Analysis of Algorithms

CS 477/677

Minimum Spanning Trees (MST)

Instructor: George Bebis

Chapter 23
 Minimum Spanning Trees
• Spanning Tree
– A tree (i.e., connected, acyclic graph) which contains
all the vertices of the graph
• Minimum Spanning Tree
– Spanning tree with the minimum sum of weights
8 7
b c d
4 9
2
a 11 i 4 14 e
7 6
8 10
g g f
• Spanning forest 1 2

– If a graph is not connected, then there is a spanning

tree for each connected component of the graph
2
 Applications of MST
– Find the least expensive way to connect a set of
cities, terminals, computers, etc.

3
 Example
8
Problem b c
7
d
4 9
• A town has a set of houses 2
and a set of roads a 11 i 4 14 e
7 6
• A road connects 2 and only 8 10
h g f
2 houses 1 2

• A road connecting houses u and v has a repair

cost w(u, v)
Goal: Repair enough (and no more) roads such
that:
1. Everyone stays connected
i.e., can reach every house from all other houses
2. Total repair cost is minimum

4
 Minimum Spanning Trees
• A connected, undirected graph:
– Vertices = houses, Edges = roads
• A weight w(u, v) on each edge (u, v)  E

Find T  E such that: 8 7

b c d
4 9
1. T connects all vertices 2
a 11 i 4 14 e
2. w(T) = Σ(u,v)T w(u, v) is
7 6
8 10
minimized h g f
1 2

5
Properties of Minimum Spanning Trees

• Minimum spanning tree is not unique

• MST has no cycles – see why:

– We can take out an edge of a cycle, and still have
the vertices connected while reducing the cost

• # of edges in a MST:
– |V| - 1

6
 Growing a MST – Generic Approach

• Grow a set A of edges (initially empty)

• Incrementally add edges to A such that
they would belong
8 7
to a MST b c d
4 9
2
a 11 i 4 14 e
7 6
8 10
– An edge (u, v) is safe for A if and only if A 
h g f
Idea: add
{(u, v)} only
is also “safe”
a subset edges
of some MST 1 2

7
 Generic MST algorithm

1. A ← 
2. while A is not a spanning tree
3. do find an edge (u, v) that is safe for A
4. A ← A  {(u, v)} b
8
c
7
d
4 9
2
5. return A a 11 i 4 14 e
7 6
8 10
h g 2
f
1
• How do we find safe edges?

8
 Finding Safe Edges
8 7 V-S
b c d
• Let’s look at edge (h, g) 4
2
9

a 11 i 14 e
– Is it safe for A initially? 4
7 6
8 10
S
• Later on: h
1
g 2
f

– Let S  V be any set of vertices that includes h but not

g (so that g is in V - S)
– In any MST, there has to be one edge (at least) that
connects S with V - S
– Why not choose the edge with minimum weight (h,g)?

9
 Definitions
8 7
b c d
4 9
2
• A cut (S, V - S) S a 11 i 4 14 e S
7 6
is a partition of vertices V- S  8 10
 V- S
h g 2
f
1
into disjoint sets S and V - S
• An edge crosses the cut
(S, V - S) if one endpoint is in S
and the other in V – S

10
 Definitions (cont’d)
8 7
• A cut respects a set A 4
b c d
9
2
of edges  no edge S a 11 i 4 14 e S
7 6
in A crosses the cut V- S  8
h g f
10
 V- S
1 2
• An edge is a light edge
crossing a cut  its weight is minimum over all
edges crossing the cut
– Note that for a given cut, there can be > 1 light
edges crossing it

