Introduction to Algorithms

6.006

Lecture 16
Prof. Piotr Indyk

Menu
Last week: Bellman-Ford
O(VE) time
general weights
Today: Dijkstra
O( (V+E)logV ) time
non-negative weights

Introduction to Algorithms

4/5/11

Single source shortest path

problem
Problem: Given a digraph G = (V, E) with nonnegative edge-weight function w, and a node s, find
(s, v)* for all v in V
Want a fast algorithm
Question: what if all edge weights are equal to 1 ?
a

*Paths can be found as well

Introduction to Algorithms

4/5/11

BFS
1

Running time: O(V+E)

Weighted graphs
7

b
3

f
5

c
8
2

5
d

g
4

Question II: what if all edge weights are

integers in the range 1W ?
Introduction to Algorithms

4/5/11

Weighted graphs
3b
3

7
5

c
8
2

5
d

4
5
Algorithm:
Create an unweighted graph by splitting each
edge with weight w into w pieces
Run BFS

Running time: O(V+WE)

Introduction to Algorithms

4/5/11

Greedy approach
IDEA: Greedy.
1. Maintain a set S of vertices whose shortestpath distances from s are known.
2. At each step add to S the vertex v V S
whose distance estimate from s is minimal.
3. Update the distance estimates of vertices
adjacent to v.

Introduction to Algorithms

4/5/11

Dijkstras algorithm
d[s] 0
for each v V {s}
do d[v]

S
QV
Q is a priority queue maintaining V S
while Q
do u EXTRACT-MIN(Q)
S S {u}
for each v Adj[u]
relaxation
do if d[v] > d[u] + w(u, v)
then d[v] d[u] + w(u, v)
step

Implicit DECREASE-KEY
Introduction to Algorithms

4/5/11

Dijkstra: Example
(,-)

initialization

3
(0,*)+

(,-)
f

7
5

c
8

(,-)

5
8

g
(,-)

(,-)
h

j (,-)

d
(,-)

(,-)
e

6
i
(,-)

Dijkstra: Example
(3,a)+

1st iteration

3
(0,*)+

(,-)
f

7
5

c
8

(8,a)

5
8

g
(,-)

(,-)
h

j (,-)

d
(5,a)

(,-)
e

6
i
(,-)

Dijkstra: Example
2nd iteration
3
(0,*)+

(10,b)
f

(3,a)+ b

c
8

(8,a)

d
(5,a)+

(,-)
e

g
(,-)

Introduction to Algorithms

(,-)
h

j (,-)

6
i
(,-)

4/5/11

11

Dijkstra: Example
3rd iteration
3
(0,*)+

(10,b)
f

(3,a)+ b

c
8

(7,d)+

d
(5,a)+

(,-)
e

g
(9,d)

(,-)
h

j (,-)

6
i
(,-)

Dijkstra: Example
4th iteration
3
(0,*)+

(10,b)
f

(3,a)+ b

c
8

(7,d)+

5
8

g
(9,d)+

Introduction to Algorithms

(,-)
h

j (,-)

d
(5,a)+

(15,c)
e

6
i
(,-)

4/5/11

13

Dijkstra: Example
(3,a)+

5th iteration

3
(0,*)+

(10,b)+
f

7
5

c
8

(7,d)+

d
(5,a)+

(15,c)
e

g
(9,d)+

(,-)
h

j (,-)

6
i
(13,g)

Dijkstra: Example
(3,a)+

6th iteration

3
(0,*)+

(10,b)+
f

7
5

c
8

(7,d)+

5
8

g
(9,d)+

Introduction to Algorithms

(16,f)
h

j (,-)

d
(5,a)+

(15,c)
e

6
i
(13,g)+

4/5/11

15

Dijkstra: Example
7th iteration
3
(0,*)+

(10,b)+
f

(3,a)+ b

c
8

(7,d)+

d
(5,a)+

(14,i)+
e

g
(9,d)+

j (19,i)

6
4

(16,f)
h

6
i
(13,g)+

Dijkstra: Example
(3,a)+

8th iteration

3
(0,*)+

(10,b)+
f

7
5

c
8

(7,d)+

5
8

g
(9,d)+

Introduction to Algorithms

(15,e)+
h

j (19,i)

d
(5,a)+

(14,i)+
e

6
i
(13,g)+

4/5/11

17

Dijkstra: Example
(3,a)+

9th iteration

3
(0,*)+

c
8

(7,d)+

6
5

g
(9,d)+

(7,d)+

d
(5,a)+

g
(9,d)+

2
j (17,h)+

(15,e)+
h
(14,i)+

j (17,h)+

i
(13,g)+

3
(0,*)+

(10,b)+
f

(15,e)+
h

d
(5,a)+
(3,a)+

(14,i)+
e

Shortest-path tree

(10,b)+
f

i
(13,g)+

Correctness Part I
Lemma. Initializing d[s] 0 and d[v] for all
v V {s} establishes d[v] (s, v) for all v V,
and this invariant is maintained over any sequence
of relaxation steps.
Proof. Recall relaxation step:
if d[v] > d[u] + w(u, v) set d[v] d[u] + w(u, v)
d[u]

s
d[v]
Introduction to Algorithms

w(u, v)

v
4/5/11

19

Correctness Part II
Theorem. Dijkstras algorithm terminates with
d[v] = (s, v) for all v V.

Proof.
It suffices to show that d[v] = (s, v) for every v V
when v is added to S
Suppose u is the first vertex added to S for which
d[u] (s, u) . Let y be the first vertex in V S along a
shortest path from s to u, and let x be its predecessor:

u
S, just before
adding u.

s
Introduction to Algorithms

y
4/5/11

20

Correctness Part II
(continued)
S
s

Since u is the first vertex violating the claimed invariant,

we have d[x] = (s, x)
Since subpaths of shortest paths are shortest paths, it
follows that d[y] was set to (s, x) + w(x, y) = (s, y) just
after x was added to S
Consequently, we have d[y] = (s, y) (s, u) d[u]
But, d[y] d[u] since the algorithm chose u first
Hence d[y] = (s, y) = (s, u) = d[u] - contradiction
Introduction to Algorithms

4/5/11

21

Analysis of Dijkstra
|V |
times

while Q
do u EXTRACT-MIN(Q)
S S {u}
for each v Adj[u]
degree(u)
do if d[v] > d[u] + w(u, v)
times
then d[v] d[u] + w(u, v)
DECREASE-KEY

Time = (V)TEXTRACT-MIN + (E)TDECREASE-KEY

Introduction to Algorithms

4/5/11

22

Analysis of Dijkstra
(continued)
Time = (V)TEXTRACT-MIN + (E)TDECREASE-KEY
Q
array

TEXTRACT-MIN TDECREASE-KEY
O(V)

binary
O(lg V)
heap
Fibonacci O(lg V)
heap
amortized

Total

O(1)

O(V2)

O(lg V)

O(E lg V)

O(1)
O(E + V lg V)
amortized worst case

Introduction to Algorithms

4/5/11

23

Tuesday: generic algorithm

d[s] 0
for each v V {s}
do d[v]

initialization

while there is an edge (u, v) E s. t.

d[v] > d[u] + w(u, v) do
select one such edge somehow
set d[v] d[u] + w(u, v)
d[u]
endwhile

relaxation
step
u

w(u, v)

d[v]

How to do it in O( (V+E)logV ) time ?

Introduction to Algorithms

4/5/11

24