are at most n contractions as each contraction

reduces the number of vertices.

The expansion can trivially be done in the same time as

needed for all contractions.

An augmentation requires time O(n). There are at most n

of them.

In total the running time is at most

n (O(mn) +O(n)) = O(mn
2
) .
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 601/604
Analysis

A contraction operation can be performed in time O(m).

Note, that any graph created will have at most m edges.

The time between two contraction-operation is basically a

BFS/DFS on a graph. Hence takes time O(m).

There are at most n contractions as each contraction

reduces the number of vertices.

The expansion can trivially be done in the same time as

needed for all contractions.

An augmentation requires time O(n). There are at most n

of them.

In total the running time is at most

n (O(mn) +O(n)) = O(mn
2
) .
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 601/604
Analysis

A contraction operation can be performed in time O(m).

Note, that any graph created will have at most m edges.

The time between two contraction-operation is basically a

BFS/DFS on a graph. Hence takes time O(m).

There are at most n contractions as each contraction

reduces the number of vertices.

The expansion can trivially be done in the same time as

needed for all contractions.

An augmentation requires time O(n). There are at most n

of them.

In total the running time is at most

n (O(mn) +O(n)) = O(mn
2
) .
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 601/604
Analysis

A contraction operation can be performed in time O(m).

Note, that any graph created will have at most m edges.

The time between two contraction-operation is basically a

BFS/DFS on a graph. Hence takes time O(m).

There are at most n contractions as each contraction

reduces the number of vertices.

The expansion can trivially be done in the same time as

needed for all contractions.

An augmentation requires time O(n). There are at most n

of them.

In total the running time is at most

n (O(mn) +O(n)) = O(mn
2
) .
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 601/604
r
i
j
q
b j
q
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 602/604
Case 2: non-empty stem
Let P
3
be alternating path from r to w. Dene M
+
= M P
3
.
In M
+
, r is matched and w is unmatched.
G must contain an augmenting path w.r.t. matching M
+
, since M
and M
+
have same cardinality.
This path must go between w and q as these are the only
unmatched vertices w.r.t. M
+
.
For M

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

cardinality.
This path must go between r and q.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 603/604
Case 2: non-empty stem
Let P
3
be alternating path from r to w. Dene M
+
= M P
3
.
In M
+
, r is matched and w is unmatched.
G must contain an augmenting path w.r.t. matching M
+
, since M
and M
+
have same cardinality.
This path must go between w and q as these are the only
unmatched vertices w.r.t. M
+
.
For M

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

+
the blossom has an empty stem. Case 1 applies.
G

has an augmenting path w.r.t. M

+
. It must also have an
augmenting path w.r.t. M

, as both matchings have the same

cardinality.
This path must go between r and q.
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 603/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
1
0
1 2 3 4 5 6
7 9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
1
0
1 2 3 4 5 6
7 9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
1
0
1 2 3 4 5 6
7 9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
3
3 4 5 6
10
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
3
3 4 5 6
10
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
3
3 4 5 6
10
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
3
3 4 5 6
10
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2 3 4 5 6
9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
1
2
0
1 2 3 4 5 6
7 9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
1
2
0
1 2 3 4 5 6
7 9 10
13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604
Example: Blossom Algorithm
2
0
1 2 3 4 5 6
7 8 9 10
11
12 13
14
15
EADS 21 Maximum Matching in General Graphs
Ernst Mayr, Harald Rcke 604/604