11
 Theorem
• Let A be a subset of some MST (i.e., T), (S, V - S) be a
cut that respects A, and (u, v) be a light edge crossing
(S, V-S). Then (u, v) is safe for A .
Proof:
• Let T be an MST that includes A
S
– edges in A are shaded
• Case1: If T includes (u,v), then
it would be safe for A
u
• Case2: Suppose T does not include
the edge (u, v)
• Idea: construct another MST T’ v
V-S
that includes A  {(u, v)}
12
 Theorem - Proof
• T contains a unique path p between u and v
• Path p must cross the S

cut (S, V - S) at least x

once: let (x, y) be that edge u y

• Let’s remove (x,y)  breaks p

T into two components. v

V-S
• Adding (u, v) reconnects the components
T’ = T - {(x, y)}  {(u, v)}
13
 Theorem – Proof (cont.)
T’ = T - {(x, y)}  {(u, v)}
S
Have to show that T’ is an MST:
x
• (u, v) is a light edge
u
 w(u, v) ≤ w(x, y) y
p
• w(T ’) = w(T) - w(x, y) + w(u, v)
v
≤ w(T) V-S

• Since T is a spanning tree

w(T) ≤ w(T ’)  T’ must be an MST as well
14
 Theorem – Proof (cont.)

Need to show that (u, v) is safe for A:

i.e., (u, v) can be a part of an MST

• A  T and (x, y)  T  S

(x, y)  A  A T’ x

• A  {(u, v)}  T’ u y
p
• Since T’ is an MST
v
V-S
 (u, v) is safe for A
15
 Prim’s Algorithm
• The edges in set A always form a single tree
• Starts from an arbitrary “root”: VA = {a}
• At each step: 8 7
b c d
– Find a light edge crossing (VA, V - VA) 4 9
2
– Add this edge to A a 11 i 4 14 e
7 6
– Repeat until the tree spans all vertices 8 10
h g 2
f
1

16
 How to Find Light Edges Quickly?
Use a priority queue Q: 8 7
b c d
• Contains vertices not yet 4 9
2
11
included in the tree, i.e., (V – VA) a i 4 14 e
7 6
– VA = {a}, Q = {b, c, d, e, f, g, h, i} 8 10
h g 2
f
1
• We associate a key with each vertex v:
key[v] = minimum weight of any edge (u, v)
connecting v to VA

Key[a]=min(w1,w2) w1

w2
a

17
 How to Find Light Edges Quickly?
(cont.)
• After adding a new node to VA we update the weights of all
the nodes adjacent to it
e.g., after adding a to the tree, k[b]=4 and k[h]=8
• Key of v is  if v is not adjacent to any vertices in VA

8 7
b c d
4 9
2
a 11 i 4 14 e
7 6
8 10
h g 2
f
1

18
 Example
  
b
8
c
7
d 0  
4 9
 2  Q = {a, b, c, d, e, f, g, h, i}
11
a i 4 14 e VA = 
7 6
8 10 Extract-MIN(Q)  a
h g 2
f
1
  

4
  
8 7 key [b] = 4  [b] = a
b c d
4 9 key [h] = 8  [h] = a
 2 
a 11 i 4 14 e
4 8
7 6
8 10 Q = {b, c, d, e, f, g, h, i} VA = {a}
h g f
1 2 Extract-MIN(Q)  b
8  

19
 Example
4 8
  key [c] = 8  [c] = b
8 7
b c d key [h] = 8  [h] = a - unchanged
4 9
 2 
8 8
a 11 i 4 14 e
7 6 Q = {c, d, e, f, g, h, i} VA = {a, b}
8 10
h g f Extract-MIN(Q)  c
1 2
8  
key [d] = 7  [d] = c
4 8 
7
8 7 key [f] = 4  [f] = c
b c d
4 9 key [i] = 2  [i] = c
2 2
 
