Approximation Theory - Mathematical Tripos - Paper, Exams of Mathematics

This is the Past Paper of Mathematical Tripos which includes Combinatorics, Kruska Katona Theorem, Vertex Boundary, Harper’s Inequality, Usual Compression Operator, Construction of Cellular Homology, Intersection Pairing, Differentiable Map etc. Key important points are: Approximation Theory, Jackson Operator, Second Modulus of Smoothness, Bernstein Polynomial, Sequence of Polynomials, Integral Coefficients, Chebyshev Alternation Theorem, Polynomials of Degree

Typology: Exams

2012/2013

Uploaded on 02/27/2013

lakshanya
lakshanya 🇮🇳

4.4

(25)

98 documents

1 / 5

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
MATHEMATICAL TRIPOS Part III
Monday, 4 June, 2012 9:00 am to 12:00 pm
PAPER 70
APPROXIMATION THEORY
Attempt no more than FOUR questions.
There are SIX 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.
pf3
pf4
pf5

Partial preview of the text

Download Approximation Theory - Mathematical Tripos - Paper and more Exams Mathematics in PDF only on Docsity!

MATHEMATICAL TRIPOS Part III

Monday, 4 June, 2012 9:00 am to 12:00 pm

PAPER 70

APPROXIMATION THEORY

Attempt no more than FOUR questions. There are SIX 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.

Let jn be the Jackson operator, i.e., for a 2π-periodic function f from C(T),

jn(f, x) :=

∫ (^) π

−π

f (x − t)Jn(t) dt, Jn(t) :=

2 πn(2n^2 + 1)

sin^4 nt 2 sin^4 2 t

∫ (^) π

−π

Jn(t) dt = 1.

Prove that, for any f ∈ C(T), we have the estimate

‖jn(f ) − f ‖ 6 c ω 2 (f, (^) n^1 ) ,

where ω 2 (f, t) is the second modulus of smoothness of f.

Hence, derive that if f is twice continuously differentiable, then

En(f ) 6 c 1 n^2 ‖f ′′‖C(T).

a) For f ∈ C[0, 1], write down the definition of the Bernstein polynomial Bn(f ) of degree n, and prove that ‖Bn(f )‖∞ 6 ‖f ‖∞.

b) For a function f ∈ C[0, 1] that takes integer values at x = 0 and x = 1, and for the sequence of polynomials

B n∗(f, x) :=

∑^ n

k=

n k

f

k n

xk(1 − x)n−k^ ,

prove that ‖Bn(f ) − B∗ n(f )‖∞ → 0 as n → ∞. Here, ⌊t⌋ is the largest integer not bigger than t.

c) Hence show that a function f ∈ C[0, 1] is approximable by polynomials with integral coefficients if and only if f (0) and f (1) are integers.

Part III, Paper 70

Let (Ni) and (Mi) be the B-splines bases of degree k − 1 with L∞- and L 1 - normalization, respectively, defined on a knot sequence ∆ = (ti)ni=1+k ⊂ [0, 1].

Given f ∈ C[0, 1], let

PS (f ) := s∗^ =

∑^ n

j=

aj Nj

be the orthogonal projection of f onto S := span (Nj ) with respect to the ordinary inner product (f, g) =

0 f^ (x)g(x)^ dx. Then^ PS^ is also well defined as an operator from^ C[0,^ 1] onto C[0, 1].

a) Show that the max-norm of PS satisfies the inequality

‖PS ‖∞ 6 ‖G−^1 ‖ℓ∞ ,

where G = (gij ) is the Gram matrix with the elements gij = (Mi, Nj ).

b) For linear splines (k = 2) and equidistant ∆, with ti+1 − ti = h for all i, compute the entries of G.

c) Using the fact that G is totally positive, or otherwise, prove the estimate

‖G−^1 ‖ℓ∞ 6 3 , k = 2.

[You may use any appropriate theorems on the inverse of certain matrices if correctly stated.]

Part III, Paper 70

a) Define a multiresolution analysis of L 2 (R) with a generator φ and explain how it is related to existence of an orthonornal wavelet ψ.

b) Prove that the following properties of φ

  1. φ(x) =

n

anφ(2x − n), 2) {φ(· − n)}n∈Z is an orthonormal sequence

are equivalent to

1 ′) f (2t) = m(t)f (t) , m(t) = (^12)

n

ane−int,

|f (t + 2πk)|^2 ≡ 1 a.e.

where f is the Fourier transform of φ, i.e., f (t) = φ̂ (t) =

R φ(x)e

−ixt (^) dx.

c) Verify that conditions 1′) − 2 ′) are fulfilled for the function f = φ̂ defined as

f =

1 , t ∈ [−π, π) 0 , otherwise.

Using the inverse Fourier transform or otherwise, determine the corresponding generator φ.

END OF PAPER

Part III, Paper 70