Topics in Abstract Algebra Herstein Solutions
No ratings yet
Topics in Abstract Algebra Herstein Solutions
93 pages
The Role of Input and Output in Second L
100% (1)
The Role of Input and Output in Second L
14 pages
Maximum Matching Bipartite Graph
No ratings yet
Maximum Matching Bipartite Graph
25 pages
Matchings and Covers in Bipartite Graphs
No ratings yet
Matchings and Covers in Bipartite Graphs
5 pages
(Lynn E. Roller) in Search of God The Mother The (BookFi) PDF
100% (4)
(Lynn E. Roller) in Search of God The Mother The (BookFi) PDF
401 pages
Matching and Hall's Theorem
No ratings yet
Matching and Hall's Theorem
3 pages
Bipartite Matching: Chapter 26 of CLSR
100% (1)
Bipartite Matching: Chapter 26 of CLSR
19 pages
Graph Theory: Maximum Flows and Min Cost Flows
No ratings yet
Graph Theory: Maximum Flows and Min Cost Flows
13 pages
Notes On Kate Chopin - The Story of An Hour
100% (1)
Notes On Kate Chopin - The Story of An Hour
22 pages
Free Crochet Pattern Lion Brand Tweed Stripes Tweedy Mitered Afghan
No ratings yet
Free Crochet Pattern Lion Brand Tweed Stripes Tweedy Mitered Afghan
4 pages
Matching
No ratings yet
Matching
71 pages
Nafs and Rizq PDF
No ratings yet
Nafs and Rizq PDF
4 pages
Matching (Graph Theory)
No ratings yet
Matching (Graph Theory)
6 pages
Ezra Pound Literary Essays PDF
50% (2)
Ezra Pound Literary Essays PDF
2 pages
A Day at The Farm! Exercises Worksheet
No ratings yet
A Day at The Farm! Exercises Worksheet
1 page
Lesson 2
No ratings yet
Lesson 2
2 pages
Matching: Algorithms and Networks
No ratings yet
Matching: Algorithms and Networks
52 pages
Where The Forest Meets The Sea Sample Lesson Plan
100% (2)
Where The Forest Meets The Sea Sample Lesson Plan
29 pages
43 - Päges - Maximum Matching Algorithm
No ratings yet
43 - Päges - Maximum Matching Algorithm
43 pages
Breadth-First Search of Graphs
No ratings yet
Breadth-First Search of Graphs
13 pages
Even Path
No ratings yet
Even Path
6 pages
Discussions About Vandanam and Vanakkam
100% (1)
Discussions About Vandanam and Vanakkam
7 pages
2024 CS1200 Lec13 Details
No ratings yet
2024 CS1200 Lec13 Details
8 pages
Hopcroft Karp Bipartite Matching
No ratings yet
Hopcroft Karp Bipartite Matching
7 pages
Or766 ch7 PDF
No ratings yet
Or766 ch7 PDF
77 pages
CS6234 L1 Matching
No ratings yet
CS6234 L1 Matching
68 pages
Graph Matching: Shmuel Wimer
No ratings yet
Graph Matching: Shmuel Wimer
71 pages
Matching in Planar Graphs
No ratings yet
Matching in Planar Graphs
27 pages
Lecture 11: Matchings: Definitions. Hall'S Theorem and Sdrs. Perfect Matchings
No ratings yet
Lecture 11: Matchings: Definitions. Hall'S Theorem and Sdrs. Perfect Matchings
50 pages
Edmond Blossoms Algorithm
No ratings yet
Edmond Blossoms Algorithm
11 pages
All Pair Shortest Path: IOI/ACM ICPC Training June 2004
No ratings yet
All Pair Shortest Path: IOI/ACM ICPC Training June 2004
61 pages
Discrete Optimization Lecture Notes 4
No ratings yet
Discrete Optimization Lecture Notes 4
4 pages
Kavathekar Scribe PDF
No ratings yet
Kavathekar Scribe PDF
10 pages
Sec 13
No ratings yet
Sec 13
11 pages
Design and Analysis of Algos
No ratings yet
Design and Analysis of Algos
137 pages
Lecture 16
No ratings yet
Lecture 16
9 pages
CSE 202: Design and Analysis of Algorithms: Instructor: Kamalika Chaudhuri
No ratings yet
CSE 202: Design and Analysis of Algorithms: Instructor: Kamalika Chaudhuri