a 11 i 4 14 e
7 6 7 4 8 2
8 10
h g f
Q = {d, e, f, g, h, i} VA = {a, b, c}
1 2
8   Extract-MIN(Q)  i
4
20
 Example
4 8 7 key [h] = 7  [h] = i
8 7
b c d key [g] = 6  [g] = i
4 9
2 2  7 46 8
a 11 i 4 14 e
6
Q = {d, e, f, g, h} VA = {a, b, c, i}
7
8 10
h g f
Extract-MIN(Q)  f
1 2
87 
6 4
4 8 7 key [g] = 2  [g] = f
b
8
c
7
d key [d] = 7  [d] = c unchanged
4 9
2 2 10 key [e] = 10  [e] = f
a 11 i 14 e
4 7 10 2 8
7 6
8 10 Q = {d, e, g, h} VA = {a, b, c, i, f}
h g 2
f
1
7 6 4 Extract-MIN(Q)  g
2 21
 Example
4 8 7 key [h] = 1  [h] = g
8 7
b c d 7 10 1
4 9
2 2 10 Q = {d, e, h} VA = {a, b, c, i, f, g}
a 11 i 4 14 e
7 6 Extract-MIN(Q)  h
8 10
h g 2
f
1
7
1 2 4 7 10
4 8 7
8 7
Q = {d, e} VA = {a, b, c, i, f, g, h}
b c d
4 9 Extract-MIN(Q)  d
2 2 10
a 11 i 4 14 e
7 6
8 10
h g 2
f
1
1 2 4
22
 Example

4 8 7
8 7
b c d
9 9
key [e] = 9  [e] = f
4
2 2 10
9
a 11 i 4 14 e
7 6 Q = {e} VA = {a, b, c, i, f, g, h, d}
8 10
h g f Extract-MIN(Q)  e
1 2
1 2 4 Q =  VA = {a, b, c, i, f, g, h, d, e}

23
 PRIM(V, E, w, r)
1. Q← 
Total time: O(VlgV + ElgV) = O(ElgV)
2. for each u  V
3. do key[u] ← ∞ O(V) if Q is implemented
as a min-heap
4. π[u] ← NIL
5. INSERT(Q, u)
6. DECREASE-KEY(Q, r, 0) ► key[r] ← 0 O(lgV)
Min-heap
7. while Q   Executed |V| times
operations:
8. do u ← EXTRACT-MIN(Q) Takes O(lgV) O(VlgV)

9. for each v  Adj[u] Executed O(E) times total

10. do if v  Q and w(u, v) < key[v] Constant O(ElgV)

11. then π[v] ← u Takes O(lgV)

12. DECREASE-KEY(Q, v, w(u, v))

24
 Using Fibonacci Heaps

• Depending on the heap implementation, running time

could be improved!

25
 Prim’s Algorithm
• Prim’s algorithm is a “greedy” algorithm
– Greedy algorithms find solutions based on a sequence
of choices which are “locally” optimal at each step.

• Nevertheless, Prim’s greedy strategy produces a

globally optimum solution!
– See proof for generic approach (i.e., slides 12-15)

26
 A different instance of the
generic approach
S
(instance 1)

(instance 2)
u

tree1
v
V-S

• A is a forest containing connected u

components
– Initially, each component is a single
vertex v
tree2
• Any safe edge merges two of
these components into one
– Each component is a tree 27
 Kruskal’s Algorithm
• How is it different from Prim’s algorithm?
– Prim’s algorithm grows one
tree all the time
– Kruskal’s algorithm grows tree1

multiple trees (i.e., a forest)

at the same time.
u
– Trees are merged together
using safe edges
v
– Since an MST has exactly |V| - 1 tree2
edges, after |V| - 1 merges,
we would have only one component
28
 Kruskal’s Algorithm
8 7
• Start with each vertex being its b c d
4 9
own component 2
a 11 i 4 14 e
• Repeatedly merge two 7 6
8 10
components into one by h g f
1 2
choosing the light edge that
We would add
connects them edge (c, f)
• Which components to consider
at each iteration?
– Scan the set of edges in
monotonically increasing order by
weight

