You are on page 1of 4

Graph Theory Theorems and Exercises

Shourya Pandey
December 13, 2019

Here is a list of miscellaneous graph theory problems you might enjoy; these
might not all be relevant from an olympiad perspective.

1 Enumerative Graph Theory


Question 0. (Cayley’s Formula) Prove that the number of labelled spanning
trees of Kn is nn−2 .

Question 1. Prove that the number of labelled spanning trees in Kr,s is


rs−1 · sr−1 .

Question 2. As n → ∞, what is the expected fraction of leaves in a labelled


tree on n vertices, where each tree is chosen uniformly at random? (Hint: Use
Linearity of Expectation.)

Question 3. The ”wheel graph” Wn is a graph on n+1 vertices V = {1, 2, · · · , n, n+


1} defined as follows. The vertices 1, 2, · · · , n form a cycle in that order. Also,
the vertex n + 1 is connected to each of 1, 2, · · · , n.
Prove that the wheel graph has
√ !n √ !n
3+ 5 3− 5
+ −2
2 2

spanning trees.

Question 4. Prove, combinatorially, that the number of trees tn with the


vertex set [n] satisfies the recurrence
n−1  
X n−2
tn = k tk tn−k
k−1
k=1

Note that this proves the non-trivial identity


n−1
X n − 2
n−2
n = k k−1 (n − k)n−k−2
k−1
k=1

1
Try proving the identity via algebraic methods.

Question 5. Let G be a simple, undirected graph with n vertices labelled


{1, 2, · · · , n}, and m edges.
3(n−1)
(a) Prove that if m > 2 , then there exists an edge e ∈ E(G) belonging
to two different cycles.

(b) Prove that for all odd n ≥ 3, there exists a configuration with exactly
3(n − 1)/2 edges such that each edge belongs to at most one cycle. Such a
configuration is called a triangular cactus.

(c) Let n ≥ 3 be an odd integer. Prove that the number of distinct triangular
cacti is

(n − 1)!
· n(n−3)/2
2(n−1)/2 · ((n − 1)/2)!

Note that the vertices are distinct, so graphs that look the same after removing
the labelled but are different with the labels are to be treated as different.

2 Algorithmic Graph Theory


Question 6. In this setting, we consider a connected graph G(V, E) on n
vertices labelled {1, 2, · · · , n}, and each edge e ∈ E(G) has an integral weight
associated with it, described by the weight function w : E → Z. Our aim is
to find a spanning tree of G with the smallest possible weight; such a spanning
tree is called a Minimum Spanning Tree (MST). Consider the following ”greedy”
procedure to find an MST:
1. Sort the edges of G in non-decreasing order of their weights.
2. Pick the smallest edge. Check if it forms a cycle with the spanning tree
formed so far. If cycle is not formed, include this edge. Else, discard it.
3. Repeat step 2 until we have already included n − 1 edges in our graph.
Prove that this procedure actually finds a minimum spanning tree. This proce-
dure is called Kruskal’s Algorithm.

Question 7. Let G(V, E) be a graph. A vertex cover is a subset S ⊆ V


such that for any edge uv ∈ E, either u ∈ S or v ∈ S (therefore the vertices in
S ”cover” all edges). A minimum vertex cover is a vertex cover of the smallest
possible size. No good algorithm is known for finding the size of the minimum
vertex cover (if you find such an algorithm, you get a million dollars).

Let us restrict ourselves to trees. If T is a tree, prove that there exists a

2
minimum vertex cover of T that does not contain any leaf of T . Using this, find
an efficient algorithm to find a minimum vertex cover of T .

Question 8. The problem of efficient checking whether a graph can be coloured


using 3 colours is also open (and will also fetch you a million dollars). Instead,
find an efficient algorithm that determines whether a graph G can be coloured
using 2 colours (that is, the algorithm determines whether a graph is bipartite
or not).

3 Turán’s Theorem
Question 9. (Mantel’s Theorem) Prove, via induction on n, that a graph
2
on n vertices that avoids a K3 has at most n4 edges. In the induction step, go
from n to n + 2.

Question 10. Suppose G is a graph on n vertices that has no K4 . We want to


2
show that G has at most n3 edges.

1. Prove that the statement is true if n = 1, 2 or 3.


2. Suppose the statement is true for n. We prove it for n + 3 now. Let G be
a graph on n + 3 vertices that has no K4 . If G had no K3 , then by the
previous question, we are done. Otherwise, suppose G has a K3 . Prove
that the number of edges across this K3 and the remaining n vertices is
at most 2n.
3. Complete the proof by induction.
Question 11. (Turan’s Theorem) Prove, via induction on k, the following
theorem: A graph G on n vertices that avoids a Kk , where k ≥ 3, has at most
 2 
n 1
1−
2 k−1

edges. Also show that there exists a graph for which equality holds.

4 Graph Colourings
Question 13. Consider the infinite graph G whose vertices are lattice points
on the Euclidean Plane. Two vertices (a, b) and (c, d) are adjacent iff the line
segment joining them has no other lattice point on it. What is the chromatic
number of G?

Question 14. Let G be a graph with chromatic number k, and let T be a


tree with k vertices. Suppose the vertices of T are coloured with k colours, so
that each vertex receives a distinct colour. Prove that in any k-colouring of

3
G, there exists a ’coloured copy’ of T . More precisely, there is a one-to-one
function f from V (T ) to V (G) such that v and f (v) have the same colour for
all v ∈ V (T ), and f (u)f (v) is an edge in G for all edges uv ∈ T .

Question 15. Let G be a graph with vertex set [n], and let P (k) denote
the number of proper k-colourings of G. Show that P (x) is a polynomial in
x of degree n, called the chromatic polynomial of G. Give examples of two
non-isomorphic graphs whose chromatic polynomials are the same.

Question 16. A total k-colouring of a graph is an assignment of one of k


colours to each vertex and each edge of a graph, such that any two adjacent or
incident elements have distinct colours. In other words, adjacent vertices must
have distinct colours, edges with common endpoint must have distinct colours,
and the colour of an edge must be different from that of its endpoints. An old
conjecture states that every graph with maximum degree ∆ has a total (∆ + 2)-
colouring. Show that this is true for ∆ = 2 and 3.

You might also like