12 pages
Competitive Programming: Maximum Bipartite Matching
No ratings yet
Competitive Programming: Maximum Bipartite Matching
19 pages
Maximum Uniquely Restricted Matching in Permutation Graphs
No ratings yet
Maximum Uniquely Restricted Matching in Permutation Graphs
9 pages
Algorithm Design by Jon Kleinberg, Eva Tardos (1) - 396
No ratings yet
Algorithm Design by Jon Kleinberg, Eva Tardos (1) - 396
1 page
Mathematical Foundations of Computer Science Lecture Outline
No ratings yet
Mathematical Foundations of Computer Science Lecture Outline
4 pages
Lec 2
No ratings yet
Lec 2
5 pages
Finding Adjacency Matching
No ratings yet
Finding Adjacency Matching
4 pages
MIT6 046JS15 Recitation7
No ratings yet
MIT6 046JS15 Recitation7
6 pages
Maximum Matchingekar Scribe
No ratings yet
Maximum Matchingekar Scribe
10 pages
L7 L8 L9 Matching Main
No ratings yet
L7 L8 L9 Matching Main
51 pages
7.5 Bipatite Matching
No ratings yet
7.5 Bipatite Matching
12 pages
Lecture 17
No ratings yet
Lecture 17
9 pages
Matching
No ratings yet
Matching
13 pages
Graph Theory 6
No ratings yet
Graph Theory 6
30 pages
Lecture 20
No ratings yet
Lecture 20
5 pages
Maximum Flow Continued: Edmonds-Karp Algorithm Maximum Bipartite Matching
No ratings yet
Maximum Flow Continued: Edmonds-Karp Algorithm Maximum Bipartite Matching
17 pages
Notes Chapter 3
No ratings yet
Notes Chapter 3
13 pages
5 Weighted Matching
No ratings yet
5 Weighted Matching
27 pages
CS261: A Second Course in Algorithms Lecture #6: Generalizations of Maximum Flow and Bipartite Matching
No ratings yet
CS261: A Second Course in Algorithms Lecture #6: Generalizations of Maximum Flow and Bipartite Matching
8 pages
20 Flow3
No ratings yet
20 Flow3
73 pages
07maxflow Applications
No ratings yet
07maxflow Applications
51 pages
Lecture 18
No ratings yet
Lecture 18
3 pages
Advanced Algorithms Course. Lecture Notes. Part 5: Speeding Up Maximum Flow Computation
No ratings yet
Advanced Algorithms Course. Lecture Notes. Part 5: Speeding Up Maximum Flow Computation
5 pages
J2EE Architecture
No ratings yet
J2EE Architecture
49 pages
Sec Augmenting Path Algorithms
No ratings yet
Sec Augmenting Path Algorithms
152 pages
Lecture Notes On Bipartite Matching
No ratings yet
Lecture Notes On Bipartite Matching
12 pages
Advanced Algorithms Lecture 03,1
No ratings yet
Advanced Algorithms Lecture 03,1
7 pages
ch2 PDF
No ratings yet
ch2 PDF
4 pages
Maximum Flow: - Maximum Flow Problem - The Ford-Fulkerson Method - Maximum Bipartite Matching
No ratings yet
Maximum Flow: - Maximum Flow Problem - The Ford-Fulkerson Method - Maximum Bipartite Matching
32 pages
Matching Graph
No ratings yet
Matching Graph
4 pages
Lecturer: Michel X. Goemans Scribe: Nick Harvey
No ratings yet
Lecturer: Michel X. Goemans Scribe: Nick Harvey
3 pages
Matching Graphs
No ratings yet
Matching Graphs
4 pages
Matchings
No ratings yet
Matchings
7 pages
HTH622B - OPERATORS MANUAL TimberRiteä 30H Control and Measuring System OMF071051 Issue 04 DIC06 (ENGLISH) AJUSTES
No ratings yet
HTH622B - OPERATORS MANUAL TimberRiteä 30H Control and Measuring System OMF071051 Issue 04 DIC06 (ENGLISH) AJUSTES
180 pages
Jason Weston Reasoning Alignment Berkeley Talk
No ratings yet
Jason Weston Reasoning Alignment Berkeley Talk
106 pages
Hindu Law Notes and Study Material
No ratings yet
Hindu Law Notes and Study Material
17 pages
