0% found this document useful (0 votes)
29 views3 pages

2018 Iiipaper 110

This document contains instructions for a mathematical exam with 4 questions on extremal graph theory. It provides the requirements for the exam, including that candidates should attempt no more than 3 of the 4 questions. The questions cover topics such as the Erdos-Stone theorem, extremal functions related to numbers of copies of graphs, Szemerédi's regularity lemma, and bounds related to the Ramsey number of a complete graph.

Uploaded by

Kaustubh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
29 views3 pages

2018 Iiipaper 110

This document contains instructions for a mathematical exam with 4 questions on extremal graph theory. It provides the requirements for the exam, including that candidates should attempt no more than 3 of the 4 questions. The questions cover topics such as the Erdos-Stone theorem, extremal functions related to numbers of copies of graphs, Szemerédi's regularity lemma, and bounds related to the Ramsey number of a complete graph.

Uploaded by

Kaustubh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

MATHEMATICAL TRIPOS Part III

Friday, 1 June, 2018 9:00 am to 11:00 am

PAPER 110

EXTREMAL GRAPH THEORY

Attempt no more than THREE questions.


There are FOUR questions in total.
The questions carry equal weight.

STATIONERY REQUIREMENTS SPECIAL REQUIREMENTS


Cover sheet None
Treasury Tag
Script paper

You may not start to read the questions


printed on the subsequent pages until
instructed to do so by the Invigilator.
2

1
State the Erdős-Stone theorem.
Let t, r ∈ N be given and let (Gn ) be a sequence of graphs, such that Gn has order n
n
and size (1−1/r+o(1)) 2 . Show that if Kr+1 (t) is not a subgraph of Gn then Gn contains
an r-partite subgraph H of minimum degree (1 − 1/r + o(1))n.
Show that if χ(F ) = r +1 and G is extremal for F then G itself has minimum degree
(1 − 1/r + o(1))n, where n = |G|.
Let r, s ∈ N be fixed. Show that, for large n, the unique extremal graph for sKr+1
(s disjoint copies of Kr+1 ) is Ks−1 + Tr (n − s + 1) (that is, Ks−1 with every vertex joined
to every vertex of a Turán graph on the remaining vertices).
[Hint. As usual, assign each vertex of G to the class in H in which it has fewest
neighbours. Consider the cases (a) some vertex of G has more than o(n) neighbours in its
own class, applying induction on s, (b) some class contains s independent edges and (c)
each class has a set of 2(s − 1) vertices meeting all edges in its class.]

2
Let G be a graph of order n and let kp (G) be the number of copies of Kp in G. Let
c ∈ R and let f (G) = k2 (G) − ck3 (G). Show that, amongst graphs of order n, the function
f (G) takes its maximum on some complete multipartite graph.
Deduce that, for 0 6 θ 6 1, if G has (1 − θ)k2 (T2 (n)) + θk2 (T3 (n)) edges then it
contains at least θk3 (T3 (n)) triangles, where Tp (n) is the p-partite Turán graph of order n.
Suppose now that n > 4 is even, and that G has n2 /4 + 1 edges. Show that G has
at least n/2 triangles.
[Hint. Consider the case where every edge is in some triangle, and the other case.
For the latter, apply induction on n: if uv is in no triangles, how many edges can meet uv?]

3
State and prove Szemerédi’s Regularity Lemma. [You may assume that if U ′ ⊂ U
and W ′ ⊂ W satisfy |U ′ | > (1 − δ)|U | and |W ′ | > (1 − δ)|W | then |d(U ′ , W ′ ) − d(U, W )| 6
2δ, and also any quantitative form of the Cauchy-Schwarz inequality that you need.]
Let G be a graph of order n with 51n2 /200 edges. Explain why there exists some
c > 0 such that V (G) contains disjoint sets U1 , U2 , U3 with |Ui | > cn, 1 6 i 6 3, and with
each pair (Ui , Uj ), 1 6 i < j 6 3, being 10−3 –uniform and having density at least 10−3 .

Part III, Paper 110


3

4
Recall from the lectures that c(t) = inf{c : e(G) > c|G| =⇒ G ≻ Kt }.

Show that there exists some constant β > 0 such that c(t) > βt log t for large t.
[Standard probabilistic facts may be assumed if stated clearly.]

Show that c(t) 6 7t log t for large t.
[You may assume that, for every integer k, if e(G) 6 11k|G| then G ≻ H where
|H| 6 11k + 2 and 2δ(H) > |H| + 4k − 1.]
Show that, if n > 4 and G has n vertices and 2n − 2 edges, then G ≻ K4 .
[Hint. Apply induction on n, considering the case where every edge lies in at least
two triangles and the other case.]
Does the same statement hold with 2n − 2 replaced by 2n − 3?

END OF PAPER

Part III, Paper 110

You might also like