1. This document contains 5 math problems related to applied and computational mathematics. The problems cover topics like polynomial approximation, finite element methods, iterative methods for solving linear systems, integer factorization, and similarity solutions to partial differential equations.
2. Test-takers are instructed to solve 4 out of the 5 problems presented. The problems vary in difficulty and include proving theorems, analyzing algorithms, describing numerical methods, and finding similarity solutions.
3. The document provides a group test for a math contest, containing 5 challenging applied math and computational math problems for participants to solve 4 of. The problems cover diverse topics and require different proof-based and computational skills.
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 ratings0% found this document useful (0 votes)
31 views3 pages
Applied 2012 Team
1. This document contains 5 math problems related to applied and computational mathematics. The problems cover topics like polynomial approximation, finite element methods, iterative methods for solving linear systems, integer factorization, and similarity solutions to partial differential equations.
2. Test-takers are instructed to solve 4 out of the 5 problems presented. The problems vary in difficulty and include proving theorems, analyzing algorithms, describing numerical methods, and finding similarity solutions.
3. The document provides a group test for a math contest, containing 5 challenging applied math and computational math problems for participants to solve 4 of. The problems cover diverse topics and require different proof-based and computational skills.
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
GROUP TEST
S.-T YAU COLLEGE MATH CONTESTS 2012
Applied Math. and Computational Math.
Please solve 4 out of the following 5 problems.
1. If the function u(x) is in C k+1 (has continuous (k + 1)-th deriva-
tive) on the interval [0, 2], and a sequence of polynomials pn (x) (n = 1, 2, 3, ...) of degree at most k satisfies C 1 (1) |u(x) − pn (x)| ≤ k+1 ∀ 0≤x≤ , n n where the constant C is independent of n, prove C̃ 1 2 |u(x) − pn (x)| ≤ k+1 ∀ ≤x≤ , n n n with another constant C̃ which is also independent of n. 2. Consider the one-dimensional elliptic equation d2 − u(x) = f (x), 0 < x < 1, dx2 with homogeneous boundary condition, u(0) = 0 and u(1) = 0, f ∈ L2 (0, 1). (i) Describe the standard piecewise linear finite element method for this boundary value problem. (ii) Is this method stable and convergent? If so, what is the order of convergence? (iii). In this case, the linear finite element method has a super con- vergence property at the nodal point xj (j = 1, 2, ..., N ), i.e. uh (xj ) = u(xj ), here uh is the finite element solution and u is the exact solution. Could you explain why? 3. Let A = (aij ) ∈ MN ×N (C) be strictly diagonally dominant, that is, N X |aii | > |aij | for all 1 ≤ i ≤ N, j=1,j6=i
Assume that A = I + L + U where I is the identity matrix, L and U
are the lower and upper triangular matrices with zero diagonal entries. 1 2
Now, we consider solving the linear system Ax = b by the following
iterative scheme: (∗) xk+1 = (I + αΩL)−1 [(I − Ω) − (1 − α)ΩL − ΩU )]xk + (I + αΩL)−1 b where Ω := diag(ω1 , ...ωN ) and 0 ≤ α ≤ 1. (When α = 1, it gives the SOR method.) (1) Prove that the linear system Ax = b has a unique solution. (2) Prove that the necessary condition for the convergence of (*) is N Y |1 − ωi | < 1 i=1
(3) Let M = (I + αΩL)−1 [(I − Ω) − (1 − α)ΩL − ΩU )]. Prove that
the spectral radius ρ(M ) of M is bounded by: |1 − ωi | + |ωi |(|1 − α|li + ui ) ρ(M ) ≤ max i 1 − |ωi α|li P whenever |ωi α|li < 1 for all 1 ≤ i ≤ N where li = j<i |aij | P and ui = j>i |aij |. (4) Using (c), prove that the sufficient condition for the convergence of (*) is 2 0 < ωi < for all 1 ≤ i ≤ N 1 + li + ui 4. The famous RSA cryptosystem is based on the assumed difficulty of factoring integers N = pq (called RSA integers) which are products of two large primes p and q which should be kept secret. Currently p and q are chosen to be about 500 bits long, that is, p, q ≈ 2500 . Assume someone uses the following algorithm to find secret n-bit primes p and q to form an RSA integer N = pq: • Choose a random odd 500-bit integer s. • Test the odd numbers s, s + 2, s + 4, etc. for primality until the first prime p is found (note the primality testing is very easy nowdays). • Continue testing p + 2, p + 4, p + 6, etc. for primality until the second prime q is found. • Compute and publish N = pq, but keep p and q secret. How secure is this procedure? Can you suggest an algorithm to factor an RSA integer N = pq generated this way? Note that there are about x/ log x primes up to x, where log x is the natural logarithm. This means that the expected gap between two consecutive n-bit primes is log 2n = n log 2 ≈ 0.69 · n. 3
5. The solution h(r, t) of the following Boussinesq equation describes
the hight of a circular drop of fluid spreading on a dry surface h = 0: µ ¶ ∂h 2 1 ∂ ∂(h2 ) = ∆r (h ) = r , r > 0, t > 1 ∂t r ∂r ∂r with Z ∞ ∂h ¯¯ 1 ¯ = 0, h(r, t)rdr ≡ ∂r r=0 0 64 The solution is positive on a finite range 0 ≤ r ≤ r∗ (t) with h(r∗ (t), t) = 0 defining a moving “edge” position with no fluid outside of the droplet. For r > r∗ (t) truncate the solution beyond the edge to be zero ( h ≡ 0 for r > r∗ (t)). (a): Show that this problem is scale invariant by finding relations h(r, t) = H(T )h̃(r̃, t̃), r = R(T )r̃, t = T t̃ so that the problem for h̃(r̃, t̃) is identical to the original problem. (b): Determine the ODE for the similarity function Φ(η) with h(r, t) = tα Φ(η), r = ηtβ . (c): Determine the explicit solution for Φ(η) and then use h(r, t) = tα Φ(η) R ∞to find r∗R(t)r for t ≥ 1. Hint 0 hrdr = 0 ∗ hrdr.