29
 Example
1. Add (h, g) {g, h}, {a}, {b}, {c}, {d}, {e}, {f}, {i}
8 7
b c d 2. Add (c, i) {g, h}, {c, i}, {a}, {b}, {d}, {e}, {f}
4 9
2 3. Add (g, f) {g, h, f}, {c, i}, {a}, {b}, {d}, {e}
a 11 i 4 14 e 4. Add (a, b) {g, h, f}, {c, i}, {a, b}, {d}, {e}
7 6
8 10 5. Add (c, f) {g, h, f, c, i}, {a, b}, {d}, {e}
h g 2
f 6. Ignore (i, g) {g, h, f, c, i}, {a, b}, {d}, {e}
1
7. Add (c, d) {g, h, f, c, i, d}, {a, b}, {e}
1: (h, g) 8: (a, h), (b, c) 8. Ignore (i, h) {g, h, f, c, i, d}, {a, b}, {e}
2: (c, i), (g, f) 9: (d, e) 9. Add (a, h) {g, h, f, c, i, d, a, b}, {e}
4: (a, b), (c, f) 10: (e, f) 10. Ignore (b, c) {g, h, f, c, i, d, a, b}, {e}
6: (i, g) 11: (b, h) 11. Add (d, e) {g, h, f, c, i, d, a, b, e}
7: (c, d), (i, h) 14: (d, f) 12. Ignore (e, f) {g, h, f, c, i, d, a, b, e}
{a}, {b}, {c}, {d}, {e}, {f}, {g}, {h}, {i} 13. Ignore (b, h) {g, h, f, c, i, d, a, b, e}
14. Ignore (d, f) {g, h, f, c, i, d, a, b, e}
30
 Implementation of Kruskal’s Algorithm

8 7
• Uses a disjoint-set data 4
b c d
9
2
structure (see Chapter a 11 i 4 14 e
7 6
21) to determine whether 8 10
h g 2
f
1
an edge connects
We would add
vertices in different edge (c, f)

components

31
 Operations on Disjoint Data Sets
• MAKE-SET(u) – creates a new set whose only
member is u
• FIND-SET(u) – returns a representative element
from the set that contains u
– Any of the elements of the set that has a particular
property
– E.g.: Su = {r, s, t, u}, the property is that the element be
the first one alphabetically
FIND-SET(u) = r FIND-SET(s) = r
– FIND-SET has to return the same value for a given set
32
 Operations on Disjoint Data Sets
• UNION(u, v) – unites the dynamic sets that
contain u and v, say Su and Sv
– E.g.: Su = {r, s, t, u}, Sv = {v, x, y}

UNION (u, v) = {r, s, t, u, v, x, y}

• Running time for FIND-SET and UNION depends

on implementation.
• Can be shown to be α(n)=O(lgn) where α() is a
very slowly growing function (see Chapter 21)
33
 KRUSKAL(V, E, w)
1. A← 
2. for each vertex v  V
O(V)
3. do MAKE-SET(v)
4. sort E into non-decreasing order by w O(ElgE)
5. for each (u, v) taken from the sorted list O(E)
6. do if FIND-SET(u)  FIND-SET(v)
7. then A ← A  {(u, v)} O(lgV)
8. UNION(u, v)
9. return A
Running time: O(V+ElgE+ElgV)=O(ElgE) – dependent on
the implementation of the disjoint-set data structure
34
 KRUSKAL(V, E, w) (cont.)
1. A← 
2. for each vertex v  V
O(V)
3. do MAKE-SET(v)
4. sort E into non-decreasing order by w O(ElgE)
5. for each (u, v) taken from the sorted list O(E)
6. do if FIND-SET(u)  FIND-SET(v)
7. then A ← A  {(u, v)} O(lgV)
8. UNION(u, v)
9. return A
- Running time: O(V+ElgE+ElgV)=O(ElgE) O(ElgV)
- Since E=O(V2), we have lgE=O(2lgV)=O(lgV)
35
 Kruskal’s Algorithm
• Kruskal’s algorithm is a “greedy” algorithm
• Kruskal’s greedy strategy produces a globally
optimum solution
S
• Proof for generic approach
x

applies to Kruskal’s
u y
algorithm too

v
V-S

36
 Problem 1
• (Exercise 23.2-3, page 573) Compare Prim’s algorithm
with and Kruskal’s algorithm assuming:

(a) sparse graphs:

In this case, E=O(V)

Kruskal:

O(ElgE)=O(VlgV)

Prim:
- binary heap: O(ElgV)=O(VlgV)
- Fibonacci heap: O(VlgV+E)=O(VlgV)
37
 Problem 1 (cont.)

