2010/11/17

Weighted Graphs
22c:21 Computer Science II Data Structures
In a weighted graph, each edge has an associated numerical value, called the weight of the edge Edge weights may represent, distances, costs, etc. Example:
In a flight route graph, the weight of an edge represents the distance in miles between the endpoint airports

SFO

ORD LGA

PVD

Shortest Paths
HNL LAX DFW

MIA
2

Shortest Path Problem

Given a weighted graph and two vertices u and v, we want to find a path of minimum total weight between u and v.
Length of a path is the sum of the weights of its edges.

Shortest Path Properties

Property 1:
A subpath of a shortest path is itself a shortest path

Example:
Shortest path between Providence and Honolulu

Property 2:
There is a tree of shortest paths from a start vertex to all the other vertices

Applications
Internet packet routing Flight reservations Driving directions

Example:
Tree of shortest paths from Providence

SFO HNL

ORD LGA

PVD SFO HNL

ORD LGA

PVD

LAX

DFW

LAX

DFW

MIA
3

MIA
4

Dijkstras Algorithm
The distance of a vertex v from a vertex s is the length of a shortest path between s and v Dijkstras algorithm computes the distances of all the vertices from a given start vertex s Assumptions:
the graph is connected the edges are undirected the edge weights are nonnegative

Edge Relaxation
Consider an edge e = (u,z) such that
u is the vertex most recently added to the cloud z is not in the cloud s d(u) = 50 u e

We grow a cloud of vertices, beginning with s and eventually covering all the vertices We store with each vertex v a label d(v) representing the distance of v from s in the subgraph consisting of the cloud and its adjacent vertices At each step
We add to the cloud the vertex u outside the cloud with the smallest distance label, d(u) We update the labels of the vertices adjacent to u

d(z) = 75
z

The relaxation of edge e updates distance d(z) as follows:

d(z) min{d(z),d(u) + weight(e)} s d(u) = 50 u e

d(z) = 60
z

2010/11/17

Example
0 8 2 B 2 8 E 7 3 C 2 9 F 1 D 5 4 B 2 8 5 E 7 3 A 4 8 2 C 2 9 F 1 8 D 5 2 3 7 B 5 E 7 3 A 0 4 8 2 2 C 9 A 0

Example (cont.)
4 1 8 F 3 D 5 0 8 4 2 C 3 E 9 F 1 8 3 D 2 5
7

0 8 2 8 B 2 5 E 7 3 2 C 9 1 11 F 3 D 5 B 2 5 7 7 A 4 8 2 A

0 B

A 2 7 5 E 3 2 9 1 8 F

4 3

D 5

Dijkstras Algorithm
Graph operations A priority queue stores the vertices outside the cloud
Key: distance Element: vertex Algorithm DijkstraDistances(G, s) Q new heap-based priority queue for all v G.vertices() if (v == s) setDistance(v, 0) else setDistance(v, ) Q.insert(getDistance(v), v) while Q.isEmpty() u Q.removeMin() for all e G.incidentEdges(u)dge e z G.opposite(u,e) r getDistance(u) + weight(e) if r < getDistance(z) setDistance(z,r) Q.update(r, z)
9

Analysis
Method incidentEdges is called once for each vertex

Label operations
We set/get the distance and locator labels of vertex z O(deg(z)) times Setting/getting a label takes O(1) time

Priority queue operations

Each vertex is inserted once into and removed once from the priority queue, where each insertion or removal takes O(log n) time The key of a vertex in the priority queue is modified at most deg(w) times, where each key change takes O(log n) time

Locator-based methods
insert(k, v) inserts a vertex with key into Q update(k, v) changes the key of an item

We store two labels with each vertex:

Distance (d(v) label) position in priority queue

Dijkstras algorithm runs in O((n + m) log n) time provided the graph is represented by the adjacency list structure
Recall that v deg(v) = 2m

The running time can also be expressed as O(m log n) since the graph is connected and n = O(m). 10

Dijkstra's Shortest Path Algorithm

Find shortest path from s to t.
24

Dijkstra's Shortest Path S = Algorithm { }

PQ = { s, 2, 3, 4, 5, 6, 7, t }
24

2
9

3 0 s
2 6 9

s
14

18

18 14

6
30

6
30 15 5 6 20 16 11

5
11

19 15 5

19

6 20 16

44

t distance label

44

2010/11/17

Dijkstra's Shortest Path S = Algorithm { }

PQ = { s, 2, 3, 4, 5, 6, 7, t }
delmin

Dijkstra's Shortest Path S = Algorithm {s}

PQ = { 2, 3, 4, 5, 6, 7, t }
decrease key

2 0 s
9 18 14 24

3 0 s
2 6 9

X 9 2
24

18 14

6
30

X 14 6
30

5
11

5
11

19 15 5 6

19

15 5 20

6 20 16

16

44

7 X 15

44

Dijkstra's Shortest Path S = Algorithm {s}

PQ = { 2, 3, 4, 5, 6, 7, t }
delmin X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm { s, 2 }

