Graph Theory

Graph theory is the study of the

properties of graph structures. It
provides us with a language with
which to talk about graphs.
 Degree
• The degree of a vertex is the number of
edges incident upon it.
• The sum of the vertex degrees in any
undirected graph is even (twice the
number of edges).
• Every graph contains an even number of
odd-degree vertices.
 Connectivity
• A graph is connected if there is an
undirected path between every pair of
vertices.
• The existence of a spanning tree is
sufficient to prove connectivity.
• The vertex (edge) connectivity is the
smallest number of vertices (edges) which
must be deleted to disconnect the graph.
 Some terms
• articulation vertex  biconnected
• bridge  edge-biconnected
• Testing for articulation vertices or bridges i
s easy via brute force.
 Cycles
• Eulerian cycle (path): a tour which visits ev
ery edge of the graph exactly once.
• Actually, it is a circuit, not a cycle, because
it may visit vertices more than once.
• A mailman’s route is ideally an Eulerian cy
cle, so he can visit every street (edge) in t
he neighborhood once before returning ho
me.
• An undirected graph contains an Eulerian
cycle if it is connected and every vertex is
of even degree.
• A Hamiltonian cycle is a tour which visits e
very vertex of the graph exactly once.
• The traveling salesman problem asks for t
he shortest such tour on a weighted graph.
 Planer Graph
• Euler’s formula:

n-m+f=2

• n:# of vertices
• m:# of edges
• f: # of faces
• Trees: m=n-1, f=1
• Cubes: n=8, m=12, f=6
 MST
• Kruskal’s algorithm:

starting from a minimal edge

• Prim’s algorithm:

starting from a given vertex

– How about maximum spanning tree
– and Minimum Product spanning tree
 Kruskal’s Algorithm

• Algorithm Kruskal(G)
• Input ： G=(V, E) 為無向加權圖 (undirected weighted graph) ，其中
V={v0,…,vn-1}
• Output ： G 的最小含括樹 (minimum spanning tree, MST)
• T← //T 為 MST ，一開始設為空集合
• while T 包含少於 n-1 個邊 do
• 選出邊 (u, v) ，其中 (u, v)E ，且 (u, v) 的加權 (weight) 最小
• E←E-(u, v)
• if ( (u, v) 加入 T 中形成循環 (cycle) ) then 將 (u, v) 丟棄
• else T←T(u, v)
• return T
Kruskal’s Algorithm -Construct MST
 Prim’s algorithm

• Algorithm Prim(G)
• Input ： G=(V, E) 為無向加權圖 (undirected weighted graph) ，其中 V={v
0,…,vn-1}
• Output ： G 的最小含括樹 (minimum spanning tree, MST)
• T← //T 為 MST ，一開始設為空集合
• X←{vx} // 隨意選擇一個頂點 vx 加入集合 X 中
• while T 包含少於 n-1 個邊 do
• 選出 (u, v)E ，其中 uX 且 vV-X ，且 (u, v) 的加權 (weight) 最小
• T←T(u, v)
• X←X{v}
• return T
 8 7 8 7
b c d b c d
4 9 4 9
2 2
a 11 i 4 14 e a 11 i 14 e
6 4
6
8 7 10 7
8 10
h g f h g f
1 2 1 2

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

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

8 7
b c d
4 9
2
a 11 i 4 14 e
6
8 7 10
h g f
1 2
 One-to-all shorted path

• Dijkstra 演算法 : Dijkstra 演算法屬於求取單一來源 (source) 至所有目標 (destination)

頂點的一至多 (one-to-all) 最短路徑演算法
 All-pair shortest path

• Floyd-Warshall 的所有頂點對最短路徑 (all-pair shortest path) 演算法：