Single-Source Shortest Paths
Shortest-path problem
• In a shortest paths problem, we are given a weighted,
directed graph G = (V, E) with weight function
w : ER mapping edges to real valued weight.
The weight of path p = { v0, v1, …….vk} is the sum of
the weights of its constituent edges:
w(p)= ∑ w ( vi-1, vi ).
i = 1 to k
2
Shortest-path problem
We define the shortest-path ( SP ) weight from u to v by
δ (u , v) = min{w( p ) : u → v}, ∃a path from u to v.
p
∞, otherwise.
A SP from vertex u to vertex v is then defined
as any path with weight w ( p ) = δ ( p ).
3
Variations in SPs
Single-source shortest-paths problem:
Given a graph G = ( V, E , W ), we want to
find shortest path from a given source
vertex s ε V to every vertex v ε V.
4
Variations in SPs
Single-destination shortest-paths problem:
Given a graph G = ( V, E , W ). Find a SP to
given destination vertex t from every vertex
v. By reversing the direction of each edge
in the graph, we can reduce this problem to
a single-source problem.
5
Variations in SPs
Single-pair shortest-paths problem:
Find a SP from u to v for given
vertices u and v. If we solve the
single-source shortest path
problem with source vertex u, we
solve this problem also.
6
Variations in SPs
All-pairs shortest-paths problem:
Find a SP from u to v for given vertices u
and v. This problem can be solved by
running a single-source algorithm once
from each vertex; but it can usually be
solved faster, and its structure is of
interest in its own right.
7
Negative Weight Edges
In some instances of single-source shortest-path
problem, there may be edges whose weights are
negative. If the graph G = ( V, E ) contains no
negative-weight cycles reachable from the source s,
then for all v ∈ V, the shortest-path weight δ ( s, v )
remains well defined, even if it has a negative value.
8
Negative Weight Edges
If there is a negative cycle reachable from s,
to a vertex on the cycle can be a shortest
path a lesser-weight path can always be
found that follows the proposed “shortest”
path and then traverses the negative weight
cycle. If there is a negative cycle on some
path from s to v, we define δ ( s, v ) = - infinity9
Shortest-path tree rooted at s
s s
10 10
1 1
10
1 x w 1 x w
t t
1 10 10 1 10
u
10 y y
v 1 10 v
1 1
z z
Original Graph G Shortest-path tree rooted at s
10
Predecessor sub-graph
During the execution of a shortest-paths algorithm,
however, the π values need not indicate shortest
paths. As in BFS, we shall be interested in the
predecessor sub-graph Gπ= (Vπ, Eπ) induced by
the π values. Here again, we define the vertex set
Vπ to be the set of vertices of G with non-NIL
predecessors, plus the source s:
Vπ = { v ∈ V: π [ v ] = NIL} U { s }
The directed edge set
π [ v ], v ) ∈ E : v ∈ Vπ- { s }}.
Eπ = { (π 11
Predecessor graph
π[s] π[t] π[u] π[v] π[w] π[x] π[y] π[z]
NIL s NIL t s t x v
s s
10 10
1 1
10
1 x w 1 x w
t t
1 10 10 1 10
u
10 y y
v 1 10 v
1 1
z z
Original Graph G Shortest-path tree rooted at s 12
Initialize-Single-Source (G, s)
Initialize-Single-Source(G,s)
{ for each vertex v ∈ V[G]
do d[v] ∞
π[v] NIL // end for
d[s]0
}
13
Relaxation
Relax(u,v,w)
{ if d[v]>d[u]+w(u,v)
then d[v]d[u]+w(u,v)
π[v]u
}
14
Relaxation
Relax(u,v,w) Relax(u,v,w)
s
sv
if w(u,v)=2 (<3) π[v]u
s
4 6
u w(u,v) v
4 7 s
u w(u,v) v sv
if w(u,v)=4 (>3)
4 7
u w(u,v) v
15
The Bellman-Ford algorithm
Bellman-Ford(G,w,s)
{ Initialize-Single-Source(G,s)
for i = 1 to |V-1|
do for each edge (u,v)∈ E
do Relex(u,v,w)
for each edge (u,v) ∈ E
do if d[v]> d[u]+w(u,v)
then return false
return true }
16
The Bellman-Ford algorithm
5 5
(a) ∞ -2 ∞ (b) 6 -2 ∞
6 6
8 -3 8 -3
s 0 7 s 0 7
-4 -4
7 7
∞ 2 ∞ 7 2 ∞
9 9
Each pass relaxes the edges in lexicographic order: (u, v), (u, x), (u, y), (v, u), (x, v),
(x, y), (y, v), (y, z), (z, u), (z, x).