PQ = { 3, 4, 5, 6, 7, t }
3 0 s
9 18 14

X 9 2
24

X 14 6
30

X 14 6
30

5
11

5
11

19 15 5 6

19

15 5 20

6 20 16

16

7 X 15

44

7
16

44

X 15

Dijkstra's Shortest Path S = Algorithm { s, 2 }

PQ = { 3, 4, 5, 6, 7, t }
decrease key X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm { s, 2 }

PQ = { 3, 4, 5, 6, 7, t }
X 33 3 0 s
9

X 9 2 delmin
18 14 24

X 33 3

X 14 6
30

X 14 6
30

5
11

5
11

19 15 5 6

19

15 5 20

6 20 16

16

7
17

44

7 X 15

44

X 15

2010/11/17

Dijkstra's Shortest Path S = Algorithm { s, 2, 6 }

PQ = { 3, 4, 5, 7, t }
32 X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm { s, 2, 6 }

PQ = { 3, 4, 5, 7, t }
32 X 9 2 0 s
9 18 14 24

X X 33 3

X X 33 3

X 14 6
30

X 14 6
30

44 X 5

11

19 15 5 6

44 X 5

11

19

15 5 20

6 20 16

16

7 X 15

44

7 X 15 delmin

44

Dijkstra's Shortest Path S = Algorithm { s, 2, 6, 7 }

PQ = { 3, 4, 5, t }
32 X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm { s, 2, 6, 7 }

PQ = { 3, 4, 5, t }
X X 33 3 0 s
9 18 14

delmin 32

X 9 2
24

X X 33 3

X 14 6
30

X 14 6
30

44 35 X X

11

19 15 5 6

44 35 X X

11

19

15 5 20

6 20 16

16

7 X 15

44

t 59 X

7 X 15

44

t 59 X

Dijkstra's Shortest Path S = Algorithm { s, 2, 3, 6, 7 }

PQ = { 4, 5, t }
32 X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm { s, 2, 3, 6, 7 }

PQ = { 4, 5, t }
32 X 9 2 0 s
9 18 14 24

X X 33 3

X X 33 3

X 14 6
30

X 14 6
30

44 35 34 XX X

11

19 15 5 6

44 35 34 XX X

11

19

15 5 20

6 20

16

delmin
44

16

7 X 15

44

t 51 59 X X

7 X 15

t 51 59 X X

2010/11/17

Dijkstra's Shortest Path S = Algorithm { s, 2, 3, 5, 6, 7 }

PQ = { 4, t }
32 X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm { s, 2, 3, 5, 6, 7 }

PQ = { 4, t }
32 X 9 2 0 s
9 18 14 24

X X 33 3

X X 33 3

X 14 6
30

X 14 6
30

44 35 34 XX X

11

45 X 4

19 15 5 6

44 35 34 XX X

11

45 X 4 delmin
6

19

15 5 20

16

20

16

7 X 15

44

t 50 51 59 X X X

7 X 15

44

t 50 51 59 X X X

Dijkstra's Shortest Path S = Algorithm} { s, 2, 3, 4, 5, 6, 7

PQ = { t }
32 X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm} { s, 2, 3, 4, 5, 6, 7

