Graphs

CSC2001F Data Structures

Jan Buys <[email protected]>

Department of Computer Science

School of IT
University of Cape Town

2024
(Based on slides from Hussein Suleman)
Graphs
 A graph is a set of nodes/vertices and a set of edges connecting
vertices
 V: set of all vertices/nodes
 E: set of all edges. Each edge (v,w) ∈E connects two nodes v and w
 G = (V, E)

 Graphs are more general than trees: no constraints on edges

department of computer science

Graph Visualized
 V={A,B,D,G,E,F}
B
A  E={(A,B),(A,D),
(A,E),(B,G),(D,B),
(D,E),(D,F),(G,F),
D G
(E,F)}
 |V|=6

E
 |E|=9
F

department of computer science

Graph Examples
 Transportation networks

https://groundup.org.za/article/which-parts-of-metrorail-are-still-closed/
department of computer science
Graph Examples
 Knowledge graphs

department of computer science

Graph Examples
 Social network graphs

https://upload.wikimedia.org/wikipedia/commons/9/9b/Social_Network_Analysis_Visualization.png

department of computer science

Why Graphs
 More problem domains:
 Internet architecture
 Task dependencies in operating systems
 Electricity distribution

 We can use graphs to solve important problems:

 Find shortest path between nodes
 Analyse community structures
 Find related information

department of computer science

Some Types of Graphs
B B
A A

D D

Directed, unweighted Undirected, unweighted

4 B 4 B
A A
2 2
5 5
D D

Directed, weighted Undirected, weighted

department of computer science

Definitions
 A node w is adjacent to v if (v,w)∈E
 A path is a sequence of nodes connected by edges
 The length of a path is the number of edges (unweighted) or the sum
of costs (weighted)
 A cycle is a path that begins and ends at the same node and has at
least 1 edge
 A directed acyclic graph (DAG) is directed and has no (directed) cycles
 A dense graph has many edges – a sparse graph has relatively few
edges
 In most applications, graphs are sparse:|E| is typically much less
than |V|2 (all possible edges)

department of computer science

 Storing Graphs: Adjacency Matrix

4 A B D E F G
B
A A 4 1 5
1
1 2 B 1
D G D 2 3 8
5
8
E 4
3
2
F
E F
4 G 2

department of computer science

 Storing Graphs: Adjacency Lists
A: (B,4) (D,1) (E,5)
4 B
A B: (G,1)
1
1 2 D: (B,2) (E,3) (F,8)
D G
E: (F, 4)
5
8 F:
3
2 G: (F, 2)
E F
4
Lists are better for sparse
graphs – use less memory
department of computer science
 Graph traversal
 Visit all the nodes in a (connected) graph
 Find the shortest path from a given node to all other nodes
 Assume an unweighted graph: weights are equal for every
edge (weights are nominally 1)
 Basic idea: Perform a breadth-first search (BFS) using a
queue to maintain list of nodes to visit
 Generalization of level order tree traversal

department of computer science

BFS: Algorithm
 Add start node to queue with 0 distance
 While queue is not empty:
 Get node v from queue
 For each edge from v to w:
 If distance to w is infinity (w not yet processed)
 Distance to w = distance to v + 1
 Path to w goes through v
 Add w to queue

department of computer science

 BFS: Example 1/5
0,null Queue: A
B
A
Distance to A set to 0
D G
A has no previous node
All other distances set to
infinity
E F

department of computer science

 BFS: Example 2/5
0,null Dequeue A; Update distances
1,A
B for B=1, D=1, E=1; previous
A
node in each case is A;
enqueue B/D/E;
D 1,A G

Queue: B, D, E

E F
1,A

department of computer science

 BFS: Example 3/5
0,null Dequeue B; Update distance
1,A
B for G=2; previous node is B;
A
enqueue G
2,B
D 1,A G
Queue: D, E, G

E F
1,A

department of computer science

 BFS: Example 4/5
0,null Dequeue D; Update distance
1,A
B for F=2; previous node is D;
A
enqueue F
2,B
D 1,A G
Queue: E, G, F

E F 2,D
1,A

department of computer science

 BFS: Example 5/5
Dequeue E
0,null 1,A
B (already processed – not infinity)
A
Dequeue G
2,B
(already processed – not infinity)
D 1,A G
Dequeue F
(already processed – not infinity)
No more items to process
E F 2,D
1,A

department of computer science

BFS: Code

From Weiss

department of computer science

BFS: Complexity
 O(|V| + |E|)
 Why:
 We process each vertex once.
 We process each edge once.

 If the graph is dense, then this could become O(|V|2)

because |E| is then close to |V|2.

department of computer science