17
The Bellman-Ford algorithm
5 5
(c) 6 -2 4 (d) (e) 2 -2 4
6 6
8 -3 8 -3
s 0 7 s 0 7
-4 -4
7 7
7 2 2 7 2 -2
9 9
18
The Bellman-Ford algorithm
The Bellman-Ford algorithm runs in time O(VE), since
the initialization in line 1 takes O(V) time, each |V|-1
passes, over the edges in lines 2-4 takes O(E) time, and
the for loop of lines 5-7 takes O(E) time.
19
Single-source shortest paths in DAGs
DAG-Shortest-Path(G,w,s)
{ Topologically sort V[G]
Initialize-Single-Source(G,s)
for each u taken in topological
order
do for each v ∈ adj[u]
do Relax(u,v,w)
20
Single-source shortest paths in DAGs
DAG-Shortest-Path(G,w,s)
{ Topologically sort V[G]
Initialize-Single-Source(G,s)
for each u taken in topological
order
do for each v ∈ adj[u]
do Relax(u,v,w)
21
Single-source shortest paths in DAGs
• Topological sort needs O(|V|+|E|)
• Initialize-Single-Source requires O(|V|)
• Nested for loops embedded in the
algorithm takes O(|V|+|E|)
22
DAG-Shortest-Path
(a) 6 1
5 s 2 7 -1 -2
∞ 0 ∞ ∞ ∞ ∞
3 4
2
(b) 6 1
5 s 2 7 -1 -2
∞ 0 ∞ ∞ ∞ ∞
3 4
2
23
DAG-Shortest-Path
(c) 6 1
5 s 2 7 -1 -2
∞ 0 2 6 ∞ ∞
3 4
2
(d) 6 1
5 s 2 7 -1 -2
∞ 0 2 6 6 4
3 4
2
24
DAG-Shortest-Path
(e) 6 1
5 s 2 7 -1 -2
∞ 0 2 6 5 4
3 4
2
(f) 6 1
5 s 2 7 -1 -2
∞ 0 2 6 5 3
3 4
2
(g) 6 1
5 s 2 7 -1 -2
∞ 0 2 6 5 3
3 4
2
25
Dijkstra’s algorithm
Q: Priority queue with d as the key
Dijkstra(G,w,s)
{ Initialize-Single-Source(G,s)
Q = V[G]
while Q is not empty
do u = Extract-Min(Q)
for each v ∈ adj[u]
do Relax(u,v,w)
}
26
Dijkstra’s algorithm
(a) 1 (b) 1
10 ∞ ∞ 10 10 ∞
9 9
s 2 3 6 s 2 3 6
0 4 0 4
7 7
5 5
∞ ∞ 5 ∞
2 2
27
Dijkstra’s algorithm
(c) 1 (d) 1
10 8 14 10 8 11
9 9
s 2 3 6 s 2 3 6
0 4 0 4
7 7
5 5
5 7 5 7
2 2
28
Dijkstra’s algorithm
1 1
(e) (e)
10 8 9 10 8 9
9 9
2 3 2 3
s 6 s 6
0 4 0 4
7 7
5 5
5 7 5 7
2 2
29
Dijkstra’s algorithm
• Linear array O(|V|2) when G = V ( V, E, W)
is represented as adjacency matrix.
• Binary heap O(|E|log|V|)
• Fibonacci heap O(|E|+|V|log|V|)
30