PQ = { t }
32 X 9 2 0 s
9 18 14 24

X X 33 3

X X 33 3

X 14 6
30

X 14 6
30

44 35 34 XX X

11

45 X 4

19 15 5 6

44 35 34 XX X

11

45 X 4

19

15 5 20

6 20 16

16

7 X 15

44

t 50 51 59 X X X

7 X 15

44

t delmin 50 51 59 X X X

Dijkstra's Shortest Path S = Algorithm t } { s, 2, 3, 4, 5, 6, 7,

PQ = { }
32 X 9 2 0 s
9 18 14 24

Dijkstra's Shortest Path S = Algorithm t } { s, 2, 3, 4, 5, 6, 7,

PQ = { }
32 X 9 2 0 s
9 18 14 24

X X 33 3

X X 33 3

X 14 6
30

X 14 6
30

44 35 34 XX X

11

45 X 4

19 15 5 6

44 35 34 XX X

11

45 X 4

19

15 5 20

6 20 16

16

7 X 15

44

t 50 51 59 X X X

7 X 15

44

t 50 51 59 X X X

2010/11/17

Extension
Using the template method pattern, we can extend Dijkstras algorithm to return a tree of shortest paths from the start vertex to all other vertices We store with each vertex a third label: parent edge in the shortest path tree In the edge relaxation step, we update the parent label
Algorithm DijkstraShortestPathsTree(G, s) for all v G.vertices() setParent(v, ) for all e G.incidentEdges(u) { relax edge e } z G.opposite(u,e) r getDistance(u) + weight(e) if r < getDistance(z) setDistance(z,r) setParent(z,e) Q.replaceKey(getLocator(z),r)
31

Why Dijkstras Algorithm Works

Dijkstras algorithm is based on the greedy method. It adds vertices by increasing distance. Suppose it didnt find all shortest distances. Let F be the first wrong vertex the algorithm processed. When the previous node, D, on the true shortest path was considered, its distance was correct. But the edge (D,F) was relaxed at that time! Thus, so long as d(F)>d(D), Fs distance cannot be wrong. That is, there is no wrong vertex.
0 8 2 B 2 7 5 E 7 3 C 2 9 1 8 F D 5 3 A 4

32

Why It Doesnt Work for NegativeWeight Edges

Dijkstras algorithm is based on the greedy method. It adds vertices by increasing distance.
0 A 6 7 B 2 5 E 7 0 5 C -8 1 9 F 4 D 5 4

Bellman-Ford Algorithm
Works even with negativeweight edges Must assume directed edges (for otherwise we would have negative-weight cycles) Iteration i finds all shortest paths that use i edges. Running time: O(nm). Can be extended to detect a negative-weight cycle if it exists How?
Algorithm BellmanFord(G, s) for all v G.vertices() if v = s setDistance(v, 0) else setDistance(v, ) for i 1 to n-1 do for each e G.edges() { relax edge e } u G.origin(e) z G.opposite(u,e) r getDistance(u) + weight(e) if r < getDistance(z) setDistance(z,r)
34

If a node with a negative incident edge were to be added late to the cloud, it could mess up distances for vertices already in the cloud.

Cs true distance is 1, but it is already in the cloud with d(C) = 5

33

Bellman-Ford Example
Nodes are labeled with their d(v) values
8 -2 -2 7 3 9 5 -2 1 8 7 3 0 4 8 -2 -2 9 5 1 4 0 4

DAG-based Algorithm
Works even with negative-weight edges Uses topological order Doesnt use any fancy data structures Is much faster than Dijkstras algorithm Running time: O(n+m).
Algorithm DagDistances(G, s) for all v G.vertices() if v = s setDistance(v, 0) else setDistance(v, ) Perform a topological sort of the vertices for u 1 to n do // in topological order for each e G.outEdges(u) // relax edge e z G.opposite(u,e) r getDistance(u) + weight(e) if r < getDistance(z) setDistance(z,r)
36

8 -2 5 8 1 -2 6 7 3

0 1 9 9

4 -1

8 -2 4 5 5 -2 7 3 1

0 1 9 4 9

-2

-2

-1 5
35

2010/11/17

DAG Example
Nodes are labeled with their d(v) values
1 1

All-Pairs Shortest Paths

0 -2 4 -2 9
6 1 5

8 -2
3

0 7
2

4 1 9 5
5 4 3

8 -5 8 7 3
2

-5

3
6 1

8 -2
3

0 7 1
6 2

4 1 9 7
5 4

8 -2 4 5 -1
3

0 7 0 1 3
6 2

4 1 9 7 4
5 37 4

8 -5

-2

5 -5

-2

-1 5

Find the distance between every pair of vertices in a weighted directed graph G. We can make n calls to Dijkstras algorithm (if no negative edges), which takes O(nmlog n) time. Likewise, n calls to Bellman-Ford would take O(n2m) time. We can achieve O(n3) time using dynamic programming (similar to the Floyd-Warshall algorithm).

Algorithm AllPair(G) {assumes vertices 1,,n} for all vertex pairs (i,j) if i = j D0[i,i] 0 else if (i,j) is an edge in G D0[i,j] weight of edge (i,j) else D0[i,j] + for k 1 to n do for i 1 to n do for j 1 to n do Dk[i,j] min{Dk-1[i,j], Dk-1[i,k]+Dk-1[k,j]} return Dn
i

Uses only vertices numbered 1,,k (compute weight of this edge)

j k

Uses only vertices numbered 1,,k-1

Uses only vertices numbered 1,,k-1

38