Tutorial (Advanced Programming) Worksheet 6:: Assignment 1: Heat Equation
No ratings yet
Tutorial (Advanced Programming) Worksheet 6:: Assignment 1: Heat Equation
3 pages
Identifying and Remediating Reading Difficulties
No ratings yet
Identifying and Remediating Reading Difficulties
17 pages
Spelling Patterns Chart: Pattern: - Cei
No ratings yet
Spelling Patterns Chart: Pattern: - Cei
2 pages
PT - Science 6 - Q2
No ratings yet
PT - Science 6 - Q2
5 pages
Unit 12 Ok
No ratings yet
Unit 12 Ok
5 pages
Basic Mathematical Tools: I 1 I I 1 I I J
No ratings yet
Basic Mathematical Tools: I 1 I I 1 I I J
6 pages
Crossword Puzzle Jadi
No ratings yet
Crossword Puzzle Jadi
2 pages
NMR Vorlesung SS 2013 3
No ratings yet
NMR Vorlesung SS 2013 3
20 pages
English 3: Could-Couldn'T
No ratings yet
English 3: Could-Couldn'T
20 pages
Bcse328l Cryptocurrency-Technologies TH 1.0 70 Bcse328l
No ratings yet
Bcse328l Cryptocurrency-Technologies TH 1.0 70 Bcse328l
2 pages
Tutorial Presentation 8
No ratings yet
Tutorial Presentation 8
15 pages
Basic Mathematical Tools For Imaging and Visualization: Dr. Tobias Lasser
No ratings yet
Basic Mathematical Tools For Imaging and Visualization: Dr. Tobias Lasser
15 pages
Java Programming 2 Syllabus
No ratings yet
Java Programming 2 Syllabus
3 pages
Examen 2 Parte
No ratings yet
Examen 2 Parte
3 pages
Free Knitting Pattern Lion Brand Landscapes Diagonal Furrows Scarf
No ratings yet
Free Knitting Pattern Lion Brand Landscapes Diagonal Furrows Scarf
3 pages
Solution01 Wise2011
No ratings yet
Solution01 Wise2011
5 pages
CD Unit 3
No ratings yet
CD Unit 3
14 pages
Tutorial (Advanced Programming) Worksheet 8:: Assignment 1: Lists
No ratings yet
Tutorial (Advanced Programming) Worksheet 8:: Assignment 1: Lists
2 pages
Effiziente Algorithmen Und Datenstrukturen I: Aufgabe 1 (10 Punkte)
No ratings yet
Effiziente Algorithmen Und Datenstrukturen I: Aufgabe 1 (10 Punkte)
8 pages
St. Mary's Educational Institute: Online Learning 1
No ratings yet
St. Mary's Educational Institute: Online Learning 1
7 pages
Q3W7
No ratings yet
Q3W7
12 pages
SE Patterns
No ratings yet
SE Patterns
15 pages
Simulation and Animation: Computer Graphics & Visualization
No ratings yet
Simulation and Animation: Computer Graphics & Visualization
15 pages
Tutorial Presentation 1
No ratings yet
Tutorial Presentation 1
17 pages
Tutorial Presentation 2
No ratings yet
Tutorial Presentation 2
19 pages
Week 1: Required Readings and Videos/Homework
No ratings yet
Week 1: Required Readings and Videos/Homework
5 pages
G-Stomper 5 - Pattern Sequencer
No ratings yet
G-Stomper 5 - Pattern Sequencer
83 pages
Step by Step Guide LyncDebugTools - Snooper 2013
No ratings yet
Step by Step Guide LyncDebugTools - Snooper 2013
13 pages
Main Memory: Bibliographical Notes
No ratings yet
Main Memory: Bibliographical Notes
2 pages
Activities Guide and Evaluation Rubric - Unit 2 - Task 4 - Suprasegmentals
No ratings yet
Activities Guide and Evaluation Rubric - Unit 2 - Task 4 - Suprasegmentals
5 pages
Write A Short Story About A Strange Dream You Had
No ratings yet
Write A Short Story About A Strange Dream You Had
2 pages
A-level Maths Revision: Cheeky Revision Shortcuts
From Everand
A-level Maths Revision: Cheeky Revision Shortcuts
Scool Revision
3.5/5 (8)
A Short Course in Automorphic Functions
From Everand
A Short Course in Automorphic Functions
Joseph Lehner
No ratings yet