(b) dense graphs

In this case, E=O(V2)

Kruskal:

O(ElgE)=O(V2lgV2)=O(2V2lgV)=O(V2lgV)
Prim:
- binary heap: O(ElgV)=O(V2lgV)
- Fibonacci heap: O(VlgV+E)=O(VlgV+V2)=O(V2)

38
 Problem 2

(Exercise 23.2-4, page 574): Analyze the

running time of Kruskal’s algorithm when
weights are in the range [1 … V]

39
 Problem 2 (cont.)
1. A← 
2. for each vertex v  V
O(V)
3. do MAKE-SET(v)
4. sort E into non-decreasing order by w O(ElgE)
5. for each (u, v) taken from the sorted list O(E)
6. do if FIND-SET(u)  FIND-SET(v)
7. then A ← A  {(u, v)} O(lgV)
8. UNION(u, v)
9. return A
- Sorting can be done in O(E) time (e.g., using counting sort)
- However, overall running time will not change, i.e, O(ElgV) 40
 Problem 3
• Suppose that some of the weights in a connected
graph G are negative. Will Prim’s algorithm still
work? What about Kruskal’s algorithm? Justify
your answers.
– Yes, both algorithms will work with negative weights.
Review the proof of the generic approach; there is no
assumption in the proof about the weights being
positive.

41
 Problem 4
• (Exercise 23.2-2, page 573) Analyze Prim’s
algorithm assuming:

(a) an adjacency-list representation of G

O(ElgV)

(b) an adjacency-matrix representation of G

O(ElgV+V2)

42
 PRIM(V, E, w, r)
1. Q← 
Total time: O(VlgV + ElgV) = O(ElgV)
2. for each u  V
3. do key[u] ← ∞ O(V) if Q is implemented
as a min-heap
4. π[u] ← NIL
5. INSERT(Q, u)
6. DECREASE-KEY(Q, r, 0) ► key[r] ← 0 O(lgV)
Min-heap
7. while Q   Executed |V| times
operations:
8. do u ← EXTRACT-MIN(Q) Takes O(lgV) O(VlgV)

9. for each v  Adj[u] Executed O(E) times

10. do if v  Q and w(u, v) < key[v] Constant O(ElgV)

11. then π[v] ← u Takes O(lgV)

12. DECREASE-KEY(Q, v, w(u, v))

43
 PRIM(V, E, w, r)
1. Q← 
Total time: O(VlgV + ElgV+V2) = O(ElgV+V2)
2. for each u  V
3. do key[u] ← ∞ O(V) if Q is implemented
as a min-heap
4. π[u] ← NIL
5. INSERT(Q, u)
6. DECREASE-KEY(Q, r, 0) ► key[r] ← 0 O(lgV)
Min-heap
7. while Q   Executed |V| times
operations:
8. do u ← EXTRACT-MIN(Q) Takes O(lgV) O(VlgV)
9. for (j=0; j<|V|; j++) Executed O(V2) times total
10. if (A[u][j]=1) Constant
11. if v  Q and w(u, v) < key[v]
12. then π[v] ← u Takes O(lgV) O(ElgV)
13. DECREASE-KEY(Q, v, w(u, v)) 44
 Problem 5
• Find an algorithm for the “maximum” spanning
tree. That is, given an undirected weighted
graph G, find a spanning tree of G of maximum
cost. Prove the correctness of your algorithm.
– Consider choosing the “heaviest” edge (i.e., the edge
associated with the largest weight) in a cut. The
generic proof can be modified easily to show that this
approach will work.
– Alternatively, multiply the weights by -1 and apply
either Prim’s or Kruskal’s algorithms without any
modification at all!

45
 Problem 6
• (Exercise 23.1-8, page 567) Let T be a MST of
a graph G, and let L be the sorted list of the
edge weights of T. Show that for any other MST
T’ of G, the list L is also the sorted list of the
edge weights of T’

T, L={1,2} T’, L={1,2}

46